Waveform analysis of geographic patterns recorded on visible and infrared imagery

PAGE 67

57 Conrac monitor were connected to the video tape-recorder as the imagery tape was replayed (see Fig. 13) . The waveform monitor vras adjusted appropriately and a Textronix oscilloscope was attached to its display tube (see Fig. 14). This oscilloscope camera is equipped with a Polaroid film back capable of handling the roll film used to record electronic traces such as waveforms. Polaroid Type 3000 film (rated 3000 ASA) was used in this introductory experiment, but proved to be unsatisfactory because the tine required for the exposure was greater than one con^lete cycle of the field rate (30 scans per second) « producing a blurred trace. An additional problem was the length of tine required for the recording of one waveform. This arose when the waveform magnification feature of the waveform monitor was tested. In the IX waveform magnification position, the entire waveform is conqpressed in horizontal-zucis length to fit the five-inch display tube (see Fig. 14) . In that position, the entire waveform could be seen at one time, but it was sufficiently coiiQ>ressed to render it uninterpretable. The instrument also includes capability for 5X and 25X magnifications of waveforms. The 23X increase in size proved to be too large, but the 5X magnification allowed recording

PAGE 68

58 < UJ < o UJ a. o o (/) o o o CO O U. U > < CD O q: o o LJ •H O CO O § :5 (0 g > •H ^ O O (U U M-4 O c o •H +J fO >H +J m 3 rH H •H U H +J ro g CO

PAGE 69

59 Fig. 14, Arrangement of components used in oscilloscope recording of waveforms displayed on Tektronix Model 529 Waveform Monitor. Camera is mounted on hinges and is swung out for inspection of waveform monitor display tube. Waveform can be viewed through upper part of camera while film is being exposed.

PAGE 70

60 of one entire expanded waveform on four or five photographs, unfortunately, this requires a considerable period of tiaa« at least 20 minutes, for the recording of one single waveform. Because of the film speed and recording cycle time problems, it was decided to test the capability of another type of waveform monitor, called the Colorado Video Model 302 Video Analyzer (see Fig. 13) . This was a new product capable of simultaneously displaying on a conventional monitor the image, the scan line selected, and the resultant waveform superimposed on a changeable and positionable grid (see Fig. 16). in short, this video Analyzer promised to be the answer to equipment selecticm problems, since the brightness of the waveform trace and its san^ling grid made it possible to record the waveform with conventional films (see Figs. 17 and 18) . Various films were tested to determine the best alternative for recording waveforms produced on the Model 302 (see Figs. 19 and 20) . All of these films proved to have sufficient resolution to record the image, scan line, grid, and waveform. In fact, their resolution was so great that it was discerned that the waveform produced by the Model 302 was not really a line but a series of dots simulating a line. It was also discovered that the

PAGE 71

61 Fig, 15. Various components used to present and record waveforms produced by the Colorado Video Model 302 Video Analyzer, Note the arrangement of image, waveform, and grid on the Conrac Monitor.

PAGE 72

62 Fig, 16. A 35-millimeter photograph of monitor face. Note scan line vertically presented on left side of image, and waveform indicating gray-tone levels intersected by scan line. Horizontal line at top of image is electronic "straight edge" used for correlating specific locations on scan line and waveform. These elements have been generated by the Model 302 Video Analyzer.

PAGE 73

63 ill I -J < UJ z o (T O O UJ a: I o Io CO o u. UJ > < 5 UJ < c m VI +» I rH (0 0) M-l O u o o V o o c 1^ u p a 3 p M ^k^ a; o E 0»

PAGE 74

64 Q LU CL < II O LU Q O a: o LiJ > < a: < o o >LU a: u: uj o < o X 0. G a 9 10

PAGE 75

65 I #1 f H Fig. 19, A 35-millimeter camera being used to photograph simultaneously presented waveform, image, and sampling scan line on monitor, while sampling location and other pertinent data are noted. Instrument in right lower foreground is Colorado Video Model 302 Video Analyzer. Controls on front of Model 302 are for choosing location and width of scan line.

PAGE 76

66 Fig, 20. A 4" x 5" Speed Graphic with Polaroid Sheet Film Pack being used in exposure test of Polaroid 55 P/N film. Note grid superimposed on monitor test pattern by Colorado Video Model 302 Video Analyzer.

PAGE 77

67 Instrtuoent was very hard to adjust* especially in the calibration of a 100 percent black-to-'whlte ratio. These factors together would not have negated the use of the Model 302 « since It was very ccMivenient to have the Image, scan line* and waveform slmultemeously presented and easily photographable. The one factor vrhlch finally made this Instrument unusable was the very excessive rise time needed to record gray-tone fluctuations. The time required was so great that only the peaks of excursions v^re considered valid samples of gray-tone level. During the period In which the Blodel 302 Video Analyzer was being evaluated, attention was drawn to a new faster speed oscilloscope-trace recording film, called Polaroid Polascope Type 410 (see Pig* 21) . This film has a 10,000 ASA speed and has proven capable of recording waveforms at faster exposure times than the normal television field rate. Because of this film's capability and the much shorter rise times characteristic of the Tektronix Model 329 Waveform Monitor, It was decided to return to the use of that instrument for the monitoring and recording of the waveforms produced by the scan lines previously selected. At the same tine, a 35-milllmeter camera was used to photograph the position of the scan lines as they actually tra-

PAGE 78

68 Fig, 21. Oscilloscope camera being used to record section of waveform using Polaroid Type 410 film. Photograph has just been removed from camera back.

PAGE 79

69 versed the Images, providing a permanent record of their location (see Fig. 22) . Gray-Tone Level Control Since the gray-tone level finally displayed in a waveform can be altered at many different previous points in the instrumentation* it was necessary to evolve a procedure for maintaining consistent gray-tone levels through the electronic system. Zn any rescaling and/or reproduction of the imagery prior to %raiveform monitoring « an automatic exposure timer was used in conjunction with the photographic enlarger* and a film chip densitoaeter provided the final control on graytone. This alleviated any problem that might arise from inconsistent gray-tones appearing on the imagery. Before each real-time or video-tape-recording session began « the television camera and/or video-tape-recorder was calibrated for normal contrast brightness and focus using a conraercial broadcast test pattern. The camera and recorder controls were then left constant during the session. Each cc»nplete set of imagery was scanned and its waveforms recorded during one session* eliminating any Inconsistencies engendered by having to recalibrate the camera or taperecorder.

PAGE 80

70 Fig. 22, A 35-milliineter photograph used for locating traverse of scan line. Image is nine-lens multiband photograph. Frame 42, Lens 2, Line 2,

PAGE 81

71 Gray-tones presented to the conventional monitor are not altered by any adjustment of that monitor. Consequently, It Is possible to adjust the brightness and ccxitrast of the monitor to provide the best picture for selecting and recording the location of the scan line (see Pig. 22). Because of Its primary testing function* the Tektronix Model 529 has several controls capable of changing the character and gray-tone level of a v^veform. Primary of these Is the calibrator switch, which provides a onevolt peak-to-peaX reference voltage for calibrating a 100 percent black-to-whlte contrast signal generated by a tendlvlslon black-to-whlte "stair-step" test pattern. This signal Is then posltlonable on the 100 percent grid scale Illuminated graticule placed In frcmt of the display cathoderay tube. A frequency-response switch allows the selection of several conventional response characteristics. In the "flat" response position, considerable signal noise Is Included In the waveform. Ninety percent of this noise, however. Is removed In the "IEEE" (Institute of Electrical and Electronic Engineers) response position, producing a waveform which Includes a meuclmum amount of gray-tone Information

PAGE 82

72 and a minirauro amount of signal noise. Other controls of the wavefonn monitor include the magnification feature mentioned previously, a trace-illumination control, and a control for changing the lateral position of the waveform in order to view successive portions of it when either of the magnification positions is used. Variations in exposure timing of the 08cillo8C(^>e camera do not influence the gray-tone levels presented on the waveform monitor. Once a usaible exposure time and lens setting are found, they are maintained throughout the recording. Future Instrumentation Additions Since one of the major dra%tfbacks in the waveform analysis procedure is the recording of the actual waveforms using an oscilloscope camera, it is planned to include an X-Y axis strip-chart recorder (see Figs. 23 and 24) among the experimental components. This will permit direct graphical recording of the traces displayed on the waveform monitor cathode-ray tube, without having to resort to the laborious process of putting the waveform recorded on Polaroid film into analyzable graphical form. Additional ins t rumen taticm plans also include the use of television cameras, monitors, and waveform monitors

PAGE 83

73 ill < CO tr

PAGE 84

74 Q LU < 9 1 > o u. o UJ > < o UJ > q: cr LU ctr < X o I q: CD < pj™

PAGE 85

75 having several different scem-llne rates, such as 729, 875, emd 945 lines, and higher frequency response, such Instrumentation sets will permit the comparison of images using systans having Increasingly higher resolution capability. Graphical Presentation of Waveforms Since each of the original polaroid photographs (Fig. 21) included only part of the waveform magnified to 5X, it vras necessazy to mosaic these photographs in order to have a oonposite record of each waveform. This problon had been anticipated before the photography was begun, and hence sufficient overlay between successive Polaroid photographs was retained so that accurate linear mosalclng might be accomplished. After each mosaic was completed, it was projected onto a sampling grid using a Saltzman Auto-Focus Projector. During this process, the waveform was enlarged approximately 3X so that a 100 percent gray-tone fluctuation occupied a five-inch vertical interval on the sampling grid (see Fig. 25) • After each waveform had been enlarged, the scan lines traversing each image were segmented according to the specific terrain types crossed. The distances representing these individual terrain-type occurrences were then scaled to the enlarged waveform size, and the specific parts of the wave-

PAGE 86

76 0> t&J UJ < £ lAl > < IaJ Id 4J !3 g > o t) U P O in t'

PAGE 87

77 form corresponding to the terrain-type occurrences were identified and isolated by vertical lines (see Figs. 26 through 37) . The sanpling grid used for the enlarged waveforms is standard graph paper* with ten divisions per inch along both X and Y axes. This type of graph paper permits the accurate sampling of 1 percent gray-tone level chemges along the y-axis and 0.05-inch distance along the x-aucis. Statistical Presentation of Gray-Tone Levels In any analysis of waves which are not regular mathematical functions such as simple periodic sinusoidal waves, the conventional wave equations are not applicable. Even Fourier Analysis has been applied to certain waveforms composed of higher order harmonics. This tool, however, does not appear to be applicable to waveforms representative of remote-sensor imagery, which are apparently too 4 con$>lex to be resolved into hazmonic con^}onents. Any number of measurements of waveform parameters and subsequent calculations could be made from waveforms, but all such measurements would depend to some degree on the roost 4 Hubbs, op. cit ., p. 109.

PAGE 88

78

PAGE 89

79 9-(SSVU9) ai3U N3dO 1-0*0)1 QBAWNn 5-(SS»«OI0">3IJ N3<)0 -(U3A03 ONI an3ld N3dO »-|SSTII9l 013li N3dO CM Ui < IT Z-OVOH OBAW 101 ONIXUTd a3AW z-issvMs) ansid N3dO i-avoH osAw -(SSVM3) anaj N3dO c

PAGE 90

80 C ro ••2 iH E CO c 0) rH I 0) c •H C CP C •H 4-» c 0) CO ^ (1) CO > C ro Qj ? hJ m » i 0) u (0 re I CD (%) n3A3T 3N0i-AVMS

PAGE 91

81 9-(SSV(l9) ai3IJ N3dO i-avou asAWNn S-ISSVdS) 013li N3o: UJ 1 o z < m »-(SSVII9) 013li N3d0 Z-OYOM a3/»i ~t~
PAGE 92

82 13

PAGE 93

83 § 8 (%) n3A3n 3N0i-AVd9 G m rH i (0 C <0 iH I Q) C •H c +* C
PAGE 94

84 (1) m •rt +> m i Xi u c ro C •r4 Q) fli
PAGE 95

85 u •H 4J (0 s c (0 &« +J c (1) (0 0) u e u o m ra o c o •H p (0 4) m o •H C cu < h •' PL4 OU T3A3n 3N01-AV(i9

PAGE 96

86 o •H 4J (0 § U ,c o c ro a. ^
PAGE 97

87 e u c •H c (0 V4 o > O C o •w +» m +J c
PAGE 98

88 •H 4J C Q) m 6

PAGE 99

89 E u Q) -P ty> C •H P C 0) (0 d) u n E u o > o c o +j ro +J C
PAGE 100

90 conventicmal of all variables dealing with waves. These variables are amplitude and frequency. Amplitude of a vmve can be simply stated as the distance away from the base line or reference line of any point on the wave as measured alcmg the Y-axis. In this case, the amplitude of any point on the waveform is in reality the gray-tone level of that point, expressed as a percentage of white (100 percent) . On Figure 25, for example, gray-tone level, as san^lcKl every 0.10 inch, would be for Terrain Type "A" (reading left to right) t 44 percent, 38 percent, 32 percent, etc. Frequency, on the other hand, is simply the number of fluctuations occurring per unit length of the curve. A description of the derivation and significance of the statistical parauneters pertaining to waveforms which have evolved during this research will occupy several of the following pages. The Number of Samples (n) A certain number of gray-tone level sauries are taken from each of the terrain-type occurremces. In Figure 25, for exan^le, in Terrain Type "A" n would equal 20 samples. Bach 0.1-inch minor grid division is used as a sampling point, and all such divisions crossed or intersected (see left-hand edge of Terrain Type "A") are counted as

PAGE 101

91 sampling points. (N.B, Note that each individual terraintype occurrence is coded by a nixmber appearing after lt« both on the large graphs « i.e.. Figures 26 through 37 « and on all tables appearing in the Appendices.) The Summation of Gray-Tone Levels (^x) The letter x is used to denote any specific graytone level* and the parameter Xx is simply the total of all gray-tone levels in any individual terrain-type occurrence. The Average Gray-Tone Level (x) This parameter is simply the total of the gray-tone levels sampled in any terrain-type occurrence (^^x) divided by the niaaber of seui^les (n) • It denotes the average* or Man* gray-tone level occurring in any particular terraintype occurrence which has been traversed by a scan line. The unit used with x is percent. The Number of Excursions (E) As previously noted* a waveform "excursion" is a positive or negative peaking (see Pig. 25) of the curve and represents a local maximum gray-tone level. The symbol B represents the total nuznber of excursions appearing in any waveform segment representing a discrete terrain type. Inflection points* or those positions on the curve where the

PAGE 102

92 rotation of a tangent to the cnurve chemges direction, are counted as one-half of a peaked excursion, since they do represent significant gray-tone fluctuation* although not sufficient to produce a positive or negative peak. In Figure 25, for example. Terrain Type "B" exhibits 27*5 excursions . The Waveform Segment (L ) The symbol L represents the length of the waveform segment. In Inches, corresponding to an Individual terraintype occurrence, as measured along the X-axls. These X-axls waveform-segment distances can be converted to represent actual ground-scale distances, as determined from the Image. When necessary, these segment distances could also be converted to lengths of time, since a constant amount of tine Is required to traverse the entire length of scan line. This length Is measured to the nearest 0.05 Inch. The Scale Waveform Segment (L ) In order to reduce L to actual Imageiry scale. It must be multiplied by a constant factor. This constant represents the number of feet of Imagery scale corresponding to one Inch on the enlarged waveform. For the nine-lens nultlband can^ra Imagery, this has been determined to be

PAGE 103

93 94.3 ft./ln. For the panchromatic and thermal infrared images used, it is 694 ft. /in. The Excursion Frectuency (f) The excursion frequency is calculated by dividing the number of excursions (E) occurring in any particular terrain-type occurrence by the length of the corresponding waveform segment. Since there is normally a variation of distance between the individual excursicms* this parameter is in reality an average frequency calculation, and thus represents the average X-axis distance occupied by each excursion. The calculation of excursion frequency provides a means for determining the spacing between gray-tone level fluctuations appearing on the imageiry. The unit for this measure is reciprocal distance, or inches The Scale Excursion Frequency (f ) The excursion frequency f is converted to scale excursion frequency f by multiplying f by some constant appropriate to the scale of the imagery being amalyzed. This scale constant represents the ratio of the number of excursions per mile on the original imagery to the nunft>er of excursions per inch on the ifiraveform as it is presented on the sampling grid. For the nine-lens multiband camera imagery.

PAGE 104

94 this constant has been calculated to be 17. 95. For the panchromatic and thermal Infrared image* it was found to be 3.70. The multiplication of the excursion frequencies by these factors reduces f to a scale parameter, f , which can be compared between multispectral sets. The unit of the nsasure thus beccxnes excursions per mile. The Excursion wavelength ( X ) This parameter is calculated by determining the reciprocal of the excursion frequency (f ) . It represents the actual average distance in inches between individual excursions. The Scale Excursion Wavelength ( X„ ) As was the case with the conversion of L to L « X X s is converted to A by using the scale constant of each s imagery set, 94.3 ft. /in. for the nine-lens imagery and 694 ft. /in. for the panchromatic and thermal infrared set. This measure, then, represents the actual average distance in feet between individual excursions. |||ean Deviation of Gray-Tone Level (x-x) This statistical measure is simply the summation of the absolute (algebraic sign disregarded) departure of each gray-tone level percentage sample from the average gray-tone

PAGE 105

95 level Xt divided by the niimber of siunples n. Fully exIx-xl pressed* it v^ould be • This calculation determines the average dispersion of the santpled data about the mean. In this waveform analysis, it provides an estimation of the relative size of the exctirsions, i.e., the degree of graytone fluctuation within any terrain-type occurrence. Standard Deviation of Gray-Tone Level (s) /, _-j2\ 1/2 This variable, fully denoted by I -*2iz2£2— J ^ ^g a conventional statistical variable which also gives an indication of the dispersion of the data. It expresses, however, percentages of the data %^ich exist within certain ranges from the mean of the data. For example, a standard deviation of 5.00 would indicate, when applied to the gray-tone level data, that, for one standard deviation, approximately 68.3 percent of the data would lie between ± S.OO percentage points of the mean; for two standard deviations, 95.4 percent %#ould lie between ± 10.00 percentage points of the mean; and for three standard deviations, 99.7 percent of the data would be included within ± 15.00 percentage points of the mean. This provides a more precise measure of the dispersion and relative size of the excursions.

PAGE 106

CHAPTER IV EXPERIMENTAL RESULTS This chapter presents the results amd conclusions derived fron the waveform analysis of the nine-lens multiband camera imagery and the panchromatic and thermal infrared images. The calculations for each of the scan lines are reviewed and the significant findings presented. The terrain-type occurrences are then paired* where possible, and waveform-analysis results compared for each individual type of sensor. Wjivef orm Analysis Results t Nine-Lena Multiband Camera Imacfery Primary Data The primary seuqpling data and calculations for the nine-lens multiband camera imagery are summarized in Appendix I. In that section, each specific terrain-type occurrence is listed and denoted by the code number appearing inreediately after the occurrence, e.g., open field-1, open field-2, etc. These notations will also be found concurrently on the waveform graphs (Figs. 26 through 37) . These 96

PAGE 107

97 calculations are sunmarized by Individual scan lines so that •ach data group reprssenting one scan line nay be oonpared directly with its corresponding waveform graph. Also for the sake of clarity, calculations for both scan lines from each of the nine-lens images are included on each data page. Compiled Dat^ The primary data has b«en reorganized in i^pendix III by grouping all occurrences from both lines together l^ terrain type and either totaling or averaging the data where appropriate; e.g., n has been totaled for each terrain type %^ile X has been averaged. The mean deviation and standard deviation, because of their mathematical character, must be weighted before being averaged. This is acconqplished by multiplying the mean deviation and standard deviation for each occurrence by n« the number of gray-tone level samples, and then calculating the average of the accximulated products . The light over-all visual impression of the Lens 2 image (see Fig. 4) is manifested in the average gray-tone level calculations. A close grouping of the lightest graytones, 68.6 percent for paved road, 70.3 percent for open field (no cover), and 70.6 percent for paved parking lot, substantiates this visual grouping of lighter tones. Inter-

PAGE 108

98 mediate gray-tones are characteristic of unpaved road (62.6 percent) and the construction site (6S.4 percent) , while the darkest tones on the image are characteristic of open field (grass) « which returns 51.6 percent. Excursion frequencies, and similarly excursion wavelengths, show little differentiation betvreen terrain types on the Lens 2 image, providing little opportunity for legitimate discrimination using these two measures related to spacing of textural differences. Even though the two foregoing parameters indicate similar excursion spacing, the standard deviations for each terraintype group do correspond to texture differences. The roost diverse of the gray-tone assemblages are associated with the construction site, whose gray-tone level san^led data shows a standard deviation of 7.03. This is much higher than any other Lens 2 standard deviation, and is demonstrated on Figure 26 by the diverse types of excursion patterns and gradients. Paved road and unpaved road are well separated with standard deviations of 3.07 and 2.58, respectively, but open field (grass) , open field (no cover) , and the paved parking lot are grouped together with 3.63, 3.06, and 3.08, respectively, again providing little differentiation. Identification of these terrain types would have to be accomplished in conjunction with a graphical inspection of the

PAGE 109

99 waveform gradients which represent the rapidity with which contrasts between terrain types change. Figures 26 and 27 illustrate this well. The construction-site waveform segment is a gradual change from its neighbors, while open field (grass) and paved parking lot are sharp departures from their adjacent terrain types « as is illustrated by the nearly vertical excursions. As was mentioned in Chapter III« the ZiCns 2 image is excellent for shape measureoMnt and boundary determination, but inferior for general identification purposes. This conclusion has been amplified by the foregoing discussion of calculated gray-tone characteristics. The Lens 4 calculations exhibit greater diversity among the different terrain types (see Figs. 5, 28, and 29). Again, c^en field (grass) is the darkest of the terrain types, with an average gray-tone level of 22.0 percent. By contrast, open field (no cover) has the lightest tone (36.3 percent) . Paved road and paved parking lot have similar gray-tone levels, 50.3 imd 48.3 percent, respectively, indicating the use of similar construction materials, and are well differentiated from unpaved road (39.2 percent). Excursion frequency is also highest for the construction site (3.42/in.). unpaved road and paved road are %#ell discriminated by frequencies of 3.24/in. and 1.90/in.,

PAGE 110

100 but again open field (grass) (2.86/ln.), open field (no cover) (3.01/ln.), and paved parking lot (3.08/ln.) are not differentiated. Standard deviations provide a better measure of graytone assemblage. The high score of 22.65 for the construction site Illustrates the great diversity of gray-tcmes associated with such a complex cultural feature. The rough surface of open field (no cover) also produces a standard deviation (8.60) characteristic of a complex cultural feature. The relatively smooth open field (grass) , unpaved road, paved road« and paved parking lot can be grouped together with characteristically low scores of 5.35, 3.92, 5.56, and 4.87. The low gray-tone level of the background function performed by open field (grass) In this Image plus the relatively higher average gray-tone levels of the other terrain types give Lens 4 the distinction of having the gray-tone diversity necessary for both measurement and Identification. This conclusion Is also borne out by the calculations. The Lens 9 Image exhibits the light (40.7 percent) open field (grass) gray-tone characteristic of Infrared film (see Figs. 6, 30, and 31). With the exception of open field (no cover) (41.6 percent), this large background area

PAGE 111

101 has the lightest return on the image. All other terrain types are darker « with the cultural construction site (16.4 percent) « paved parking lot (21.0 percent)* and paved road (28.6 percent) being characteristically the darkest features, unpaved road (37.2 percent) is expectedly similar to open field (no cover) (41.6 percent). Paved road again exhibits the lowest excursion freguency« 1.88/in.« but the construction site does not exhibit the diversity of gray-tone so characteristic of it in Zionses 2 and 4, and thus produces a low f also* 2.03/ln. Am might have been anticipated, because of the general dark tones and lack of terrain-type contrasts, the standard deviations associated with the Lens 9 image allow little discrimination betvreen terrain types; all values for s lie between 7.73 for the construction site and 3.68 for the open field (no cover) , with no significant cultural or natural groupings. Corrected Data After all calculations had been made using all sampling points on the grids, 2Ui attenpt was made to reduce the discrepancies among the data by correcting the calculations for each waveform segment for the inaccuracies produced on the graphs by rise time or fall time. Those

PAGE 112

102 sampling points repres^itlng transitional gray-tones produced by rise time were eliminated from the san^llng data and the entire range of calculations performed again. This corrected data for the nine-lens multlband camera Imagery is compiled in Appendix V* As can be seen from an inspection of the Lens 2 image corrected data* the elimination of rise-time inaccuracies produces the same basic analytic patterns* but amplifies these patterns by increasing the range of the data for each statistical variable. The difference between highest and lowest, or range of average gray-tone levels has increased from 19.0 to 19.9 percent, excursion frequencies from 0.61/in. to 0.72/in., and standard deviation from 4.43 to 3.02, providing greater differences between calculations for individual terrain types. Substantiating the hypothesis that the elimination of gray-tone samples generated by risetime discrepancies is the observation that the range of calculations for each waveform segment Included with a terraintype group, such as the eleven waveform segments representing open field (grass) , is lower for the corrected data. In effect, this means that the grouping of the data around the various averages is smaller, and is demonstrated by the lower standard deviations in all cases where rise-time corrected waveform segments occur.

PAGE 113

103 The Lens 4 and Lens 9 calculations also exhibit the effects of: (1) separating the data by Increasing the range between terrain types, and (2) decreasing the range of data included within each group of terrain-type occurrences. This serves to enhance the use of the Lens 4 image as the roost useful for interpretation and measurement, but does not improve the Lens 9 image's position as the least useful in these general matters. Sensor Comparison Totals » Nine-Lens Multiband Camera Images In order to coo^are the statistical trends for each terrain type as exhibited by the various film-filter combinations sampled, the totals and averages for each terraintype occurrence group have been conpiled for all lenses in Tables 3 and 4. Table 3 includes the data %«hich has not been corrected for rise time. Since no sampling points have been renoved, an inspection of the various totals for n« L , and L will reveal the ability of the waveform monitor to display consistently a vraveform which can be dimensionally con^ared with another waveform from a c<»i^arable image. The slight discrepancies existing between these dimension variables for each terrain type are composite errors from several sources.

PAGE 114

104 TABLE 3 SENSOR COMPARISON TOTALS BY TERRAIN TYPEt MULTIBAND CAMERA IMAGERY NINE-LENS Terrain Type Lens No. Zx Construction site Open field (grass) Paved road Unpaved road Open field (no cover) Paved parking lot 74

PAGE 115

lOS TABLE 3. — Extension

PAGE 116

106 TABLE 4 SENSOR COMPARISON TOTALS BY TERRAIN TYPEx NINE-LENS MULTIBAND CAMERA IMAGERY (CORRECTED FOR RISE TIME) Terrain Type Lens No. Zx Construction site Open field (grass) Paved road Ukipaved road Open field (no cover) Paved parking lot 74

PAGE 117

107 TABLE 4. — gxtansion Statistical Parameters

PAGE 118

108 They are undoubtedly minor dimensional differences between the images displayed before the television camera. There are also minor differences between the waveforms of each image resulting from the determination of the edge of the imagery on the waveform. Minor discrepauicies also result from small errors in display placement positions between subsequently scemned images « and, finally, there are small errors engendered in the linear nosaicing of the Polaroid prints to produce the composite waveform. All of these errors totaled average 0.9 of one percent of the total dimensions recorded throughout the iraagery-television camerawaveform monitor-oscilloscope camera-roosaic-enlarged waveform progression. The totals c^i^iled in Table 3 reveal some patterns not previously noticed. There is a general decrease in average gray-tone level from Lens 2 through Lens 9, with the exception of the open field (grass) category. In scmie cases, this amounts to a con^lete reversal of gray-tone, as in paved parking lot, and in others it is only a slight shift (unpaved road, for example) . Some terrain-type categories exhibit remar)cably consistent excursion frequencies between images, such as open field (grass), unpaved road, open field (no cover) , and paved parking lot. others

PAGE 119

109 •xhlbit widely divergent values for f , such as the construction site. Categories showing consistent excursion frequencies also show consistent excursion wavelengths, because of the reciprocal mathematical relationship between these two variables. With the exception of the construction site and open field (no cover) categories « there is a general increase in standard deviations frcxn Lens 2 through Lens 9. Most likely this is the result of the wider range of %iravelengths sampled by the Lens 9 film-filter conbinaticm* which* in •ffect« produces greater variation within terrain types than between them. These same trends are exhibited in Table 4, which presents the compiled data after correction for rise-time discrepancies. The significant difference between the tvro sets of data lies in the standard deviations. The removal of a few oaiqpling points has resulted in a substantial decrease in most average standard deviations, again illustrating the value of this method of data refinwnent. Waveform Analysis Results i Panchrgnatic and Thermal Infrared Images Primary Datai Panchromatic Image The primary sampling data and calculations for the panchromatic image are presented in Appendix II. As was the

PAGE 120

no case with the nine-lens multiband camera imagery presented in Appendix 1, the data ia listed by specific terrain-type occurrences* and each occurrence is denoted by the code number appearing immediately after the occurrence. In Appendix II, the data has been grouped for each individual scan line. Where possible* the abbreviations PAN for panchromatic and IR for infrared will be used for the sake of simplicity throughout this discussion. Compiled Data I Panchromatic Image The primary sampling data and calculations have been reorganized in Appendix IV by grouping all occurrences of each specific terrain type together frcxn each of the three scan lines across the image. This data has been totaled and/or averaged in the same manner described in the treatment of the nine-lens multiband camera data. The panchromatic image (see Fig. 8) and its associated waveforms (see Figs. 32, 33, and 34) illustrate the complexity of gray-tone patterns and land uses which influenced the choice of this image as representative of panchrcxnatic aerial photography. Because it is a photographic record of all wavelengths of visible light, it is more difficult to separate out individual terrain types by such stringent gray-tone ranges as was accomplished in the narrow

PAGE 121

Ill bandwidth photography generated by the nine-lens raultiband camera* Using all of the grid saunpling points « the following ranking of gray-tone levels by terrain type was produced: Cleared right-of-way 59.0% open field 57.9% Road 57.2% Shoal yntex 54.0% Swan^) 50.5% Op«n woods (deciduous) 46.0% Tree row 41.9% Dense %«oods (coniferous) 39.3% Dense woods (deciduous) 37.3% Lake 36.2% Open woods (coniferous) 34.8% As might be expected, the three lightest tones « if grouped together « represent much the same type of terrain as seen by an interpreter. Only the linearity of most roads serves to distinguish them from other similar terrain types. On the PAN image (see Fig. 8) itself, it will be noticed that where roads do cross open fields and/or the cleared right-of-way extending vertically through the left center of the photograph* they occasionally become almost in^erceptible except for linearity. It is also interesting to notice the graytone similarity between swamp and shoal water (where bottom is visible) in the large lakes. With the exception of graytones belonging to the deeper parts of the large lakes, the previous ranking shcM^s all of the lowest gray-tone values to

PAGE 122

112 be associated with forested areas. Of significance here are the similarities between open woods (deciduous) and tree rows, those forested areas generally lining roads or acting as property lines. The slight difference between these two values suggests that they might possibly have been group>ed together In the original san^llng categorization. While there appears to be slgnlflcauit differentiation between all terrain types on the basis of gray-tone level, the following ranking of excursion frequencies may add additional Inf ormatl<»i t Open woods (coniferous) 4.12/ln. Cleared right-of-way 4.00/ln. Tree row 3.66/ln. Open field 3.52/ln. Dense woods (deciduous) 3.45/ln. Open woods (deciduous) 3.25/ln. Lake 3.24/ln. Dense woods (coniferous) 3.23/ln. Swanp 2.61/ln. Road 2.35/ln. Shoal water 1.79/ln. This ranking again points out the similarities between open field, tree row, and cleared right-of-way. While It might appear that open woods (coniferous) should also be Included In such a grouping, the recollection of discrete trees with small crowns characteristic of a young pine forest will serve to explain the high excursion frequence noted. The mature deciduous forests exhibit similar frequencies while

PAGE 123

113 the even-crovmed nature of the planted coniferous stands produces a slightly lower frequency. The diverse character of each of the various terrain types is demonstrated even further in the ranking of standard deviations! Cleared right-of-way 12.18% Shoal water 8.21% Open woods (deciduous) 7.13% Open vfoods (coniferous) 6.15% Open field 5.76% Road 5.70% Dense woods (coniferous) ...... 5.12% Dense woods (deciduous) 4.99% Swan^ 4.85% Lake 3.13% The extreme range of data included in the calculation for the cleared right-of-way is produced by many different types of terrain which lie adjacent to and which are crossed by the cleared strip, undoubtedly many of these diverse graytone levels will be removed during correction for rise time. The high standard deviation associated with shoal water results frcMn the gray-tone gradient associated with it. As the water deepens cmd gray-tone decreases* a wide range of values is recorded for each occurrence. Both vegetative types of open woods exhibit similar ranges of gray-tone values « although they are widely separated by average graytone. This is indicative of small clearings associated with such forest patterns. Open fields and roads show similar

PAGE 124

114 gray-tone ranges, and both types of dense woods have similar ranges of gray-tone, just as they have nearly identical average gray-tone levels. It is not extraordinary that the gray-tone ratnge of swai^p approximates that of dense woods (deciduous) , since a considerable portion of the poorly drained lands have Just that vegetative cover. The lakes, finally, exhibit their characteristically even-toned return, amplified by the lowest standard-deviation calculation. Corrected Data» Panchromatic Image The primary data was corrected for rise time using the same technique as discussed during the correction of the nine-lens multiband camera data. Approximately the sane general ordering of average gray-tone values is exhibited when the corrected data is ranked as follows: Cleared right-of-way 67. S% Road 39.6% Open field 38.0% Shoal water 34.4% swamp 30.3% Open woods (deciduous) 43.4% Dense woods (coniferous) 38.3% Tree row 36.3% Lake 36.0% Open woods (coniferous) 34.8% Deaae woods (deciduous) ..33.6% In this ranking, the trio of cleared right-of-'way, road, and open field still dcnninates the lighter gray-tones, although

PAGE 125

115 their relative position Is somewhat reversed. The remaining ranking remains generally the same, although relative positions have also changed somewhat among the forest-type occurrences. The general ordering of the terrain types with respect to corrected excursion frequ^icles also seens to have followed the same pattern as average gray-tone value. The basic ranking remains the same, but relative changes have taken place In conjunction with an Increase In the remge of values. The ranking of corrected excursion frequencies follows! Cleared right-of-way 6.67/ln. Tree row •••. 4.41/ln. Open woods (coniferous) 4.12/ln. Open field 3.64/ln. Dense %«ooda (deciduous) 3.61/ln. Open iiTOOds (deciduous) 3.38/ln. Dense woods (coniferous) 3.38/ln. Lake 3.33/ln. Road 2.86/ln. Swamp 2.80/ln. Shoal water 1.96/ln. Even though several different terrain types now are listed as having the highest frequencies. It must be pointed out that these types have very few samples Included In their calculations. This Is especially critical In calculation of excursion frequencies, since one or two additional samples could produce much la%#er or higher frequencies. It Is seen, nevertheless, that the basic ordering observed In the rank-

PAGE 126

116 Ing of uncorrected data Is basically the same, although minor relative changes have occurred. In the general ordering of the stzmdard deviations of the corrected data, there has been an extreme contraction of the range of this variable, and a considerable reorganization. The ranking of the corrected standard deviations is as follows t Open woods (deciduous) 6.47% Shoal water 6.24% Open woods (coniferous) 6.13% Open field 5.34% Dense woods (deciduous) 4.36% 9wanp 4.33% Dense woods (coniferous) 4.11% Road 3.93% Tree row 2.81% Cleared right-of-way 2.51% Lake 2.26% The roost significemt change in this ordering is, of course, the relative inovement of the cleared right-of-way calculation from highest in uncorrected data to next to lowest in the corrected data. This again illustrates the mathematical instability of a calculation coR^>rising only several samples. This ranking, as does the ranking of the corrected data, illustrates the wide range of gray-tone values contained in waveform segments of both vegetative types of open woods, the intermediate ranges of both types of dense woods, and the very low range of gray-tone values associated with segments of lake traverses.

PAGE 127

117 Primary Datat Thermal Infrared Image The primary sampling data and calculations for the thenaal infrared image are also included in Appendix II. The same listing method and coding of waveform segments applies to this image as applied to the nine-lens multiband camera imagery and the panchromatic image. Coiqoiled Data; Thermal Infrared Image As with the previous imagery « the data and calculations reorganized by terrain type are included in Appendix IV. The same totaling and averaging methods have been used in the analysis of this image. The infrared image (see Pig. 9) and its waveforms (see Figs. 35« 36, and 37) exhibit the great diversity of gray-tone patterns illustrated by the comparable PAN image. Those graytones presented here« however* do not in^ly differences in reflected light as do the PAN gray-tones « but instead represent relative heat differences, with white and black representing relatively %/armer and colder areas, respectively. It is interesting to note that three patterns discernible on the PAN image, shoal water, tree row, and cleared right -of -^way, are not discernible on the IR image. The shoal water is not seen on the IR image because of the more complete mixing of warm waters within the lakes set up

PAGE 128

118 during atimroer convectional movements. The tree rows are not seen because they are undoubtedly smaller than the instantaneous £leld-o£-vlew (resolution) of this unclassified sensor, and the cleared right-of-way, while vaguely visible on the image itself, was not noticeable on any of the waveforms because of the boundary positions and contrasts in those areas where scan lines traversed it. By contrast, however, the river is one of the prominent features of the infrared image, while it goes unnoticed on the PAN waveforms because of its similar gray-tone level and position within forested areas in the vicinity of its scan-line traverses. Another terrain-type category is also recognized on the ZR image. On Line 1, a small stand o£ deciduous vegetation is found Just to the left of Little Portage Lake. Since it lies on poorly drained soils, and is therefore different from the other deciduous occurrences, it has been called "s%^aii^ hardwoods," and its totals separated from both deciduous and swoop occurrences in the Appendices. Consequently, after omitting three terrain types and adding another, ten types are recognized on the IR image. Re-emphasizing the nighttime origin of this infrared image, and keeping in mind the thermal significance of

PAGE 129

119 the gray-tone level « the following ranking of average graytone levels as calculated from the primary sampling data provides a significant distributions River 68.7% Lake 67.4% Dense woods (coniferous) 61.6% Road 57.8% Dense %«oodB (deciduous) 52.1% 8w«np hardwood 51.7% Open woods (deciduous) 51.5% Open field •••••...•••.. 37.3% Open woods (coniferous) 35.2% Swamp 31.8% This distribution illustrates emphatically the meteorological phenixienon usually referred to as the "differential heating and cooling of land and water." It is seen that the vrater surfaces returned much higher temperatures at this late evening time than did the land surfaces. Interesting to note are the slightly higher temperatures in the Huron River. This is to be expected because of the lack of convectional circulation c(»iaon to the lakes, allowing the river water to heat to a higher temperature during the long summer days. One terrain type, dense woods (coniferous) , appears to retain its higher temperature even thoiigh surrounding terrain types have cooled off considereibly. A good example of this is seen on Line 2 of the IR image, occur rence-3. The high temperatures at this and similar locations occur

PAGE 130

120 because the air apace beneath the dense crowns of such vegetation acts as a heat sink during the day and as an Insulator at night, creating the very even distribution of high temperatures recorded by the infrared scanner. The closely grouped average gray-tone levels of dense woods (deciduous) , swamp hardwoods , and open woods (deciduous) Illustrate the similar physical character of each of these subtypes. They act as heat sinks also, but to a lesser degree because of the vertical stratification common to such stands, whic^ permits faster loss of heat at night. Also physically grouped together are the coolest terrain types, open field, open woods (coniferous) , and swamp. The similar heating and cooling characteristics of these types result from their lack of insulating vertical stratification, plus easy access by advection currents* which import regional temperatures. Excursion frequency, when calculated for a thermal Infrared image, represents the local distribution of tenperature fluctuations within a specific waveform segment. By this means, it gives an indication of the homogeneity of temperatures within a specific terrain type. It does not, however. Indicate the magnitude of the fluctuation. The

PAGE 131

121 following array of ranked excursion frequencies illustrates this significance! iS^ ' ' \ 3.76/in. Open woods (coniferous) 3.64/in, Dense woods (deciduous) 3.45/in! ^^ 3.27/in! Open field 3.20/in. ^^^«' 3.16/in. Opmn ifoods (deciduous) 3.03/in. Oonae woods (coniferous) 2!86/in! 8wup hardwoods 2.83/in! "^ 2.69/in. This close grouping of excursion frequencies points out the frequent convectional teniperature differences in the lakes, open woods (coniferous) , and dense woods (deciduous) terrain types. It also illustrates the homogeneous teaperatures characteristic of dense woods (coniferous), swanp hardwoods, and swamp, the latter two low frequencies being produced by the ameliorating effects of considerable soil moisture. The magnitude of these temperature fluctuations is illustrated by the following order of standard deviations of gray-tone level f luctuations i Swamp hardwoods 14.60% Dense woods (deciduous) ' lo!98% Qp«n woods (deciduous) [ 8!543C 8vmap 7.57X Open field 724X ^^ • • .'.'." 6.6356 Dense woods (coniferous) 5.99% Open woods (coniferous) 5 92% f^y®'^ s'.iax ***• 3.50%

PAGE 132

122 This ranking lllxistrates the larger temperature fluctuations associated with the three deciduous subtypes. In the case of swaiTip hardwoods* It becomes appar^it that while a low excursion frequency exists, when fluctuations do occur they are very significant, pointing out the wet or dry characteristic exp€»cted of such a forest stand. Both coniferous terrain types show low magnitude of tenf>erature fluctuation, emphasizing the even crown and homogeneous heights of planted coniferous forest plots. As would be theorized, both water types have the least range of ten^erature. Corrected Data: yhermal Infrared Image The ccHnplled data was corrected for rise time using the same techniques concerning such corrections as previously dlsctissed. An Inspection of the following ranking of corrected average gray-tone levels Illustrates characteristics of such corrected data as were previously noted: River 68.79; Lake 68.1% Dense woods (coniferous) 63.09^ Road 60.3% Dense vroods (deciduous) 52.1% Swanq? hardwoods 51.7% Open woods (deciduous) 51.3% Open field 37.2% Open woods (coniferous) 33.9% Swai!^ 30.2% A slight decrease In the range of the average gray-tone values Is noted, but the order remains the same. This

PAGE 133

123 latter characteristic indicates the stability of the relative therxnal levels associated with each individual terrain type. The ranking of excursion frequencies « as follows* however* shows minor changes t Open woods (coniferous) 4.00/in. Road 3.90/in. Lake 3.85/in. Dense woods (deciduous) 3.53/in. Open field 3.30/in. River 3.16/in. Dense woods (coniferous) 3.12/in. Open woods (deciduous) ... 3.03/in. twaanp 3.01/in. Swan^ hardwoods 2.83/in. The minor changes here involve a slight increase in the range of the data and a reordering of open woods (coniferous) and road terrain types. As was mentioned with reference to the PAN image* however* these two types involve few saaqpling points, and thus significant frequency changes are expected in conjunction with the elimination of only one or two samples. The remaining part of the array* however* is constant. The elimination of a few sampling points from those terrain types having few primary ssunples also produces a reorganization of the array of standard deviations of graytone levels s Swamp hardwoods 14.60% Dense i^oods (deciduous) 10.73%

PAGE 134

124 Open woods (deciduous) 8»54% Open field 6.03% River 5.78% swamp • 5.40% Open woods (coniferous) 4.08% Dense woods (coniferous) 3.38% Road 3.21% Lake 1.81% The road and river terrain types have exchanged approximate locations « as have open emd dense woods (coniferous). The roost notable change in the stamdard deviations, however « is the near-nullifying effect of lake-tenperature fluctuation with the removal of several boundary rise-time-induced values . sensor Comparison Totals t Panchrcanatic and Theinnal Infrared Images In order to compare the utility of each image in discerning certain gray-tone patterns* the compiled data and corrected data have been groi:^ed together for each type of sensor. Only the terrain types which appear simultaneously on both images are presented. Table 5 presents the compiled data v^ich has not been corrected for rise-time inaccuracies. The comparative interpretation of the average gray-tone levels depends on an overwhelming number of factors « such as time of day* season* cliroatological factors, etc. A large backlog of data would be needed to isolate any particular combinaticxi of variables

PAGE 135

125 TABLE 5 SENSOR COMPARISON TOTALS BY TERRAIN TYPEi PANCHROMATIC AND THERMAL INFRARED IMAGES (PAIRED OCCURRENCES)

PAGE 136

126 TABLE 5. — Extension

PAGE 137

127 dealing with gray-tones. As an exaziqc>le« v^lle there is a complete gray-tone reversal for the lakes between the PAN and ZR InageSf their gray-tones would be approximately similar if the infrared scanner had been used during the daytime. For this reason* it appears that the usefulness of gray-tone level comparisons is more restricted to simultaneously procured imagery. Of great importance, however, are the similarities noted between the scale excursion frequencies. In all cases except for roads, the only cultural terrain-type category in this multispectral set, there appears to be some suggestion of natural terrain frequencies being demonstrated by the similar excursion frequencies. with respect to the standard deviation of gray-tone levels, however, there is a general increase in the range of the gray-tone seniles from PAN to ZR. This suggests an increase in the amount of information available from the uncorrected gray-tone data on the infrared image. Table 6 presents the corrected coR^>arisons of the PAN and ZR images. The gray -tone data still illustrates its dependence on a variety of factors, but the differences between the scale excursion frequencies have decreased with the elimination of rise-tine inaccuracies, increasing their

PAGE 138

128 TABLE 6 SENSOR COMPARISON TOTALS BY TERRAIN TYPE: PANCHROMATIC AND THERMAL INFRARED IMAGES (PAIRED OCCURRENCES f CORRECTED FOR RISE TIME) Terrain Type Sensor n

PAGE 139

129 TABLE 6. — Extension

PAGE 140

130 usefulness. A reduction in the ranges of data as presented by the standard deviations of the IR gray-tone samples, with the exception of dense woods (deciduous) , has produced similar values for that variable for both sensors. Analysis of Scale and Resolution Differences The nine-lens multiband canara image was originally recorded at a scale of 1x10,000. An inspection of one of these images reveals that resolutions on the order of one foot are recorded. For example, notice that single tire tracks are reproduced on the Lens 4 image (see Pig. 5) . This appears to be quite smaller than the resolving power of the television camera-waveform monitor system. The smallest measurement made was nine feet. Of course, as the size of the pattern being traversed increased, the percentage of error introduced through waveform measurement became smaller. As an example, measurement of the width of various roads crossed varied widely up to 100 percent, but the measurement of the paved parking lot varied only between 613 and 623 feet. It must be pointed out, however, that roads were traversed at all angles, and not always at right angles across minimum widths. Originally photographed at contact scale of 1:20,000,

PAGE 141

131 the photomosaic scaled to lt24,400 reveals resolutions on the order of five or six feet. This can be verified by the identification of single isolated crowns of young conifers on the photograph (see Pig. 8) . This resolution could not be matched by the video system, however, with approximately 20 feet being the minimum distance recorded. Because of rise-time inaccuracies and the IEEE frequency response selected, however, approximately 100 feet is the minimum distance reproducible on the waveform graph. Although rescaled to the same size as the PAN image, the original IR image had a scale of 1«46,000. Because of security restrictions on the resolving capability of this instrument, only ground-scale resolutions on the order of 40 feet vrere possible. This dimension is approximately the width of a paved road, and, as can be seen from the IR image (see Pig. 9) , these roads can be identified only when they are oriented to the general direction of the scanning function. i.e., from top to bottom, in other cases, roads are distinguished by their function of field separations, etc. in no case did the minimum distance reproducible on the waveform monitor decrease to the resolution capability of the imagery. This may eventually necessitate the use of the "flat" frequency response previously described.

PAGE 142

CHAPTER V CONCLUSION Suninary This study is the most recent in a series o£ continuing investigations into the feasibility and methodology associated with the objective measurement of geographic patterns and distributions using electronic instrumentation. The research objectives described at the onset of this report included t (1) the application and refinement of an electronic scanning technique; (2) the development of a wavefoinn discrimination methodology; (3) the evaluation of this methodology; (4) an investigation of the utility of such techniques; and (3) the continuation of a long-range research program. A summary of the results of each of these investigations is included in the following pages. The Scan-Line Technique After the imagery to be included in the study was chosen* which included multlspectral sets of nine-lens multiband caunera imagery and panchromatic and thermal infrared 132

PAGE 143

133 images, the problems of display* orientation, and consistent scan-line sampling were defined emd solved. A consistent display methodology was developed by constructing templates capeible of positioning the image at the same location and with the same orientation before the television camera. A variety of different electronic components were tested, with the advemtages and disadvantages of each being presented. The best method of recording the scan-line waveform was also investigated, with various cameras, films, and exposures being tested. An experimental arrangement utilizing a conventional oscilloscope-type waveform monitor and conpatible oscilloscope camera was finally chosen as the best combination. The particular waveform monitor was chosen because of its high-frequency response and brief rise-time characteristics, and the camera chosen because of its use of extremely fast film, permitting the recording of the smallest trace fluctuation. Waveform Discrimination Methodology A technique for identifying the particular waveform segments associated with specific occurrences of various types of geographic phen<»aena was developed. The results of this technique were transferred to a san^ling framework and the average gray-tone level associated with each terrain-type

PAGE 144

134 occiirrence was calculated, other calculations Included the frequencies of waveform excursions and the mean and standard deviations of the sanf>led gray-tone data. These calculations were reviewed for each individual scan-line recorded* and the significant results presented. It was pointed out that* with respect to the three nine-lens images san9>led« the Lens 4 image provided the best results for identification and measurement of patterns recorded. Average gray-tone levels were indicative of different terrain types on each of the images, with excursion frequencies amd gray-tone standard deviations providing backup information for terrain-type discrimination purposes. The advantages of correcting the priroairy and compiled data for discrepancies introduced by rise-time differences were also reviewed and a comparison of the results and trends between different film-filter combinations was discussed. Pattern discrimination on each of the separate panchromatic emd theirmal infrared images was also accomplished by calculating average gray-tone level. The advantages of ccnnbining this data with the results of excursion-frequency and standard-deviation calculaticxis was also discussed. The problems concerning comparative waveform analysis of multisensor imagery procured at different times and under different conditions were also presented.

PAGE 145

135 Methodology Evaluation The accuracy of each individual technique and measurement was discussed in the respective section in which it was described and presented. In all techniques, however, an effort was made to retain diaiensional and gray-tone stability and record these characteristics in an unbiased manner so that the most objective results could be achieved. This author feels that these characteristics were maintained to the degree possible with the instrumentation and techniques which were used. The basic agreement of the graytone calculations with visual inspection and the coaparable excursion frequencies encountered between sensors indicate that this objectivity has been achieved. The significamt gap b«twe«n methodology and usable suialytic tool became apparent, however, %i^en it was mentioned that a Isurge mass of data must be accumulated before non-simultaneous multispectral scmsor imagery can be validly oonpared. The following paragraph contains sobm of the ideas relevant to this problem* Continuation of the Research The analysis of the data and calculations presented in the various chapters of this report indicate the need for continuing research along several significant lines. First,

PAGE 146

136 an effort must be made to procure significant amounts of representative simultaneous multispectral imagery which can be subjected to the same sort of waveform analysis. Second « different Instruments capable of higher resolving and recording of phenomena displayed on such imagery must be tested and utilized. Third « additional techniques must be developed for recording spatial characteristics of distributions. These might include development of different sampling patterns, rotation of the imagery, or utilization of mathematical techniques such as line integrals. Additional research must also include the use of strip-chart recorders to alleviate the degradation of the waveforms by enlargement and recopying. Potential Utilization of Waveform Analysis in Geographic Research The most significant and ionediate utilization of the techniques and methodology presented above might possibly be in the identification and inventory of outstanding geographic distributions such as forested areas, water bodies, urban areas, etc. After sufficient data has been accumulated, seasonal changes and annual trends of such distributions and their physical characteristics will be available for analysis.

PAGE 147

137 Such techniques as have been developed will be used In the near future in an analysis of the instrumentation to be used in the resource-observing and measuring satellites. This analysis will include an exaaination of the resolution characteristics and gray-tone asseoiblages produced by the sensors to be used. This analysis will aid in the evolution of resolution and gray-tone pattern keys which can be used in conjunction with the video-type information which will be returned from these satellites. Any number of potential utilisations of geographic pattern analysis can be imagined using such multispectral imagery interpretation techniques. This author hopes that such techniques as have been discussed and evaluated will prove helpful in extensive sampling and comparison of different earth cunvironments.

PAGE 148

APPENDICES

PAGE 149

APPENDIX I PRIMARY DATAt NINE-LENS MULTIBAND CAMERA IMAGERY

PAGE 150

140 TABLE 7 PRIMARY DATA: FRAME 42 « LENS 2, LINES 1 AND 2 Terrain Type Waveform Segment Line 1 Construction site Open field (grass) Paved road Unpaved road Line 2 Open field (no cover) Paved parking lot Open field (grass) Unpaved road Paved road -1 -1 -2 -3 -4 -5 -1 -2 -1 -2 -3 -1 -1 -1 -1 -2 74 4,842 65.4 9

PAGE 151

141 TABLE 7. — Extension Statistical Parameters '•X 'S ^ ^8 ^ ^ » 1''-='' 22.0

PAGE 152

142 TABLE 8 PRIMARY DATA: FRAME 42, LENS 4, LINES 1 AND 2 Terrain Type Line 1 Construction site Open field (grass) Unpaved road Paved road Line 2 Open field (no cover) Paved parking lot unpaved road Paved road Open field Waveform Segment -1 -1 -2 -3 -4 -5 -1 -2 -3 -1 -2 73 3,258 44,6 10

PAGE 153

143 TABLE 8. — Extension Statistical Parameters ^c ^ f fg X Xa ^^"^' 25.0 7.30 688 3.42 61.39 0.29 27 18.93 22.65 1.5

PAGE 154

144 TABLE 9 PRIMARY DATA: FRAME 42 « LENS 9, LINES 1 AND 2 Terrain Type f«Eiveform Segment Zx Line 1 Construction site Open field (grass) Ohpaved road Paved road Line 2 -1 -1 -2 -3 -4 -5 -1 -2 -3 -1 -2 75

PAGE 155

TABLE 9. — Extension 145 Statistical Parameters > X x-x 15.0 7.40 698 2.03 36.44 0.49 46 6.26 7.73 1.0

PAGE 156

APPENDIX ZI PRIMARY DATA: PANCHROMATIC AND THERMAL INFRARED IMAGES

PAGE 157

147 TABLE 10 PRIMARY DATA* PANCHROMATIC IMAGE, LINE 1 Terrain Type Waveform segment Dense woods (deciduous) Open woods (deciduous) Open field Swan^ Shoal %^ter Lake -1 -2 -1 -2 -3 -1 -2 -3 -4 -5 -1 -2 -1 -2 -3 -4 -5 -1 -2 -3 -4 8

PAGE 158

148 TABZiE 10 » "— Bxteislon

PAGE 159

149 TABLE 11 PRIMARY DATA: PANCHROMATIC XJMGE, LINE 2 Terrain Type Dense woods (deciduous) Dense woods (coniferous) Open woods (deciduous) SWBxnp Tree row Open field Road Cleared right-of-way Waveform Segment -1 1

PAGE 160

150 TABLE 11. — Extension

PAGE 161

151 TABLE 12 PRIMARY DATAt PANCHROMATIC IMAGE* LINE 3 Terrain Type Dense woods (deciduous) Dense woods (coniferous) Open woods (deciduous) Open vroods (coniferous) Swamp Tree row Open field Wavaform Segment ^x -1

PAGE 162

152 TABLE 12. — Bxtenaion

PAGE 163

153 TABLE 13 PRIMOtY DATA: THERMAL INFRARED IMAGE « LINE 1 Terrain Type

PAGE 164

154 TABZ
PAGE 165

155 TABLE 14 PRIMARY DATA: THERMAL INFRARED IMAGE, LINE 2

PAGE 166

156 TABLE 14, — Extension

PAGE 167

157 TABLE 13 PRIMARY CATAt THERMAL INFRARED IMAGE « LINE 3 Terrain Type

PAGE 168

158

PAGE 169

APPEiraiX III TERPAIN-TYPE COMPILATIONS « NINE-LEHS MULTIfiAND CAMERA IMAGERY

PAGE 170

160 TABLE 16 TERRAIN-TYPE COMPILATIONS: FRAME 42 « LENS 2, LINES 1 AND 2 Terrain Type Waveform Segment Construction site Open field (grass) Totals Paved road Totals UDpaved road Totals Open field (no cover) Paved parking lot Ll-1 74 4«842 65.4 Ll-1

PAGE 171

161 TABLE 16. — Extenaion Statistical Parameters 22.0 7.35 693 3.00 53.85 0.33 31 5.71 7.03 3.0

PAGE 172

162 TABLE 17 TERRAIN-TYPE COMPILATIONS t PRAME 42, LENS 4, LINES 1 AND 2 Terrain Type Wiavefom Segnent rx Construction site Open field (grass) Ll-1 73 3,258 44.6 Totals ttapaved road Totals Paved road Totals Open field (no cover) Paved parking lot Ll-1

PAGE 173

X63 TABLB 17. — Extension Statistical Parancters EL L f f X X \x-x\ a X s s a ' 25.0 7.30 688 3.42 61.39 0.29 27 18.93 22.65 1.5

PAGE 174

164 TABLE 18 TERRAIN-TYPE COMPILATZONS s PRAMS 42, LESS 9, LZHBS 1 AND 2 Terrain Type Waveform Segnent rx Construction site Open field (grass) Ll-1 75 1,234 16.4 Totals Uhpaved road Totals Paved road Totals Open field (no cover) Paved parking lot Ll-1

PAGE 175

165 TABLE 18. — Extension Statistical Parameters •x 's ' ^s >^ ^a I''-'" 15.0

PAGE 176

APPENDIX IV TERRAIN-TYPE COMPILATIONS: PANCHROMATIC AND THERMAL INFRARED IMAGES

PAGE 177

167 TABLE 19 TERRAIN-TYPE COMPILATIONS t PANCHROMATIC IMAGE, ALL LINES Terrain Type Open field Totals Shoal water Totals Lake wiavefom

PAGE 178

168 TABLE 19. — Extension Statistical Parameters EL L f f A X |x-x| s X s s s 2.0 0.30 208 6.67 24.70 0.15 104 1.38 1.64 1.0 0.60 416 1.67 6.18 0.60 416 2.78 3.20 11.5 3.15 2,186 3.65 13.50 0.27 187 7.13 9.17 1.0 0.40 278 2.50 9.25 0.40 278 5.76 6.76 3.5 1.40 972 2.50 9.25 0.40 278 5.68 7.08 14.0 3.20 2,221 4.38 16.20 0.23 160 2.43 2.92 3.0 0.90 625 3.33 12.30 0.30 208 1.14 1.32 5.5 1.85 1,284 2.97 11.00 0.34 236 3.24 5.29 1.0 0.25 174 4.00 14.80 0.25 174 5.50 5.50 1.0 0.35 243 2.86 10.58 0.35 243 4.50 5.17 3.0 0.85 590 3.53 13.07 0.28 194 10.67 13.28 5.5 1.75 1,214 3.14 11.60 0.32 222 3.80 4.23 3.0 1.25 868 2.40 8.83 0.42 291 7.05 8.62 5.0 0.95 659 5.26 19.45 0.19 132 1.48 1.81 11.0 2.55 1,770 4.31 15.95 0.23 160 2.56 3.43 3.0 0.75 520 4.00 14.80 0.23 174 1.92 2.77 5.0 1.25 868 4.00 14.80 0.25 174 1.44 2.06 1.5 0.55 382 2.73 10.10 0.37 257 5.17 6.13 2.0 0.50 347 4.00 14.80 0.25 174 4.48 5.99 4.5 1.70 1,180 2.65 9.81 0.38 264 6.32 7.46 1.0 0.40 273 2.50 9.25 0.40 278 6.00 7.13 8.0 2.35 1,631 3.40 12.58 Q.29 201 7.96 9.62 96.0 27.25 18,914 3.52 13.02 0.28 194 4.61 5.76 1.0 0.55 382 1.82 6.74 0.55 382 5.11 6.09 1.0 0.35 243 2.86 12.95 0.35 243 2.50 2.50 0.5 0.35 243 1.43 5.29 0.70 486 3.38 4.15 1.5 0.75 520 2.00 7.40 0.50 347 6.00 7.80 1.0 0.80 555 1.25 4.63 0.80 555 10.15 13.03 5.0 2.80 1,943 1.79 6.62 0.56 339 6.52 8.21 8.5 2.05 1,423 4.15 15.35 0.24 167 2.46 2.74 6.0 1.80 1,249 3.33 12.30 0.30 208 2.24 3.23 13.5 4.50 3,123 3.00 11.10 0.33 229 1.82 3.33 1.5 0.75 520 2.00 7.40 0.50 347 2.41 2.76 29.5 9.10 6,315 3.24 11.99 0.31 215 2.10 3.13

PAGE 179

169 TABLE 19. —Continued Terrain Type Waveform Segment Dense woods (coniferous) PAN 2-1 -2 -3 PAN 3-1 Totals Open vroods (coniferous) pan 3-1 -2 Totals Tree row

PAGE 180

170 TABLE 19.— Extension (Continued )

PAGE 181

171 TABLE 19. — Continued Terrain Type Wiivefozn Segment Open woods (deciduous) Totals Swaiqp Totals PAN

PAGE 182

172 TABLE 19. — Extension (Continued ) Statistical Parameters E

PAGE 183

173 TABLE 20 TERRAIN-TYPE COMPILATIONS s THERMAL INFRARED IMAGE, ALL LINES Terrain Type

PAGE 184

174 TABLE 20. — Extension

PAGE 185

175 TABLE 20. — Continued Terrain Type Waveform Segment Open field Totals Dense woods (deciduous) IR

PAGE 186

176 TABLE 20. — Extension (Continued ) Statistical Parameters 2.5 0.70 486 3.57 13.20 0.28 194 7.47 8.68 6.0 1.65 1,145 3.64 13.45 0.27 187 7.78 9.38 1.5 0.85 590 1.76 6.51 0.57 396 1.70 2.11 8.0 1.95 1,353 4.10 15.15 0.24 167 7.07 8.07 1.5 0.80 555 1.88 6.95 0.53 368 8.69 10.39 3.5 1.00 694 3.50 12.95 0.28 194 2.60 2.90 3.0 0.95 625 3.15 11.65 0.32 222 1.94 2.62 1.0 0.45 312 2.22 8.22 0,45 312 4.48 5.04 6.0 1.80 1,249 3.33 12.30 0.30 208 5.21 7.62 1.0 0.20 139 5.00 18.50 0.20 139 2.00 2.00 3.0 0.80 555 3.75 13.87 0.27 187 4.97 6.87 3.0 0.80 555 3.75 13.87 0.27 187 8.69 9.43 3.5 1.45 1,006 2.41 8.93 0.41 285 9.74 11.45 I.O 0.35 243 2.86 10.58 0.35 243 1.11 1.25 1.0 0.35 243 2.86 10.58 0.35 243 6.75 8.20 5.5 1.45 1,006 3.79 13.65 0.26 181 1.89 2.73 2.5 1.30 902 1.92 7.11 0.52 361 8.77 10.09 5.5 1.55 1,076 3.55 13.15 0.28 194 7.56 9.84 5.0 1.55 1,076 3.23 11.92 0.31 215 5.00 7.00 3.5 1.40 971 2.50 9.25 0.40 278 4.76 7.78 7.0 2.20 1,527 3.18 11.77 0.31 215 9.12 10.65 8.5 2.40 1,666 3.54 13.10 0.28 194 2.74 3.69 83.0 25.95 17,974 3.20 11.84 0.31 215 5.81 7.24 3.0 0.80 555 3.75 13.88 0.27 187 2.62 3.59 1.5 0.70 486 2.14 7.92 0.47 326 4.45 5.46 1.0 0.50 347 2.00 7.40 0.50 347 2.17 2.36 3.0 0.90 625 3.33 12.30 0.30 208 4.30 5.97 5.5 1.90 1,319 2.90 10.70 0.34 236 7.67 8.99 4.5 1.35 937 3.33 12.30 0.30 208 8.86 11.65 21.0 4.55 3,158 4.62 17.10 0.22 153 12.11 14.09 6.5 2.45 1,700 2.65 9.80 0.38 264 11.42 13.77 5.0 1.65 1,145 3.03 11.20 0.33 229 10.57 12.20 51.0 14.80 10,272 3.45 12.76 0.29 201 9.19 10.98

PAGE 187

177 TABLE 20. — Continued

PAGE 188

178 TABLE 20. — Extension (Continued ) Statistical Pareuneters \ \ ^ ^B >-^ S "'-''I 1.0

PAGE 189

APPENDIX V TERBAIN-TYPE COMPILATIONS CORRECTED FOR RISE TIMEt NINE-LENS MUIAIBAND CAMERA IMAGERY

PAGE 190

180 TABLB 21 TERRAIN-TYPE COMPZIATIONS CORRECTED FOR RISE TIMEt FRAME 42, LENS 2, LINES 1 AND 2 Terrain Type Haveform Segment Zy: Construction site Open field (grass) Ll-1 74 4,842 65.4 Totals Paved road Totals unpaved road Totals Open field (no cover) Paved parking lot Ll-1

PAGE 191

181 TABLE 21, — Extension Statistical Parameters E L L f f ^ A „ x-3c| 8 X S 8 S 22.0 7.35 693 3.00 53.85 0.33 31 5.71 7.03 3.0

PAGE 192

182 TABLE 22 TERRAIN-TYPE COMPILATIONS CORRECTED FOR RISE TIMEt FRAME 42, LENS 4, LIMES 1 AMD 2 Terrain Type Waveform Segment Construction site Open field (grass) Totals Uhpaved road Totals Paved road Totals Open field (no cover) Paved parking lot U-1 73 3,258 44.6 Ll-1

PAGE 193

183 TABLE 22, — Extension Statistical Parauneters EL L f f X X |x-xl s X s s 25.0 7.30 688 3.42 61.39 0.29 27 18.93 22.65 1.5

PAGE 194

184 TABI£ 23 TERBAIN-TYPE COMPXIATIONS CORRECTED FOR RISE TZMEt FRMffi 42« LENS 9« LINES 1 AND 2 Terrain Type iteveform Segnent n Zx Construction site Open field (grass) Ll-1 75 1,234 16.4 Totals unpaved road Totals Paved road Totals Open field (no cover) Paved parking lot Ll-1

PAGE 195

185 TABIiB 23, ' -Bxtension Statistical Parameters E

PAGE 196

APPENDIX VI TERRAIN-TYPE COMPILATIONS CORRECTED FOR RISE TIME: PANCHROMATIC AND THERMAL INFRARED IMAGES

PAGE 197

187 TABLE 24 TERRAIN-TVPE COMPILATIONS CORRBCTED FOR RISE TIME: PANCHROMATIC IMAGE, ALL LINES Terrain Tyi>e HHvefonn Segment Tx Open field Totals Shoal water PAN

PAGE 198

188 TABLE 24. — Extension

PAGE 199

189 TABLE 24.— Continued

PAGE 200

190 TABLE 24. — Extenaion (Continued )

PAGE 201

191 TABLE 24. — Continued

PAGE 202

192 TABLE 24. — Extension (Continued)

PAGE 203

193 TABI£ 25 TERRAIN-TYPE COMPZIATIONS CORRECTED FOR RISE TIME: THERMAL INFRARED IMAGE « ALL LINES Terrain Typ« Waveform Segment Lake Totals River Totals Open woods (deciduous) Totals Omnip Totals IR 1-1 -2 -3 IR 3-1 IR

PAGE 204

TABLE 25. — Extension 194 Statistical Parameters

PAGE 205

195 TABLE 25. — Continued Terrain Type Wavefona Segment Open field Totals Danae woods (deciduous) IR 1-1 -2 -3 -4 -5 ZR 2-1 -2 -3 -4 -5 -7 -8 -9 -10 IR 3-1 -2 -3 -4 -5 -6 -7 IR 1-1 -2 -3 -4 IR 2-1 -2 IR 3-1 -2 -3 Zk 7

PAGE 206

196 TABLE 25. — Extension (Continued) Statistical Parameters '•x '•« ^ ^. A X, IK-Kl 2.5

PAGE 207

197 TABLE 25.— Continued

PAGE 208

198 TABLE 25, — Bxtenaion (Continued )

PAGE 209

SELECTED BIBLIOGRAPHY Books American Society of Photogranmetry. Manual of Photographic Interpretation . Menasha* Wis.t George Banta Co., Inc., 1960. Conn, G. K. T. , and D. Q. Avery. Infrared Methods . New Yorlct Academic Press, 1960. Eastman Kodak Ccopany. Infrared and Ultraviolet Photography . Advanced Data Book M-3. Rochester, N. Y. j Eastman Kodak Company, 1963. . Kodak wratten Filters for Scientific and Technical Use . Scientific and Technical Data Book B-3, 22nd ed. Rochester, N. Y. t Eastman Kodak Company, 1965. Hackforth, H. L. Infrared Radiation . New Yorkt McGrawHill Book Co., Inc., 1960. Holter, M. R., et al . Fundamentals of Infrared Technology . New Yorkt The Macmillan Co., 1962. Jamieson, John A., Raymond H. McFee, Gilbert N. Plass, Robert H. Grube, and Robert G. Richards. Infrared Physics and Engineering . New Yorkt McGraw-Hill Book Co., Inc., 1963. Kruse, P. w. , L. D. McGlauchlin, and R. B. McQuistan. Elements of Infrared Technology . New Yorkt John Wiley and Sons, 1962. Lattman, L. H. , and R. G. Ray. Aerial Photographs in Field Geology . New Yorkt Holt, Rinehart, and Winston, 1965. 199

PAGE 210

200 Simon, Ivan. Infrared Radiation . Princeton: D. Van Nostrand Co., Inc., 1966. Strahler, A. N. The Earth Sciences . New York: Harper and Row, 1963. Reports Alexeenko, E. J. Dirt Roads as an Indicator in the Aerial Photo Interpretation of Soils and Terrain . Translated from the Russian. Washington: U.S. Array Engineers Research and Development Laboratories, 1966. Altschaeffl, A. G. Effect of Soil Moisture and Other Natural Variables on Aerial Photo Gray Tones , unpublished Master's thesis, Purdue university, 1955. Avery, Gene. Forester's Guide to Aerial Photp Interpreta tion . Occasional Paper No. 156, Southeastern Forest Experiment Station, U.S. Department of Agriculture, 1960. . Identifying Southern Forest Types on Aerial Photographs . Station Paper No. 112, Southeastern Forest Experiment Station, U.S. Department of Agriculture, I960. Baer, Ferdinand, and William Kaxma. Numerical Analysis of Tiros Radiation Observations . Technical Paper No. 67, Department of Atmospheric Science, Colorado State university, 1965. Bandeen, W. R. , H. I&lev, and I. Strange. Radiation Clima tology in the Visible and Infrared from Tiros Meteorological Satellites . Technical Note D-2534, National Aeronautics and Space Administration, 1965. Bell, E. E., et al . Infrared Technigues and Measurements . Final Engineering Report, Contract AF 33 (96160) -3312, Ohio State university Research Foundation, 1957. Brown, P. E. , and A. M. O'Neal. The Color of Soils in

PAGE 211

201 Relation to Organic Matter Content . Research Bulletin No. 75, Iowa Agricultural Experiment Station, 1923. Cook, John C. Research on the Electr
PAGE 212

202 Greaves, J. R., R. Wexler, and c. J. Bowley. The Feaaibillty of Sea Surface Tempera tiare Determination Using Satel lite Infrared Data * Final Report, Contract NASW1157. Aracon Geophysics Co., 1965. Holter, Marion R., and Fabian C. Polcyn. Comparative Multi spectral Sensing . Report No. 2900-484-R. Ann Arbor t Institute of Science and Technology, university of Michigan, 1965. Kern, Clifford D. Evaluation of Infrared Emission of Clouds and Ground as Measured by Weather Satellites . Environmental Research Paper No. 155, U.S. Air Force Cambridge Research Laboratories, 1965. Latham, James P. The Distance Relations and Some other Characteristics of Cropland Areas in Pennsylvania ? An Experiment in Methodology . Technical Report No. 4, NR 339-055. Washington i Office of Naval Research, 1958. . Electronic Measurement and Analysis of Geo graphic Phenomena . Final Report, NR 387-023. waishington: Office of Naval Research, 1964. . Methodology for Instrumented Geographic Analysis . Technical Report No. 2, NR 387-023. Washington: Office of Naval Research, 1962. . Possible Applications of Electronic Scanning and Cong>uter Devices to the Analysis of Geographic Phenomena . Technical Report No. 1, BIB 387-023. Washington I office of Naval Research, 1959. . A Study of the Application of Electronic Scan ning and Con^uter Devices to the Analysis of Geo graphic Phenomena . Final Report, NR 387-023. Washington: Office of Naval Research, 1959. Latham, James P., and Richard E. Witmer. Comparative Wave form Analysis of Multisensor Imagery . Technical Report No. 3, NR 387-034. Washington: Office of Naval Research, 1967. (To be published in Photograraroetric Engineering , July, 1967.)

PAGE 213

203 LlRiperl8« T.« and D. G«orge. ElactrcMnagnetic Field signatures in the Optical Spectrum . Ann Arbor t Willow Run Laboratories, Institute of Science and Technology, university o£ Michigan (no date) . Ludlum, R. Ground-Truth Research for an Airborne Multisensor Survey Program . Texas Instrtunents , Inc., 1966, Ludlum, R. , and J. R. Van LopiX. A Remote Sensing Survey of Areas in Central Coastal Louisiana (Part I — Discussion) . Texas Instruments, Inc., 1966. Marschner, P. J. Land Use emd Its Patterns in the united States . Agricultural Handbook No. 133. Washington* U.S. Department of Agriculture, 1959. Marshall, Ernest W. Air Photo Interpretation of Great Lakes Ice Features . Special Report No. 23. Ann Arbors Great Lakes Research Division, Institute of Science and Technology, university of Michigan, 1966. Mathers, Boyd L. Relative Effectiveness of Different View ing Devices for Photo Interpretation . Technical Research Note 179. Washington: U.S. Army Personnel Research Office, 1966. Merifield, P. M. , and J. RansMlkanp. Terrain in Tiros pictures . Lockheed-California Co., 1964. Moxham, R. M. , D. R. Crandell, and w. Marlett. Thermal Features at Mount Rainier, Washington as Revealed by Infrared Surveys . Professional Paper 52 5d. Washington: U.S. Geological Survey, 1965. National Aeronautics and Space Administration. Geographic Data from Space . Proceedings of the Conference on the Use of Orbiting Spacecraft in Geographic Research, Houston: Manned Spacecraft Center, January 28-30, 1965. Olson, Charles E,, Jr., et al . An Analysis of Measurements of Light Reflectance from Tree Foliage Made during 1960 and 1961 . Agricultural Experiment Station, university of Illinois, 1964.

PAGE 214

204 Parker, D. , D. Fisher, C. Miller, and J. Morgan. Remote Sensing of Environment (Final Report) . Report No. 4864~6-F. Ann Arbor t Institute of science and Technology, university of Michigan, 1963. Saari, J. M. , R. W. Shorthlll, and T. K. Deaton. Infrared and visible Images of the Eclipsed Moon of Decem ber 19, 1964 . U.S. Air Force Cambridge Research Laboratories, 1966. Swet, C. J. Line Scan Television for Earth Observation Satellites . Baltimore: implied Physics Laboratory, The Johns Hopkins university, 1966. U.S. Air Force. Tactical Air Reconnaissance . Manual 55-6. Washington: U.S. Air Force, August, 1964. U.S. Army. Basic and Advanced Infrared Technology . Redstone Arsenal, Ala.: U.S. Army, April, 1965. Electronics Applied In Aerial Surveying . Fort Belvolr, Va. : U.S. Anny, 1965. (Translated from Chinese. Author: Kuo, Wel-hslng.) . Transients and waveforms. Department of the Aintny Technical Manual 11-669. Washington: U.S. Army, November, 1951. U.S. Naval Photographic Interpretation Center. Photographic Interpretation Keys: Military Geology . U.S. Naval Photographic Intelligence Series, November, 1956. Photographic Interpretation Manual: Photographic Metrics . U.S. Naval Photographic Intelligence series, April, 1955. Willow Run Laboratories. Peaceful Uses of Earth-Observation Spacecraft . Volume I: Introduction and Summary . Volume II: Survey of Applications and Benefits . Volume III: Sensor Requirements . Ann Arbor: Institute of Science and Technology, tjnlversity of Michigan, 1966. Wilson, R. A. The Evaluation of an Airborne Infrared Mapper as a Tool for Detecting and Measuring Fires . Research

PAGE 215

205 Paper INT-25. Intennountaln Forest and Range Experiment Station « 1966. Articles and Periodicala Abraham « V. "Relative Geometric Strength of Frame, Strip, and Panoramic Cameras , " Photogrammetrlc Engineering , December, 1961, pp. 753-766. Adams, F. L. "The Design of Meteorological Sensors," in Proceedings of the Second Symposium on Remote Sensing of Environment . Ann Arbors University of Michigan, 1963, pp. 435-452. Alexander, Robert H. "Geographic Data from Space," Professional Geographer , November, 1964, pp. 1-5. Astheimer, R. w. "An Infrared Radiation Air Thermometer," in Proceedings of the second Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1963, pp. 375-402. Avery, Gene, and Cirilo B. Serna. "Photo Interpretation Keys for Identifying Vegetation," Proceedings of the Society of American Foresters , 1965, p. 159. Badgley, Peter c. "The Applications of Remote Sensors in Planetary Exploration," in Proceedings of the Third Syaposium on Remote Sensing of Environment . Ann Arbors university of Michigan, 1965, pp. 9-28. Badgley, Peter C. , and W. L. vest. "Orbital Remote Sensing and Natural Resources , " photogrammetrlc Engineering , September, 1966, pp. 780-790. Bally, Everett H. , and William R. Parish. "Portable Low Cost Detector for Latent Forest Fires , " in Proceedings of the Third Symposium on Remote Sensing of Environ ment . Ann Arbors University of Michigan, 1965, pp. 649-666. Ballard, S., euid W. Vtolfe. "Recent Developments in Infrared Technology," Applied Optics , September, 1962.

PAGE 216

206 Bandeen, W. R. # et al . "Infrared and Reflected Solar Radiation Measurements from the TIROS II Meteorological Satellite*" Journal of Geophysical Research , October, 1961, pp. 3169-3185. Bang, B. A. "High sensitivity Television as an Aid to Low Level Photographic Recording, " Journal of the Society of Motion Picture and Television Engineers , September, 1961. Barnes, R. B. "Thermography of the Human Body," Science , Volume 140 (1963), pp. 870-877. Barr, E. Scott. "Historical Survey of the Early Development of the Infrared Spectral Region," American Journal of Physics , Volume 28, No. 42 (1960). Barringer, A. R. "The Use of Multi-Parameter Remote Sensors as an In^>ortant New Tool for Mineral and Wkter Resource Evaluation," in Proceedings of the Fourth Symposium on Remote sensing of Environment . Ann Arborz university of Michigan, 1966, pp. 313-326. Bell, E. E., and I. L. Eisner. "Infrared Radiation from the White Sands at White Sands National Monument, New Mexico," Journal of the Optical Society of America , Volume 46 (1956) , pp. 303-304. Beller, William S. "IR Photos Yield Data on Natural Resources," Technology Week , August 15, 1966, pp. 2430. Beuttner, Konrad J. "The Consequences of Terrestrial Surface Infrared Emissivity, " in Proceedings of the Third Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1965, pp. 549562. Biberman, L. "Problems in Near Real Time Reconnaissance," in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbors university of Michigan, 1966, pp. 101-110. Bird, J. B«, and A. Morrison. "Space Photography and Its Geographical Applications," Geographical Review , October, 1964, pp. 464-486.

PAGE 217

207 Brody« R. H. , and J. R. Ermllch. "Fourier Analysis of Aerial Photographs," in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbor: University of Michigan, 1966, pp. 375-392. Cantrell, J. L. "Infrared Geology," Photograncaetric Engineer ing , November, 1964, pp. 916-922. Carneggie, David M. , and Donald T. Lauer. "Uses of Multiband Remote Sensing in Forest and Range Inventory, " Photogrammetr ia , August, 1966, pp. 115-141. Carr, D. o. , and R. F. Blakely. "TMoperature Variation at a Depth of 30 cm. in clay Till and Outwash Sand and Gravel," in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1966, pp. 203-214. Chalfin, G. T.f and w. B. Ricketts. "3.2 Mm Thexnaal Imaging SiQ>erinents , " in Proceedings of the Fourth Symposium on ReiBOte Sensing of Environment . Ann Arbor: university of Michigan, 1966, pp. 859-865. Colwell, Robert N. "Aerial Photographs Show Range Conditions," California Agriculture , December, 1961, pp. 12-13. . "Aerial Photography — A Valuable Sensor for the Scientist," American Scientist, March, 1964. "Aerial Photography of the Earth's Surface; Its Procurement and use," Applied optics , June, 1966, pp. 883-892. . "Multiband Sensing," in Proceedings of the First symposium on Remote Sensing of Environment . Ann Arbor t university of Michigan, 1962, pp. 69-70. . "Platforms for Testing Multi-Sensor Equipment," in Proceedings of the Second Symposium on Remote Sensing of Environment . Ann Arbor: University of Michigan, 1963, pp. 7-49. . "Sone Practical Applications of Multispectral Reconnaissance," American Scientist, March, 1961.

PAGE 218

208 Colwell, Robert N. "A Systematic Analysis of Some Factors Affecting Photographic Interpretation," Photogrammetrlc Engineering , May, 1954, pp. 433-453. "The Taking of Helicopter Photography for Use In Photogrammetric Research and Training," Photograuwmetrlc Engineering , September, 1956, pp. 613-621, . "Uses and Limitations of Multlspectral Remote Sensing, " In Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbors university of Michigan, 1966, pp. 71-100. Colwell, Robert N. , et al . "Basic Matter and Energy Relationships Involved In Remote Reconnaissance," Photo grammetric Engineering , September , 1 96 3 . Colwell, Robert N., and D. C. Olson. "Thermal Infrared Imagery and Its use In Vegetation Analysis by Remote Aerial Reconnaissance," In Proceedings of the Third Symposlinn on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1965, pp. 607-622. Cook, R. N. "Preliminary Findings on a New Method of Measurement of Photographic Resolving Power," In Proceedings of the Second Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1963, pp. 139-144. Cooper, Charles F. "Potential Applications of Remote Sensing to Ecological Research," in Proceedings of the Third Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1965, pp. 601606. Coulson, K. L. , G. M. Bourlcus, and E. L. Gray. "Optical Reflection Properties of Natural Surfaces," Journal of Geophysical Research , September 15, 1965, pp. 4601-4611. Crawford, N. C. , and R. W. Peplles. "Remote Sensor Returns as Historical Documents," in Proceedings of the Fourth Syngaosium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1966, pp. 599604.

PAGE 219

209 Davis, Charles M. "Geographic Application Program for Data from Remote Sensors in Aircraft and Spacecraft," Professional Geographer , September, 1966, p. 318. Derenyi, Eugene E., and Gottfried Konecny. "Geometry of Infrared Imagery," Canadian Surveyor , September, 1964, pp. 279-290. "Infrared Scan Geometry," Phot ograraroe trie Engineering , September, 1966, pp. 773-778. DeWaard, R., and E. M. wormser. "Description and Properties of Various Thermal Detectors," Proceedings of the Institute of Radio Engineers , Volume 47, No. 1508 (1959). El-Ashby, M. R., and H. R. vthnless. "Shoreline Features and Their Changes," Photograwntetric Engineering , February, 1967, pp. 184-189. Elms, D. G. "Mapping with a strip Camera," Photogrammetric Engineering , September, 1964, pp. 638-653. England, G., and J. o. Morgan. "Quamtitative Airborne Infrared Mapping," in Proceedings of the Third Symposium on Remote Sensing of Environment . Ann Arbor t university of Michigan, 1965, pp. 681-690. Estes, John E. "Some Applications of Aerial Infrared Imagery, " Annals of the Association of American Geographers . December, 1966, pp. 673-682. . "Some Geographic jl^plications of Aerial Infrared Imagery, " in Proceedings of the Fourth symposixan on R«BDte Sensing of Environment . Ann Arbors Uhiversity of Michigan, 1966, pp. 173-182. Feder, A. M. "Programs in Remote Sensing of Terrain," in Proceedings of the Second Symposium on Remote Sens ing of Environment . Ann Arbor » university of Michigan, 1963, pp. 51-63. Fischer, W. A. "Geologic Applications of Remote Sensors," in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbor t University of Michigan, 1966, pp. 13-20.

PAGE 220

210 Fischer* W. A., et al . "Infrared Surveys of Hawaiian Volcamoes," science . Volume 146, No. 6 (1964), pp. 733-742. Fritz, s. "The Variable Appearance of the Earth from Satellites," Monthly weather Review , October-December, 1963, pp. 613-620. Frost, Robert E. "Aerial Photography; A Reappraisal of the Technology," In Proceedings of the First symposium on Remote Sensing of Environment . Ann Arbor t university of Michigan, 1962, pp. 61-68. Fujlta, T. "Evaluation of Errors In the Graphical Rectification of Satellite Photographs," Journal of Geophysi cal Research . Volume 70, No. 24 (1965), pp. 59976007. Fujlta, T., and W. Bandeen. "Resolution of the Nimbus High Resolution Infrared Radiometer," Journal of Applied Meteorology , Volume 4, No. 4 (1965), pp. 492-503. Gates, o. M. "The Energy Environment In Which We Live," American Scientist , September, 1963. Gates, D. M. , et al . "Spectral Properties of Plants," Applied Optics , April, 1965, pp. 11-20. Gates, M. "Characteristics of Soil and Vegetated Surfaces to Reflected and Emitted Radiation," in Proceedings of the Third Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1965, pp. 573-600. Glever, P. H. "Needs for Remote Sensing Data in the Field of Air and water Pollution Control," in Proceedings of the Fourth symposium on Remote Sensing of En vironment . Ann Arbors university of Michigan, 1966, pp. 21-24. Glmbarzevsky, Philip. "Land Inventory Interpretation," Photograraraetrlc Engineering , November, 1966, pp. 967-976. Goldbert, I. L. , et al . "Nimbus High Resolution Infrared Measurements , " in Proceedings of the Third symposium

PAGE 221

211 on RwBote Sensing of Eavironinent . Ann Arbor: university of Michigan, 1965, pp. 141-152, Goodman, R. E. "Photo/Pield Prospecting," photograinroetric Engineering , March, 1960, pp. 100-105. Harpe, E., Jr. "Anthropology and Remote Sensing," in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbori university of Michigan, 1966, pp. 727-730. Harris, David E., and Caspar L. Woodbridge. "Terrain Mapping by use of Infrared Radiation," Photogrammetric Engineering . January, 1964, pp. 134-139. Hawkins, J. K. , and C. J. Munsey. "Automatic Photo Reading," Photogrammetric Engineering . July, 1963, pp. 627-637. Hirsch, s. N. "Applications of Rcanote Sensing to Forest Fire Detection and Suppression," in Proceedings of the Second Syngaosium on Remote Sensing of Environment . Ann Arbor t University of Michigan, 1963, pp. 295308. • "Preliminary Experimental Results with Infrared Line Sceuiners for Forest Fire Surveillance," in Proceedings of the Third Symposium on Remote Sensing of Environment . Ann Arbor: University of Michigan, 1965, pp. 623-648. Hoffer, R. M. , et al . "Vegetable, Soil, and Photographic Factors Affecting Tone in Agricultural Remote Multispectral Sensing," in Proceedings of the Fourth Sym posium on Remote Sensing of Environment . Ann Arbori university of Michigan, 1966, pp. 113-134. Holter, M. R. , and w. c. Wolfe. "Optical -Mechanical Scanning Techniques," Proceedings of the Institute of Radio Engineers , September, 1959, pp. 1540-1330. Horwitz, L. P., and G. L. Shelton, Jr. "Pattern Recognition Using Autocorrelation," Proceedings of the Institute of Radio Engineers . January, 1961, pp. 175-185.

PAGE 222

212 Hubbs, John C. "The New Pulse," Journal of the Precision Meaaurera^itB Association (Instruments and Control Systems) , November, 1966, pp. 109-116. Jastrow, R., and A. G. w. Cameron. "Spaces Highlights of Recent Research," Science , September 11, 1964, pp. 1129-1139. Jensen, Herbert A., and Robert N. Colwell. "Panchromatic versus Infrared Minus-Blue Aerial Photography for Forestry Purposes in California," Photogrammetric Engineering , March, 1949, pp. 201-223. Johnson, P. L. "A Consideration of Methodology in Photo Interpretation," in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbor t university of Michigan, 1966, pp. 719-726. Jones, R. Clark. "Noise in Radiation Detectors," Proceedings of the Institute of Radio Engineers . Volume 47, No. 1481 (1959). , "phenomenological Description of the Response and Detecting Ability of Radiation Detectors," Proceedings of the Institute of Radio Engineers . Volume 47, No. 1495 (1959). Kedar, Yehuda. "The Use of Aerial Photographs in Research in Physiographic Conditions and Anthropogeographic Data in Various Historic Periods," Photograunroetric Engineering . July, 1938, pp. 584-587. Kern, C. D. "Desert Soil Ten^eratures and Infrared Radiation Received by TIROS III," Journal of Atmospheric Science . Volxjme 20 (1963) , p. 175. Ketchum, R. D., Jr., and W. I. wittman. "Infrared Scanning the Arctic Pack Ice," in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbort university of Michigan, 1966, pp. 636-656. Kiefer, Ralph W. "Landform Features in the united States," Photograrametric Engineering . February, 1967, pp. 174-182. Kinsman, Frank E. "Some Fundamentals in Non-Contact Electro-

PAGE 223

213 magnetic Sensing for Geosdence Purposes « " In Proceedings of the Third synposlum on Remote Sensing of Environment . Ann Arbor x university of Michigan* 1965, pp. 495-516. Klntzlnger, P. R. "Geothermal Survey of Hot Ground near Lordsbury, New Mexico*" science , Volume 124 (1956), p. 629. Kllnm, Lester E. "Regional Description Based on Texture and Pattern of unit Areas" (Abstract) , Annals of the Association of American Geographers . Sept«iiber« 1956, p. 256. Konecny, Gottfried, and Eugene E. Derenyl. "Geometrical Considerations for Mapping from scan Imagery," in Proceedings of the Fourth Symposium on Remote Sens ing of Environment . Ann Arbor: university of Michigan, 1966, pp. 327-338. Lacate, D. S. "Wlldland Inventory and Mapping," Forestry Chronicle , June, 1966, pp. 184-194. LaFord, Charles D. "IR Mapping System Readied for Market," Technology Week , February 27, 1967, pp. 30-33. Lancaster, Charles w. , and Allen M. Feder. "The Multisensor Mission," Photograunmetr ic Engineering , May, 1966, pp. 484-494. Latham, James P. "The Distance Relations of Cropland Areas in Pennsylvania" (Abstract) , Annals of the Associ ation of American Geographers , September, 1958, p. 277. . "Geographic Analysis and Remote Sensing Capability, " in Proceedings of the Second Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1963, pp. 65-79. . "Remote Sensing of Environment," Geographical Review , April, 1966, pp. 288-291. . "Resume of the Special Session on Remote Sensing Held at the 1965 AAAS Meeting," in Proceedings of the Fourth Symposlvm> on Remote Sensing of Environ -

PAGE 224

214 ment. Ann Arbors university of Michigan* 1966, pp. 539-546. Lattman, L. H. "Geologic Interpretation of Airborne Infrared Imagery, " in Proceedings of the Second Symposivun on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1963, pp. 289-293. Leestma, R. A. "Applications of Air and Spacebome sensor Imagery for the study of Natural Resources , " in Proceedings of the Fourth Symposium on Remote Sens ing of Environment . Ann Arbor: University of Michigan, 1966, pp. 111-118. Legault, R. R. "The Motivation for Multispectral Sensing," in Proceedings of the Third Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1965, pp. 71-78. Legault, R. R., and F. C. Polcyn. "investigations of Multi-Spectral Image Interpretation," in Proceedings of the Third symposium on Remote Sensing of Environ ment . Ann Arbor: university of Michigan, 1965, pp. 813-821. Leistner, G., and P. P. Dieter. "Some Characteristics of Panoramic C2unera8 for Aerial Surveillance, " Photo graphic Science and Engineering , Volume 5, No. 3, pp. 257-262. Leonardo, Earl S. "Capabilities and Limitations of Remote Sensors," Photogrammetric Engineering , November, 1964, pp. 1005-1010. Leuder, O. R. "Airphoto Interpretation as an Aid in Mineral Reconnaissance and Development," Photogrgumnetric Engineering , Deceaiber, 1953, pp. 819-830. Liang, Ta. "Airphoto Interpretation of Engineering Soils in Tropical Environments," in Proceedings of the Third Symposium on Panote Sensing of Environment . Ann Arbor: university of Michigan, 1965, j>p. 531536. Limperis, T. "Target and Background Signature Study," in Proceedings of the Third Symposium on Remote Sens-

PAGE 225

215 Ing of Environment . Ann Arbor t university of Michigan, 1965, pp. 423-434. Livingston, w. c. "Resolution Capability of the ImageOrthicon Camera Tube under Nonstandard Conditions , ** Journal of the Society of Motion Picture and Television Engineers , October, 1963. Lohman, S. w. , and c. J. Robinove. "Photographic Description and Appraisal of Water Resources , " photograun metria. Volume 19, No. 3 (1964) . Lowe, D. S., and John G. N. Braithwaite. "A Spectrum Matching Technique for Enhancing Image Contrast, " Applied Optics , June, 1966, pp. 893-898. Lowe, D. s., and F. C. Polcyn. "Multispectral Data Collection Program, " in Proceedings of the Third symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1965, pp. 667-680. Lyon, R. J. P., and E. A. Bums. "Analysis of Rocks and Minerals by Reflected Infrared Radiation, " Econcwiic Geology , Volume 58 (1963) , p. 274. Lyon, R. J. P., and J. w. Patterson. "Infrared Spectral Signatures — A Field Geological Tool," in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbor t University of Michigan, 1966, pp. 215-230. McAlister, E. D. "Infrared-Optical Techniques Applied to Oceanography Measurement of Total Heat Flow from the Sea Surface," Applied Optics , May, 1964. McClellan, W. D., J. P. Meiners, and D. G. Orr. "Spectral Reflection studies on Plants," in Proceedings of the Second symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1963, pp. 403414. McLerran, James H. "Airborne Crevasse Detection," in Pro ceedings of the Third Symposium on Remote Sensing of Environment . Ann Arbor: Uhiversity of Michigan, 1965, pp. 801-802.

PAGE 226

216 McLerran, James H. "Infrared Sea Ice Reconnaissance," in Proceedings of the Third syny>osiura on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1965, pp. 789-800. McLerran, James H. , and Joseph O. Morgan. "Thermal Mapping of Yellowstone National Park," in Proceedings of the Third Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1965, pp. 517530. Marble, D. F., and E. N. Thomas. "Some Observations on the Utility of Multispectral Photography for Urban Research," in Proceedings of the Fourth Synqx)sium on R«aote Sensing of Environment . Ann Arbor: university of Michigan, 1966, pp. 135-144. Marcus, Leslie P., and Robert N. Colwell. "Determining the Specifications for Special Purpose Photography," Photograrametric Engineering , September, 1961, pp. 618-626. Meier, M. P., R. H. Alexander, and W. J. Campbell. "Multispectral Sensing Tests at South Cascade Glacier, Washington," in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1966, pp. 143-160. Meyer, M. A. "Remote Sensing of Ice and Snow Thickness," in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbor: University of Michigan, 1966, pp. 183-192. Meyer, M. P., and D. W. French. "Forest Disease Spread," Photogrammetric Engineering , September, 1966, pp. 812-814. Miller, Lee D. "Location of Anoooalously Hot Earth with Infrared Imagery in Yellowstone National Park," in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1966, pp. 751-769. Miller, William C. "Uses of Aerial Photographs in Archeological Field Work," American Antiquity , January, 1957, pp. 46-62.

PAGE 227

217 Mollneux, Carlton E. "Aerial Reconnaissance of Surface Features with the Multiband Spectral System, " in Proceedings of the Third Symposium on Remote Sens ing of Environment . Ann Arbor: University of Michigan, 1965, pp. 399-422. . "Air Force R«note Sensing Programs , " in Pro ceedings of the First Symposixan on Remote Sensing of Environment . Ann Arbor: University of Michigan, 1962, pp. 71-74. Morgan, Joseph. "Infrared Technology," in Proceedings of the First Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1962, pp. 45-60. Morrison, A., and B. J. Bird. "Photography of the Earth from Space and Its Non-Meteorological Applications," in Proceedings of the Third syng)OSium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1965, pp. 357-376. Moxheun, R. M., and A. Alcraz. "Infrared Surveys at Taal Volcano, Philippines," in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1966, pp. 827-844, Narva, M. A., and F. A. Muckler. "Visual Surveillance and Reconnaissance from Space Vehicles," Human Factors , 1963, pp. 295-315. Meuhauser, R. G., et al . "The Design and Performamce of a High-Resolution Vidicon, " Journal of the society of Motion Picture and Television Engineers , November, 1962. Norberg, W., et al . "Preliminary Results of Radiation NMiaureaents from the TIROS III Meteorological Satellite," Journal of Atmospheric Science , Volume 19 (1962), pp. 20-30. Ockert, D. L. "Satellite Photography with Strip and Frame Cameras," photogr am—trie Engineering , September, 1960, pp. 562-568. Olson, C. E., Jr. "Elements of Photographic Interpretation

PAGE 228

218 Comnion to Several Sensors," Photogrammetric Engineering , September, 1960, pp. 651-656. Olson, C. E. , Jr. "The Energy Plow Profile in Remote Sensing," in Proceedings of the Secx>nd Symposium on Remote Sensing of Environment . Ann Arbor i university of Michigan, 1963, pp. 187-199. . "Infrared Sensors and Their Potential Applications to Forestry," Papers of the Michigan Academy of Science, Arts, and Letters , Volume 30 (1965) , pp. 39-47. Olson, C. E., Jr., and R. E. Good. "Seasonal Changes in Light Reflectance from Forest Vegetation," Photogrammetric Engineering , March, 1962, pp. 107-114. Ory, Thomas R. "Line-Scanning Reconnaissance Systems in Land Utilization and Terrain Studies," in Proceedings of the Third Symposium on Remote Sensing of Environ ment . Ann Arbor: University of Michigan, 1965, pp. 393-398. Paijmans, K. "Typing of Tropical Vegetation by Aerial Photographs and Field Sampling in Northern Papua," Photogrammetr ia , February, 1966, pp. 1-25. Parker, Dana. "Sane Basic considerations Related to the Problem of Ranote Sensing," in Proceedings of the First symposium on Remote Sensing of Environment . Ann Arbori university of Michigan, 1962, pp. 7-18. Poulin, A. O., and T. A. Harwood. "Infrared Mapping of Glacier Thermal Anomalies," Proceedings of the Sym posium on Glacier Mapping , Ottawa, Ontario, Canada, September, 1965. Poulin, A. O. , and J. W. Patterson. "Infrared Imagery in the Arctic under Daylight Conditions," in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1966, pp. 231-242. Pourciau, L. L. , M. Altman, and C. A. Washburn. "A HighResolution Television System," Journal of the Society

PAGE 229

219 of Motion Picture and Television Engineers , February, 1960. Prentice, Virginia L. "Remote Sensing of Environment," Professional Geographer , January, 1963, pp. 20-21. Proceedings of the Institute of Radio Engineers , special Issue on Infrared Physics and Technology, septenber, 1959, pp. 1420-1647. Reeves, Dache. "Aerial Photography and Archaeology," Ameri can Antiquity , Volume 2 (1936), pp. 102-107. Robinove, c. J. "Infrared Photography and Imagery in Water Resources Research," American Water works Associatioi Journal, Volume 57, No. 7 (1965), pp. 834-840. . "Remote-Sensor Applications in Hydrology," in Proceedings of the Fourth syngvosiuro on Remote Sensing of Environment . Ann Arbor t University of Michigan, 1966, pp. 25-32. Rosenfeld, Azriel. "An Approach to Automatic Photographic Interpretation, " Photogrammetric Engineering , Septenber, 1962, pp. 660-665. . "Automatic Recognition of Basic Terrain Types from Aerial Photographs," Photogrammetric Engineering , March, 1962, pp. 115-132. Saunders, P. M. , and c. H. wilkens. "Precise Airborne Radiation Thermometry," in Proceedings of the Fourth Symposium on Remote sensing of Environment . Ann Arbor t university of Michigan, 1966, pp. 815-826. Schneider, willieun J. "Water Resources in the Everglades," Photogranroetric Engineering , Mdvember, 1966, pp. 958-966. Schulte, D. W. "The Use of Panchromatic, Infrared, and Color Aerial Photography in the Study of Plant Distribution," Photogrammetric Engineering , December, 1951, pp. 688-712. Shay, J. R. "Some Meeds for Expanding Agricultural Remote Sensing Research," in Proceedings of the Fourth

PAGE 230

220 Sympoaiiaa on Remote Sensing of Environment . Ann Arbors University of Michigan* 1966, pp. 33-36, Sherman, J. C. "Accumulation of Geographic Data through Remote Sensing Techniques," in Proceedings of the Second Syiqposium on Remote Sensing of Environment . Ann Arbor: University of Michigan, 1963, pp. 427430. Siraonett, D. S. "Present and Future Needs of Remote Sensing in Geography, " in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbors university of Michigam, 1966, pp. 37-48. Steiner, D. , and H. Haefner. "Tone Distortion for Automated Interpretation," Photogrammetric Engineering , March, 1965, pp. 269-280. Stone, Kirk R. "A Guide to the Interpretation and Analysis of Aerial Photos," Annals of the Association of Awrican Geographers , September, 1964, pp. 318-328. Strandberg, Carl H. "Water Quality Analysis," Photogram metric Engineering , March, 1966, pp. 234-248. Strangeway, D. W. , and R. C. Holmer. "Infrared Geology," in Proceedings of the Third SympoaiiMn on Remote Sensing of Environment . Ann Arbors University of Michigam, 1965, pp. 293-320. Suits, G. R. "Declassification of Infrared Devices," Photograumnetric Engineering , November, 1966, pp. 988-992. . "The Nature of Infrared Radiation and ways to Photograp'i It," Photogrammetric Enc^ineering , December, 1960, pp. 763-772. Tarkington, R. G., and A. L. Sorem. "Color and False-Color Films for Aerial Photography," Photograutaaetric Engineering , January, 1963, pp. 88-95. Teleki, Geza. "Aerial Photography and Sea-Ice Forecasting," Naval Research Reviews , March, 1962, pp. 1-8. Tewinkel, G. C. "Water Depths from Aerial Photographs," Photogrammetric Engineering , November, 1963, pp. 1037-1042.

PAGE 231

221 Theurer* Charles. "Color aad Infrared Eaqjerimental Photography for Coastal Mapping," Photogranimetric Engineering , September, 1959, p. 368, Thoren, Ragner. "Photointerpretation in Military Intelligence," Photogrananetric Engineering . April, 1952, pp. 428-451. Udall, s. L. "Resource Understanding — A Challenge to Aerial Methods," Photogramaetric Engineering . January, 1965, pp. 63-75. Van Lopik, J. R. , and L. A. Ylarborough. "Connnents on Remote Sensing Needs in Geoscience Engineering and Explorations , " in Proceedings of the Fourth Symposium on Remote Sensing of Environment . Ann Arbor: university of Michigan, 1966, pp. 49-54. Wark, D. Q. , G. Ykmamoto, and J. H. Lienesch. "Methods of Estimating Infrared Flux and Surface Temperature from Meteorological Satellites," Journal of Atmo spheric Science , September, 1962, pp. 369-384. Wickens, G. C. "The Practical Application of Aerial Photography for Ecological Surveys in the Savannah Regions of Africa," Photogrammetria , August, 1966, pp. 33-41. Wilson, R. C. "Forestry Applications of Remote sensing," in Proceedings of the Fourth symposiviro on Remote Sens ing of Environment . Ann Arbor: university of Michigan, 1966, pp. 63-70. Winkler, E. M. "Relationship of Airphoto Tone Control and Moisture Content in Glacial Soils," in Proceedings of the second Symposium on Remote Sensing of Environ ment . Ann Arbor t university of Michigan, 1962, pp. 107-118. Wolvin, John H. "Multi -Channel Photo Spectrum Recording," in Proceedings of the Second Symposium on Remote Sensing of Environment . Ann Arbor: University of Michigan, 1963, pp. 81-88.

PAGE 232

222 Pap>era Atkinson* John H. "Optimizing Photography through the Atmosphere," paper presented at 1966 A.S.P. Semi -Annual Conventlc»i, Los Angeles, California, September 2730, 1966. Heller, R. C. , G. E. Doversplke, and R. C. Aldrlch. "Identification of Tree Species on Large-scale Panchromatic and Color Photographs , " paper presented at 1963 Annual Meeting of American Society of Photograametry, Vteshington, D. C. , March 25-28, 1963. Hdlter, N. R. "Infrared and Multlspectral Sensors," paper presented at the 61st Annual Meeting of the A.A.6.« Colximbus, Ohio, April 18-22, 1965. "Sensing Instruments," paper presented at the Conference on Planetology and Space Mission Planning, Mew York, N. Y. , November, 1965. Latham, James P. "Machine Evaluation of Images for Regionalization Problems," paper presented to the Commission on the Interpretation of Aerial Photographs, International Geographical Union, Ottawa, Canada, March 16, 1967. . "Role of Instrumentation in Geographic Research," paper presented to the Annual Meetings of the American Association for the Advancement of Science, Philadelphia, Pennsylvania, December 29, 1962. Numbower, Leonard E. "Extraction of Socio-Economic Information from Aerial Photographs," paper presented at 1966 A.S.P. Semi -Annual Convention, Los Angeles, California, September 27-30, 1966. weber, Philip A. " Photogrammetric Research Wdrk on the Gulf Stream," paper presented at 1966 A.S.P. SemiAnnual Convention in Los Angeles, California, September 27-30, 1966.

PAGE 233

223 Miscellaneous Brunnschweiler, Dieter. Physiographic Elements of the Dexter Test Area Relevant to the Interpretation of Remote Sensing Records * unpublished report, NSF Summer Conference on Remote Sensing of Environment « university of Michigan « June, 1966. university of Michigan* School of Natural Resources. stinchfield Woods (Map). Ann Arbor i university of Michigan, 1961. (Scalex 1*11,520.)

PAGE 234

BIOGRAPHICAL SKETCH Richard Everett wltmer was born in swarthmore* Pennsylvania « on July 3, 1942. He moved to Miami, Florida, in 1944, where he received his elementary and secondary education. He entered the University of Florida in September, 1958, where he received the degree Bachelor of Science with a major in physics in 1962. Shortly thereafter, he entered the Graduate School of the University of Florida in the Department of Geography and pursued graduate study toward the degree Master of Science, which was awarded in December, 1964. During 1964 and 1965, he continued graduate studies in the Departments of Geography and Geology at the University of Colorado, and returned to the University of Florida in 1965 to continue studying for the degree Doctor of Philosophy. In June, 1965, he became Instructor of Geography at Florida Atlantic University and Research Assistant for an Office of Naval Research contract. In July, 1967, he became Research Assistant Professor of Geography and Research Associate for a National Aeronautics and Space Administration/Geological Suirvey research contract. The author is a member of the Association of American 224

PAGE 235

225 Geographers and Its Southeastern Division, the Florida Society of Geographers, the American Society of Photograroetry, and the American Association for the Advancement of Science. »

PAGE 236

This dissertation was prepared under the direction of the chairman of the candidate's supervisory committee and has been approved by all members of that conmlttee. It was submitted to the Dean of the College of Arts and sciences and to the Graduate Council, and was approved as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December 19$ 1967 Dean, Co Sciences SUPERVISORY COMMITTEE: Chairman 2^ £^ r>^^' (J^-'^^^^-^di.U^frrf^ Dean, Graduate School