sponses associated with significant range pheno- 
mena. Also, he uses other image characteris- 
tics, including size, shape, texture, pattern, and 
associated features, to aid in detecting, iden- 
tifying, and evaluating directly, or by conver- 
gence of evidence, most discernible features. 
Ultimately, the image analyst can check the 
accuracy of his interpretations by either (a) 
comparing them with ground information ob- 
tained at the time of the overflight, or (b) by 
taking the remote sensing data into the field 
and comparing it with ground features. 
These “tried and true” visual interpretation 
techniques for extracting information from an 
image are well established and are adequate 
for analysis of relatively small land areas. (A 
complete treatment of photo interpretation 
may be found in “Manual of Photographic In- 
terpretation” (American Society of Photo- 
grammetry 1960). 
However, for analysis of large areas where 
the need for rapid and detailed information is 
urgent, image interpretation using visual tech- 
niques may be too time consuming. 
A means for increasing the rate and accuracy 
of information extraction from remote sensing 
data—using visual interpretation—is by a dou- 
ble-sampling procedure. Such a procedure in- 
volves procurement of imagery of an area from 
two levels. For example, a generalized, synop- 
tic view of a large land area can be provided 
by high-flying aircraft, or from earth-orbiting 
vehicles. At the same time, a very detailed 
view of small sample sites within such a large 
area could be obtained using a 70-mm. camera 
system capable of procuring large-scale photos. 
Thus, an image analyst could derive detailed 
information by studying a few images of 
sample sites and extending this information 
over a much larger land area as provided by 
the synoptic view. 
To satisfy the increased demand for rapid 
information about our resources, the day may 
come when several remote sensors will be oper- 
ated at frequent intervals (as from an earth- 
orbiting vehicle.) Then, the photo interpreter 
left to his own means may suddenly become 
deluged with remote sensing data. Recognition 
of this problem has spurred research efforts 
for more rapid techniques for extracting infor- 
mation from the remote sensing data. 
The photo interpreter’s judgment is not 
likely to be replaced by automation, but many 
encouraging techniques yet in the rudimentary 
stages have been recognized as being capable 
of assisting him and expediting information 
flow (Colwell 1968): 
(1) Image enhancement by additive color 
techniques (described in detail by Yost and 
Wenderoth 1968). Briefly, this technique per- 
mits an interpreter to analyze simultaneously 
many black-and-white images such as those in 
figures 4 and 6. The black-and-white images 
theoretically having the same geometry are 
projected through various colored filters and 
are superimposed in common registry on a 
screen. A single color composite formed in this 
manner brings together in various colors the 
tone signatures unique to many spectral bands. 
FIGURE 8 (Color Plate II). Photo pairs, A, B, and C (corresponding Anscochrome color and Ektachrome 
Infrared Aero photos) show enlarged portions of large-scale 70-mm. aerial photographs. Photo pair A, taken in 
July, shows a sharp ecotone between a big-sagebrush and wet meadow plant community (the small-scale view 
is seen at the arrow in color plate I). At this large-scale ( = 1/600) one can estimate shrub density and cover. 
Cattle droppings can be discerned in the meadow. Note also the rodent paths in the meadow. Photo pair B, 
taken in July, illustrates the feasibility of identifying range plants. Bitterbrush (Purshia tridentata) at 1 (a 
palatable species for cattle and deer) is readily distinguished from big-sagebrush (Artemisia tridentata) at 2 
(important because of its abundance, but not a preferred species by either cattle or deer) on the Ektachrome 
Infrared Aero photo. Buckwheat plants (Hriogonum sp.), are readily identified at this phenological stage by 
yellow flowers, at 3. Photo pair C, taken in late August, shows range vegetation in the Black Mesa area of 
Colorado. Helenium hoopseii at 4 and Geranium fremontii at 5 are distinguishable from other species on the Ekta- 
chrome Infrared photo; this again illustrates the feasibility of identifying important range plants at the appro- 
priate phenological stage (see text by Reppert and Driscoll in these Proceedings for further detail). 
Photo D is an Ektachrome Infrared photo taken in early May 1967, with an RC-9 camera (scale 1/84,000) ; 
California annual grassland, east of Berkeley, Calif., is shown. At this date, the dense, healthy range vegetation 
associated with deeper soils appears reddish, whereas the range vegetation associated with shallow upland sites 
and south-facing exposures has dried and appears straw colored. Variations in the reddish appearance of the 
healthy annual grass may be attributed to species composition, vegetation density, degree of utilization, and 
vigor or health. 
Photo E is also an Ektachrome Infrared Aero photo which shows a portion of the same annual grassland 
seen in photo D, but at a relatively larger scale, 1/33,500) and at a different seasonal stage (August). Observe 
that in August when all the annual vegetation is dry, there is little advantage to using color infrared photos 
to evaluate range conditions. 
173 
