691 
Ing 
ts 
flectance, 
Pattern recognition as applied to automatic multispectral processing 
is discussed in terms of a statistical description of features where each 
Pattern or picture element (pixel) is considered as a vector in n-dimensional 
space whose components are the simultaneous response in each spectral band 
from one ground resolution element. The goal of the recognition system is 
to define partitions in this space such that each region can be identified 
with a class of patterns, through the techniques of statistical decision 
nents from 
s to 
ch they 
all be as 
s is 
i learning 
advantage 
human 
y be 
the 
ion and 
theory. 
Pattern recognition is accomplished by comparing information derived 
from an input pixel with similar data derived from known sample patterns 
(called signatures, paradigms, or prototypes) which come from training sets. 
The specification of these signatures is accomplished utilizing a learning 
algorithm. The important ansatz is that these signatures are truly 
"representative" of the class and all of the class dispersion. Based on 
these comparisons, a decision is made as to the nature of the input pattern. 
The key to multispectral recognition is invariance. For example, it is 
desirable that the classification assigned to an object or pattern of interest 
be independent of the position of that object in the field of view, the aspect 
aature. 
a material 
aatures 
sequence 
at which it is viewed; the background against which it is seen, partial 
obscuration of the object, minor changes within a class, and changes in 
illumination or atmospheric condition. It is not too difficult to provide any 
one of these invariances. To provide all of the desired invariances with a 
practical amount of hardware, however, requires that the preprocessing and 
feature extraction mechanism extract the essence of the classes to be 
tion that 
actral 
some 
tical in 
probability 
tatistical 
used to 
identified. 
ADVANCES IN PROCESSING METHODOLOGY 
The automatic processing R&D objective is simply stated as follows: To 
make large area and small area MSS-based Earth Resources Survey (and land use 
and pollution monitoring) informations systems practical by: 
a. increasing accuracy of information extraction (mensuration, location, 
ze that 
es be 
ultaneously, 
the 
, the 
□f error 
present 
Photo 
and correct classification) 
b. decreasing cost (through powerful processing techniques which can 
use reduced ground observations and fast processing on low cost 
equipment) 
c. developing means to disseminate information in user applications 
terms (e.g., volume of production not simply area planted to a crop) 
d. decreasing time for information extraction to preserve information 
value from decay 
e. demonstrating utility in user applications of simplest technique to 
meet desired performance and documenting performance achieved with 
various techniques to allow design of operational systems 
issivity 
procedures 
als of 
f. employing multistage area sampling and bias correction wherever 
accuracy requirements can be met through this cost-saving approach 
g. relaxing costly constraints on data acquisition imposed solely by 
weak processing techniques.