4. Applications and Results 
In this section only a brief indication is given of the applications 
to which these techniques or their earlier counterparts have been applied 
and the type of results achieved. 
It is the benefits in user applications which is the raison d^ etre 
of the entire development of this technology. As shown in the table below 
the potential of multispectral sensing and automatic processing has been 
amply demonstrated. 
1. Many applications in a variety of user disciplines have 
been demonstrated to be feasible under limited conditions 
(that is the discrimination is sufficiently accurate). 
Operational-prototypes may be feasible currently in some 
applications from both technical and cost aspects. 
No operational use is yet being made of information 
derived in this manner. 
A large number of user applications of multispectral earth resources 
and land use information systems have been demonstrated to be feasible in 
scaled down programs. Some applications are further advanced than others 
and operational prototypes should be exercised next for some of these. 
However, to my knowledge no operational use is yet being made of any 
information system on earth resources employing multispectral sensing. 
I believe this is largely because they must become cost effective first. 
One of the primary goals of the efforts being undertaken in remote 
sensing is the development of techniques which will enable large-area 
crop surveys without the need for expending a significant amount of 
manpower gathering ground information. If the amount of necessary ground 
information can be reduced, cost-effective remote crop survey systems 
will become a reality. To accomplish this goal, the effectiveness of 
spectral signatures must be extended in time and space. 
The discussion which follows by Nalepka [33] describes a relatively 
successful attempt at applying object class spectral signatures derived 
from one data set to another set of data gathered on a different day at 
a different location some 200 kilometers distant. In particular, data 
from Segment 203 of the Corn Blight Watch Intensive Study Area were 
processed using signatures from Segment 212. 
The data from Segment 203 were prepared and preprocessed for 
feature enhancement. The preprocessing included the stabilization of 
the data by the sun sensor signal and the elimination of angle effects. 
Since both data sets included a similar distribution of objects having 
the same basic spectral properties, the differences in the angle- 
corrected signal levels were used to quantify differences in scene 
irradiance and atmospheric transmittance at the two locations. This 
information was then used to adjust the spectral signatures determined 
from Segment 212 data so that they could be applied to the Segment 
203 data set.