The next step in the system is extractive processing, designed to extract 
information pertinent to the user from the data. For example the user may wish 
To discuss these points in more detail, first refer to Figure 1, which is 
a familiar block diagram of a typical earth resources survey system. The 
system begins with a sensor viewing phenomena of the terrain or ocean. 
The observed phenomena are the reflected or emitted spectral radiance of the 
scene. Sensor data is collected from remote platforms, so some method of 
data storage and telemetry are required. The next step in the system is 
geometric and radiometric preprocessing to permit production of a map-like 
rendition of the true radiance of the terrain. This step is an absolute 
necessity. Both geographic reference and radiometric feature invariance 
are sine qua non to successful systems. The goal is to reduce the distortions 
of geometry and of radiometric fidelity to negligible levels for further 
processing and analysis. This step in the procedure may require ancillary 
information (e.g., knowledge of spacecraft attitude and ground "control points" 
for performing geometric corrections and knowledge of atmospheric properties 
for correction of radiometric errors). The accuracy and scanner sensitivity 
to which these corrections must be performed are dependent on the application 
being addressed, and may vary widely. 
ANCILLARY DATA ANCILLARY DATA 
The purpose of signal conditioning, or "preprocessing" is to provide data 
preparation and handling, to provide geometric and radiometric correction, to 
provide a convenient input format, to emphasize or enhance aspects of the input 
signal which are deemed important, to provide in many cases a reduction in the 
dimensionality of the input data, and most importantly, to provide invariance. 
An almost universal approach to recognition is to extract properties or 
features from the original signal, and to perform the recognition on the 
feature profile of the input signal. This serves several functions. First, by 
reducing the input pattern to its essential features, the memory required for 
storing the signatures is reduced. Secondly, by reducing the input pattern 
to unique independent features, a considerable amount of invariance to exact 
form is obtained. Finally, a degree of invariance to noise and background 
may be achieved.