to know how many lakes there are in Minnesota (claims of 10,000 not with 
standing) , and extractive processing could be used to map all water bodies 
and count them. 
Most decision mechanisms are based on multivariate discriminate analysis 
that partitions measurement space on the basis of the training set signatures 
which then allows a decision to be made for an appropriate classification for 
each input pixel. 
The function of display is attendant with each and every one of the other 
functions of a multispectral processing system as one facet of the man-machine 
interaction. Interactive controls and commands allow the intervention of the 
operator to direct that certain things be done which would not otherwise be 
done automatically. 
The user model is the next step in the system. It is a crucial step 
because it translates the information extracted from the remote sensing data, 
adds ancillary information that a user may have at his disposal, and creates 
a product which can help a user make a decision. For example, suppose the 
previously mentioned user wanted to estimate the migratory water-fowl produced 
in Minnesota. Because these birds nest in or near water, the amount of water 
present is an important but not the only variable. The number of waterfowl 
present to breed is important, as is the food supply predator information. 
These ancillary variables are required to calculate the number of migratory 
waterfowl. 
Carrying this example a bit further, the user may want to know the number 
of waterfowl to be able to set hunting limits in the fall. He will need to 
consider other ancillary information, such as what neighboring states are doing, 
before making a decision. This is the user’s decision model. 
The impression that user and decision models are necessarily formal, 
mathematical models implemented on computers is not true. Rather these models 
are more nearly well defined procedures that managers follow to arrive at 
conclusions or to convert earth resources processed data to a form they can 
use. The trend is to increasing formalism and mathematics in user models 
where inputs and outputs are usually quantifiable. Decision models, with their 
typical socio-economic ancillary inputs, probably will seldom be formalized to 
the stage where computer implementation is feasible. 
With this perspective on an earth resources system, it is perhaps 
pertinent to point out that with increasing development of high data rate 
sensors (by NASA) and of user interest in outputs from the system in more 
timely fashion (to have maximum impact on decisions, information must be 
timely), the squeeze is on the middle of the system (preprocessing, extrac 
tive processing, and user model areas.) In the remainder of the paper we 
consider some recent advancements in extractive processing and preprocessing 
and in user model development which we feel begin to close the gap between 
the sensor’s abilities to collect data and the user’s desires to digest it, 
and the ability of extractive processing systems to keep up with the data.