277 - 
the 
: be 
ilation 
1 solution 
certain 
some feasible characteristics of the sensors (e.g. field angle, wave length, 
signal/noise ratio, coverage frequency for the area, how can sun be tilted? 
etc.). Some criteria for selection of channels was analysed in /9/; a 
general analysis of the systems are given also in 10-11/. 
s problem 
and 
In other 
hat a 
TECHNICAL OPTIMIZATION OF PHOTOINTERPRETATION- 
REMOTE SENSING SYSTEMS AND PROCESSES 
pl-exity 
The technical optimization for every component as well as for 
the whole process means: /a/ the maximum improvement of performances as 
resolution, detectability, recognizability, signal/noise ratio, linear 
sensivity, dynamic field; /b/ the carrying out of some correlation among 
different components of the system or systems in view of the best use of 
roaching 
. That 
eir 
designed 
ss. 
their performances which are abviously unequal; /c/ the analysis and 
interpretation of component data which are rather unhomogeneous even after 
the establishing of the initial correlation named in item /b/; /d/ the 
design of a system in the broadest meaning of optimization - composed by 
hybrid links and materialized by a path made up from discreet elements or 
especially composed from pathways having the features most flexible and 
suitable to cope with phenomena which are rather heterogeneous and must 
be considered in their dynamics. 
till now 
Adoption of methodology and basic techniques is necessary, for 
a rational approach. In spite of its inherent difficulties, we have chosen 
to use the information theory, that has been adopted - at least generally - 
by the International Society of Photogrammetry. This general frame would 
c analysis 
f the 
not 
deserve a special discussion. 
For the purpose of dealing with the proposed problem we shall 
consider several characteristic elements: /a/ the volume of quantitative 
information; /b/ the calibration-homogenization-correlation of data within 
amenta1 
1 aspect 
a system as much as unitary; /c/ the risk strategy; /d/ the models design 
and their finishing at every stage. 
mainly 
r as a 
note 
/6/. 
/a/ Volume of quantitative information 
Starting from the known formula (1) and from the flow chart and 
programme given in /2/: 
N R (R) NR N (R) 
zes are 
I = P E E P log P -PE E P log 2 PEP (1) 
11 N N 1 1 N IN 
^chnical 
nulti- 
)ect with 
the levels of information volume at every stage can be compared with the 
necessary information volume to separate some certain objects and phenomena. 
/b/ Calibration, homogenization and correlation of data 
These were presented in a previous paper /12/. Finally, even 
after these calibrations, the data still have some unhomogeneity. Thus