209 
geointormation structures. 
Mathematic model of composition. On the basis of the 
presented theoretical propositions a mathematic model of 
composition for analysis and differentiation of 
polymorphic-isomerical structures of geoinformation has been 
worked out. . In this model probability estimation of 
interdependency between components within its structure is 
used. Consider the structure of the model. Let each point 
(pixel) of the remotely sensed object, of be represented with 
the set X- ' L = 1 , 2, ,k> of values of intensity of 
refi ec t: ed rad iati on 'si goal s > 
each point of the 
Ject 
>et of compositions of various signals 
f 
Z J * 
k a > is realised. As a measure of relationship between values 
of intensities in composition os will use probability of 
their joint realisatierc To describe probability structure 
c.*t 
composi t i. on s. 
1 et.'s 
appi y 
family óf 
a ws o - 
: stribution of probabilities C Pj (kj)1 , j = 1 
intensities, which gives a full description 
structure of composit ions of 1 oformation field 
the composition model of the image the values of 
in composilions are fixed by means of intervals 
the type of intordependency between them. 
of i nn&r 
(image). In 
intensities 
given, and 
wh iCh i S 
d e t er m i n ed according to 
t=j 
P- (kj ) n P. (X; >, j=2, 3, ...,n(l>, 
J J L =2 L 
may change. Therefore, the difference between information 
structures of relations between the components it is 
composed of, and which is coded by the value of the type of 
interdependency between them in accordance with (1). Three 
situations are estimated which mean correspondence, 
norrcorrespondence arid independence between components, and 
are coded with signs +, 0. The composition code is 
constructed as follows. All the k of intensities are 
examined and the value of the criterion is estimated, then 
are sequentially eyainined combinations Ck - 1), (k — 2) , 
, e ,, 2 of intensities, and the values of the criterion are