€ classified with respect to selected object or phenomenon 
u(x, y, )=4 v(x, jJ )ev, the dynamics of which we are going to investigate. See 
figure 1 as an example. 
n n The reasonable range of temporal layers used to 
Ys =0 SS =0 one comparative analysis is about 3-8, taking into 
i ! consideration that the result has to be distinguished for 
i=l 
human perception. In the case n = 4 layers it means we 
have 16 temporal states of object but for n = 5 layers it is 
n 32 states to express. 
i=l 
: fi 
= d We propose to define 
S Aseo Sg x0 
jzl j=1 x | 
i I,—a T, * a;T, | 
s y, =0 S y,=0 I, =b,T, + b,T, + b,T, 
i=l 
We decided to use this transform. I; C T, tC T, t6, T. 
where 7,,7,,7,, T,,T; are values of one pixel in 
3. COMPARATIVE POINT TRANSFORM 
each of the layer we suppose the binary images, "n ‚LG 
  
Let us assume we have temporal data set of registered and +, are values of the same pixel of the red, green 
images I, (x.y) k=12,...,n. and blue component of the result color synthesis. 
Coefficients a, ,b, and c, for n=4,5 are 
; presented in table 1. | 
Then we are able to gain temporal layers T, (x. y) zT | der 
On the 
possible 
When ( 
compos 
  
[== 
3 
  
n 
  
4 
& 
  
  
3.1 Stat 
  
The cor 
develop 
       
  
  
v4 T 
L,LBEZDEKOV 
  
year 
  
forest dynamics 
  
  
  
Tab. 2: 
Fig. 1 Five temporal layers of observed area years ir 
296 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B5. Vienna 1996