-i 
cl 
A smaller value 8 as shown in figure 3 makes the process 
more unquiet, for it gives the random variable n, more 
influence on r,- Proper values for the process parameter 8 
should be derived in advance by analysing recordings from 
practical flights (e.g. flights of the "Flugzeugmesspro- 
gramm (FMP)"). 
The time dependent model mentioned at the beginning of this 
chapter can now be formulated. The collinearity equations 
read in the general form as 
> 
tt 
0 o 9 9 
X(X,Y,Z,c,O*Au,9*A0, FtAK,X &AX. , Ye AY 2442 ) (5a) 
o o 0 o o o 
n 
o o o 
o 0 
y = y(X,Y,Z,c,0+Aw, $+A6,K+A_,X+AX ,Y4AY, ,Z*AZ,) — (5b) 
The image coordinates x and y are expressed as functions 
of the ground coordinates X,Y,Z,the calibrated focal 
length c and the exterior orientation parameters Q* Ao, 
$449, Sene Xeax 9 av Z+aZ,- The deviations Aw, A¢, Ax, 
AX,, AY, and aZ, from the nominal values 8,0,2,3,.2 2 
are modeled by second order Gauss Markov processes. 
o 
Au, = 28450, 1 - Bau, 2 + ox (6a) 
Boy = 28p88, yt BEAN Lp + GK (6b) 
Ak, > 06404 -1 ^7 Blanc, os EB, (6c) 
AX uk 7 2848X,4 «17 RÍAX 2 * PE (6d) 
ak t P kel REAY ooo + Av; k (6e) 
AL, = gal, yy 7 PES 2 * PLE (6f) 
In equations(6) index k denotes the line number of the 
remote sensing image. The process parameter 8 may be 
different for the various orientation parameters.