a) rms-x «R,G,B» [pix] 
0.6 
  
  
  
AN 
  
ll ( 
  
  
  
  
  
  
  
  
  
  
11 21 31 41 51 61 
(c) rms-y <R,G,B> [pix] 
  
0.6 
A blue 
  
  
Au 
Li 
  
ANS / 
  
eI 
0.2 NT SC 
  
  
  
  
  
  
  
  
  
  
11 21 31 41 Hi 61 
(e) mean.diff-x «R,G,B,mc» [pix] 
0.04 
0.02 
20:02 
-0.04 
-0.06 
-0.08 
11 21 31 41 51 61 
  
236 
(b) rms-x «mc,L Y» [pix] 
0.6 
  
  
  
  
= 
f. 
  
NS 
NOS 
& 
ZA 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
0.1 => 
11 21 31 41 51 61 
(d) rms-y «mce,L,Y» [pix] 
0.6 
0.5 
0.4 7 
0.3 7 
0.2 A 
9:1 X Rs 
1 
11 21 31 41 51 6 
(f) mean.diff-y «R,G,B,mc» [pix] 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
o.04H—A ie / 
0.02 UN >> M 
i VN =< 
-0.02 M. x ne 
-0.04 / P 
-0.06 
-0.08 
70.445 21 31 41 51 61 
Figure 2: Empirical precision of matching in the Schorndorf project. Left column: single colour channels R,G,B. 
Right column: multichannel mc, intensities I, Y. Horizontal axis: width of the quadratic matching window. (a) 
and (b) rms in x, (c) and (d) rms in y, (e) and (f) mean difference in x and y. 
IAPRS, Vol. 30, Part 5W1, ISPRS Intercommission Workshop “From Pixels to Sequences”, Zurich, March 22-24 1995