4.3. Accuracy Tests for Radar Imagery 
  
A comprehensive review of recently known radar imagery tests is reported in /66/. For comparison 
purposes of restitution methods a partial tabulation is contained in table 3. 
Author radar type terrain scale no. of no. of method of 
control check restitution 
points points 
  
Brit. Royal flat 159 linear con- 
Radar ESI 1:250 OOO formal transé. 
  
piecewise 
polyn. 
parametric 
(piecewise 
lin. f.) 
  
weighted 
arith. meas 
  
lin. least 
Sq. int. 
Booman Brit. Royal 
et al Radar ESI 
2m 
Konecny Westinghouse mountainous 
1972 1: 250 ooo 80 single image 
linear confor- 
mal transf. 
  
stereo opposite 
linear conformed 
transf. 
  
Stereo same 
side linear 
conf. transf. 130 
Konecny GEMS 1000 mountaineous single image 
1974 Phoenix area 1: 400 ooo 23 linear conf. 
transf. 
Clerici GEMS 1000 mountaineous 
Kubik Phoenix area 1: 400 ooo 17 33 single image 
1974 polynomial 
Deremyi 314 conformed 
1974 GEMS 1000 flat part 300 Stereo affine 
Phoenix area 1 : 400 000 318 300 polynomial 
Table 3 : Accuracy tests for radar imagery 
The accuracies achieved by the most appropriate restitution methods fall into the range of 3 
times the nominal resolution. 
Comparing accuracies of standard aerieal photography restitutions of aerial triangulation to 
1/4 the resolution, of satellite scanners to 1/2 the pixel size, of radar to 3 times the pulse 
resolution and of scanners to 2 to about 12 times the resolution with the appropriat method 
of restitution one will have to conclude, that the knowledge of platform orientation parameters 
can greatly increase the mapping potential of scanners, since the accuracy looks to be propor- 
tional to the platform height for dynamic sensors.