Figure 2: Sub-images of JERS-1 SAR image pair: SAR-1 (left) and SAR-2 (right). 
2. GEOMETRIC MODELLING 
2.1. Geometric Accuracy Assessment 
Parametric mapping models have been determined for 
the SAR as well as for the optical image data. A-priori 
parameter information was obtained from given and/or 
known tracking parameters of the sensor. These a-priori 
mapping models were further optimised by means of least 
squares parameter adjustment techniques being based 
on ground control point (GCP) measurements. The 
accuracy of the achieved mapping models can be 
assessed by statistical parameters of the adjustment 
residuals like RMS, minimum or maximum values. 
These statistical parameters are summarised in Table 1 
for the optical and for the SAR image data, respectively, 
showing both residuals in along and across track (x/y) as 
well as nominally in East and North (E/N). Besides, the 
length of the residuals, denoted by |xy| and |EM, is given. 
It can be seen from Table 1, that for the optical data an 
RMS-accuracy of less than 1 pixel in along and across 
track and about 1 pixel in length is achievable, the latter 
being equivalent to some 20 meters on ground with 
reference to the pixel resolution of the images. 
  
  
  
  
  
  
  
  
  
OPS-2 |RMS; 08. .09 12. 171; 135 218 
MING 18-210 01] -386 265 14 
MAX: 15 1.9) 2.3 353: 264: 4141 
SAR-1./ RMS! 16. 171 23| 1941 200! 277 
MIN : -3.9 -3.8. 0.4| -38.5: -397: 50 
MAX: 30 35 44| 468 472 528 
SAR-2 |RMS; 16; 16) 22| 189; 195) 272 
MIN: 38 31: 0.1] 404 447 09 
  
  
  
  
37.00 479: 51.4 
  
Table 1: Statistics on set up accuracy for the JERS-1 
image data, given in image pixels as well as nominally in 
meters on ground. 
672 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996