Paulo, 
e main 
shed of 
grating 
data on 
mental 
ANALYSIS OF ERS-1/SAR IMAGES QUALITY 
Laura Frulla 
Julio Jacobo-Berlles 
Alejandro C. Frery 
Marta Mejail 
CAERCEM-CONICET 
Julián Alvarez 1218 
Buenos Aires 1414, Argentina 
ISPRS Commission VII / Working Group 2 
ABSTRACT 
synthetic Aperture Radar (SAR) images are used increasingly in applications of remotely sensed data, for 
land use investigation and in the recognition of geological features. Despite of this fact, certain textural 
patterns should be found out easily, and it does not occur normally. In side looking radars, the incident power 
on the illuminated area is nos constant, varying as we move from near range to far range. This can result in an 
increment in the pixel values and, therefore, generating an erroneous classified scene by some pattern 
recognition algorithm based on this feature. This can be seen as a non-stationarity of the mean grey level 
value. On the other hand, a N-look SAR image is obtained by synthesizing N passes with different pointing 
angles from one scene and one set of raw data, and then by averaging the gray values of these resulting 1-look 
images, so obtaining the moltilook image. The equivalent number of looks can be determined from two 
different methods. One of them consists of calculating the variation coefficient over homogeneous and 
extended areas. The other method is an adjustment to polynomial regression models. In this paper a method 
for correcting the antenna pattern of images over different areas of Argentina is applied. This correction is 
developed before evaluating the equivalent number of looks of the ERS-1/SAR image being used, and this 
number is verified against the nominal values. 
023