Noise estimation and removal in detected SAR images

Inhalt:
The knowledge of noise, either thermal and quantization, is fundamental in SAR processing, calibration and image interpretation. In this paper, we propose an approach to estimate the noise floor, basing on the raw data, either ScanSAR or STRIPMAP. This approach is pretty general, although it is applied to ENVISAT-ASAR mission, and validated by means of the noise measures in the source packets made by the sensor. The accurate estimate of noise leads to its removal from detected SAR images, gaining contrasts, radiometric accuracy and improving quality in the mosaic, as it is shown in the results achieved from WSM and AP data.