Wavelet based ScanSAR Image Compression Minimizing Block Effects

Abstract

ScanSAR images are important products of modern spaceborne SAR systems. A wide swath is covered during one single data take. Usual processing of ScanSAR images with a high number of looks not only reduces speckle but also leads to data compression. However, the amount of data can be further reduced, if an appropriate ScanSAR data compression is applied. In our paper, we propose a new wavelet based compression technique, which adapts to the ac energy distribution in ScanSAR images and minimizes artifacts due to compression. Special attention is paid to reduce block effects. Our technique is tested on ScanSAR data of RADARSAT and the SIR-C space shuttle mission. Compressing the SIR-C ScanSAR scene of Chickasha/Oklahoma, USA with factor 1:8, we achieve an improvement in the overall signal-to-distorsion-noise ratio (SDNR) from 29.7 dB to 48.03 dB.