Study of Synthetic Aperture Radar Data Compression and Encoding. Part 2. SAR Image Speckle Suppression and Data Compression Algorithms

摘要

Algorithms for reducing speckle in SAR images and the ensuing compression of these images are discussed. Speckle is suppressed by an adaptive local-variant two dimensional filter in the time domain, operating on blocks of 5 x 5 pixels. Compression is achieved by a transform coding algorithm. Monospectral treatment of SAR imagery is emphasized. Integration of the algorithms proves impossible. It is shown that data compression methods using an appropriate decorrelating transform on the optimally filtered channels are optimum for multispectral SAR-data compression. Compression methods exploiting clustered outcomes in multichannel imagery are suboptimal for compression of multispectral SAR data and have a lower performance than methods based upon transform coding, i.e., Chaturvedi algorithm.