SAR image compression based on image decomposing and targets extracting

With the bandwidth restriction in airborne synthetic aperture radar (SAR)-based human-in-the-loop applications, the acquired SAR images should be compressed with loss to overcome the conflict of image quantity and the response time. In this paper we described a framework of SAR image compression. The SAR image is decomposed into two components, namely structural and textural components. The target region mask is used to retain the important target information by allocating more bits during compression, while less bits is allocated for the background region. The obtained result showed that the compressed image using the proposed algorithm has better visual effect under the same bit rate compared with JPEG2000 algorithm.