A High Compression Algorithm for Synthetic Aperture Sonar and Radar Images
Andreas Arnold-Bos · 2025
We present a novel compression algorithm that is suitable for coherent imaging systems such as sonar or radar images. The algorithm is tested on real data and shown to greatly outperform JPEG, JPEG2000, WebP, SPIHT for several metrics, easily reaching a hundredfold compression factor and breaking the consensus that coherent images are hard to compress. Our approach is based on a statistical representation of speckle in a wavelet pyramid, allowing aggressive compression while retaining realistic speckle.