SAR Data Compression with Respect to Polarimetric and Interferometric Applications

In this paper SAR raw data and SAR image data compression is addressed. For SAR raw data compression we suggest a modified FFT-BAQ. Significant advantages in performance are achieved compared to the standard BAQ and integration in the processing chain is feasible to reduce computational requirements. SAR image compression is mainly performed in wavelet domain. Depending on the application a conversion from cartesian format to polar format is performed prior to the transform. The algorithms are applied on data sets of airborne SAR instruments (E-SAR) and spaceborne systems (X-SAR). The performance is analysed in detail. It is given in terms of signal-to-quantization-noise ratio, phase error, scene dependency and image classification ability. The achievable classification accuracy is a measure for image quality, which is not common but hopefully increases the acceptance of compressed SAR data by the user community.