Performance evaluation of SAR raw data compression methods in presence of non ideal channel characteristics

We have investigated the effect of non ideal channel characteristics on the performance of some Synthetic Aperture Radar (SAR) raw data compression methods for satellite based SAR systems. The non ideal characteristics of satellite downlink communication channel have been modelled in terms of the varying bit error rate which depends on the constellation type and various noise levels. Block Adaptive Quantization ( BAQ), Amplitude Phase Wavelet Transform(APWT) and one dimensional Discrete Cosine Transform combined with Entropy coding(1D_DCT+E) have been tested and compared to elaborate how robust they behave in exploiting the limited satellite downlink capacity in presence of different error levels when transmitting the compressed SAR data bitstream to ground station. The results show that in the case of BAQ compression method, the lower compression gain (more compressed bits) option is more susceptible to channel induced errors than other options. APWT method shows different robustness behaviour. For this method, the lower compression gain options always perform better than those with higher compression gains. The degradation rate in APWT method is also smaller than that of the BAQ method which reflects better robustness for APWT in comparison with BAQ. The robustness property of 1D_DCT+E method is the worst among these three methods as is evident from the early degradation of SNR criteria.