Reduction of SAR digital processing time using vector quantization

Abstract

A procedure is presented to directly process synthetic-aperture radar (SAR) phase history data in a vector-quantized form. The procedure is particularly valuable when SAR has previously been vector-quantized to reduce bandwidth for downlink or other purposes. The vector quantization (VQ) process necessarily degrades the image, the degradation being generally a direct function of VQ compression ratio. However, the individual quality measures are affected differently. The computation compression technique (CCT) described is faster because only addition is required, as contrasted with both multiplication and addition for the FFT. The fact that preprocessed segments are overlapped and summed lends itself to parallel processing. Parameters for using CCT are the size of each vector, whether 1-D or 2-D, the number of entries in the codebook, and the address bits. These parameters influence the amount of time savings, image degradation, and amount of required memory.<>