Parallel processors for synthetic aperture radar imaging

Synthetic aperture radar (SAR) signal processing poses a significant challenge due to its very large computation and data storage requirements. The paper compares various design approaches for parallel SAR processors capable of processing high resolution satellite SAR data at a rate of 1/10 real-time. Sample designs are presented for parallel processors based on general purpose microprocessors, general purpose digital signal processors (DSP), and special purpose DSPs. In each case, the number of processing elements required and the expected performance is predicted. A heterogeneous vector/scalar architecture composed of general and special purpose DSP devices is proposed that offers a good balance of performance and flexibility