Range-Dimension Processing Development For Airborne SAR Based On Multiple DSP Processors

Real-time imaging processing is a key problem in the engineering implementation of airborne synthetic aperture radar(SAR). It include range-dimension processing(involving pulse compression and motion compensation) and azimuth(cross-range) dimension processing. A method for implementing SAR range- dimension processing using multiple DSP processors is developed according to the requirements of a airborne real-time SAR imaging radar. In this paper, the task, difficulties and the algorithms to be used from the principles of SAR are analyzed. The hardware implement of range-dimension processing is discussed in detail. A multi-processor system composed of multiple SHARC chips in which each processor is linked via LiNK port is presented. The hardware implementation block diagram of this multi-processor system is given. The software implementation range-dimension processing is discussed in detail. The key issues of hardware and software implementation, such as topological structure, communication method, buffer area partitioning, task partitioning, timing sequence assigning, data pack, are discussed. The main control flow process diagram is presented. This real-time imaging processing processor has been demonstrated through the flight trial of the airborne SAR. This airborne SAR has implemented real-time imaging with 3m X 3m resolution first in China. Its flight trial results win high praise of national ?63?project抯 experts group due to the high quality imagery it produces. This prove that the hardware and software design of range-dimension processing is correct and reliable.