Acceleration of synthetic aperture radar imaging via subaperture chirp-scaling approach based on heterogeneous graphics-processing-unit–central-processing-unit architecture

Abstract

For large-scale synthetic aperture radar (SAR) data, it is difficult to achieve SAR imagery in a timely manner using traditional imaging algorithms. Subaperture architecture is introduced to a chirp-scaling algorithm to process large-scale SAR data. By means of the subaperture partition, the real data can be processed in parallel based on heterogeneous graphics processing unit–central processing unit architecture. The principle of subaperture partition is not only discussed, but also the closed form of the overlap rate for a subaperture combination under the mode is first theoretically presented, which provides a foundation for airborne real-time data processing. Finally, simulation and real data experiments validate the proposed approach.