Subaperture Real-time Imaging Algorithm Based on GPU

Abstract:
Rapid deformation measurement of geological bodies requires radar imaging in real time. The traditional CPU platform is difficult to meet the real-time requirements of high data rate deformation monitoring. In this paper, a GPU-based realtime subaperture imaging method for orbit GB-SAR is proposed. Several methods are used to optimize the key steps of imaging. Generating corresponding parameters and temporarily saving intermediate results on GPU to avoid data transmission during processing. The verification experiment based on the measured data shows that the proposed algorithm has good focusing performance and acceleration performance, and the acceleration ratio can reach 18 to 27, and the relative error of amplitude and phase is close to 0. With the increase of detection distance, the acceleration effect will further improve, and finally the acceleration ratio tends to be saturated at a certain distance.