针对合成孔径声纳(SAS)成像过程运算量大,处理时间长,难以满足实时成像要求的问题,论文提出了采用GPU实现SAS成像算法并行处理,并对距离多普勒成像算法(RDA)进行了并行化处理方法研究。仿真实验结果表明:基于CUDA的SAS并行成像算法极大提高了计算效率,并在保持成像结果精度的同时,满足了SAS实时成像的要求。
Since the large amount of computation,the long processing time and difficulty in meeting requirements of realtime imaging exist in Synthetic Aperture Sonar(SAS)imaging process,a SAS parallel imaging algorithm based on GPU is proposed.The powerful GPU parallel computing ability is used to implement range-Doppler algorithm in parallel.The simulation results show thatSAS CUDA-based parallel imaging algorithm improves the efficiency greatly,which keeps the high accuracy of the imaging result and is suitable for real-time SAS image processing.