Research on SAR Image Processing Performance Based on OpenMP and CUDA Parallel Model

In the field of remote sensing image processing, Synthetic Aperture Radar(SAR) image has a large amount of data and takes a long time to process. Especially in the military field and disaster detection, conventional serial processing methods cannot meet the requirements of real-time performance. To solve this problem, the high-performance computing advantages of parallel processing are utilized. The optimization effect of SAR image processing performance was studied based on OpenMP shared memory parallel processing method and CUDA parallel model. Combined with the use of OpenCV computer vision library, basic image processing tasks and parallel processing framework were designed. According to the comparison of the execution time between parallel processing and serial processing, the SAR image preprocessing speed is increased by 2 to 16 times. It is concluded that the parallel computing model is an effective fast SAR image processing optimization scheme.