Asynchronous parallel synthetic aperture radar backward projection imaging method

Abstract

The invention relates to an asynchronous parallel synthetic aperture radar backward projection imaging method. According to the method, block partition processing is performed on echo data of a host end in the azimuth; and the transmission and processing of the data are carried out simultaneously by means of the asynchronous parallel function of a CUDA flow; and therefore, the resources of a graphics card can be fully utilized. Since the block partition processing is performed on the echo data in the azimuth, large-scale data can be processed; the asynchronous parallel function is realized through the CUDA flow, so that data transmission and data processing can be carried out simultaneously, so that transmission time of the data between a host end and an equipment end can be hidden; data transmission at the host end and the equipment end adopts a function cudaHostAlloc to allocate the page-locked memory of the host end, so that the memory can always reside in a physical memory; and since a CUDA function cudaMemcpyAsync is used in a data transmission process, data transmission between the host end and the equipment end can be realized through a DMA technology, and the intervention of a CPU is not required.

Description

technical field [0001] The invention belongs to the technical field of radar, in particular to a synthetic aperture radar (Synthetic Aperture Radar, SAR) imaging technology. Background technique [0002] SAR is a high-resolution microwave imaging system with all-weather, all-weather capabilities that cannot be achieved with infrared sensing, optical imaging, and the ability to penetrate some coverings. These advantages make SAR very suitable for earth observation, and has a wide range of practical applications in military and civilian applications. [0003] The Back Projection (BP) algorithm is a SAR imaging algorithm based on time-domain coherent processing, which can be applied to SAR imaging of various motion trajectories. The basic idea is to calculate the length from each pixel in the imaging area to the aperture The distance history between the internal SAR antenna platforms, and then the corresponding time-domain echo signals are coherently accumulated to reconstruct...

AI Technical Summary

Problems solved by technology

However, limited by the size of the video memory, when the amount of data to be processed is large, the parallel back projection algorithm between pixels must process the data in blocks

Therefore, when the current inter-pixel parallel back projection algorithm processes data, the data and transmission engine of the GPU cannot run at the same time, resulting in the hardware resources of the graphics card not being fully utilized, so that when the amount of SAR data is large, the signal processing efficiency cannot be effectively improved

Images