基于MapReduce的合成孔径雷达后向投影快速成像方法

A Fast Imaging Method by Synthetic Aperture Radar Back Projection Based on MapReduce

针对雷达大场景高分辨率高精度快速成像的应用需求,提出一种基于MapReduce的合成孔径雷达后向投影快速成像方法,将方位向成像任务划分成若干个成像单元,进行分布式并行化方位向成像计算,最后将所有成像单元的计算结果进行相参累加。该方法对每个脉冲数据标上天线阵元的位置信息,使得各个脉冲数据可以并行补偿相位;采用相等脉冲数划分成一个数据块的方式提高计算效率和实现负载均衡;设置Combiner函数对成像单元内的计算结果进行提前聚合,解决后期聚合时间较长的问题。实验验证了该方法的有效性,在保证成像准确的前提下,该方法的方位向成像在4台物理计算机搭建的分布式计算平台中进行,其计算速度是单机计算的后向投影方位向成像方法的3.7倍,可见该方法可以实现合成孔径雷达大场景高分辨率高精度快速成像。

Aiming at the application requirements of fast imaging of large radar scenes, high-resolution and high-precision, a fast imaging method of synthetic aperture radar back projection based on MapReduce was proposed. The azimuth imaging task was divided into several imaging units for distributed parallel computing, and finally reduced the results of all the imaging units by coherent accumulation. The method marked each pulse data with the position information of the antenna array element, so that each pulse data could be compensated the phase in paralleling. The method of dividing the equal number pulses into a data block improved the calculation efficiency and achieved load balancing. The Combiner function was set up to reduce the calculated results in the imaging unit in advance, solving the time-consuming problem in the late stage. The effectiveness of the proposed method is verified by experiments. Under the premise of ensuring accurate imaging, the azimuth imaging of the proposed method is carried out in a distributed computing platform built by 4 physical computers, and its calculation speed is 3.7 times that of the backward projection azimuth imaging method calculated by a single machine. It can be seen that the proposed method can achieve fast imaging of large scenes, high-resolution and high-precision of synthetic aperture radar.