基于最优区域划分的子块快速因子分解后向投影算法

An improved Sub-Image Fast Factorized Back Projection Algorithm Based on Optimal Regional Partition

  • 摘要: 后向投影(Back Projection,BP)算法具有精确聚焦、完美运动补偿等优点,适合于机载超宽带合成孔径雷达(Ultra Wide Band Synthetic Aperture Radar,UWB SAR)成像,但是巨大的计算量限制了它的实际应用。子块快速因子分解后向投影算法(Sub-Image Fast Factorized Back Projection, SIFFBP)算法大幅度减小了BP算法的计算量,提高了BP算法的实用性。本文通过分析SIFFBP算法区域划分的约束条件,提出了一种基于最优区域划分的改进算法,解决了传统SIFFBP算法在小波束积累角时加速性能下降的问题。相比于传统算法,当波束积累角小于90度或成像区域长宽相差较大时,改进算法在保证成像精度的同时进一步减小了计算量。仿真和实测P波段SAR数据的成像结果验证了改进算法的性能。

     

    Abstract: Back Projection (BP) algorithm is suitable for airborne Ultra Wide Band Synthetic Aperture Radar (UWB SAR) imaging for its advantages such as perfect focusing and motion compensation. However, its application will be limited by the big computation load. The Sub-Image Fast Factorized Back Projection (SIFFBP) algorithm, which substantially reduces the computational load, can improve the practicability of the BP algorithm. In this article, an improved SIFFBP algorithm based on optimal regional partition is proposed by analyzing the regional division constraints. It provides a way to solve the acceleration-decline problem of the conventional SIFFBP algorithm when the SAR system has a small integration angle. Compared with the conventional algorithm, the proposal can further reduce the computational load when the integration angle is smaller than 90 degree or when the length of the imaging region is much larger or smaller than the width, While at the same time, it retains the same imaging precision. The performance of the proposal is demonstrated by using simulated data, as well as real SAR data collected by a P band airborne SAR.

     

/

返回文章
返回