基于照度的SAR图像均衡算法研究  

Research on SAR Image Equalization Algorithm Based on Illumination

作　　者：梁福来[1] 宋千[1] 张汉华[1] 周智敏[1] 

机构地区：[1]国防科技大学电子科学与工程学院,长沙410073

出　　处：《电子与信息学报》2012年第10期2458-2465,共8页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(60972121)资助课题

摘　　要：受气流和飞行稳定性等的影响,轻小型无人机常常存在剧烈的位置和姿态抖动。姿态抖动能够导致机载合成孔径雷达(SAR)系统对地面的照射能量不均匀,引起图像增益不均衡。为剔除成像模型对图像增益分布的影响,还原目标散射特征,该文提出"照度"的概念以表征单位面积上雷达的照射能量,同时给出了照度的计算公式并推导得到成像幅度可视为目标散射强度与照度的乘积。针对传统图像均衡方法将照度和目标散射强度混淆,使得均衡结果无法反映真实目标散射特征的问题,提出一种新的基于照度的图像均衡方法。该方法在后向投影(BP)累加过程中计算成像区域中各像素点的照度,依据照度对成像结果进行加权处理。仿真和实测数据处理结果表明,该方法能够有效抑制载机位置和姿态误差引起的图像增益不均衡现象,准确还原图像真实散射强度信息。Due to wind turbulence and flight instability, the flight trajectories of drones often exhibit highly variety in position and attitude. The deviation from linear stable paths causes non-uniform illuminating energy of Synthetic Aperture Radar （SAR） system on targets, and also the fluctuant SAR image amplitude. In order to suppress the effect of imaging model and revert the scattering features of targets, in this paper the concept of ＇illumination＇ is introduced to denote the radar illumination energy per unit area. Then the expression of illumination is deduced, and shows that image amplitude can be considered as the product of illumination and scattering intensity of targets. Traditional image equalization methods confuse the difference between illumination ununiformity and amplitude of targets, thus the equalization results can not show the real scattering intensity of targets. A novel image equalization method based on illumination is presented to overcome these conditions unconsidered by previous algorithms. In this method the illumination is computed during the Back-Projeetion （BP） accumulation and the image result is weighted by the illumination. The computer simulation and real data results demonstrate the effectiveness of this approach in suppressing the image amplitude ununiformity caused by variety in position and attitude and resuming the scattering feature of targets.

关 键 词：合成孔径雷达 后向投影成像 图像均衡 照度 

分 类 号：TN957.52[电子电信—信号与信息处理]