一种基于相位补偿的干涉图自适应滤波算法  

An Adaptive Interferogram Filtering Algorithm Based on Phase Compensation

作　　者：谭宏伟 周芳[1,2] 贺丰收 TAN Hongwei;ZHOU Fang;HE Fengshou(School of Computer and Information,Hefei University of Technology,Hefei 230009,China;Anhui Province Key Laboratory of Industry Safety and Emergency Technology,Hefei 230601,China;Avic Lehua Electronic Technology Research Institute,Wuxi,Jiangsu 214063,China)

机构地区：[1]合肥工业大学计算机与信息学院,合肥230009 [2]工业安全与应急技术安徽省重点实验室,合肥230601 [3]中国航空工业集团公司雷华电子技术研究所,江苏无锡214063

出　　处：《遥感信息》2021年第4期55-62,共8页Remote Sensing Information

基　　金：国家自然科学基金项目(61701156);航空科学基金项目(2019200P4001)。

摘　　要：为了提高干涉图滤波算法的去噪效果并能充分保持相位细节,在传统干涉图相位补偿滤波的基础上提出一种新的基于相位补偿的自适应滤波算法。首先,采用较小窗口的坡度相位补偿滤波或多视滤波方法对干涉图预滤波,在保持相位结构的同时初步降低噪声对干涉图频谱分布的影响;其次,对预滤波干涉图进行频谱截断的Goldstein平滑处理,有效提高相位补偿滤波中地形轮廓相位的提取精度;然后,基于相位噪声标准差与相干系数以及多视视数间的函数关系,结合滤波后的噪声标准差值、滤波的最大视数以及最小视数三个参数,对残余相位进行自适应像元数目的多视滤波,从而避免低相干区域滤波不足或高相干区域滤波过度。对仿真和实测干涉图进行滤波实验。结果表明,相比于对比算法,该算法能更加有效地去除噪声并保持相位的细节信息,滤波效果最好。For improving the denoise effect and keeping the phase details of interferogram filtering algorithm,this paper proposes a new adaptive filtering algorithm based on phase compensation.Firstly,the slope phase compensation filter or the multi-look filter with small window is used to pre-filter the interferogram,so as to reduce the influence of noise on the spectrum of the interferogram while maintain the phase structure.Secondly,the spectrum truncated Goldstein filtering is performed on the pre-filtered interferogram,which can effectively improve the accuracy of terrain contour phase extraction in the phase compensation filter.Finally,based on the functional relationship between the standard deviation of the phase noise,the coherence coefficient and the multi-look number,the proposed algorithm combines three parameters which are the standard deviation of the phase noise after filtering,the maximum multi-look number and the minimum multi-look number to perform the multi-look filtering with adaptive pixel number for the residual phase,so as to avoid insufficient filtering in the low coherence region or excessive filtering in the high coherence region.The results show that compared with several comparison algorithms,the proposed algorithm can remove noise more effectively meanwhile maintain the detailed information of the phase well,and the filtering effect is the best.

关 键 词：干涉图滤波 相位补偿 噪声统计特性 相干系数 自适应滤波 

分 类 号：TN957.52[电子电信—信号与信息处理]