一种基于Non-Local Means的地基差分干涉雷达相位滤波改进方法  

An Improved Phase Filtering Method for Ground-based Differential Interferometer Radar Based on Non-Local Means

作　　者：冯丽源 邓云开 聂祥飞[1] 杨鸿 FENG Liyuan;DENG Yunkai;NIE Xiangfei;YANG Hong(College of Electronic and Information Engineering,Chongqing Three Gorges University,Chongqing Key Laboratory of Geo-environment Monitoring and Disaster Warning of Three Gorges Reservoir Area,Chongqing 404000,China;School of Information and Electronics,Beijing Institute of Technology,Beijing 100081,China)

机构地区：[1]重庆三峡学院电子与信息工程学院,三峡库区地质环境监测与灾害预警重庆市重点实验室,重庆404000 [2]北京理工大学信息与电子学院,北京100081

出　　处：《信号处理》2022年第1期100-108,共9页Journal of Signal Processing

基　　金：国家自然科学基金(61971037,61427802);重庆市自然科学基金(cstc2020jcyj-msxmX0608,cstc2020jcyj-jqX0008);重庆市教育委员会科学技术研究计划青年项目资助项目(KJQN20210122)。

摘　　要：地基差分干涉雷达相位图中,往往含有大量相位噪声,严重影响相位解缠和形变测量结果。有鉴于此,本文提出一种改进的自适应非局部均值(Non-Local Means)组合滤波算法。该算法首先利用相干系数构造出可自适应的平滑参数模型,有效改善了Non-Local Means算法在滤波参数选择上的固定性。其次,利用维纳滤波可对空变噪声进行有效滤除的优点,本文将自适应Non-Local Means算法与维纳滤波进行有效结合,更好的保持了相位连续性,改善了图像边缘模糊度。实测数据表明,相比于自适应中值滤波算法、Goldstein算法和传统NonLocal Means算法,本文算法在抑制相位噪声、保持相位连续性以及改善图像边缘模糊度方面具有明显优势。The phase pattern of ground-based differential interferometer radar usually contains a lot of phase noise,which seriously affects the phase unwrapping and deformation measurement results.In order to obtain a higher image quality,an improved Non-Local Means combined filtering algorithm is proposed in this paper.Firstly,the algorithm uses the coherence coefficient to construct an adaptive smoothing parameter model,which effectively improves the limitations of the Non-Local Means algorithm in the selection of filtering parameters.Secondly,since Wiener filtering can effectively filter the space-variant noise,this paper effectively combines adaptive Non-Local Means with Wiener filtering to better maintain the phase continuity and improve the image edge ambiguity.The measured data show that,compared with the adaptive median filtering algorithm,Goldstein algorithm and traditional Non-Local Means algorithm,the proposed algorithm has obvious advantages in suppressing phase noise,maintaining phase continuity and improving image edge ambiguity.

关 键 词：地基差分干涉雷达 相位图噪声 平滑参数模型 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]