一种基于权重自适应相似性度量改进的InSAR-BM3D算法  

An Improved InSAR-BM3D Algorithm Based on Weight Adaptive Similarity Measure

作　　者：李兴[1] 王志勇[1] 王子豪 吴斌[1] LI Xing;WANG Zhiyong;WANG Zihao;WU Bin(College of Geodesy and Geomatics,Shandong University of Science and Technology,Qingdao,Shandong 266590,China)

机构地区：[1]山东科技大学测绘与空间信息学院,山东青岛266590

出　　处：《遥感信息》2023年第2期71-78,共8页Remote Sensing Information

基　　金：国家自然科学基金项目(41876202)。

摘　　要：针对InSAR-BM3D算法在三维成像雷达高度计干涉图近距端纹理密集区域、远距端低信噪比区域干涉条纹细节保持及去噪能力不足的问题,提出了一种基于权重自适应相似性度量改进的InSAR-BM3D算法。以皮尔森相关系数和余弦相异性为基础,构建了一种根据干涉条纹密集程度和噪声水平变化而对皮尔森相关系数和余弦相异性进行权重分配的相似性度量,进一步保证了相似块搜索过程的准确性;在最终估计阶段,对基础估计图像中噪声残留区域进行了进一步滤波处理,提高了算法的去噪能力。利用模拟数据和机载三维成像雷达高度计影像数据进行了实验,并与均值滤波、Goldstein滤波、NL-Means滤波以及各向异性扩散滤波等算法进行了比较,结果证明该算法能够在有效滤除噪声的同时进一步保留干涉条纹的细节。In order to solve the problem of InSAR-BM3D algorithm’s insufficient effect of fringe detail preservation and denoising in the dense texture area at the near distance and low SNR area at the far distance of 3D imaging radar altimeter interferogram,an improved InSAR-BM3D algorithm based on weight adaptive similarity measure is proposed.Based on Pearson correlation coefficient and cosine dissimilarity,a similarity measure is proposed to assign the weight of Pearson correlation coefficient and cosine dissimilarity according to the density of interference fringe and the change of noise level,which further guarantees the accuracy of similar block search process.In the final estimation stage,the residual noise area in the basic estimation image is further filtered to improve the denoising ability of the algorithm.Experiments are carried out using simulated data and airborne 3D imaging radar altimeter image data.Compared with mean filtering,Goldstein filtering,NL-Means filtering and anisotropic diffusion filtering,the results show that the proposed algorithm can effectively filter noise while further preserving the details of interference fringes.

关 键 词：干涉相位去噪 InSAR-BM3D 相似性度量 三维成像雷达高度计 非局部滤波 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]