面向多基线相位解缠的改进小波聚类算法  

An Improved WaveCluster Algorithm for Multibaseline Phase Unwrapping

作　　者：王正国 袁志辉 陈立福[1,2] 邢学敏 WANG Zhengguo;YUAN Zhihui;CHEN Lifu;XING Xuemin(School of Electrical and Information Engineering,Changsha University of Science and Technology,Changsha 410114,China;Institute of Radar Remote Sensing Applications for Traffic Surveying and Mapping,Changsha 410114,China;School of Traffic and Transportation Engineering,Changsha University of Science and Technology,Changsha 410114,China)

机构地区：[1]长沙理工大学电气与信息工程学院,长沙410114 [2]交通测绘雷达遥感应用研究所,长沙410114 [3]长沙理工大学交通运输工程学院,长沙410114

出　　处：《遥感信息》2025年第1期149-155,共7页Remote Sensing Information

基　　金：湖南省自然科学基金(2024JJ5014);长沙理工大学研究生科研创新项目(CXCLY2022078)。

摘　　要：相位解缠对于地表高程反演和形变监测具有重要意义。针对传统小波聚类算法应用于多基线相位解缠时,在类边界处出现较多错误聚类及连通区域识别不准确的问题,提出了一种改进小波聚类算法。该算法在传统小波聚类算法的基础上,融入了自适应网格划分技术,并与深度优先搜索算法相结合。通过对特定密度区间的网格进行拆分后判定,可以减少仅采用等间隔网格划分在类边界上产生的错误类;利用深度优先搜索算法识别显著网格组成的连通区域,解决了固定邻域模式识别到的小类数量过多的问题。仿真数据和Tandem-X测量的真实数据实验表明,改进小波聚类算法不仅显著提高了类边界处的相位解缠精度,而且减少了小类的数量,改善了相位解缠的效率。Phase unwrapping is important for surface elevation inversion and deformation monitoring.The traditional WaveCluster algorithm encounters challenges when applied to multibaseline phase unwrapping,leading to numerous errors in cluster boundaries and inaccurate identification of connected regions.To address this issue,an improved WaveCluster algorithm is proposed.Building upon the traditional WaveCluster algorithm,this algorithm incorporates adaptive grid partitioning and integrates it with depth-first search algorithm.By splitting grids within specific density ranges,it reduces errors generated by using equidistant grid partitioning at cluster boundaries.Additionally,employing depth-first search algorithms aids in identifying connected regions formed by significant grids,thereby resolving the problem of excessive small clusters recognized by fixed-neighborhood pattern.Experiments with simulation data and real data measured by Tandem-X show that the improved WaveCluster algorithm can effectively reduce the number of small clusters,enhance the efficiency of phase unwrapping,and significantly improve the accuracy of phase unwrapping at cluster boundaries.

关 键 词：合成孔径雷达干涉测量 相位解缠 小波聚类 自适应网格划分 深度优先搜索 

分 类 号：TN959.3[电子电信—信号与信息处理]