一种改进的SAR干涉图综合滤波算法研究  

Research on an Improved Integrated Filtering Algorithm of SAR Interferogram

作　　者：宋瑞[1] 刘广[2] PERSKI Zbigniew 郭华东[2] 

机构地区：[1]北京航空航天大学仪器科学与光电工程学院,北京100191 [2]中国科学院对地观测与数字地球科学中心,北京100094 [3]Polish Geological Institute National Research Institute,Carpathian Branch,Poland 31560

出　　处：《国土资源遥感》2012年第1期28-35,共8页Remote Sensing for Land & Resources

基　　金：科技部中国与波兰政府间科技合作项目(编号:3417);国家自然科学基金资助项目(编号:41001264);中国科学院对地观测与数字地球科学中心数字地球科学平台重大项目(编号:DESP01-04-10)共同资助

摘　　要：SAR干涉图作为相位信息的载体,其质量直接影响对研究区域形变状况的进一步分析,采取有效的滤波算法能抑制干涉图相位噪声,提高干涉测量精度。在获得的干涉相位图中,由于矿区开采而造成的地表沉降会体现出近环状相位条纹的特征。针对这一特点,对传统的基于梯度的滤波算法做出了改进,并结合Goldstein频域滤波和改进的梯度自适应滤波,提出了一种适用于矿区沉降形成的SAR干涉相位模式滤波方法。选取河北峰峰煤矿的PALSAR干涉相位图作为实验数据,对该滤波方法做出了详细的性能评价和对比。结果表明,采用本文提出的综合滤波方法在显著降低实验区SAR干涉图相位噪声的同时,也很好地保持了相位分辨率,使由于矿区沉降而造成的形变相位环的边缘形态更加清晰。The interferogram with SAR phase information is the key factor that directly affects the further analysis of the deformation in the experimental area.Effective filter algorithm can suppress phase noise in the interferogram and improve the precision of the interferometry.In the obtained phase images,the characteristics of nearly annular stripe show the surface subsidence caused by coal mining.The authors therefore improve the original gradient-based filter algorithm firstly,and then combine Goldstein filter with the improved gradient-based adaptive filter,making it perform more effectively in SAR interference phase mode from mining subsidence.In the experiment the authors selected the PALSAR interferogram data of the Fengfeng coal mine in Hebei Province to evaluate the detailed performance of such filtering methods.Experimental results show that the edge of deformation stripe in phase image is clearer,suggesting the validity of the integrated filtering algorithm which is characterized by good denoising effect and nice preservation of phase resolution.

关 键 词：SAR干涉图 Goldstein滤波 改进的梯度自适应滤波 矿区地表沉降 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]