SAR干涉图滤波与相位解缠算法比较研究  

COMPARISON AMONG METHODS OF FILTERING AND PHASE UNWRAPPING FOR SAR INTERFEROGRAM

作　　者：李陶[1] 张诗玉[1] 周春霞[2] 

机构地区：[1]武汉大学卫星导航定位技术研究中心,武汉430079 [2]武汉大学测绘学院中国南极测绘研究中心,武汉430079

出　　处：《大地测量与地球动力学》2007年第1期59-64,共6页Journal of Geodesy and Geodynamics

基　　金：国家自然科学基金(40474002;40604002);武汉大学地球空间环境与大地测量教育部重点实验室基金(905152533-05-08)

摘　　要：在LP范数框架下,对经典的相位解缠算法的数学模型进行了研究,将解缠算法分为3类,并利用多种解缠方法对伊朗Bam地区的地形SAR干涉图进行了实验分析。结果表明:Goldstein滤波方法有效地减少了残点和枝切线的分布,提高了干涉图的视觉效果和信噪比。基于网络流的L1范数方法可以得到最优的全局解,但运算效率较低;L2范数方法也能得到较好的全局解,运算效率较高;L0范数方法不能得到很好的解缠结果,存在较多的断点和不确定性,但是运算速度极快。On the basis of the framework of L^P-norm, the mathematic model of classic phase unwrapping methods is studied. Phase unwrapping methods are classed into three types. The topographic SAR interferogram in Barn region of Iran is analyzed with these phase unwrapping methods. The results show that the Goldstein faltering method can improve effectively the distribution of residuals and branch cuts and raise the visual effects of the interferogram as well as the ratio of signal to noise. The L^1-norm methods based on network currents can achieve an optimal resolving, but the efficiency is rather low, the L^2-norm method can obtain better optimal resoloring and has better efficiency, L^0-norm algorithm performed best in efficiency, but with discontinuity and uncertainty.

关 键 词：雷达干涉测量 滤波 相位解缠 L^P范数 枝切法 

分 类 号：P225.1[天文地球—大地测量学与测量工程]