(InSAR矿区形变监测的边缘保持−Goldstein组合滤波方法)
摘 要: 介绍InSAR技术用于监测矿区地表形变的干涉图滤波方法，为避免对密集条纹强噪声干涉图重复进行Goldstein等方法过度滤波而导致的条纹信息丧失，提出一种边缘保持−Goldstein组合滤波方法。采用具有强条纹保持能力的边缘保持滤波方法进行滤波，根据其伪相干值及相位标准偏差改进Goldstein方法的滤波因子，从而达到更有效滤波的目的。结果表明：边缘保持−Goldstein组合滤波比单次Goldstein滤波残差点改善率提高了 23.7%，边缘保持指数高达0.74，效果优于二次Goldstein滤波效果。此法结合频域和空域滤波的优点，对于高密度条纹强噪声干涉图不仅具有较强的噪声控制能力，而且具有很强的条纹细节保持能力。
关键字: INSAR；矿区；干涉图；伪相干值；边缘保持；Goldstein滤波
（1. School of Earth Science and Info-physics , Central South University, Changsha 410083, China;
2. Key Laboratory of Precise Engineering Surveying and Deformation Hazard Monitoring of Hunan Province,
Changsha 410083, China;
3. Institute of Investigation and Surveying of Jiangmen City, Jiangmen 529040, China）
Abstract:The filters for InSAR interferograms used in monitoring mining area deformation were introduced. The edge-preserved Goldstein filter was presented in order to avoid the loss of fringe detail of InSAR interferograms with strong noise and dense fringes caused by over filtering when processed repeatedly by traditional methods such as Goldstein filter. The interferogram was filtered by edge-preservation filter with strong edge-preserving performance. The result was further smoothed by the improved Goldstein filer, whose filtering parameter was determined by pseudo-coherence and phase standard deviation derived from the result. The experiment results demonstrate that the new filter is 23.7% higher than that of once Goldstein filter in noise reduction while its edge preservation is 0.74, superior to twice Goldstein filter. This method performs well not only in noise reduction but also in detail preservation when applied to InSAR interferogram with strong noise and dense fringes.
Key words: InSAR; mining area; interferogram; pseudo-coherence; edge-preservation; Goldstein filter
