自适应伪相干值改进的InSAR干涉图Goldstein滤波算法

Improvement of Goldstein Filtering Algorithm for In SAR Interferogram Based on Adaptive Pseudo-correlation

In SAR技术作为当前反演DEM及变形监测的重要手段之一,其干涉图的滤波效果将直接影响着最终获取结果的精度。在In SAR干涉相位图滤波算法中,Goldstein滤波是一种最经典的频率域滤波算法。针对其滤波参数选取的不确定性及全局不合理性问题,提出了一种自适应改进的Goldstein干涉图滤波算法。该算法将自适应伪相干值作为滤波参数,使其能够进一步凸显干涉图的质量,对于干涉图质量较好的部分进一步减弱滤波程度以更好地保留相位信息,对于干涉图质量较差的部分进一步增强滤波程度以有效抑制相位噪声。采用Terra SAR-X/TanDEM-X数据对改进滤波算法进行了定性及定量分析,结果表明:改进滤波算法在噪声抑制及相位信息保留方面相对于Goldstein算法及其改进算法具有明显的优势;此外,枝切树相位解缠及DEM反演分析也进一步验证了改进滤波算法的可靠性。

Interferometric synthetic aperture radar（In SAR） technique is one of the important methods for inversion of digital elevation model（DEM） and deformation monitoring,the filtering effect of In SAR interferogram will directly affect the accuracy of the final result.Goldstein filtering algorithm represents the most classical frequency domain filtering algorithm for In SAR interferogram.Considering the uncertainty and global irrationality of the filtering parameter selection,a new improvement Goldstein filtering algorithm based on adaptive pseudo-correlation is proposed.According to the basic principal of the improved algorithm,the adaptive pseudo-correlation is taken as the filtering parameter,so as to highlight the quality of interferogram.The filtering strength of the interferogram with high quality can be reduced,so as to preserve the phase information effectively,meanwhile,the filtering strength of the interferogram with low quality can be improved to inhibit the phase noise.The qualitative analysis and quantitative analysis of the performance of the improved algorithm are done based on Terra SAR-X/Tan DEM-X data.The study results show that the performance of noise suppression and phase fringe detail information preserving of the improved algorithm is superior the one of Goldstein filtering algorithm and its improved algorithms,beside that,the reliability of the improved algorithm is further demonstrated by the analysis of branch-cut phase unwrapping and DEM inversion.