InSAR处理中预滤波方法的适用性研究

Study on the Applicability of Pre-filtering Methods in InSAR Processing

在InSAR处理中,频谱偏移是影响干涉图质量的主要因素之一,预滤波的作用是去除干涉像对中频谱偏移引起的去相干。首先对预滤波原理进行了介绍,然后以河北省张北地区和长江三峡地区的ERS SLC数据以及青海玉树地区和山西省临汾地区的ENVISAT ASAR SLC数据为例,分别对方位向和距离向的频谱偏移进行了预滤波实验,并通过相干性直方图和相干性平均值对影响预滤波效果的因素进行了着重分析。实验表明,在主辅图像多普勒中心频率差值较大时,方位向预滤波效果明显;时间基线太长、剧烈地表形变或植被覆盖较密引起的主辅图像去相干性较大,或空间基线较短均会导致距离向预滤波效果的明显降低。

In InSAR processing, spectral shift is one of the factors leading to a deterioration of the quality of interferogram. The role of the pre-fihering is to remove the decorrelation caused by the spectral shift of interferometric data. Firstly, the pre-fihering principle is introduced, and then the pre-filtering experiments on the spectral shift in azimuth and range direction are done with ERS SLC data of Zhangbei County, Hebei Province and of the Yangtze Three Gorges around, with ENVISAT ASAR SLC data of Yushu County, Qinghai Province and of Linfen County, Shanxi Province. In addition, by means of the coherence histograms and the average value of coherence, the factors affecting the pre-filtering effectivity are specially analyzed. The test shows that azimuth pre-filtering would be more effective if the difference of the center Doppler frequency between the master and slave is larger. Meanwhile, the range pre-fihering effect would be less effective if the decorrelation of coherence are caused by too long temporal baseline, severe surface deformation, or densely covered vegetation, or if spatial distance is too short.