Edge Detection-based Adaptive Approach for Interferometric SAR Image Noise Reduction
基于边缘检测的自适应干涉SAR降噪方法

Thermal and "Speckle"noise is an obstacle to get digital elevation model(DEM) from the interferogram by 2-D unwrapping of interferogram phase. An adaptive filtering method for synthetic aperture radar(SAR)interferometric phase images is presented in this paper. Based on sixteen directional masks, which are the enlarged Prewitted edge masks, the fringe direction can be determined. Then those pixels on the same direction is filtered by average phase filtering method to generate filtered phase value. In fact, the change of interferogram phase is a gradual change. The pixel's direction can be selected from the neighbouring directions of the preceding pixels in order to avoid too many computation about the directions of the present pixel. This method is useful, especially for tightly packed fringes. The method can be carried out easily. Using interferograms generated by SIR-C/X-SAR repeat-pass, its effectiveness is tested.