Research On Filtering Methods Of InSAR Interferogram

Interferometric synthetic aperture radar (InSAR) was developed in1960's. It makes use of the phase information extracted from the SLC data of SAR to get the three dimension information and elevation change of the terrain. It applies the interferometry technology into SAR and expends the functions of SAR.One important use of InSAR is to obtain the digital elevation model (DEM) of the terrain. The quality of InSAR interferogram determines the precision of the derived DEM and it can influence the complexity of phase unwrapping and other procedures in sequence. Due to the presence of noises in the interferogram, the filtering of the interferogram is an important step in the data processing of InSAR.Good researches on the development history and fundamentals of InSAR are done in this paper. The procedures of getting an interferogram and the characteristics of textures and noises in an interferogram are also analyzed in-depth.In this paper, deep researches on filtering methods of InSAR Interferogram are done. The periodic pivoting mean filtering method and Goldstein-adaptive filtering method are programmed based on the understanding of part of the filtering methods in existence and then comparative tests are carried out with practical data.The experimental results indicate that the two methods both have good effect with the data in the tests. But the periodic pivoting mean filtering method performs badly in preserving the sharp sawtooth profile of the interferogram although it can remove most of the noises. The goldstein-adaptive filtering method has good performance in preserving the sharp sawtooth profile of the interferogram, and it does better than the periodic pivoting mean filtering method as a whole. But if we take the calculation work and complexity of the two methods into account, we'll find that if we need to process the large amount data of the whole image in the tests, the periodic pivoting mean filtering method will also be a good choice.