Research On InSAR Phase Processing And Evaluation Method In Complicated Scene

Interferometric Synthetic Aperture Radar(InSAR)system obtains the surface features by extracted the phase difference from two coherent SAR complex images.The spatial target information obtained by the SAR radar remote sensing technology is extended from the two-dimensional high-resolution image to the three-dimensional high-resolution digital elevation model.InSAR promotes rapid development of the SAR field in the new branch.As the InSAR system technology becames mature gradually,the system observation object has transformed from the simple target to the complex target.In undulate terrain such as high mountains,buildings and etc of the complex scene,the phenomenons of layover,shadow and low coherence lead to the undersampling problem of the acquired interferometric phase,which arising a large number of ‘block discontinuous' area.It brings the great difficult to the traditional InSAR processing.In this paper,the interferometric phase processing and evaluation method in complex scene were taken as the research objectives,and some robust and efficient phase filtering,unwrapping and evaluation methods were proposed.The concrete contents have summarized as follows:In the second chapter,the paper introduced the basic principle and processing flow of the InSAR measurement in detail.The SAR signal model,image disjunction source and interferometric phase statistical characteristics of the InSAR measurement were deduced and analyzed later.In the third chapter,the four robust interferometric phase filtering algorithms in complex scene: weighted mean filtering algorithm,SpInPHASE filtering algorithm,WFF filtering algorithm and MPLOW filtering algorithm were introduced in this paper firstly.The advantages and disadvantages of the four algorithms were analyzed.Then,the traditional evaluation indexes of the filtering algorithm were summarized and analyzed.It was found that the existing filtering algorithm can distinguish the noise suppression performance and the phase detail perseving performance of the filtering algorithm in the complex scene barely.In this paper,two new indexes of noise suppression performance and detail loss performance were deduced based on the definition of SPD index.Finally,the simulation experiment were carried out,and the results were analyzed by combining the characteristics of different algorithms in complex scene to verify the effectiveness of the two indexes.In the fourth chapter,the paper studied the robust combined long-short baselines unwrapping technique.Firstly,the single baseline InSAR phase unwrapping technique was summarized.It was found that the single-phase phase unwrapping is highly dependent on the phase gradient information.In the case of the phase difference between adjacent pixels beyond in complex scene,the unwrapped result was not ideal.Secondly,the robust unwrapping technique based on the long-short baselines in complex scene was proposed.The technique marked the phase discontinuity regions by the continuous detection of the long-short baselines unwrapped phases,then the long baseline unwrapped phase was corrected by the short baseline unwrapped phase in each region.The paper proposed a new method based on the morphological image segmentation technique improved by the traditional marking method.Finally,based on the unwrapped phase evaluation method and the simulation in the actual complex scene,the effectiveness of the robust combined long-short baselines unwrapping technique and the improvement of the new marking method were verified.