Change Detection In PolSAR Images Based On HLT Difference Image And Triple Markov Field

The Polarimetric Synthetic Aperture Radar(Pol SAR)system benefits from the mechanism of alternately receiving and transmitting signals of different polarization modes.It has allweather and all-weather observation capabilities.Compared to traditional SAR systems,it can obtain richer feature information,which is more suitable for environmental monitoring in complex scenes.Therefore,the change detection of Pol SAR images has become an important research branch of SAR remote sensing image analysis and understanding.The change detection of the Pol SAR image refers to the technology of processing and analyzing the image data obtained in the same area at different times,and obtaining the feature change information according to the difference characteristics between the data.Pol SAR systems also introduce coherent speckle noise during imaging,reducing the credibility of the change detection results.Therefore,the suppression of coherent speckle noise while maintaining accurate positioning of change detection is the focus and difficulty of research in this field.In this paper,the main research and innovation work are as follows:In Chapter 3,for the accurate construction of difference images and the positioning accuracy of change detection,a change detection algorithm for polarization SAR images(as HLTTMF algorithm for short)based on HLT difference images and triple Markov fields(TMF)is proposed.The algorithm first introduces the Hotelling-Lawley Trace(HLT)statistic operator,and uses the bilaterally-improved HLT operator to accurately construct the difference image of two-phase Pol SAR image change detection.Then on the HLT difference image,the TMF model is used for modeling and analysis.The triple Markov field model introduces an auxiliary field U to model heterogeneous regions in the image,which has better spatial correlation description and noise suppression capabilities.Therefore,the accurate location of the change detection can be more effectively performed on the HLT difference image.The effectiveness and feasibility of the algorithm in this chapter are proved by analyzing the detection results of the four sets of measured polarization SAR image data.In the fourth chapter,for the problem of noise suppression in the process of polarization SAR image change detection,a Pol SAR image change detection algorithm based on HLT difference image and wavelet domain TMF(HLT-WTMF algorithm for short)is proposed. After the HLT difference image is obtained,the algorithm uses stationary wavelet transform to perform 4-layer wavelet decomposition on the HLT difference image to calculate the lowfrequency components corresponding to each scale decomposition.For the low-frequency components of each layer,the TMF algorithm is used to perform change detection,obtain four low-frequency component change detection results of each scale.Small-scale decomposition noise is larger but change detection edge positioning is high,but large-scale is the opposite;Finally,the multi-scale decomposition of low-frequency component change detection result map uses multi-scale fusion strategy for noise suppression to obtain the final change detection result.The simulation experiments and analysis on the four sets of measured Pol SAR image data prove that the algorithm has good noise suppression ability,and the change detection results obtained are accurate.