Research On The Methods Of Maritime Ship Detection Based On Spaceborne Copmact Polorimetric SAR Imagery

Compact polarimetric Synthetic aperture radar(CP SAR)is a new kind of polarimetric SAR system.The system has large potential in application and development for that it can maintain good full polarimetric information while reducing the complexity of the system design.Compared with full polarization,CP SAR has a simple system,a wider width of image and inherent advantages in ocean surveillance.Detection of sea ship targets based on spaceborne CP SAR is a research point in polarimetric SAR sea remote sensing nowadays.Therefore,it has important theoretical significance and wide practical value in studying on the CP SAR image ship detection technology.Aiming at CP SAR image ship detection,the paper focus on the study of the reconstruction of the fully polarimetric SAR data,ship target characteristic polarization parameters and ship target detection.The main works and innovation points are as follows:1.Study and accomplish two typical full polarimetric information reconstruction method based on CP SAR data.The reconstruction performance is analyzed based on analysis of the comparison with measured data.The three modes of scattering vector and covariance matrix in the compact polarization are analyzed in detail firstly.Then,the three modes of compact polarimetric data are constructed by Souyris reconstruction model and Nord reconstruction model respectively.From the comparisons with measured data,the reconstruction results of two reconstruction methods in thickly forested area are accurate,but in the area of ocean are not very good.2.Study the feature extraction method of CP SAR,and the different polarization feature parameters' effect on ship detection performance based on comparison experiments.The paper researches on the common polarimetric feature parameters in ship target detection,and extracts full and CP polarimetric feature parameters.Because the "sensitivity" of ship target is different for different polarimetric parameters,in this paper,the histogram analysis,normalized distance,similarity comparison,CFAR detection result were compared and analyzed towards the selected 8 kinds of polarimetric parameters.After the comparison,the co-polarized coefficient,polarimetric entropy,and polarization can result in a better detection performance.However it is hard to get the ideal detection result for a single polarimetric parameter.Thus it is necessary to make full use of the polarization characteristics for the CP SAR image and utilize various polarimetric feature parameters jointly to detect targets.3.Present a CP SAR image ship detection method based on weighted feature vector SVM.The experimental results demonstrate that the method can gain good performance through wiping out false alarms and fuzzy noise.Specifically,the method extracts polarimetric parameters in compact polarimetric data to construct weighted feature vector firstly.Then the detection of ship targets in compact polarimetric SAR images based on the weighted SVM classifier.Finally,the different scattering mechanisms of the three components after the m-? decomposition were removed with false alarm.Based on the NASA / JPL AIRSAR and Radarsat-2 polarimetric SAR data,we simulate CTLR mode abbreviated polarization data.Experimental results show that,this method can detect ship targets,and remove the false alarm and fuzzy noise effectively at the same time.