Research On Data Processing And Information Extraction Of Multi-aspect Polarimetric Sar

Multi-aspect Polarization Synthetic Aperture Radar(MA-PolSAR)can not only provide scattering information for different azimuth targets day and night,but also provide polariza?tion scattering information under different polarization modes in the case of clouds,fog,and rain.So,the target's information become more timely and more comprehensive.There-fore,MA-PolSAR is one of the frontiers in the field of SAR.It is of great significance for applications such as object recognition and feature extraction.Since the related exploration of MA-PolSARis immature,the design of the sensor,imaging methods,and the applications of the back-end need to be further researched.This article aims at the problem of SAR radiative equalization in MA-PolSAR appli-cations,the problem of multi-aspect polarization information changes in the target,and the characterization of the characteristics of MA-PolSAR features.Three studies have been conducted on the data using the four-direction full-polarization SAR image acquired by the Beijing Institute of Radiometrics-designed airborne C-SAR in Baodi Experimental Area in Tianjin.(1)The radiation differences in SAR image result from the side-view observation mode and antenna pattern.It seriously affects the quality and application of the SAR image.Using Experiment Mode Decomposition(EMD)technology with adaptive,self-driven,multi-scale and other advantages,proposed the method which based on SAR image of adaptively EMD to balance the radiation differences.Through the simulation data of known radiation differences and the equalization processing of airborne SAR image data of different groups from different polarization modes,experiments have proved that this method can effectively and adaptively balance the radiation differences of SAR images and extract radiation difference curves.(2)The polarization decomposition of the target is the essence of PolSAR feature ex-traction.Different polarization directions will be obtained for different incident directions.In this paper,the polarisation target decomposition technique is used to extract the polar-ization characteristics of different aspects of typical objects such as water,crops and trees in the study area,including scattering type and scattering mechanism characteristics;Then we analyzed the difference of polarization characteristics of different ground objects under different aspects,and formed a polarization characteristic matrix,which provides abundant ground object scattering information for MA-PolSAR data analysis.(3)Due to the discontinuity of spatial sampling of MA-PolSAR,the angular spectrum of the different feature cannot be obtained directly.Based on the time series idea and the time series model,an Auto-regressive Integrated Moving Average(ARIMA)model for multi-angle sequences of waters and crops was established for the data samples obtained by sub-aperture decomposition of MA-PolSAR,and the prediction accuracy of the model exceeds 90%.Therefore,the ARIMA model of the angle sequence can obtain the angle spectrum of different polarization features of different targets,to predict the polarization characteristic parameters under different angles.