Research On Target Detection Of High Resolution PolSAR Images Based On Psychology Cognition

The Polarimetric Synthetic Aperture Radar(PolSAR)is widely used in military and civilian research because of its unmatched advantage,which can all-weather and all-weather work.With the development of imaging technology,resolution of PolSAR continues to improve.However,the existing methods often do not fully explore the rich information of PolSAR images in high-resolution PolSAR image interpretation,including the classification,detection and identification,restricting the application of the promotion of high-resolution PolSAR system.In this thesis,a high-resolution PolSAR image target detection method based on the professional cognitive process is studied in view of the problems in the current high-resolution PolSAR image interpretation.Psychological cognition refers to the process of information processing in order to form the final judgment of the external things of human beings acting on the sensory organs of human beings.Through the research,it is found that the professional interpreters usually carry out the image interpretation from the whole to the local,and make certain reasoning and fuzzy judgment according to the rich knowledge and experience of geography.Finally,they make fusion decision to complete the interpretation process.Based on the imaging characteristics of PolSAR images and the characteristics of psychological cognition process of PolSAR images,a high-resolution PolSAR image "logical-psychological" cognitive model(LPCM)is built with knowledge and experience full participation to fully tap the rich information of high-resolution PolSAR images in order to achieve efficient intelligent PolSAR image interpretation and typical object detection.In this thesis,we first systematically introduce the domestic and foreign research status of psychological cognition and PolSAR image processing.We also systematically expatiate and explain the basic theory and model of the application of psychological cognition in the field of image and the extraction and target interpretation of PolSAR image.Based on the existing theoretical knowledge and model,a high-resolution PolSAR image "logical-psychological" cognitive model(LPCM)is built to simulate the image interpretation process of the professional interpreters.Moreover,the corresponding mathematical construction and algorithm implementation are realized by using the theory of context,semantic features and knowledge base,using fuzzy logic,neural network and probability analysis.Finally,the experimental results obtained from high-resolution PolSAR images of two large scenes are used to validate and quantify the results.The analysis shows that the proposed algorithm has a good image interpretation effect and has a certain universality,which has important application value and significance for PolSAR image interpretation research.