Study On Intelligent CFAR Detection Algorithm Based On Fuzzy Logic And Background Statistics Model

Synthetic aperture radar(SAR)has been widely applied in aerospace,earth observation,military reconnaissance and other fields.How to quickly detect interesting targets from complex scenes is of great significance for battlefield reconnaissance,military command and so on.For the problem of difficult target detection and lack of intelligence in complex scenes,this thesis studies SAR images in complex scenes,and finally proposes an algorithm based on fuzzy logic for intelligent detection of targets in SAR images.The main research contents and results are as follows:(1)The fitting of clutter distribution under different background is studied.The 6 statistical distributions are used to fit different background images in Sandia database.The fitting methods include the traditional three fitting accuracy criteria and the proposed fitting criteria based on fuzzy clustering membership value(FCC).The best statistical models of low vegetation,high vegetation,urban buildings,sea clutter and other regions are given.Finally,a statistical distribution model library of traditional fit rule and FCC criterion is established.(2)The intelligent iterative CFAR detection algorithm(SII-CFAR)and fuzzy set theory based on statistical models are studied.After analyzing,fuzzy logic is introduced into SII-CFAR detection algorithm,and SII-CFAR detection algorithm based on fuzzy logic is proposed.First consider the uncertainties in the process of fitting,with a statistical model based on the FCC criterion is replaced by the SII-CFAR detection algorithm used in traditional statistical model fitting criterion based on;secondly,according to the traditional CFAR detector with the use of "hard decisions" and lead to the problem of low utilization of information,the local CFAR detection algorithm in SII-CFAR detector replace the fuzzy CFAR detector.The experiment shows that the SII-CFAR detector based on fuzzy logic designed in this thesis is better than the SII-CFAR detector.(3)The SAR image segmentation algorithm is studied.Combined with the background statistics model and FCC criterion,a statistical model is proposed to discriminate the region of interest(ROI)algorithm.This algorithm is based on fuzzy logic and background matching for intelligent detection area.First,focus on the image segmentation algorithm based on CV model,in order to prevent the interference of small target image segmentation,this paper proposes an improved image segmentation algorithm based on CV model in segmentation before the small target detection in a preliminary test,after the initial detection of image and the original SAR image fusion.Then the hidden part of the small target in the original image,resulting in a reduced probability of interference small target image segmentation,to further improve the effect of image segmentation.Then,from the two region of the image segmentation results in the extracted ROI region,because different regions have the best statistical background model is corresponding,thus the best statistical model of a certain region can be used as a feature of the region,so each region by the best statistical model and image segmentation of ROI region determination of the background feature matching(feature matching using fuzzy cluster membership value fitting criterion based on FCC),so as to extract the possible ROI region.After extracting the ROI area,the SII-CFAR detector based on fuzzy logic is used to detect the target detection area.The experimental results show that the detection rate has been greatly improved.