Research On Target Detection Algorithm Under Complex Ground Background

Infrared detection system has the advantages of good maneuverability,strong anti-jamming ability and full-time work,etc.Which make Infrared detection technology become a common means in target detection field,such as infrared fire detection,infrared warning and infrared guidance,etc.However,in the detection of ground targets,the ground scene has high complexity and interference,and when the target imaging distance is far away,the contrast and signal-to-noise ratio of the target imaging are reduced,which making the detection of targets in complex ground background more difficult.Therefore,the research on target detection technology under complex ground background has very important theoretical significance and engineering application value.This article has carried on the thorough research to the target detection algorithm in the complicated ground background,mainly carried out the following work.Firstly,the infrared imaging mechanism and the mathematical model of infrared image are studied and analyzed.The infrared target and background characteristics are analyzed from the target SNR,the characteristics of the local neighborhood of the target and the target scale by the real ground infrared target image,and the common noise in the infrared image is analyzed at the same timeSecond,an Improved Adaptive Singular Value Decomposition(IASVD)algorithm is proposed.The algorithm uses the singular value matrix of two adjacent images to update the singular value matrix,so that the predicted background image contains more background information.This improves the signal-to-noise ratio of the target in the differential image.Then,aiming at moving object detection in image sequence.An Improved Spatio-Temporal Context detection(ISTC)algorithm is proposed according to the related filtering theory.The algorithm makes full use of the spatio-temporal relationship in the image sequence and the gray information in the image,provides an effective template for the context,enhances the relevance of the target and the context region,and achieves effective target detection.Finally,simulation experiments are carried out on the proposed algorithm by the simulating target sequence image and the real infrared target sequence image.The result proves that the proposed algorithm can effectively achieve the detection of infrared moving targets in complex ground background.