Research On Infrared Weak And Small Target Detection And Tracking Technology

The infrared imaging system receives the infrared energy radiation from the object according to the infrared photoelectricity detection equipment,.Infrared radiation has strong penetration and can observe objects that cannot be observed by human eyes.Therefore,the detection of targets in infrared images has been widely used in national defense,military,aerospace and other fields.In this paper,the frame difference method is used to detect moving targets in infrared images,and the frame difference method is improved to avoid missing detection and false detection in low SNR and complex background infrared images.In this paper,the methods of single frame detection and inter-frame detection are compared,and the three aspects of pre-processing,detection and tracking verification are studied in depth.In the sample pretreatment stage,against the background of the infrared radiation intensity received by changes that affect testing results of this problem,remove the effect of noise,is sample frame differential method is more suitable for testing;In the stage of target detection,the principal component analysis method of tensor core norm is introduced,which is used to separate the target background of infrared image,and the multi-frame detection method of frame difference is used to detect the target of infrared image.By comparing these two methods of single frame detection and inter-frame detection,and analyzing their respective detection results.In the verification of target points,based on the trajectory continuity principle of moving targets,a method of inter-frame local gray probability verification is designed to further verify the detected targets,further eliminate the noise points that cannot be eliminated in the preprocessing stage,and achieve the purpose of reducing false alarm rate.In the target tracking stage,the tracking method of space-time context is introduced,and the coordinate information of single frame target obtained from the previous principal component analysis is combined to complete the tracking of small targets.Based on the frame difference method,this paper through the three dimensions of time and space on the handling of the samples,make full use of the information on the dimensions sample sequence,the detection effect than the original algorithm on the detection rate has been significantly promoted,and in the subsequent stage effectively reduce the false alarm rate of the algorithm;The advantages and disadvantages of frame difference method and single frame detection method are illustrated by combining different sample images.