Research And System Development Of Infrared Small Target Detection Technology Based On Image Enhancement

Infrared small target detection is extremely important in the strategic support of guidance,early warning and border spy drone detection in the military field.However,because the distance between the infrared detector and the small infrared target is often very far away,the small target captured does not have some texture features or shape features that common large targets have.Moreover,in general,the resolution of infrared images is relatively low,that is,there is no obvious geometric feature information,and there is no obvious change in the response of the surface gray information.For these reasons,a series of subjects such as actual small target detection and small target tracking are facing huge challenges.This paper focuses on the difficult research problems of detection and real-time tracking of small infrared targets under complex backgrounds such as sea,sky,buildings and forests.The main research work and innovation content are introduced as follows.(1)For the preprocessing of the infrared image,the background clutter of the infrared image is messy,and the contrast of the detected small target relative to the background is extremely low.In this paper,by analyzing the basic characteristics of the infrared small target,the small target model and the background model are established.After experiment and analysis,the difference information between the block of interest and the surrounding area is added to the top-hat,which effectively enhances the contrast;and on this basis,a target enhancement method based on difference statistics suitable for small target enhancement is proposed.After experimental analysis,this method performs well for target enhancement and background suppression.(2)Aiming at the research of infrared small target detection,this paper proposes a scheme for infrared small and small target detection based on biologically inspired model(PLMC)to enhance infrared image.First,in the preprocessing stage,the target is enhanced based on the biologically inspired model and the background clutter is suppressed;then the improved threshold is used to filter the false target and extract the real target.Experiments on multiple complex scene data show that compared with other newer small target detection methods,the proposed scheme has better detection performance and robustness.(3)Aiming at the research of infrared small target tracking,this paper proposes a small target tracking algorithm based on super-resolution enhancement and online detection DSST.First,the original image is updated based on the super-resolution reconstruction algorithm that incorporates infrared image features to enhance the weak and small targets.In addition,the enhanced image is used as the input of the DSST algorithm based on online detection to obtain the response map and estimate the target position.Experimental results show that,compared with several latest algorithms,the algorithm performs well in terms of accuracy.(4)Aiming at the application research of infrared small target related processing algorithms,this paper develops a complete set of experimental system for infrared small target research based on Python.The system can perform related experiments on image preprocessing,small target detection and small target tracking under infrared images.First,the system framework was designed through demand analysis,and then modularized development was carried out,and the usability,detection and tracking effects of the experimental system were tested.The system is designed and developed to study small target image processing,video stream processing,format conversion,small target detection,small target tracking and log recording and other practical functions.