Research On Infrared Target Detection Combined With Radiation Characteristic Analysis

Due to its exceptional benifits in the fields of target detection,precision guidance,space-based early warning,etc,infrared technology has attracted constant interest on a global scale in recent years.With the gradual advancement of infrared target detection algorithms,certain techniques based on compressed sensing and deep neural networks have demonstrated excellent detection efficiencies for typical infrared targets.Currently,the urgent issues in the field of infrared target detection include the following: First,the difficulty in missing detection of infrared targets when they are submerged in a bright radiation background;Second,the challenge of precisely recovering the original radiation form of infrared targets;Third,the restrictions of the target’s actual physical model are typically absent from most common daimage-based detection techniques.In order to present research methodologies and solutions to the aforementioned problems,this article uses radiation characteristics analysis as a methodology,combining theoretical analysis,data simulation,collaborative modeling,target detection,and other practical work to provide research methods and solutions to the above issues.The research contents of this thesis are as follows.1.This thesis investigated basic thermal radiation and infrared physics theory.This article mainly studies theories related to target radiation characteristics,including infrared target intrinsic radiation theory,fundamental infrared radiation laws,fundamental infrared radiation metrics,sensor models,and radiation transfer models in typical infrared scenes.Infrared target radiation transmission models and radiation feature descriptions are given based on summarizing the relevant physical models.2.This thesis studies radiation transfer model-based scene modeling approaches for infrared imaging.Based on the atmospheric radiation theory,this article studies the numerical simulation process of infrared scenes with and without false alarm sources using satellite remote sensing data and simulation programs.Simulation experiments are used to obtain the original radiance values of the scenes in multi-bands,as well as to statistically analyze the scenes’ pertinent infrared radiation characteristics.3.In typical remote sensing scenes,jointly modeling and simulation of infrared targets and backgrounds are studied.Based on an examination of the shortcomings of conventional target fusion methodologies,this research suggests a unique targetbackground collaborative modeling method and simulation procedure combined with the radiation transmission characteristics of remote sensing scenes.The specified experimental evaluation has confirmed the viability of the simulation process,the similarity between simulation data and actual data,the detectability of simulation targets,and the difficulty of identifying simulation data.4.Two infrared target detection methods combined with radiation characteristics analysis were studied.After analyzing the radiation characteristics of the simulation data and various algorithms,this thesis proposes an infrared target detection algorithm based on multi-band intrinsic radiation characteristics analysis.This algorithm partially resolves the issue of missing detection when the target is submerged in bright and heterogeneous background.Additionally,a target detection method for radiation morphology restoration is proposed,which shows better performance in radiation morphology restoration while maintaining a high detection rate.