Research On Evaluating The Impact Of Image Enhancement Technique On Performance Of Thermal Imaging Systems

Digital image processing technique is one of the effective techniques that highperformance thermal imaging systems depend on, and to some extent it can improvethe systems’ performance, but its impact on the systems’ performance has not beenevaluated effectively. This paper studied the image based target acquisition model ofthermal imaging systems that proposed by the researchers of NVESD in the UnitedStates deeply. By drawing lessons from the structural thought of this model, the modelfor evaluating the impact of image enhancement technique on the thermal imagingsystems’ performance was proposed and practical evaluation experiments wereperformed to verify the validity of the newly proposed model. The main work studiedin this paper are as follows:(1)The current existing performance evaluation models of thermal imagingsystems at home and abroad were studied and the description methods of imageprocessing technology in these models were analyzed.(2)Starting from the basic components and the imaging principle of thermalimaging systems, the imaging effects of thermal imaging systems with their modelingand simulation methods were studied.(3)The model for evaluating the impact of image enhancement technique on theperformance of thermal imaging systems was proposed to evaluate the impact ofimage enhancement technique on the performance of thermal imaging systemsquantitatively. This model can simulate the imaging process of thermal imagingsystems and the effect of image enhancement technique. The pristine input image andthe output image of a thermal imaging system to be evaluated were filtered by thehuman eyes, then the similarity between them was determined by the SSIM(Structual SIMilarity)method. The target acquisition probability of a target observedthrough a thermal imaging system was obtained from the confusion matrix.(4)Given a thermal imaging system to be evaluated and a target set, theimplementation process of the newly proposed model was described in detail. Theresults show that this model can characterize the impacts of various imageenhancement techniques on the performance of thermal imaging systems well, whichhas a good consistency with the results of actual perception tests by human eyes.The innovation of the work described in this dissertation is that a model forevaluating the impact of image enhancement technique on the performance of thermalimaging systems was proposed. The model proposed in this paper can provide anobjective basis for image enhancement algorithm selecting in actual system design, itcan also be helpful to analyze the target acquisition performance of thermal imagingsystems with given image enhancement techniques. The accuracy of this model needsto be corrected and improved further in the follow-up work.