Research On The Methods Of IR Imaging System Performance Evaluation

With the development of electro-optical technology, factors that affect the performance of infrared imaging systems (IRIS) have been changed. Traditional performance characterization methods or models become insufficient for the novel IRIS since they only consider a few or even none of these factors. Hence, it is necessary to improve traditional performance evaluation methods to satisfy the performance evaluation of the novel IRIS. In this paper, the following works are made.(1) Digital simulation for IRIS performance evaluation. By techniques of 3D-engine-control and display-interrelation, the interface between the 3D scene model and the IRIS model is constructed. And the digital simulation system for IRIS performance evaluation is set up, which is all inclusive of the components involved in the signal transmission chain, including the targets, the background, the atmosphere and the IRIS. And then, IRIS performance evaluation is performed upon this simulation system.(2) Objective Triangle Orientation Discrimination (TOD) for IRIS performance evaluation. Based on the characteristics of signal transmitting and information processing of human eye/brain systems, and the mechanism of TOD method, an objective discrimination model (ODM) is constructed. Comparison experiments are held for TOD performance data of ODM and civilian observers. Experimental results show that ODM can predict the TOD performance of civilian observers well within the common used angle frequency range. Thus, ODM can be used for TOD measurements instead of civilian observers, which results in a better IRIS performance evaluation by eliminating the artificial influence.(3) Background clutter characterization for IRIS performance evaluation. Background clutter has been gradually developing into the limiting factor of IRIS targets acquisition performance. Thus, it is important to find out how and in which way clutter affects IRIS performance. Based on the procedure of signal transmitting and processing of IRIS, the nature and characteristics of background clutter are analyzed, together with the mechanism it affects IRIS performance. Then, two clutter metrics are developed for clutter quantification, which are the Target Structure Similarity Metric (CMTSSM) and the CRLB Metric. Characteristics of the two clutter metrics are analyzed. With real IR image data, experiments are held for the relation between the clutter metrics outputs and the Automatic Target Recognition (ATR) system performance. Experimental results show that the trends of the clutter level varying with the target range are in good consistency with those of the ATR false alarm probability, which indicates the proposed clutter metrics are strong predictors of the background clutter level and can be used for ATR performance evaluation.