Development Of IR Imaging System Simulator And Its Application In Performance Evaluation

With the continuous development of infrared imaging technology, the performance of an infrared imaging system(IRIS) has been greatly improved. Meanwhile, the factors that affect the performance of an IRIS have changed with the use of new technologies. Therefore, the performance evaluation method of an IRIS should change correspondingly. To overcome the disadvantages of the traditional semi-physical simulation and injection simulation equipment, a low-cost and reconfigurable IRIS simulator, which can simulate the realistic physical process of infrared imaging, is proposed to test and evaluate the performance of the IRIS. The main content of this research is summarized as follows.(1) Based on the detailed analysis of the infrared imaging procedure, the theoretical simulation framework of an IRIS is constructed by using the full-chain simulating method. And the physical effects affecting the imaging quality are modeled accurately for the IRIS.(2) According to the theoretical simulation framework and the theoretical models of the IRIS, the architecture of the IRIS simulator is constructed. The 3D scenes are generated and the infrared atmospheric transmission effects are simulated using Object-Oriented Graphics Rendering Engine(OGRE) technology in real-time on the computer. The physical effects of the IRIS are classified as the signal response characteristic, modulation transfer characteristic and noise characteristic, and they are simulated on the single-board signal processing platform based on the core processor Field Programmable Gate Array(FPGA) in real-time using high-speed parallel computation method. The eight-channel low-noise digital-to-analog conversion circuits are designed to simulate the analog output characteristics of the IR detector for testing and evaluating the performance of the signal processor of the IR detector. In addition, a high pixel-quantization image display module is designed to present more grayscale of the infrared image with high pixel quantization accuracy, which is suitable to the performance evaluation of the IRIS. Finally, the simulated image of the IRIS simulator and the actual image of a thermal imager are compared using the image similarity metrics to verify the confidence of the IRIS simulator. The test results show that the designed IRIS simulator is reasonable and effective. Hence, it can be used for the performance testing and evaluating of the IRIS.(3) The human visual perception characteristic test platform is constructed based on the developed IRIS simulator to test the human visual contrast threshold using the non-periodic triangle patterns. The bias-free psychophysical measurement procedure and 4AFC observer task are used in the experiment. And the discrimination results of 20 observers are obtained for the triangle patterns with seven different spatial frequencies in three different background luminance levels. Then, the experimental data is fitted using Weibull function, and the curves of the human visual contrast threshold for triangle pattern is obtained. Based on the detailed analysis of the test data, the Triangle Orientation Discriminate Contrast Threshold Function(TODCTF) fitting model is proposed to characterize the human visual contrast threshold for triangle pattern. Since the characteristic of the triangle pattern is similar to the real nonperiodic target, the performance model of IRIS based on TODCTF show better agreement with the IRIS viewing characteristic of the real target.(4) The performance evaluation model of IRIS based on TODCTF is established by using the modeling methodology of the NVTherm IP model. The testing experiment for evaluating the operating range of an IRIS is carried out using the simulated images output by the IRIS simulator. The predicting results of the proposed model are compared with the experimental results and the predicting results of the NVTherm IP model. The results show that the performance model of an IRIS based on TODCTF can predict the performance of IRIS correctly, which indicates the proposed model is reasonable and effective.