Research On The Simulation Of Infrared Optical Properties Of Deep Space Targets

Various infrared images with different setups are needed for repeat testing and evaluating while developing algorithms of deep space target recognition and detection. However, the resourcs are limited to perform the launch trials, so there is not enough real infrared images to test the algorithms. Under this circumstance, infrared imaging simulation technology is brought up to solve this problem efficiently.In this paper, infrared characteristics of deep space targets are deeply studied. And a simulation system of infrared characteristics of deep space targets is designed and implemented. A lot of experiments are performed in this system.Based on the law of energy conservation and the law of angular momentum conservation, a new iterative method is developed to calculate the track of the target. The whole track can be calculated by iterating step-by-step from the initial point. This method is applied to calculate track and posture adjustment process. For particular track, the iterative method is compared with classic method, so the accuracy and efficiency can be proved. Thus it provides great convenience for real-time calculation.Infinite-volume method is used to calculate the temperature field of surfaces of the target. And the infrared radiation field is also analysed. An earth-sun system is created to simulate the shadow zone and sunshine zone, and the change of infrared radiation is calculated when enter one zone from the other. The IRAS Point Source Catalog is used to simulate the space objects in the universe. And due to the large number of objects, the celestial sphere is partitioned with a quadtree, so that the render efficiency and accuracy can be improved.The infrared camera is simulated with several layers. And all the effects are simulated respectively, including motion trail of target and space objects, blurring, dispersion, blind elements, and noise. Especially, the noise is simulated with random noise and the Perlin noise together, thus relatively real noise image can be produced. The GPU programming is used to perform the whole process. Compared to the real image of the infrared camera, the simulated images are evaluated with the histogram, the trend of intensity, and the SNR.