Study On Infrared Image Synthesis For Typical Target Scenes

Infrared(IR) imaging guiding have been an important role for developing modern military equipment with precision guiding, because it has incredible ability at correct identifying targets and anti-jamming in bad weather. Lots of various IR images are needed during the development of weapon with IR imaging guiding. However, real images from IR camera can't satisfy all the demand in testing because we can't provide with all possible IR images in different meteorologic conditions. And taking real images needs more times, lots of manpower and material resources. IR imaging simulation is an effective and economical solution for the previous problem. It plays a key role during developing and manufacture of weapon with IR imaging guiding.This paper focuses on thermal model and infrared image synthesis for typical target scenes, including building, bridge, water and railway. We simulate the modulating process of display scenes with high dynamic range. The mechanism of air transmission and infrared imaging system is also simulated for imroveing the realistic synthesis. Differing from the traditional methods, we integrate the knowledge of heat transfer, infrared physics and computer graphics. Thermal model of simulation and realistic rendering in IR band is focused on. And modern graphic hardware (GPU) is used for accelerating the high-quality rendering process.In simulation of building scene, we establish a thermal equilibrium equation according to conservation of energy at first, which is solved using finite volume method. So the thermal model is build. Second we produce an infrared texture image from a corresponding visible image based on segment of visible image and infrared material library. Finally we propose a shadow rendering algorithm based on occlusion interval using GPU. Occlusion interval is used to pre-calculate the information of visibility. Owing to the main contribution of sun, we focus on the radiance from sun which is considered as a linear light source running at fixed orbit.In simulation of bridge and water scene, we firstly build the thermal model for bridge based on analyzing its thermal process and finite volume method. Through additive source item and local linearization processing, the linear equations with unified form is obtained. Alternating direction implicit(ADI) method is used to solve the equations, the temperature distribution of bridge is gained. Then we render the IR image of bridge according to the temperature distribution. Secondly, we put forward an improved temperature calculation model for water surface. The direction of effective emissivity and quality of heat stored (or released) by water body is considered in this model. Then the surface height field of water is computed with the JONSWAP power spectral formulation. Finally a GPU-accelerated rendering scheme based on persistent grid mapping(PGM) is designed to simulate the scene including bridge and water.In simulation of railway scene, a simple numeric model is proposed to calculate the temperature distribution of railway. We mainly focus on the emissivtiy of steel track, which is small and vary with its surface status. Then a control method based on the concept of temperature window is employed to simulate the auto gain and offset of IR imaging system. So image with high dynamic range is displayed adaptively. The attenuation and blurring effect of atmosphere is simulated by combined the pre-calculate using PcModWin with modulation transfer function(MTF). Finally the effect of optic system and detector is simulated by their MTEAt last, we draw a conclusion of this paper and propose some opportunities in future work.