Research On Generating Of Realistic Infrared Texture And Resampling Methods

With the development and wide application of infrared technology, infrared scene and image simulation system has a increasing used in military and civilian fields. Infrared optoelectronic devices usually require the performance of infrared imaging system for testing and evaluation, field trials can get the most realistic test data, but this method takes enormous human and financial resources, as well as many uncontrollable environmental factors, this makes the test much more difficult. Therefore, high realistic simulated infrared scene is of great significance.According to the infrared simulation technology, a realistic infrared texture and the way of scene sampling will influence the reliability of the scene.In order to generate high fidelityinfrared textures and choose a realistic rendering samples method, the work of this paper is as follows:1.According to the theory of infrared imaging, we studied the framework of infrared scene simulation based on OGRE(Oriented Graphics Rendering Engine). Considered the influence of atmospheric attenuation, atmospheric radiation and infrared imaging system on the simulation results, the infrared scene simulation framework contains infrared scene modeling, atmospheric transmission model, the effect of infrared imaging system, infrared scene real-time rendering module and an image or data display output.2.The disadvantage of mean temperature method damaging texture information based on the infrared signature prediction model was analyzed. And we discussed temperature changing phenomenon caused by emissivity difference of material in IR short wave. Then, based on the RGB (Red Green Blue) information of the visible image colour, we modulated the temperature field modulation of the texture, which increased the details of texture, and generated a high realistic infrared texture.3.Aiming at the distortion of infrared scene simulation caused by long distance sampling based on OGRE, we established a high precision sampling model, and contrast with the default nearest point sampling method and mipmap method integrated in Rendering Engine, finally, main cause of sampling distortion was found.Studied the method based on global zoom, which can effectively solve the problem of unreasonable distribution of infrared energy due to the long-range sampling.4.For the loss of far-small target when rendering, firstly,we analyzes the reasons due to rasterization, then studied the method based on local zoom, and finally verified the rationality of this method. Applying this method to infrared scene simulation, we can effectively solved the problem of far-small target losing in infrared scene rendering.