The Research Of Dynamic Infrared Scene Simulation System Based On Augment Reality Technology

With the application of infrared detection and early warning system more and more widely,aiming at different needs of research and development,the system requires a lot of performance evaluation.Due to time,human and financial costs lead to lack of outfield test data,but infrared scene simulation system can provide full-scale experimental data,greatly reducing the development cycle,so it has been vigorously developed.This paper base on augmented reality technology,3D modeling of significant scenes in real infrared scene is carried out by binocular infrared stereo vision system.According to the principle of infrared imaging and computer graphics,the virtual target simulation model is established,and the infrared scene rendering engine is developed to complete seamless integration between the virtual target and true scene,improve the authenticity of the simulation image based on real-time simulation system.The main research work in this paper is as follows:1.Studies the theory and technology of binocular stereoscopic vision,starting from the theory of stereo matching algorithm,on the basis of the mainstream matching algorithm in the field of visible light,aiming at the problem that the texture of the infrared image is not obvious and the edge is blurred,a stereo matching algorithm based on pyramid + graph segmentation is designed.Fusion of Census and TAD multi-feature calculates the initial matching cost,on the multi-scale image obtained by the Gaussian pyramid,using the graph segmentation algorithm to weighted aggregation matching cost,completes the calculation of dense disparity for infrared scene,according to the disparity consistency of the divided blocks,the disparity plan parameters are calculated,It is estimated and optimized that the disparity of the misplaced point due to occlusion,weak texture,etc.Through the evaluation of indoor and outdoor infrared scene,the algorithm designed in this paper can effectively obtain the scene disparity map,the scenery profile is more obvious,the mismatched region is less,the algorithm real-time is better.2.The geometric model,the motion model and the radiation model of the infrared virtual target are studied.According to the idea of augmented reality,developes infrared scene simulation real-time rendering engine,use the stereo bounding box method to remove invisible faces,quickly achieve geometric consistency rendering,achieves infrared radiation consistent rendering combined with radiation calibration,the algorithm of rasterized rendering of triangular facets is completed.According to the angular resolution of the detector,set the fine sampling plane,multiple sampling pixel gray value,based on the scene depth information,complete the mutual occlusion processing between the virtual target and the real scene,generate virtual reality merge image of the infrared scene simulation.Experimental results show,virtual target and real scenes can be seamlessly smooth fusion,the sense of reality is higher.3.The experimental prototype of dynamic infrared scene simulation system based on Augmented Reality is designed and developed.The hardware platform of shortwave infrared binocular stereoscopic vision system is built,and complete high-speed acquisition,pre-processing of binocular images,developes stereo vision software,realtime optimize software algorithm.Use stereoscopic vision to obtain the depth of the significant scene in the infrared scene.Using the infrared scene real-time rendering engine,complete the virtual reality merge and deal with Occlusion in the simulation image.Tests show that the experimental prototype can realize function of 3D modeling of the infrared scene and infrared scene simulation,output real-time dynamic infrared scene simulation images,the frame rate of system is up to 15 frames per second.System versatility is better,and system can be flexible for functional expansion and application promotion.