Research And Optimization Of Collision Detection Based On Image Space

Virtual power plant leverage Virtual Reality(VR)technology to simulate threedimensional power plant scenes in proportions and formulate character models moving freely in VR power plant to roam and browse scenes.In order to enhance the sense of roaming immersion,real-time and accurate collision detection must be performed on the character model and the power plant model.After understanding and analyzing various collision detection algorithms,this article uses character models and power plant models as the research objects to study image space collision detection algorithms.The main research contents are as follows:First research the collision detection processing flow,and divide the image space collision detection algorithm into three stages: preprocessing,preliminary collision detection and precise collision detection,to achieve efficient and accurate collision detection.Secondly in the preprocessing stage,the model format is studied,the suitable model format for collision detection is selected,and the power plant model is simplified to reduce collision detection calculations.In the preliminary collision detection stage,this thesis conduct research on parallel computing,and use Open CL to implement a cross-platform character model OBB(Oriented Bounding Box)bounding box and the power plant model AABB(Axis-aligned Bounding Box)hierarchical bounding box parallel computing algorithm,which improves the efficiency of collision detection.And pass the preliminary collision model to the precise collision detection stage for further precise detection.In the precise collision detection stage,this study will obtain preliminary collision model data,and use the depth buffer,template buffer,and color buffer information retained in the image space when the model is drawn to realize the precise collision detection algorithm.VAO(Vertex Arrary Object)VBO(Vertex Buffer Object)processing method will be applied to reduce the time consumption of data transmission.FBO(Frame Buffer Object)frame buffer technology will be acquired to achieve off-screen rendering of model drawing and accurately detect collision parts while stabilizing the frame rate of model drawing.Finally,the image space collision detection algorithm will be applied and tested.Experimental tests show that the image space collision detection algorithm for the character model and the power plant model can speed up the collision detection efficiency while improving the collision detection accuracy,which has practical applicability.