Virtual Reality Collision Detection Based On The OSG And GPU Parallel Acceleration

Although many researchers had done much meaningful work about collision detection, but along with the computer software and hardware and network technology fast developing, especially the GPU parallel computing technology, real-time and accurate collision detection problem with the different scale scenes has became a hot issues.For collision detection problems of large scale and complicated scenes, we firstly organized a virtual campus roaming scene of Hebei university using the function characteristics of new 3d graphics rendering system after in detailed understanding the organized way and rendering process of OSG rendering engine. Then through deeply studying of the collision detection algorithm based on image, designing a improved based on image collision detection with the designated planar projection, stencil test and depth test, realizing the collision detection between roaming role model and scene models of the virtual campus roaming system. Meanwhile introducing the GPU parallel computing, improving the efficiency of collision detection.For the virtual surgery simulation which more focused on collision detection accuracy. In this paper, after researched the collision algorithm based on distance, we competed through the scanning line to compute segmentation vertexes, adaptive surface dividing, constructing dividing triangles, computing the distance between segmentation vertexes and dividing triangles to judge whether skull model and bone saw had collided. This algorithm can not only get the colliding position, but also can confine the collision regional in some dividing triangles. Also we continue to introduce GPU parallel computing technology, optimize the collision detection algorithm, and shorten the computation cycle.