一种基于深度纹理的碰撞检测

摘要

This paper presents a GPU-based image-space collision detection method. The depth values are rendered directly to the texture using depth texture technology of OpenGL and its extended function FBO. And then the GPGPU is employed to process the data in the textures to accomplish collision detection. Simulation experimental results show that this method takes full advantage of graphic hardware performance, and can return correct collision detection results.%提出一种基于GPU的图像空间碰撞检测方法.利用OpenGL的深度纹理技术和其扩展功能帧缓存对象(FBO)实现将深度值直接渲染至纹理,然后采用通用计算图形处理器(GPGPU)技术对纹理中的数据进行处理实现碰撞检测.仿真实验显示,该方法充分利用了图形硬件GPU的性能,并能够正确地返回碰撞检测结果.