基于图像空间的复杂模型碰撞检测算法  

Image-space collision detection algorithm between complex models

作　　者：朱连章[1] 庄华[1] 

机构地区：[1]中国石油大学计算机与通信工程学院,山东东营257061

出　　处：《计算机工程与设计》2007年第15期3675-3677,3681,共4页Computer Engineering and Design

摘　　要：提出一种使用图形硬件用于复杂模型间的快速的碰撞检测算法。算法是基于CULLIDE的执行GPU可见性查询来减少物体模型间没有邻近特征的子集,描述了一个分类方案计算物体潜在碰撞集和碰撞自由子集,提高了裁减的性能。为了减少CPU的负载,利用GPU的可编程性,在GPU上进行精确的物体相交计算。A fast collision detection algorithm is presented between complex models using graphics hardware. The algorithm is based on CULLIDE which using GPU to eliminate a subset of geometric primitives that are not in close proximity. A novel visibility-based classification scheme is described to compute potentially-colliding and collision-free subsets of objects and primitives, which considerably improves the culling performance. For balancing the load between CPU and GPU, precise intersection is calculated on GPU.

关 键 词：碰撞检测 可编程图形单元(GPU) 通用计算 潜在碰撞集(PCS) 流计算 CG语言 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]