中文摘要：

针对虚拟手术中软组织碰撞检测的实时性问题,提出了一种基于延迟更新策略的快速检测方法。该方法应用层次包围盒（BVH）的思想并根据拓扑结构特征来表示手术环境中的对象,同时使用延迟更新的策略来降低更新的计算量。首先用不同的层次包围盒来表示手术工具和手术对象;然后利用包围盒的相交测试快速排除不相交部分;再对可能发生碰撞的部分使用更为精确的三角面片相交测试来确定碰撞信息,最后使用延迟更新的策略来完成软组织变形时层次包围盒的更新。实验表明,该方法相较于现有算法在保持检测精度的同时能够有效地提高碰撞检测的速度。

英文摘要：

To achieve realtime collision detection of soft tissue in virtual surgery, this paper proposed an efficient detection algorithm based on delay updating strategy. The main contribution of this work was to use the technology of mixed bounding volumes to represent different objects in virtual surgery accoding to their feature of topology structure, meanwhile to reduce the cost of updating computation by delay strategy. First, this algorithm represented objects in surgery environment as different bounding volume hierarchies. Then it implemented the intersection test to eliminate disjoint parts fast. After that, it used more accurate triangle collision test to determine the contact status in overlapping parts. Finally it implemented updating computation on basis of delay strategy while soft tissue deformed. Experimental results show that this algorithm attains considerable speedup in most cases with respect to the existing algorithms.