中文摘要：

为了提高虚拟手术仿真系统中碰撞检测算法的效率，基于混合包围盒碰撞检测法，提出了一种快速的碰撞检测优化算法。首先利用S-AABB上层包围盒进行粗略碰撞检测，然后利用球包围盒进行底层精确碰撞检测，并采用2种优化方法代替传统上建立层次二叉树的过程。采用分区域碰撞检测法，使每次检测只是针对某一个区域内的单元体，有效地提高了碰撞检测效率；采用预测碰撞检测法，当碰撞连续发生时，预测出即将可能发生碰撞的单元体，只是针对这些单元体进行碰撞检测。最后通过实验数据，证明了该碰撞检测算法及其优化方法的有效性及其快速性。

英文摘要：

With the virtual surgery simulation system playing an increasing role in the medical field, a fast collision detection optimization algorithm based on the hybrid bounding box was put forward. At first, the S-AABB upper bounding box was built for rough collision detection, then the bottom bounding box was built for precise collision detection , and two optimization methods were used to take the place of the traditional process of establishing a binary tree. One is regional collision detection method. It divides the collision object into different regions, the detection takes aim at the corresponding region every time, increasing the collision detection efficiency effectively;and another is forecasting collision detection method. Firstly, it forecasts the element which will be possibly collided at the next instant when collisions occurred continuously. The effectiveness and rapidity of this collision detection and optimization method is proved by the experimental data.