Research On The Collision Detection Techniques In Virtual Interventional Cardiology

In recent years, with the rapid progress of computer science and technology, computers found more and more application in society and human life. Virtual Reality is a way that we used to simulate the real world in computer environment, and the Virtual Surgery is an important application of virtual reality in modern medicine. Collision Detection is one of the fundamental problems in virtual surgery, it plays an important role in improving reality and enhance immersion for users, and the complexity and real-time of virtual surgery bring new higher requirement to collision detection.Spatial Subdivision and Bounding Volume Hierarchy are two of the technology most widely used in Collision Detection. They decrease the time complexity of Collision Detection by respectively dividing the scene space or the object model and constructing the tree-like hierarchy structure. They have the same purpose to improve the efficiency of the collision detection by culling away the clearly non-intersect primitive pairs as early as possible, so they only have to do intersection test between the remains.With the background of virtual Interventional Cardiology, this paper proposes and implements a collision detection algorithm base on local scene pruning volume constructed with the moving objects in the scene. Its main idea is to obtain the local data of the scene around the moving object, to decrease the complexity of the collision detection. In acquiring the local data of the scene, we introduce several strategies to cut down the complexity of the problem, and the validity of the simplification has been proved. The algorithm takes no requirement in geometric information and topological structure of the scene, can be used to solve the collision detection among rigid objects and deformable objects.This paper also proposes another method called hybrid collision detection to solve the more usual collision detection. It utilized the spatial subdivision and bounding volume hierarchy at the same time. In the spatial subdivision process, we proposes a primitives distributing strategy based on the node of the spatial subdivision hierarchy tree, which can reduce the times of tree traversal and the number of primitives have to be distributed, and improve the speed of collision detection. Experiments showed that, the depth of the spatial subdivision process is of great importance to the efficiency of the algorithm and either too small or too big can do harm to the efficiency of hybrid collision detection.Finally, this paper implements the collision detection module for virtual Interventional Cardiology. The experiment results shows that, it can solve the collision detection problem in virtual Interventional Cardiology very well and can fulfill the system's requirement of accurate and real-time collision detection.