Research On Real Time Dynamic Collision Detection Algorithm In 3D Scene

With the rapid development of modern computer science and technology,the virtual simulation technology in the field of 3D graphics has become more and more mature.However,the problem of real-time collision of objects in 3D scenes is the biggest problem that hinders its development speed.Because the accuracy of real-time collision detection in 3D scenes has important influence on improving the authenticity of virtual simulation environment and enhancing the immersive sense of virtual simulation environment.At the same time,due to the complexity of3 D scene object model,collision detection between objects has higher requirements.Therefore,a fast and accurate 3D collision detection algorithm is very important for 3D simulation and related research.Aiming at the above problems,this thesis proposes an adaptive real-time dynamic collision detection algorithm based on hybrid Cartesian grid multibounding body hierarchy.In this algorithm,regular Cartesian grid is combined with hybrid multi-bounding boxes.By using three-dimensional bounding bodies for simple collision detection,the three-dimensional objects that do not collide can be eliminated,so as to improve the speed.Then,self-collision and rupture can be automatically processed in the rulebased Cartesian grid.The main contents and research conclusions are as follows:First,this topic is mainly based on 3D model(solid,liquid,rigid body and flexible body)as the research object.In order to verify that the adaptive real-time dynamic collision detection algorithm of mixed Cartesian mesh multi-bounding body hierarchy is suitable for the collision detection of solid-fluid,rigid-rigid body,flexible body-solid and other objects,the corresponding 3D simulation scenes were created and the 3D models were established.Then,according to the characteristics of the 3D model,the realistic rendering of 3D model deformation is realized by reasonable use of specific illumination algorithm.Second,according to the different geometric characteristics and motion of the model,appropriate bounding boxes are created,such as AABB bounding boxes,OBB bounding boxes and SPHERE bounding boxes.The simple collision detection is carried out by using the hybrid multi-bounding box,and then the MLS-MPM algorithm is used to realize the self-collision and fracture of the object.The experimental results show that the collision detection algorithm proposed in this thesis can achieve the collision effect between solid-fluid,rigid body and flexible body respectively,which verifies that the collision detection method proposed in this paper has high accuracy and computational efficiency.