Research And Application Of Rigid Body Model Collision Detection And Fluid Terrain Collision Simulation In Virtual Scene

Collision detection(CD)is an important research content in computer graphics,and it is also an indispensable part in virtual reality(VR)technology.Virtual reality is the simulation and extension of natural laws based on the exploration of natural laws in the real world and the use of computer graphics and modeling techniques.The essence of virtual reality is a three-dimensional simulation space in a computer,providing people with visual based interactive information,and making users feel as immersive as possible.Virtual reality technology is currently widely used in education,entertainment,medicine,production and other fields,and is a development trend of human-computer interaction.As a key technology part of virtual reality technology,collision detection has very important research significance.With the rapid updating and iteration of computer hardware and software in recent years,the models of three-dimensional in virtual scenes has gradually increased,and the complexity of the environment has reached a new level.This puts forward more stringent requirements for the realtime and accuracy of object collision detection in virtual scenes.In computer games,game scenes require real-time rendering,which requires high real-time collision detection requirements.With the development of the industry,players’ requirements for virtual environment simulation are gradually increasing.Therefore,the main research goal of this article is to improve the detection efficiency of collision detection algorithms in complex scenes.Currently,in computer games using threedimensional scenes,there are two common problems: simple collision detection of human-like models,and frequent model penetration;The interactive simulation process of collision between a model and a fluid terrain scene is complex,computationally intensive,and parameter adjustment cumbersome.In this paper,the following two parts of research work have been carried out to address the above two issues:1.In complex environments,in order to improve the real-time performance of virtual scenes as much as possible,the collision detection of human-like models in computer games is relatively simple,and the collision detection effect with rigid body obstacle models in the scene is average.The problem of model penetration often occurs.This paper proposes an improved collision detection algorithm based on hybrid bounding box,which combines the motion characteristics of a humanlike model.This algorithm improves the accuracy of collision detection by constructing special OBB at key locations,saving some system resources compared to traditional hybrid bounding boxes.By optimizing the logic of collision response,the problem of model penetration when a human-like model collides with a model in the environment during deformation is solved.2.Due to the large amount of computational resources required for using particle systems to simulate fluid terrain environments,developers rarely use particle simulation methods in computer games in relatively complex environments,as this can greatly affect the real-time performance of virtual scenes.Based on the traditional method of using scene depth information,this paper improves on the method by using the methods of rendering textures and virtual height field meshes to simulate the collision interaction between the model and fluid terrain,which can more save system computing resources while satisfying basic realism.