点云数据虚拟场景图像智能展示仿真研究  

Research on Intelligent Display Simulation of Virtual Scene Images for Point Cloud Data

作　　者：蔡佳含 曹欢 CAI Jia-han;CAO Huan(Chongqing Metropolitan College of Science and Technology,Chongqing 402160,China;Chongqing University of Science and Technology,Chongqing 400000,China)

机构地区：[1]重庆城市科技学院,重庆402160 [2]重庆科技大学,重庆400000

出　　处：《计算机仿真》2024年第10期448-452,共5页Computer Simulation

摘　　要：为了提供更真实、自然的虚拟现实体验,为用户提供舒适的体验效果,提出一种基于图像点云大数据的虚拟现实场景展示方法。采用体素化网格滤波器对图像点云大数据滤波处理,基于旋转图像特征描述子改进迭代最近点算法,用于图像点云大数据配准,再通过包围盒法建立点云K-邻域,结合曲率信息和均匀网格精简点云,对场景深度加以分析,基于获取的深度信息利用Unity3D实现虚拟现实场景展示。仿真结果表明,通过所提方法得到的虚拟现实场景展示效果更为理想,避免了点云精简过度的问题,亮度和饱和度更为自然,更符合人眼对图像的要求。In order to provide a more realistic and natural virtual reality experience and provide users with comfortable experience effects.A virtual reality scene presentation method based on image point cloud big data is proposed.The voxelized grid filter is used to filter the image point cloud big data,and the iterative nearest point algorithm is improved based on the rotating image feature descriptor for image point cloud big data registration.Then,the point cloud K-neighborhood is established by the bounding box method,and the point cloud is simplified by combining curvature information and uniform grid to analyze the scene depth,based on the depth information obtained,Unity3D is used to display virtual reality scenes.The simulation results show that the virtual reality scene display effect obtained by the proposed method is more ideal,avoiding the problem of oversimplification of point clouds,and the brightness and saturation are more natural,which is more in line with the requirements of human eyes for images.

关 键 词：点云大数据 虚拟现实场景展示 迭代最近点算法 均匀精简 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]