Determination And Quantitative Description Of Hollow Body In Point Cloud

With the development of computer graphics,digital image processing,optics and other technologies,people are demanding more realistic display effect.True three-dimensional display has become an important research field developed by all countries and the volumetric three-dimensional display technology is an important research direction in true three-dimensional display technology.Our country and abroad have done many researches on volumetric 3D display system.However,when3 D point cloud data used for display contains horizontal depressed area,volumetric3 D display systems often meet display problems and the color information of the depressed area can't be correctly displayed which will affect the display effect.This paper is inspired by the research of volumetric 3D display system.In order to solve the problem of 3D display system when dealing with point cloud with depressed area,this paper proposes the concept of hollow body and gives a precise definition of it.Then this paper proposes a voxel connectivity regional-growth hollow body determination algorithm(VCRHD)to determine hollow bodies in 3D point cloud and defines a set of parameters to describe the size,shape and position information of the hollow body.The main research work of this paper is as follows:Firstly,this paper introduces the characteristics of point cloud data and methods to obtain point cloud.Secondly,this paper proposes a preprocessing algorithm based on connectivity that can achieve point cloud denoising and point cloud segmentation at the same time.The point cloud preprocessing algorithm can meet the need of data in this experiment which can quickly separate the disconnected independent objects in the point cloud data and achieve simple point cloud denoising fast.Then,this paper proposes an algorthm based on the voxel connectivity principle to determine hollow bodies in the point cloud and defines a set of parameters containing the volume ratio of the hollow body,the depth of the hollow body,and the normal of the hollow body to describe the size,shape and position of the hollow body.After that,this paper establishes several sets of point cloud data with different sizes and shapes of regular hollow bodies to determine optimal algorithm parameters through comparison between experimental results and theoretical results.The optimal voxel segmentation parameters in the algorithm can guarantee the accuracy and efficiency for processing the point cloud data containing different sizes and shapes of hollow bodies.Furthermore,this paper builds a data set of 300 point cloud data containing different hollow bodies because of the groundbreaking nature of this problem.The data set can be divided into seven categories containing test data,architecture,bridges,jewelry sculptures,hardware components,indoor furniture and natural scenery.The experimental results prove the feasibility of the algorithm and this data set can be used as a reference for subsequent research work.Finally,this paper summarizes the whole work.Then this paper analyzes the limitations of the algorithm and looks forward to the application prospect in the future.