一种基于体素的室内三维连通图自动生成算法  

An Automated Voxel-Based Algorithm for Building Three-Dimensional Connected Graph

作　　者：柯芬 熊汉江[1] 戴雪峰[1] 

机构地区：[1]武汉大学测绘遥感信息工程国家重点实验室,湖北武汉430072

出　　处：《测绘与空间地理信息》2015年第3期34-37,共4页Geomatics & Spatial Information Technology

基　　金：国家科技支撑计划课题(2013AA12A203)资助

摘　　要：为了能快速计算室内导航路径,必须使用简单的数据结构表达室内复杂的路径导航信息,室内三维连通图就是一种较好的手段。但是传统的室内精细建模重在几何模型的构建和纹理数据采集,缺乏室内三维连通图的构建。针对广泛存在室内几何模型提出一种基于体素的室内三维连通图自动生成算法,对建筑物内部进行分割和填充,将室内空间划分为离散的导航空间,通过自动语义关联提取连通关系,最终生成室内空间三维连通图。In order to quickly calculate indoor navigation path, a three-dimensional （3D） indoor environment must be represented with a simple data structure.Indoor three-dimensional connected graph serves as a goodsolution.Because the traditional fine indoor modeling is focusing on the construction of geometric model and the acquisition of texture data and lackstheconstruction of indoor 3D connected graph, an automated voxel-based algorithm for building 3D connected graph is proposed.This method could be achieved by two main steps.Firstly, using voxelization algorithm to divide the indoor space into discrete navigation spaces, thus the navigation space can be abstracted as nodes.Secondly, using automatic semantic association to build connectivity relationship between nodes. Upon which, the connected graph that represents the relationship between nodes can be finally generated.

关 键 词：室内导航 体素 语义关联 连通图 

分 类 号：P228[天文地球—大地测量学与测量工程]