基于空间无序点云的层次构造三角化方法  

Layering Triangulation Method Based on Disordered Space Points

作　　者：成宏璟[1] 王庆[1] 

机构地区：[1]西北工业大学,计算机学院,陕西省语音图像重点实验室,西安710129

出　　处：《科学技术与工程》2010年第9期2114-2117,2121,共5页Science Technology and Engineering

基　　金：国家863项目(2007AA01Z314)资助

摘　　要：描述了一种新的利用空间无序点云重构三维模型的算法,利用分层扫描的原理将无序的三维点云进行平面投影,使用矢量外边界扫描算法将点信息提取为线段信息,由于分层的有序性使得线段之间存在着良好的拓扑关系,经过在有序线段中提取拐点矢量,连接上下两层相关点,点模型就完成了三角化。该方法计算速度快,拓扑清晰,且不会出现失真(泊松表面重建)和空洞现象(区域增长法)。An algorithm that uses space disordered point cloud to reconstruct three-dimensional model is described. Layer-based scanning principle was used to project disordered three-dimensional points cloud to a plane. Outside boundary-scan algorithm based on vector direction was used to extract line information from points information. As a result of stratification,good topological relation between lines was achieved. After extracting inflection point from ordered line segments,and connecting the upper and lower layer related points,the model is complete triangulation. The algorithm is fast,good topology structure,and can avoid distortion （Poisson surface reconstruction） and empty phenomena （region-growing method）

关 键 词：无序点云 分层 扫描 三角化 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]