摘要
随着三维点云技术的不断成熟,目前许多企业将其应用到了货物管理中。但在点云获取中设备会因为外界原因(光照、温度等)导致数据不完全(点云空洞),而且获取初始点云数巨多,降低了后期点云处理效率。提出了对点云的一种预处理,既要获取完整目标物的曲面特征,也要对点云进行高效地压缩。首先对原始点云进行条件滤波和统计分析滤波去噪,然后提出了一种由统计分析和三角格化结合的点云补全算法,最后将补全的点云使用改进的曲率精简法对点云进行处理,并将方法与其他方法做了对比实验。实验结果表明,方法可以将孔洞点云补全,保留货物平面特征和边缘细节信息,降低点云存储空间,为后期对点云处理提高效率有着重要意义。其中,补全算法使补全后的点云与补全之前的点云的最近点平均距离小于0.0004 m,点云精简算法的精简率可达到90%。
With the continuous maturity of 3D point cloud technology,many enterprises are currently applying it to cargo management.However,in point cloud acquisition,devices may experience incomplete data(point cloud voids)due to external factors such as lighting and temperature,and the initial number of point clouds obtained is too large,which reduces the efficiency of later point cloud processing.This article proposes a preprocessing method for point clouds,which requires both obtaining complete surface features of the target object and efficiently compressing the point cloud.Firstly,conditional filtering and statistical analysis filtering were performed on the original point cloud to remove noise.Then,a point cloud completion algorithm combining statistical analysis and triangulation was proposed.Finally,the completed point cloud was processed using an improved curvature reduction method,and a comparative experiment was conducted between this method and other methods.The experiment proves that this method can complete the hole point cloud,retain the cargo plane features and edge details,reduce the point cloud storage space,and has important significance for improving the efficiency of cargo point cloud processing in the later stage.Among them,the completion algorithm in this article makes the average distance between the nearest point of the completed point cloud and the point cloud before completion less than O.0004 meters,and the reduction rate of the point cloud reduction algorithm in this article can reach 90%.
作者
顿启鹏
毋涛
邓魏永
Dun Qipeng;Wu Tao;Deng Weiyong(School of Computer Science,Xi'an Engineering University,Xi'an 710048,China)
出处
《国外电子测量技术》
北大核心
2023年第8期111-116,共6页
Foreign Electronic Measurement Technology
基金
国家自然科学基金(青年基金)(61806160)项目资助。
关键词
点云空洞
条件滤波
统计分析
点云精简
点云补全
point cloud voids
conditional filtering
statistical analysis
point cloud simplification
point cloud completion
