基于局部特征的位姿求解算法  

Localized Feature-Based Position Solving Algorithm

作　　者：徐海璐[1] 黄志伟 戎新萍 XU Hailu;HUANG Zhiwei;RONG Xinping(College of Mechanical and Electrical Engineering,Nanjing Tech University Pujiang Institute,Nanjing 211200,China;Center Research Institute,Kaiwo New Energy Vehicle Group Co.,Ltd.,Nanjing 211200,China)

机构地区：[1]南京工业大学浦江学院机电学院,南京211200 [2]开沃新能源汽车集团股份有限公司中央研究院,南京211200

出　　处：《组合机床与自动化加工技术》2023年第10期54-58,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：江苏省工业和信息产业转型升级专项资金项目(苏工信综合[2022]115号);南京工业大学浦江学院科研项目(njpj2021-2-02)。

摘　　要：为提高点云数据配准精度,更准确获取目标物体位姿信息,提出一种基于局部特征的多源数据匹配位姿求解方法。首先,通过3D深度相机获取模型三维点云数据,并根据梯度阈值算法从大量点云数据中得到物体表面梯度值;其次,使用L+边界提取算法提取物体边缘轮廓特征信息,再遍历实体大量点云数据求解二维轮廓特征值,从而对输出的轮廓特征数据建立特征数据库;最后,对获取的点云数据特征进行数据处理,通过Hu矩匹配算法找到相对应的数据源后再由F-ICP匹配算法获取模型位姿。结果表明基于局部特征的多源数据匹配算法可以有效对点云数据进行位姿计算,匹配精度得到提升。In order to improve the registration accuracy of the point cloud data and obtain the pose information of the target object more accurately,this paper proposes a multi-source data matching pose solution method based on local features.Firstly,the model point cloud data is obtained through the 3D depth camera,and then the surface gradient value of the object is obtained from a large number of point cloud data according to the gradient threshold algorithm.Secondly,the edge contour feature information of the object is extracted through the L+boundary extraction algorithm.Then the machine traverses a large number of target object feature database for the output contour feature data.Finally,the features of the dynamically acquired point cloud data are processed,and the data source is found by Hu corresponding moment matching algorithm,and then obtain the model position and attitude accurately by F-ICP matching algorithm.The results show that the multi-source data matching algorithm based on local features can effectively the multi-source data matching algorithm based on local features can effectively cloud data and improve the matching accuracy.

关 键 词：位姿求解 点云数据 轮廓信息 匹配算法 

分 类 号：TH112[机械工程—机械设计及理论] TG506[金属学及工艺—金属切削加工及机床]