基于SIFT和LBP点云配准的接触网零部件三维重建研究  

Application of Point Cloud Registration in 3D Reconstruction of Catenary Parts Based on SIFT and LBP

作　　者：徐建芳[1,2] 刘志刚 韩志伟[1] 耿肖[1] 

机构地区：[1]西南交通大学电气工程学院,四川成都610031 [2]江苏(丹阳)高性能合金材料研究院,江苏丹阳212300

出　　处：《铁道学报》2017年第10期76-81,共6页Journal of the China Railway Society

基　　金：国家自然科学基金(U1134205;51007074;51407147);中央高校基本科研业务费专项资金(2682015CX031)

摘　　要：现有接触网的三维重建所需时间较长,工作繁琐,为解决此类问题,本文提出采用自动重建法利用点云数据实现其零部件的三维重建。点云配准是影响重建过程准确度及效率的重要因素,而目前普遍使用的SIFT匹配算法,由于构建的关键点特征向量维数高,计算量大,导致匹配速度慢。为解决此问题,本文提出利用均匀模式LBP特征值描述SIFT关键点,获取关键点特征向量,并用向量间的距离判断关键点的相似性,以确定关键点的对应关系,完成配准和重建,得到接触网零部件的三维模型。结果表明,本文所提算法可行有效,能提高匹配速度,加速三维重建。The methods of 3D reconstruction of catenary s y s t e m currently used are t i m e - c o n s u m i n g a n d full of heavy workload. In order to address this problem, a method using optical instruments to acquire point cloud data for the automated reconstruction of catenary parts was proposed in this paper. The process of point cloud registration is crucial to the efficiency and accuracy of the entire 3D reconstruction process. The SIFT algo-rithm is known as the most widely used local feature-based matching algorithm with high performance, but the intensive computation and high vector dimension of building eigenvectors for key points affect matching speed. To solve this problem, LBP eigenvalues in uniform pattern were used to describe the SIFT key points to obtain the eigenvectors of the key points. The distance between vectors was used to determine the similarity of key points to identify the correspondence of two key points in different point clouds. Then coarse registration, fine registration and surface reconstruction were completed, and the 3D reconstruction model of catenary parts was finally finished. Experimental results show that the proposed algorithm is able to realize the objective of impro-ving the matching speed, thus speeding up reconstruction process.

关 键 词：三维重建 SIFT算法 LBP特征值 点云配准 

分 类 号：U225.4[交通运输工程—道路与铁道工程]