计量学报2024，Vol.45Issue(4)：471-479,9.DOI:10.3969/j.issn.1000-1158.2024.04.02
一种基于立体靶标定位的三维重建方法
A 3D Reconstruction Method Based on Stereo Target Location
摘要
Abstract
In order to solve the problem of large data splicing error caused by lack of feature information of line laser sensor,a method of line laser data splicing based on stereoscopic target positioning is proposed.The target point library is constructed by the spatial geometry constraint of binocular stereo vision,and the position calibration of stereo target coordinate system and linear laser coordinate system is realized.Canny edge detection,sub-pixel edge extraction,ellipse fitting and other algorithms are used to identify and locate circular feature points.The homologous points of the collected local landmarks are matched with the target point library,the real-time pose of the line laser sensor is calculated.The initial coordinates of line laser are unified with the global coordinate system by coordinate system transformation,and the three-dimensional reconstruction of point cloud is realized.The experimental results show that the average absolute error of target calibration is 0.078 mm,the average fitting error of the measured standard parts is 0.127 mm.Compared with the Coordinates Measuring Machine(CMM),the average measuring error of this method is 0.165 mm.Compared with the results of robot positioning splicing and depth camera measurement,the measurement system built has smaller average measurement error and standard deviation,which can meet the demand of real-time multi-angle scanning of complex surfaces.关键词
几何量计量/三维测量/立体靶标/双目视觉/标志点定位/数据拼接误差Key words
geometric measurement/three-dimensional measurement/stereo target/binocular vision/marker position/data splicing error分类
通用工业技术引用本文复制引用
陈媛,罗哉,杨力,江文松..一种基于立体靶标定位的三维重建方法[J].计量学报,2024,45(4):471-479,9.基金项目
国家重点研发计划(2022YFF0705704) （2022YFF0705704）
国家自然科学基金(52075511,52005471,51927811) （52075511,52005471,51927811）
浙江省大学生科研创新活动计划(2022R409A052) （2022R409A052）
