喷涂系统大视场下多目视觉手眼标定数据
收藏资源简介：
本数据集面向轨道交通及工业机器人视觉系统研究领域，主要用于工业机器人搭载多目工业相机进行“眼在手上”手眼标定的定位精度验证与性能评估，旨在为机器人视觉引导系统标定、精度分析及算法优化提供标准化、可复用的数据支撑。数据依托国家重点研发计划相关课题，由武汉理工大学汽车工程学院硕博研究生依据国家标准GB/T 45501-2025《工业机器人三维视觉引导系统通用技术要求》在武汉理工大学汽车工程学院完成，数据采集与整理工作于2023年9月27日进行，总规模为188 MB。数据采集过程中，采用ABB IBR7600工业机器人作为执行平台，搭载三台海康工业相机（型号 MV-CS050-10UC）及MVL HF1228M 6MP工业镜头，对标准标定板进行多姿态拍摄，共采集330张手眼标定图像。在标定阶段，通过对标定图像进行处理，计算主相机相对于机器人末端法兰盘的转换矩阵。在验证阶段，机器人携带工业相机拍摄标记点图像，并结合像素坐标与转换矩阵计算标记点三维坐标；同时，机器人携带触针对同一标记点进行示教测量，获得实测三维坐标。通过对比计算坐标与实测坐标的差异，分别统计 X、Y、Z三个方向的坐标偏差，并对所有偏差取平均值，得到手眼标定定位精度的最终评价指标。本数据集内容包括手眼标定图片（标定期间拍摄标定板的图像）、验证图片（标记点拍摄图像）及标定数据表。本数据集记录了手眼标定转换矩阵计算结果及每个标志点的X、Y、Z方向计算值、实测值和绝对误差，系统呈现手眼标定与定位精度测定的全过程。该数据集为工业机器人视觉引导系统的标定方法研究、精度评估及多目视觉算法优化提供了可靠的数据基础，并可用于复现实验条件和对比验证其他视觉标定方案。
This dataset is targeted at the field of rail transit and industrial robot vision system research, and is mainly used for positioning accuracy verification and performance evaluation of eye-in-hand hand-eye calibration when industrial robots are equipped with multi-eye industrial cameras. It aims to provide standardized and reusable data support for calibration, accuracy analysis and algorithm optimization of robot vision guidance systems. The dataset was developed based on relevant projects of the National Key R&D Program of China, and was completed at the School of Automotive Engineering, Wuhan University of Technology by master and doctoral students of the school in accordance with the Chinese national standard GB/T 45501-2025 General Technical Requirements for Industrial Robot 3D Vision Guidance Systems. The data collection and organization work was carried out on September 27, 2023, with a total size of 188 MB. During the data collection process, an ABB IBR7600 industrial robot was used as the execution platform, equipped with three Hikvision industrial cameras (model MV-CS050-10UC) and MVL HF1228M 6MP industrial lenses. Multi-pose shooting was performed on a standard calibration board, and a total of 330 hand-eye calibration images were collected. In the calibration stage, the transformation matrix of the main camera relative to the robot's end-effector flange was calculated by processing the calibration images. In the verification stage, the robot carried the industrial camera to shoot images of the markers, and calculated the 3D coordinates of the markers by combining the pixel coordinates and the transformation matrix. Meanwhile, the robot carried a touch probe to perform teaching measurement on the same markers to obtain the actual measured 3D coordinates. By comparing the difference between the calculated coordinates and the measured coordinates, the coordinate deviations in the X, Y and Z directions were counted respectively, and the average value of all deviations was taken to obtain the final evaluation index of the hand-eye calibration positioning accuracy. This dataset includes hand-eye calibration images (images of the calibration board taken during calibration), verification images (images of the markers taken during verification) and calibration data tables. It records the calculation results of the hand-eye calibration transformation matrix, as well as the calculated values, measured values and absolute errors of each marker in the X, Y and Z directions, systematically presenting the entire process of hand-eye calibration and positioning accuracy measurement. This dataset provides a reliable data basis for the research of calibration methods, accuracy evaluation and multi-eye vision algorithm optimization of industrial robot vision guidance systems, and can be used to reproduce experimental conditions and compare and verify other vision calibration schemes.
