| 丁伟利,杨庆,张恒.基于虚拟仿真的龙门机器人板材智能分拣系统及方法[J].高技术通讯(中文),2023,33(6):630~642 |
| 基于虚拟仿真的龙门机器人板材智能分拣系统及方法 |
| Gantry robot intelligent sorting system and method for plate based on virtual simulation |
| |
| DOI：10. 3772/ j. issn. 1002-0470. 2023. 06. 008 |
| 中文关键词: 龙门机器人； 虚拟仿真； 智能分拣； 角点检测 |
| 英文关键词: gantry robot, virtual simulation, intelligent sorting, corner detection |
| 基金项目: |
| 作者 | 单位 | | 丁伟利 | （燕山大学电气工程学院智能控制系统与智能装备教育部工程研究中心秦皇岛 066004） | | 杨庆 | （燕山大学电气工程学院智能控制系统与智能装备教育部工程研究中心秦皇岛 066004） | | 张恒 | （燕山大学电气工程学院智能控制系统与智能装备教育部工程研究中心秦皇岛 066004） |
|
| 摘要点击次数: 1249 |
| 全文下载次数: 934 |
| 中文摘要: |
| 在木门加工生产线上，针对龙门机器人对不同尺寸板材分拣的需求，本文提出了一种基于虚拟仿真的龙门机器人板材智能分拣实验系统。该系统基于Unity3D软件平台开发，搭建了龙门机器人分拣任务虚拟仿真场景，并在虚拟仿真实验场景中完成了虚拟数据集采集、相机标定以及手眼标定。针对不同尺寸的板材分拣任务，本文提出了基于板材的矩形角点势自适应识别的智能分拣算法，首先利用目标检测模型在局部区域内对板材进行角点检测，由双目成像原理恢复出的角点三维坐标计算板材尺寸和目标板材的抓取位姿，以此完成木门的分拣。该方法在相机采集图像过程中出现板材被遮挡的情况以及在板材复杂纹理和复杂背景的情况中更为鲁棒。此外，该方法完全在虚拟仿真环境下进行开发测试，不仅节约成本、安全高效而且方便进行多种测试方案的改变，对于向真实场景进行算法迁移和开发具有积极作用。实验结果表明，本文提出的方法可以完成龙门机器人板材智能分拣任务，对设计木工家具制造过程中龙门机器人智能分拣系统具有指导意义。 |
| 英文摘要: |
| On the wooden door processing line, in response to the problem of gantry robot sorting different sizes of plates, a gantry robot intelligent sorting system is proposed for plate based on virtual reality technology. The proposed system is developed by Unity3D, a virtual simulation scene of the gantry robot sorting task is set, and the virtual data set acquisition, camera calibration and hand-eye calibration are completed. In order to solve the sorting of plates of different sizes, an intelligent sorting algorithm is proposed based on the feature of the plate rectangular corner, that uses the target detection model to detect rectangular corners in a local area, and calculates the size of the plate and the grasping pose of the target plate via 3D coordinates recovered by binocular imaging principle. This method is more robust when the plate in the collecting image is blocked or there is complex texture on the plate and background. In addition, the algorithm is developed and tested completely in a virtual simulation environment, which not only lowers cost, but also facilitates the changes of multiple test schemes more safely and efficiently, so that it has a positive effect on the migration and development of algorithms to real scenarios. The experimental results show that the method proposed in this paper can complete the intelligent sorting task of the gantry robot, which has guiding significance for the design of the sorting system of the gantry robot in the woodworking furniture manufacturing process. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
