| 刘博，金守峰，宿月文，严楠，李毅.堆叠筒子纱机器人无序分拣方法[J].轻工机械,2022,40(6):14-21 |
| 堆叠筒子纱机器人无序分拣方法 |
| Robotic Unordered Sorting Method for Stacked Bobbin Yarns |
| |
| DOI：10.3969/j.issn.1005 2895.2022.06.003 |
| 中文关键词: 筒子纱 分拣机器人 点云分割 点云配准 M估计抽样一致性算法 |
| 英文关键词:bobbin yarn sorting robot point cloud segmentation point cloud alignment MSAC(M Estimation Sampling Consistency Algorithm) |
| 基金项目:陕西省科技厅重点研发计划项目(2020GY 172)；陕西省重点研发计划项目（2021GY 326）；宁波市科技创新2025重大专项（2019B10075）。 |
|
| 摘要点击次数: 186 |
| 全文下载次数: 170 |
| 中文摘要: |
| 针对筒子纱分拣机器人对作业目标堆叠时识别不准确、抓取位置不稳定的问题，课题组提出了一种堆叠筒子纱的机器人无序分拣方法。构建了具有3D视觉感知的筒子纱分拣机器人系统，利用Kinect V2相机获取堆叠筒子纱的图像信息，对采集到的筒子纱点云进行处理；通过M估计抽样一致性算法（MSAC）和直通滤波算法对原始点云进行裁剪，采用改进E R分割算法和ICP算法完成点云的分割和配准，得到筒子纱的位姿信息；最后使用机器人进行抓取实验。实验结果表明：该方法能够实现对堆叠场景下的作业目标的识别定位，系统的抓取成功率达到86%，可以满足筒子纱分拣的实际生产需求。 |
| 英文摘要: |
| In response to the problems of inaccurate recognition and unstable grasping position of the stacked operation targets by the bobbin yarn sorting robot, a disorderly sorting method by robot for stacked bobbin yarns was proposed. The robot system with 3D vision perception was built, using the Kinect V2 camera to acquire the image information of the stacked yarn, and the point cloud of collected yarn was processed. The original point cloud was cropped by MSAC (M Estimation Sampling Consistency Algorithm) and straight pass filtering algorithm, and the segmentation and alignment of the point cloud was completed by improved E R segmentation algorithm and ICP algorithm to obtain the positional information of the bobbin yarn, and finally the robot was used for verification by grasping. The experimental result shows that this method can recognize and locate the working target in stacking scenario, and the success rate of the system reaches 86%, which meets the actual production requirements of bobbin yarn sorting." |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
