基于三维点云处理的包装袋绳检测

Packaging Bag Rope Detection Based on 3D Point Cloud Processing

自动拆包环节通常需要定位包装袋上包装袋绳的位置并提供给机械手以便进行操作,为准确可靠地定位包装袋绳的位置,本文设计了一种基于三维点云处理的包装袋绳检测算法。使用三维深度相机获取点云数据,选用的边界提取算法需要对点云进行大量角度计算,对点云数据进行下采样、去噪等预处理,提高检测效率,减少噪声的影响。为解决因包装袋绳点云嵌入在包装袋点云中,无法直接提取出包装袋绳点云的问题,根据包装袋绳点云与包装袋点云曲率不同的特征,基于夹角阈值的边界提取算法先提取点云边界,再基于欧氏距离分割算法对点云边界进行分割,分割出包装袋绳点云。将分割后的点云采用重心法计算出中心位置坐标,计算结果与人工测量的中心坐标最大误差在15 mm以内,满足工业机械手臂拆包的需求。对现场采集的不同种类和位置的包装袋点云进行边界提取和分割的定位检测实验,检测结果显示本文提出的检测算法可以有效地检测并定位包装袋绳的位置,方便机械手臂准确地拆解包装袋。

Automatic unpacking process usually needs to locate the position of the bag rope on the bag and provide it to the manipulator for easy operation.In order to accurately and reliably locate the position of the bag rope,an algorithm of bag rope detection based on 3D point cloud processing is designed in our present study.The first step is to obtain the point cloud data of the bag using the 3D depth camera.The boundary extraction algorithm used in this study requires a large number of angular calculations for the point cloud.To improve detection efficiency and reduce the effect of noise,the point cloud data is pre-processed with down-sampling and de-noising.In order to solve the problem that the point cloud of bag rope is embedded in the point cloud of packaging bag,and it cannot be extracted directly from the point cloud of bag rope,according to the different curvature characteristics of the point cloud of bag rope point cloud and the point cloud of packaging bag,the boundary extraction algorithm based on the angle threshold first extracts the point cloud boundary,and then splits the point cloud boundary based on the Euclidean distance segmentation algorithm to segment the point cloud.The center position coordinates are calculated by the center of gravity method.The maximum error with the manual measurement value is within 15 mm,which can meet the needs of industrial mechanical arm unpacking.The location detection experiment of boundary extraction and segmentation was carried out for the different kinds and positions of the point clouds of packaging bags collected on site.The detection results showed that the proposed detection algorithm could effectively detect and locate the position of the bag rope,which is convenient for the mechanical arm to accurately disassemble the packaging bag.