基于机器视觉的堆叠矩形纸盒识别方法研究

Research on Recognition and Grab System of Stacked Rectangular Carton Based on Machine Vision

针对堆叠矩形纸盒的识别和抓取问题,设计了一套由SCARA四轴机器人、工业相机和真空吸取装置等组成的机械臂抓取系统。提出了一种简易的手眼标定方法,并进行理论证明和实验分析。改进和完成了一套基于轮廓提取的矩形检测算法,实现对平面上多个矩形纸盒的识别;采用PnP算法对矩形纸盒进行深度估计,从而实现了对堆叠矩形纸盒的识别和抓取。

For the identification and grabbing of stacked rectangular carton,a set of mechanical arm grabbing system consisting of SCARA four axis robot,industrial camera and vacuum suction device was designed.A simple hand eye calibration method was proposed,and theoretical proof and experimental analysis were carried out.A set of rectangle detection algorithm based on contour extraction was improved and completed to realize the recognition of multiple rectangular cartons on the plane.The PnP algorithm was used to estimate the depth of the rectangular carton,thus realizing the recognition and grasping of the stacked rectangular carton.