Aiming at the problem of low recognition and positioning accuracy of robot grasping stacked steel pipe fittings,a recognition and positioning method of stacked steel pipe fittings based on point cloud information was proposed.Binocular vision was used to obtain the disparity map of the pipe fittings.HSI weight distribution and left-right consistency verification method were used to verify the oc-clusion points of the pipe fittings,and the complete disparity map of the pipe fittings was obtained.Based on the normal vector characteristics of the point cloud of the pipe fittings,the edge points of the contact part of the pipe fittings are eliminated,and the European clustering is used to segment and extract the in-dependent point cloud of the pipe fittings.The ISS algorithm is used to obtain the key point cloud data in-formation of the pipe fittings.The experimental results show that the average recognition rate of the pro-posed method for stacked Masteel three-way pipe fittings under different working conditions is 89.5%,and the robot grasping accuracy is improved by 37%on average.
NETL
NSTL
万方数据
