Robot Grinding System Based On Depth Learning And 3D Vision

One of the important development directions of complex product processing technology is automated polishing robots.Implementing polishing automation is beneficial for improving product quality and processing efficiency,and the practicality of polishing robots can be confirmed from multiple aspects.The usage situation shows that enterprises and products have increased their investment in polishing robots in various industries,including bathroom,construction hardware,automobiles,tableware,handicraft industries,etc.,with significant improvements.New types of mechanical equipment commonly use polishing robot technology in these industries and present a variety of needs.A robot polishing system based on deep learning and 3D vision is proposed to address the issues of obvious defects,low polishing efficiency,high demand for labor costs,and low standardization of product polishing in the traditional polishing industry,as well as to improve the intelligence level of China’s manufacturing industry.The process and equipment technical solution for automatic surface polishing of parts is completed.The robot polishing system first detects defects in the parts to be polished through image processing.To improve the detection efficiency and convergence speed of the algorithm,the feature values of the gray level cooccurrence matrix are introduced into the neural network;After obtaining the defect position using the case segmentation algorithm,the depth camera is used to reconstruct the part surface in 3D.In order to extract the defect point cloud information,the image registration is used to map the image detection results with the 3D reconstructed point cloud information,and the color information is used to divide the point cloud;To address the drawback of obtaining point cloud information with outliers and a stepped shape,filter the point cloud information to obtain a smooth triangular mesh surface;Finally,the smooth point cloud surface is sliced,and the robot path planning is carried out by using the line cutting tool path.The cutter contacts at the end of the robot are sampled at intervals and the robot pose information is calculated at this time.In order to reduce the impact of subsequent robot polishing on the robot body,the joint interpolation,linear interpolation and arc interpolation are combined for motion planning.