A system and method for intelligent cutting of flat bevels based on compound vision

Abstract

The invention belongs to the field of industrial intelligence, and in particular relates to a flat bevel intelligent cutting system based on compound vision, which includes a human-computer interaction unit, an overall control system, a compound vision combination connected with the overall control system, an articulated arm robot body, and a joint The end picker and torch connected to the arm robot body; the composite vision combination includes a depth camera, a grayscale camera and a structured light vision sensor; the overall control system includes a storage module, a composite vision control processing module, a trajectory planning module, a motion A control module and an equipment control module; the storage module is used to store 3D model data and cutting process data of all parts to be cut; the composite vision control processing module includes a 3D vision control processing module, a 2D vision control processing module and a structured light vision Control processing module; the invention not only uses 3D vision technology to realize the autonomous loading and unloading of target parts, but also integrates 2D vision technology and structured light vision technology to perform accurate position correction, cutting track correction and cutting pose correction for target parts. It realizes the independent generation of the cutting program, and then realizes the fully autonomous intelligent cutting system, which ensures the ideal precision of the bevel cutting of flat parts.

Description

technical field [0001] The invention belongs to the field of industrial intelligence, and in particular relates to an intelligent cutting system and method for flat bevels based on compound vision. Background technique [0002] At this stage, most enterprises still adopt the Teach-then-Play method for flat bevel cutting, that is, for the same type of parts, the first part needs to be taught and used as a template for subsequent workpiece cutting. It is necessary to align and fix the workpiece with the template, and the robot will automatically repeat the cutting procedure, which will not only lead to cumbersome online programming process, high labor cost, low efficiency, unsafe and other disadvantages, but also the following key problems: (1) manual confirmation The cutting position and cutting accuracy of the bevel are completely determined by the visual inspection of the instructor; (2) Even if the parts meet the characteristics of large batches, small batches, and consist...

AI Technical Summary

Problems solved by technology

[0002] At this stage, most enterprises still adopt the Teach-then-Play method for flat bevel cutting, that is, for the same type of parts, the first part needs to be taught and used as a template for subsequent workpiece cutting. It is necessary to align and fix the workpiece with the template, and the robot will automatically repeat the cutting procedure, which will not only lead to cumbersome online programming process, high labor cost, low efficiency, unsafe and other disadvantages, but also the following key problems: (1) manual confirmation The cutting position and cutting accuracy of the bevel are completely determined by the visual inspection of the instructor; (2) Even if the parts meet the characteristics of large batches, small batches, and consistent dimensions, it is inevitable that cutting deformation will occur when the pre-sequence steel plate is unloaded, resulting in the deformation of the steel plate. In case of unevenness, warping, etc., the cutting trajectory will deviate from the actual contour of the workpiece, which will further amplify the size deviation of the workpiece after bevel cutting, especially affecting the accuracy of the size of the blunt edge, resulting in secondary defects in the workpiece. Product rate greatly increased

The disadvantage of the technology involved in this patent is that only the camera is used to collect the target point cloud model of the workpiece to be processed, and then the workpiece is beveled according to the point cloud model to match the target workpiece template, and the workpiece is not further accurately identified Positioning and precise correction of the cutting track, the cutting accuracy is not high

The technical deficiencies involved in this patent are: (1) Due to the influence of the blanking accuracy of sheet metal parts, even the actual size of the same type of parts will have deviations. Only using 3D vision system and 2D vision system to correct the position of parts is a problem. Insufficient, especially when the deviation of the cutting shape or the cutting angle is too large, the cutting edge melts and sags, or the corners are rounded, because the actual space contour of the workpiece cannot be obtained, the cutting trajectory and cutting pose are not corrected, which will lead to The cutting accuracy, especially the reserved size of the blunt edge is too large or too small, which cannot meet the actual welding requirements; (2) The 2D vision system is used to carry out secondary fine positioning of the workpiece to be cut before loading, and after loading the material is no longer corrected. The workpiece on the cutting workbench is accurately corrected for its position and posture. This method can improve the flatness of the cutting workbench, the calibration accuracy of various related equipment, the repeated positioning accuracy of the handling robot, and even the structural form and assembly accuracy of the end picker. All put forward higher requirements, otherwise it will be difficult to ensure the accuracy of the position coordinates of the workpiece to be cut on the cutting table, which will seriously affect the cutting accuracy of the groove of the flat part

Images