Research On Industrial Robot Labeling System Based On Vision Guide

Machine vision is known as the "eye of industry" and is an important tool for industrial robots to sense the external environment.Industrial robots equipped with machine vision can increase the level of intelligence of industrial robots in the industrial production process and maximize production benefits,therefore,the research on industrial robot technology based on machine vision guidance is of great significance.For problems of low positioning accuracy and single labeling orientation of the labeling system,this paper proposes a vision-guided industrial robotic labeling system,analyzes the labeling process requirements of typical workpieces,systematically studies each step of the labeling operation,and explains each the principle of the steps,and according to the principle to build a labeling experiment platform to verify the feasibility of the system.This article mainly completes two parts of image processing and motion control:Image processing is the key to visual recognition and positioning,which mainly includes camera calibration,template matching,and hand-eye calibration.Camera calibration discusses the imaging principle and coordinate transformation of the camera in detail,and then explains lens distortion and corrects lens distortion.The template matching introduces the principle and the key three key elements.Using the shape template matching algorithm of HALCON vision software to realize the workpiece identification and positioning,this step is also the key to industrial robot multi-position and multi-position labeling.In industrial applications,two-dimensional hand-eye calibration is generally based on a nine-point calibration.This method is easy to implement and has high accuracy,but requires the camera installation mode to be "eyesto-hand".Based on the camera installation in the labeling system,this paper proposes an improved nine-point calibration for hand-eye calibration in view of the shortcomings of the nine-point calibration.The experimental results show that the method in this paper can achieve the nine-point calibration of "eyes in hands".Motion control consists of a 6-DOF industrial robot.First,the mathematical model of the industrial robot is established according to the standard D-H parameter method and the hardware parameters of the industrial robot.Then use the algebra method to solve the kinematic formulas of the industrial robot.After that,the industrial robot is simulated in MATLAB software according to the D-H parameters to verify the accuracy of the mathematical model of the industrial robot.Finally,the simulation parameters of the kinematic formula are compared with the parameters of the industrial robot teach pendant to verify whether the kinematic formula can be applied to industrial robot motion control.Image processing and motion control are related through hand-eye calibration.Therefore,hand-eye calibration is an important factor that constitutes a visually guided industrial robotic labeling.In addition to the above work,this paper also completed hardware selection,labeling system platform construction,calibration board making,multi-directional labeling of the system,graphical user interface design and python function packaging and data transmission.Finally,build a labeling system platform for testing,and objectively analyze the test results.