Design Of Visual Positioning System For Laser Welding Platform Based On OpenCv

In laser precision welding,it is necessary to manually adjust the position of workpiece or fixture due to the influence of machining precision and installation mode of product fixture,and the limitations of manual operation are bound to include worker negligence,poor repeatability and reduced work efficiency.This can result in welding results that are far from expected,resulting in product defects.Using machine vision technology to guide welding positioning is an effective method to solve this problem.By calculating the positions of the template and the target image respectively,the compensation amount can be obtained by making the difference between the two results,and the compensation amount can be sent to the positive position of the moving platform to achieve the purpose of accurate welding.Based on the research on the key technologies of machine vision and visual positioning guidance,combined with the characteristics of the products and the actual technological conditions in the laser welding process,a two-step positioning method is proposed in this paper,that is,the image is first coarsematched,and then the actual position of the current product is identified by the local geometric feature of the image to achieve fine matching.First of all,this paper makes a strict selection of machine vision hardware system by analyzing the welding demand of products;Then,the pre-processing operations such as graying,image filtering and binarization are carried out and the process of processing is introduced;In the process of realizing the function of image rough matching,this paper proposes a relatively fast template matching algorithm based on NCC matching;In terms of the function of marker positioning,this paper USES the Canny operator to get the edge image,and then gets the coordinate of each edge point to get the set of edge points.In this paper,k-means algorithm is used to make a simple dichotomy on the set of edge points,and then two lines are fitted by RANSAC algorithm.According to the principle of point and line positioning,the measurement of the current image position is completed.Then the difference between the measurement of the template image and the target image is calculated,and the relative offset of the two products can be calculated to get the output result.Considering theactual production and application of the whole equipment in this paper,and combining with the structural characteristics of the equipment,a fast and effective calibration method is proposed,which can achieve the expected calibration accuracy.After the algorithm introduced in this paper is implemented,a simulation experiment is carried out on the realization results of this algorithm.Considering the accuracy,execution time and stability,the results all meet the expected standards and meet the actual production needs.