Research And Development Of A Machine Vision-based Online Inspection System For Plate Punching Quality

As one of the common forms of metal materials,large-size metal punching sheet is widely used in machinery,automobile,construction and other fields.It is often affected by uncontrollable factors in the punching process,resulting in defects in the punching quality of the plate,such as round hole offset,insufficient hole number and so on.In addition,manual quality inspection audit will also have the phenomenon of misjudgment,thus reducing product quality and production efficiency.In view of the above problems,this thesis proposes an online detection system for plate punching quality based on machine vision.A real-time online detection system with high detection accuracy,fast detection speed and low cost was successfully designed and developed by studying and learning machine vision,image preprocessing,image feature extraction and standard part plate model database.The main research contents and results are as follows:(1)This thesis introduces the background and significance of the subject,and analyzes the research status at home and abroad.According to the requirements of the production technical indicators of the enterprise and the on-site environment of the production workshop,the overall architecture of the detection system is designed,and the hardware parts of the industrial camera,optical lens,lighting source and embedded industrial computer of the detection system are designed and selected.(2)The related operators for image preprocessing of punching sheet are introduced and analyzed in detail.Firstly,based on the extraction of region of interest(ROI),the image smoothing effects of different filtering operators are studied and analyzed.Secondly,various image segmentation algorithms are studied and analyzed.The convolutional neural network(CNN)method based on deep learning and the automatic global threshold method based on the maximum inter-class variance(Otsu)are selected for image segmentation,and the advantages and disadvantages of the two algorithms are compared and analyzed.In order to retain the necessary image contour edge feature information,morphological processing is performed on the segmented binary image containing clutter and holes.(3)The positioning and extraction of image contour edge by coarse-fine two-stage edge positioning technology is studied and analyzed.Firstly,the related operators of coarse-fine two-level edge positioning technology are studied and analyzed,and the Canny edge detection operator is selected for coarse-level edge positioning.Secondly,the optimized and improved polynomial interpolation method in any direction is used to locate the fine edge on the basis of the coarse edge positioning,and the circular hole and image edge line in the punching plate image are successfully extracted.Then,the circle fitting technology based on Hough transform and least square method is studied and analyzed,and the fitting results of the two are compared through experiments.Finally,the circle fitting technology based on least square method is selected,and the relevant size data of punching plate are obtained by image contour feature detection.(4)A simple and practical online detection system for plate punching quality was designed and developed,and the experimental results were analyzed and verified.Firstly,this thesis selects Halcon image processing software and Visual Studio 2019 software development platform for mixed programming,and develops automatic detection system software and human-computer interaction interface to realize the convenience of operation.Then,a plate model database based on standard parts is created,and the matching value parameters are defined,so that the plate model can be called directly when the system is tested,which is convenient for employees to produce,reflecting the practicability of the detection system.Finally,the experimental results are analyzed and verified.The experimental results show that the detection system achieves the design performance in terms of real-time detection,operational stability,and measurement accuracy,and meets the actual needs of the enterprise.(5)The automatic detection system put into practical application of enterprises is introduced and explained.Firstly,the hardware and function display interface of the automatic detection system are introduced.Then,the related process of the automatic detection system from design to practical application is explained.Finally,the detection efficiency and economic benefits before and after the automatic detection system are compared and analyzed.The practical application of the factory proves that the detection system has good real-time detection,operation stability and measurement accuracy,which meets the production requirements of the enterprise.The detection system effectively improves the efficiency of product quality detection,saves labor,reduces the cost of plate punching quality detection,and improves the production efficiency of enterprises.