Research On Contour Recognition And Intelligent Cutting Strategy For Wide And Heavy Plate Based On Machine Vision

Under the background of the development of my country's manufacturing equipment industry turning to automation and intelligence,the steel industry,as the foundation of the national industry,has also changed from an extensive rapid growth model that pursues output to a lean production model that pursues quality.As a type of steel plate,the demand for Wide and Heavy Plate has been increasing in recent years.In the Wide and Heavy Plate production line,due to the limitation of the rolling process,the Wide and Heavy Plate will have head and tail deformation and side bending deformation,which cannot meet the rectangular requirements of the order.Therefore,the contract product needs to be cut in the deformation area.Shearing accuracy is the main limiting factor for the yield of Wide and Heavy Plate.How to improve the cutting efficiency and reduce the cutting edge loss is the main problem to increase the production capacity of Wide and Heavy Plate.At present,most studies at home and abroad are used to detect the amount of side bending of the steel plate,adjust the rolling parameters according to the side bending data feedback control system,and do not describe and analyze the overall profile of the steel plate,and establish a strategic model to guide the subsequent cutting steps.In this thesis,an in-depth study is conducted on existing problems and technical bottlenecks:(1)According to the production characteristics of wide and heavy plates,research on camera selection and image processing algorithms: the purpose is to identify the twodimensional shape contours of Wide and Heavy Plate in real time online.The thesis first selects related hardware equipment such as industrial cameras,laser speedometers,and simulates a domestic 4300 mm Wide and Heavy Plate production line,and builds a 430 mm experimental platform in the laboratory;then uses common preprocessing and image filtering methods to process the wide and thick plate images,And proposed a denoising effect evaluation method;then use an improved edge detection algorithm to extract the sub-pixel coordinates of the steel plate contour,in order to improve the processing speed of the system,while ensuring that the steel plate contour can be characterized,the contour points are proposed using equidistant sampling;Finally,the cubic B-spline curve fitting is introduced to obtain the overall profile of the steel plate.(2)Head and tail shape calculation and shear strategy model establishment: the purpose is to accurately identify the deformation area of the head and tail of Wide and Heavy Plate and locate the cutting line position.The scanning line method is proposed to compare the number of intersections between the steel plate image and the scanning line to classify and model the deformation.This classification method can not only accurately locate the cutting position,but also provide a data basis for subsequent production line optimization.(3)Calculation of side bend shape and establishment of shear strategy model: the purpose is to calculate the amount of side bend and establish a coarse-division strategy model.After identifying the deformation area of the head and tail,it is proposed to use the maximum distance between the transmission side and the operating side of the steel plate and the profile of the steel plate as the side deflection,and establish a rough model based on the side deflection on both sides.(4)Complete the Wide and Heavy Plate contour recognition and intelligent cutting system design,and conduct system testing in conjunction with the laboratory and the Wide and Heavy Plate enterprises.At present,due to inaccurate location of coarse division in a domestic steel plant,the shortlength rate of steel plate is 0.7%,and the return rate of uncut edge is 2.5%.Through the analysis of the on-site data of Wide and Heavy Plate,the system can meet the requirements of real-time acquisition of steel plate images and accurately describe the profile of the steel plate.The length and width of the steel plate,the shear position of the head and tail,the calculation of the amount of side bending and the coarse division method basically meet the error requirements.Finally,the cause of the error is analyzed to provide ideas for the next optimization of the system.