Accurate Camber Detection And Intelligent Optimization Based On Machine Vision

With the introduction of the "Made in China 2025" strategic thinking,the implementation of intelligent control and optimization of production,focus on the development and application of intelligent technology,the realization of information depth perception,intelligent optimization decision-making and precise coordinated control become the key to the development of China's steel industry problem.In this regard,this paper takes the camber phenomenon of the rolled piece caused by various asymmetry factors in the rolling process of a 3500mm plate mill as the research object,and aims to realize the intelligent control of the camber of the rolled piece in the production process,developed and designed a machine vision-based camber detection system,and established an integrated control strategy for the camber of the plate with the combination of feedforward control and feedback control.At the same time,the application of intelligent algorithm to optimize the camber control model is deeply studied.The paper has achieved the following research results:(1)A machine vision based camber detection system for rolling pieces was established.Based on the design idea of the two-layer network control architecture,a CCD camera is installed directly above and behind the rolling mill to establish the structure of the camber detection system,and the corresponding working principle and software and hardware design of the system are explained.At the same time,through the establishment of pinhole model and lens distortion model,the detection system was calibrated by Zhang Zhengyou calibration method,and the calibration error analysis was carried out.The calibration error was less than 0.3 pixels,which met the accuracy requirement of camera calibration.(2)A method for measuring the camber rate of high-precision rolling pieces is proposed.Based on machine vision technology,median filtering,image sharpening,mathematical morphology transformation,threshold segmentation and Canny edge detection are performed on the camber image of high-temperature rolling pieces,The third-order Bezier curve model is introduced,and the RANSAC algorithm is used to fit the edge of the rolling pieces.Then the sub-pixel coordinates of the edge of the rolled piece are extracted by the improved sub-pixel edge detection algorithm based on Zernike moment.The detection accuracy is 0.02 pixel.Then,through the least squares circle fitting,the bending curvature of the center line of the rolling piece is obtained as the high-precision camber rate of the whole rolling piece.(3)A camber control system for rolling pieces was established.By studying the influence of different rolling state asymmetry factors on the camber,the relationship between the wedge shape and the camber of the exit rolling pieces is quantitatively described,the mathematical model of the camber feedforward control setting of the rolling pieces and the mathematical model of the camber feedback control of the exit rolling pieces suitable for various influencing factors based on parameter correction are established.Combining the two models,a complete machine vision-based comprehensive control strategy for the camber of plate is established,which lays a solid theoretical foundation for solving the camber defects of wide and thin products.(4)Combined with the machine learning theory,the least squares support vector machine is combined with the mathematical model of the camber feedback control based on parameter modification to establish the PSO-LSSVM-based camber integrated leveling control model.Through the compensation of the roll gap tilt adjustment value of the traditional camber control model,the intelligent control of the camber is realized,which is obviously superior to the traditional BP neural network.And compared with the field measured value verification,the relative error is only within±12%,which is nearly 2 times higher than that of the simple mathematical model,which indicates that the model built in this paper can reliably achieve the precise control of the camber.