Aluminum Plate Hot Rolled Intermediate Slab Head And Tail Defects Machine Vision Detection

The hot rolling of the aluminum plate in the intermediate billet,due to the large amount of rolling,eliminates casting defects during the rolling process,transforms the casting structure into a deformable structure,so that its continuation product has great advantages in terms of surface quality,deep processing and precision control.The national economy and other important fields are widely used,for example,machinery manufacturing,ships,and even aerospace.In the past decade,the degree of automation of the aluminum alloy plate production process has been greatly developed,but the intelligence is still far from being realized.Most of the key technical links still rely on human experience.One of the key issues is the “crocodile mouth” defects and head warping that occur during the rolling process.The current industrial solution to the problem of “crocodile mouth” defects and head warpage is to control the shape of the slab by inserting roll edges and cutting head and tail.However,because there is no effective detection method,the removal of the head and tail of the production process and the timing of the input of the vertical roller are mainly realized by manual observation and production experience.Based on this,this paper proposes a detection and control method based on machine vision.For head and tail defects,the “crocodile mouth” defect was extracted and measured using the Blob analysis method in machine vision.In the analysis process,the advantages and disadvantages of the segmented images are compared between the fixed threshold and the dynamic threshold.Finally,the image segmentation is performed using the dynamic threshold method.After image segmentation,feature regions are extracted and transformed using morphological methods to obtain the length of the “crocodile mouth” defect.The warpage height is obtained by the method of digital image processing for head warping.Finally,a method of controlling the “crocodile mouth” and head warping height of the slab head and tail is given.For the edge defects,because there is no site to build the conditions for the environment to capture images,this paper sets up a binocular system that uses aluminum plates folded into the shape of the double drums on the edges to image acquisition.The calibration platform was used to collect the image of the calibration plate.After calibration,the camera's focal length,distortion,and other internal parameters and the relative positions of the two cameras were calibrated.Using the calibrated camera,the depth profile of the aluminum profile was generated,and the depth of the folding groove of the aluminum plate was calculated from the depth image.Finally,a method for controlling the defects of the double drum at the edge was proposed.Considering industrial applications,visual inspection software needs to be developed.This article developed a head and edge detection system using the Qt framework and the Halcon library.The functions of the system are:(1)Off-line detection and measurement of the head defect length,determination of whether the preset head-to-tail tail value is reached;(2)Binocular acquisition of edge images,calculation of depth images,and use of depth images to obtain edge depressions distance.