Research On Quality Inspection Technology Of Glass Containers Based On Vision

With the development of glass container production technology and manufacturing process,the quality inspection requirements of products are also gradually increased.Visual detection technology is widely used in the field of detection,and it has the advantages of non-contact,high accuracy,online detection and high degree of automation.This paper designs a set of quality inspection system for glass container,which can detect the size and the capacity of the glass container respectively.The rotating device achieves a multi-angle detection of the size of the glass container inside and outside diameter.The glass container capacity detection problems convert to liquid level detection and tracking issues.This paper overcomes the problem of angle parallax caused by the non-horizontal position of the camera and the liquid level.In this paper,a method of detecting the inner and outer diameters of glass containers based on bi-telecentric lens is proposed.Firstly,the bi-telecentric lens and the camera are calibrated to get the internal and external parameters of the camera.Then,with using the characteristics of the inner and outer wall of the glass container,the edge of the subpixel is extracted by Hough line detection and the least square curve fitting.Finally,the actual size of the glass container is obtained according to the internal and external parameters of the camera.For the stability and accuracy of liquid level detection,a liquid level detection and tracking method for glass container was proposed.Firstly,the KCF tracking algorithm is used to determine the liquid surface area by combining the liquid level motion and the structural features.Then,the similarity matching with the proposed method is used to match the set of the connected domain.At last,the height of the liquid level is determined by using the parabolic fitting solution.Experiments were conducted on different types of glass containers to analyze the stability and accuracy of the method.The experimental results show that the standard deviation of the inner diameter measurement algorithm is less than 0.1,and the absolute error of the detection value is less than 0.02mm when the outer diameter is within the range of 10?20mm.When the diameter of the outside diameter is 20?40mm,the absolute error is less than 0.03mm.The success ratio of liquid level detection and tracking algorithm is higher than 99.8 percent,and the mean absolute error is less than 0.2ml for the minimum unit of 1ml glass container.The mean absolute error is less than 0.03ml for the minimum unit of 0.1ml glass container.