Key Technology Research Of Screw Cap Defect Detection System Base On Vision

With the implement of Made in China 2025,machine vision is widely used as a highly automated test technology.Especially in industrial applications,it can effectively improve production efficiency and product yield,while reducing resource consumption to respond to policies of energy conservation.This article is designed to study the defect detection method of the printing surface,the adhesive surface and the claws of screw cap.The main contents of research are as follows:In order to solve the problem that Canny applied to grayscale image detection is required to set thresholds manually,we proposed a Canny color image edge detection method based on Otsu adaptive threshold.The method was based on HSI color space,and calculated the gradient,magnitude and direction of all pixels in three channels by Otsu algorithm.The maximum threshold was used to process and connect the edges.Then,the edge images of three channels were added together,and finally the edge detection results were obtained after morphological processing.The experiments have shown that method can adaptively compute the threshold and detect better image edges.The defect detection of screw cap is divided into the adhesive surface,the printing surface and the claws,and based on improved edge detection algorithm,we proposed the defect detection method for gluing,printing,and claws.The experiments showed that the method had achieved defect detection such as dirt,scratches and deformation of the adhesive surface,the printing surface stains,the printing offset,as well as too large claw,too small claw and claw offset.A system hardware platform was designed and the defect detection software was designed.After real-time online tests of a large number of bottle caps,the results showed that the system can be stable and reliable operation,and its detection accuracy and speed can meet the market demand.