Research And Application Of Detection Technology For Coiling Product Defects Based On Machine Vision

With the development of computer technology and communication technology, the technology of machine vision matures gradually. It has been an indispensable technical support for the modern manufacturing industry, which is widely used in packaging, food, national defense, chemical, building materials, electronics manufacturing, automobile manufacturing, pharmaceutical and other industries.This article analyzes the current situation and demand of winding machine detection techniques and combines with independent research and development of fully automatic winding machine. A kind of detecting system winding products defect has been designed, which is based on embedded platform. The test results in the engineering environment show that the design of the detection algorithm is feasible and the embedded detecting system is working and reliable. The rate of identification for a defective product is99%, and it takes only2.4seconds to identify each product, which can meet the requirements of the factory production. Specifically, main work of this article includes the following aspects:First do the image preprocessing of collected winding products images by using the Matlab method of winding products’circle numbers. Then combined with the real-time and accuracy requirements, it aims to develop a binaryzation arithmetic to make image segmentation, and this binaryzation arithmetic is on the basis of the gray equalization peak as its threshold. And it turns out to be clear and definite coil numbers in the image of binarization and edge is much smoother as well.Labeling in the binary image connected area, through analyzing and comparing several algorithms, it put forward the identification to the connected area by recursive method, solved the problem of the coil number identification, and highed the efficiency of its recognition. Second, design the hardware platform based on S3C2440processor, with CCD camera as image acquisition collector, and then install the linux2.6.32.2core files, and transplant the related tools of user interface interface, which are ready to test software transplant.At last, the Qt software-designed defect detecting software of winding products sets a defect detection system for winding products. Through the testing of recognition rate and response time, it turns out that the technology is suitable for coiling product defect detection, and solves the shortage of defect detecting in the winding machine. The technology shows certain reference value for the development of industrial detection as well as the machine vision technology.