The Glass Surface Defect Online Detection System Development

Along with the rapid development of science and technology, the development ofthe glass industry is also facing opportunities and challenges. Especially after China’saccession to the WTO, China’s glass products are significantly in the inferior potentialin the international competition. The main reason is that the domestic glass productshave lower product quality and the technical level of product quality testing equipmentis backward. At the same time, the cost of introduction of foreign equipment is huge,and the after-sales service of these devices cannot be guaranteed, even some of themcannot suited to the unique glass production environment of China. Therefore, there isan urgent need to increase investment of domestic glass surface quality detectiontechnology, in order to develop the glass line detection of high-performance productswith independent property rights.Combined with the existing glass production line testing equipment, the paperanalyzed the optical properties common glass defects (such as bubbles, debris,scratches, etc.) in-depth, and developed a set of glass surface defect detection systemwith independent property rights. The system uses a blue LED light source as theirradiating light source, combining with the rear transmissive illumination mode andthe CCD industrial cameras to collect the images of glass defects. By building a highperformance image processing module, this paper used image enhancement techniquesto preprocess the defect image of the glass surface, and divided the image of thedefects in glass from the glass in the image of the entire frame segmentation. Finally,for each one having been extracted out of the glass surface defects, this paper furtherextracted the parameters of the relevant characteristics of these glass defects. Byregarding these characteristic parameters as a feature set of defect recognition, thispaper constructed high-speed glass surface defect recognition method to achieve thepurpose of classification of the common glass surface defects.In the end, this paper built an experimental system of glass surface defectdetection based on single-camera. This system was successfully carried out on-site installation and commissioning in the Luoyang Longmen glassworks, and did a lot ofexperiments. Field tests showed that the system can effectively detect a variety ofsurface defects in the glass production process, and the defect recognition rate is96%.Also the system is able to meet real-time requirements of the work site, and achievedthe desired effect.