Study On The Quality Of Glass On-line Vision Detection System

With the rapid development of economy, more and more demands for glass products andquality requirements higher. Due to the production technology and process aspects, there maybe defects in glass production. These defects reduce the optical performance and the quality ofglass, reducing the quality of glass products. So in the glass production line, the defectformation in glass which output from kiln requires real-time detection and cleared in a timelymanner, to ensure the quality of glass products.The traditional glass quality detection mainly adopts artificial detection method. Manualtesting is not only large workload, but easily influenced by testing personnel subjective factors.Easily lead to missed defect on the glass surface, especially undetected flaws such as lessdeformation and distortion. Greatly reduces the surface quality of the glass, so it can notensure the efficiency and precision of detection. The machine vision detection technology tothe manual detection has the advantages of fast speed, accuracy and stable and long timeoperation etc.Glass defect inspection system introduced in this paper is used the machine visiontechnology, completed the extraction and identification of glass defection, to meet therequirements of glass surface defect detection. The system includes image acquisition, imagetransmission of Gigabit Ethernet port and image processing. Image acquisition includes thecamera of choice, the choice of light source, selection of lens, and contrast the way of lightingexperiment, adjusting the light source, camera and glass to the optimum position for imageacquisition. To extracts high quality pictures for easily post treatment. Images were collectedby using the camera of Gigabit Ethernet port, via protocol communication of Gigabit Ethernetto transmit data. Acquired image data are processed through GigE transmitted to the hostcomputer, and output the processed results to judge the glass quality. If it is not qualified thenoutput the type and size of defects, and processes in real-time software interface by coordinatesearch to identify defects. Meanwhile through the man-machine interface controls to realizethe preservations of interest defect image, make batch glass quality easily to statistics analysis.