The System Research on Automatic Defect Detection of Glasses

Abstract

A glasses defect inspection system is researched and developed according to the principle of light scattering and visual inspection method, which is based on the machine vision. In order to achieve the classification of glasses, the functions of image acquisition, simple image processing, grading and sorting of glasses are designed in this system. Operation of parallel structure is adopted in this system. Forwardlighting by low-angled-ring-LED is used to get the clear images of the defects in or on the surface of the glasses, such as speckles, impurities, feathers and so on. Image processing and glasses grading based on normalization algorithms are used to acquire the identifying information of various defects of glasses. Experiment results show that the detection system, which can detect all the kinds of defects of the glasses, has high processing speed and strong anti-interference capacity. The size of the smallest defect that can be detected by the system is 0.03mm, and the average detection time of each glass is less than 2s.

Optical Imaging Deformation Inspection and Quality Level Determination of Multifocal Glasses

“…Wang et al [23] performed a symmetric energy analysis on different color spaces to evaluate the coating quality of glass. Yao et al [24] employed low-angle LED (light-emitting diode) illumination and image normalization techniques for computer vision and lens categorization to identify glass surface defects. Karangwa et al [25] developed a visual inspection platform integrating deep learning models and semantic segmentation to detect and classify the visual defects of optical glass.…”