ABSTRACT
Automatic Optical Inspection (AOI) offers a range of solutions to meet the requirements of every production facility, attracting significant interest of manufacturers of various industries. One important AOI application in the glass industry is to detect bubble defects in spherical glass, especially those used for making high-end lenses. Nevertheless, the AOI process must make the inspection decision in a reliable manner. Another challenge is that glass bubble defects are nearly transparent and can be captured by cameras only from certain viewing angles and with the aid of a specially engineered lighting mechanism. In this paper, an Artificial Intelligence method based on AOI (AI-AOI) is proposed to address the need of the glass industry as aforementioned. In specifics, our proposed method employs (1) a specially designed back-lighting mechanism to illuminate the hardly visible glass bubble defects, (2) Otsu thresholding image-segmentation method to obtain the distortion part and the core part of the defects and eliminate the fake defects caused by dust particles, and (3) a novel AI-based bubble-defect detection method capable of capturing the bubble defects as small as a few millimeters in diameter. The initial experimental results validate the feasibility of the proposed AOI method with an accuracy of 95%. If we exclude factors such as scratches by humans or the presence of dust particles in the inspection room, our method can achieve a recognition rate of 100%.
- Chang-Hwan Oh, Hyonam Joo, and Keun-Ho Rew, "Detecting low-contrast defect regions on glasses using highly robust model-fitting estimator," In Proc. International Conference on Control, Automation and Systems, 2007, pp.2138--2141.Google Scholar
- Xiangqian Peng, Youping Chen, Wenyong Yu, Zude Zhou, and Guodong Sun, "An online defects inspection method for float glass fabrication based on machine vision," International Journal of Advanced Manufacturing Technology, vol. 39, pp.1180--1189, 2008.Google ScholarCross Ref
- Francesco Adamo, Filippo Attivissimo, Attilio Di Nisio, and Mario Savino, "A low-cost Inspection system for online defects assessment in satin glass," Measurement, vol. 42, no. 9, pp.1304--1311, 2009.Google ScholarCross Ref
- Jie Zhao, Qing-Jie Kong, Xu Zhao, Jiapeng Liu, and Yuncai Liu, "A method for detection and classification of glass defects in low-resolution images," In Proc. 6th International Conference on Image and Graphics, 2011, pp.642--647.Google Scholar
- Weon-Jae Ryu, Young-June Kang, Sung-Hoon Baik, and Shin-Jae Kang, "A study on the 3-D measurement by using digital projection moiré method," Optik, vol. 119, no. 10, pp.453--458, 2008.Google ScholarCross Ref
- Quoc V. Le, Wii Y. Zou, Serena Y. Yeung, and Andrew Y. Ng, "Learning hierarchical invariant spatio-temporal features for action recognition with independent subspace analysis," In Proc. IEEE Conf. Computer Vision and Pattern Recognition, 2011, pp.3361--3368.Google Scholar
- Minmin Chen, Zhixiangilian Xu, Kilian Q. Weinberger, and Fei Sha, "Marginalized denoising autoencoders for domain adaptation," In Proc. Int'l Conference Machine Learning, 2012, pp.1627--1634.Google Scholar
- Takeo Kanade, Computer Recognition of Human Faces, Interdisciplinary Systems Research, vol. 47, Birkhauser Verlag, Basel und Stuttgart, 1977.Google ScholarCross Ref
- Karen Simonyan and Andrew Zisserman, "Very deep convolutional networks for large-scale image recognition," In Proc. International Conference on Learning Representations, 2015, pp.1--15.Google Scholar
Index Terms
- AI-based Automatic Optical Inspection of Glass Bubble Defects
Recommendations
Enhanced detection of diverse defects by developing lighting strategies using multiple light sources based on reinforcement learning
AbstractTraditional inspection systems with a single light source are not efficient at detecting a few particular defects with a single inspection. Unlike before, the multi-light source inspection environment allows us to extract more different defects in ...
Comments