傳統的自動光學檢查（AOI）系統於處理影像時存在準確度和速度問題。本研究提出了一種基於卷積神經網絡（CNN）的三階瑕疵檢測模型，以克服傳統AOI方法的缺點，此模型並可處理多重缺陷和多鏡頭情境。在第一階段使用CNN的ResNet50模型對兩個獨立的相機執行GO / NG產品分類，準確率為98.72％和98.52％。傳統的AOI系統在生產線上的整體準確度為85％，而本研究所提出的模型優於並可取代AOI系統。第二階段的CNN使用另一個獨立的ResNet50模型將NG產品分類到相對應的瑕疵類別，兩個相機的準確度分別為97.27％和92.50％，第二階段的CNN模型可以挽救被傳統AOI誤殺的良品，並取代人工的複檢工作。第三階段模型使用YOLOv3模型進行瑕疵檢測，從第二階段CNN模型的輸出中同時檢測單個產品上的多重缺陷，其模型mAP分別為84.54％和81.31%。通過三階分道策略，用於AOI系統的三階段CNN模型實現了對高速生產線的高精度和高速度要求。

Conventional rule-based automatic optical inspection (AOI) systems by image processing suffer from precision and velocity issues. This study presents a three-phase model for defect detection based on convolutional neural network (CNN) to overcome the disadvantages of conventional AOI methods. The proposed model deals with multiple-defect and multiple-lens situations. The phase I CNN model using ResNet50 performs GO/NG classification of products with accuracies of 98.72% and 98.52% for two independent cameras. Moreover, the proposed model outperforms and replaces the rule-based AOI system, the accuracy of which is 85%. The phase II CNN is another independent ResNet50 that classifies NG products into defect categories with accuracies of 97.27% and 92.50% for two cameras. Moreover, the phase II CNN model rescues good products mistakenly killed and replaces the re-inspection labors. The phase III CNN model is a YOLOv3-based CNN to simultaneously detect multiple defects and their positions in a single product. The mAPs are 84.54% and 81.31% for the two cameras. By using the three-phase CNN strategy for AOI, the proposed model achieves high precision and velocity, reduces labor power, and facilitates quality assurance for high-speed production line.

Acciani, G., Brunetti, G., & Fornarelli, G. (2006). Application of neural networks in optical inspection and classification of solder joints in surface mount technology. IEEE Transactions on Industrial Informatics, 2(3), 200-209.
Ahn, N., Kang, B., & Sohn, K. A. (2018). Fast, accurate, and lightweight super-resolution with cascading residual network. In Proceedings of the European Conference on Computer Vision (ECCV) (pp. 252-268).
Albawi, S., Mohammed, T. A., & Al-Zawi, S. (2017, August). Understanding of a convolutional neural network. In 2017 International Conference on Engineering and Technology (ICET) (pp. 1-6). Ieee.
Bejnordi, B. E., Blankevoort, T., & Welling, M. (2019). Batch-shaping for learning conditional channel gated networks. arXiv preprint arXiv:1907.06627.
Bethea, R. M., & Rhinehart, R. R. (2019). Applied engineering statistics. Routledge.
Bock, S., Goppold, J., & Weiß, M. (2018). An improvement of the convergence proof of the ADAM-optimizer. arXiv preprint arXiv:1804.10587.
Byeon, Y. H., & Kwak, K. C. (2017, July). A performance comparison of pedestrian detection using faster RCNN and ACF. In 2017 6th IIAI International Congress on Advanced Applied Informatics (IIAI-AAI) (pp. 858-863). IEEE.
Chen, H., Pang, Y., Hu, Q., & Liu, K. (2020). Solar cell surface defect inspection based on multispectral convolutional neural network. Journal of Intelligent Manufacturing, 31(2), 453-468.
Cho, J., Lee, K., Shin, E., Choy, G., & Do, S. (2015). How much data is needed to train a medical image deep learning system to achieve necessary high accuracy? arXiv preprint arXiv:1511.06348.
Dai, W., Mujeeb, A., Erdt, M., & Sourin, A. (2020). Soldering defect detection in automatic optical inspection. Advanced Engineering Informatics, 43, 101004.
Everingham, M., Van Gool, L., Williams, C. K., Winn, J., & Zisserman, A. (2010). The pascal visual object classes (voc) challenge. International Journal of Computer Vision, 88(2), 303-338.
Feliciano, F. F., Leta, F. R., & Mainier, F. B. (2015). Texture digital analysis for corrosion monitoring. Corrosion Science, 93, 138-147.
Girshick, R. (2015). Fast r-cnn. In Proceedings of the IEEE international conference on computer vision (pp. 1440-1448).
Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2014). Rich feature hierarchies for accurate object detection and semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 580-587).
Han, L., Yu, C., Xiao, K., & Zhao, X. (2019). A new method of mixed gas identification based on a convolutional neural network for time series classification. Sensors, 19(9), 1960.
He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770-778).
Huang, R., Pedoeem, J., & Chen, C. (2018, December). YOLO-LITE: a real-time object detection algorithm optimized for non-GPU computers. In 2018 IEEE International Conference on Big Data (Big Data) (pp. 2503-2510). IEEE.
Huang, Y., Qiu, C., Wang, X., Wang, S., & Yuan, K. (2020). A compact convolutional neural network for surface defect inspection. Sensors, 20(7), 1974.
Jiang, D., Qi, G., Hu, G., Mazur, N., Zhu, Z., & Wang, D. (2020). A residual neural network based method for the classification of tobacco cultivation regions using near-infrared spectroscopy sensors. Infrared Physics & Technology, 111, 103494.
Kim, P. (2017). Convolutional neural network. In MATLAB deep learning (pp. 121-147). Apress, Berkeley, CA.
Kim, S., Kim, W., Noh, Y. K., & Park, F. C. (2017, May). Transfer learning for automated optical inspection. In 2017 International Joint Conference on Neural Networks (IJCNN) (pp. 2517-2524). IEEE.
Kingma, D. P., & Ba, J. (2014). Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980.
Komatsuzaki, A. (2019). One epoch is all you need. arXiv preprint arXiv:1906.06669.
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems, 25, 1097-1105.
Li, D., Zhang, X., Jiang, T., & Zhang, M. (2013, December). Application of improved wavelet threshold de-noising algorithm in work piece inspection. In 2013 6th International Congress on Image and Signal Processing (CISP) (Vol. 1, pp. 256-260). IEEE.
Li, Y., & Lyu, S. (2018). Exposing deepfake videos by detecting face warping artifacts. arXiv preprint arXiv:1811.00656.
Liao, H. C., Lim, Z. Y., Hu, Y. X., & Tseng, H. W. (2018, July). Guidelines of automated optical inspection (AOI) system development. In 2018 IEEE 3rd International Conference on Signal and Image Processing (ICSIP) (pp. 362-366). IEEE.
Liu, Y. H. (2018, September). Feature extraction and image recognition with convolutional neural networks. In Journal of Physics: Conference Series (Vol. 1087, No. 6, p. 062032). IOP Publishing.
Jian, C., Gao, J., & Ao, Y. (2017). Automatic surface defect detection for mobile phone screen glass based on machine vision. Applied Soft Computing, 52, 348-358.
Narin, A., Kaya, C., & Pamuk, Z. (2021). Automatic detection of coronavirus disease (covid-19) using x-ray images and deep convolutional neural networks. Pattern Analysis and Applications, 1-14.
Nguyen-Meidine, L. T., Granger, E., Kiran, M., Dolz, J., & Blais-Morin, L.-A. (2020). Joint progressive knowledge distillation and unsupervised domain adaptation. arXiv preprint arXiv:2005.07839.
Parakontan, T., & Sawangsri, W. (2019, June). Development of the Machine Vision System for Automated Inspection of Printed Circuit Board Assembl. In 2019 3rd International Conference on Robotics and Automation Sciences (ICRAS) (pp. 244-248). IEEE.
Pathak, A. R., Pandey, M., & Rautaray, S. (2018). Application of deep learning for object detection. Procedia computer science, 132, 1706-1717.
Qian, N. (1999). On the momentum term in gradient descent learning algorithms. Neural Networks, 12(1), 145-151.
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 779-788).
Redmon, J., & Farhadi, A. (2018). Yolov3: An incremental improvement. arXiv preprint arXiv:1804.02767.
Ren, G., Cao, Y., Wen, S., Huang, T., & Zeng, Z. (2018). A modified Elman neural network with a new learning rate scheme. Neurocomputing, 286, 11-18.
Ren, S., He, K., Girshick, R., & Sun, J. (2015). Faster r-cnn: Towards real-time object detection with region proposal networks. arXiv preprint arXiv:1506.01497.
Sadykova, D., Pernebayeva, D., Bagheri, M., & James, A. (2019). IN-YOLO: Real-time detection of outdoor high voltage insulators using UAV imaging. IEEE Transactions on Power Delivery, 35(3), 1599-1601.
Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.
Smith, L. N. (2018). A disciplined approach to neural network hyper-parameters: Part 1--learning rate, batch size, momentum, and weight decay. arXiv preprint arXiv:1803.09820.
Tian, Y., Yang, G., Wang, Z., Wang, H., Li, E., & Liang, Z. (2019). Apple detection during different growth stages in orchards using the improved YOLO-V3 model. Computers and Electronics in Agriculture, 157, 417-426.
Tsai, Y. H., Lyu, N. Y., Jung, S. Y., Chang, K. H., Chang, J. Y., & Sun, C. T. (2019, July). Deep Learning Based AOI System with Equivalent Convolutional Layers Transformed from Fully Connected Layers. In 2019 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM) (pp. 103-107). IEEE.
Vithu, P., & Moses, J. (2016). Machine vision system for food grain quality evaluation: A review. Trends in Food Science & Technology, 56, 13-20.
Wan, H., Chen, L., Song, H., & Yang, J. (2017, November). Dorsal hand vein recognition based on convolutional neural networks. In 2017 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) (pp. 1215-1221). IEEE.
Wang, W. C., Chen, S. L., Chen, L. B., & Chang, W.-J. (2016). A machine vision based automatic optical inspection system for measuring drilling quality of printed circuit boards. IEEE Access, 5, 10817-10833.
Wibowo, F. S., Wahyudi, Y. R., & Lin, H. I. (2020, August). Development and Implementation of Novel Six-Sided Automated Optical Inspection for Metallic Objects. In 2020 International Conference on Advanced Robotics and Intelligent Systems (ARIS) (pp. 1-6). IEEE.
Yadav, A., Verma, V. K., Pal, V., & Singh, S. (2021, March). Automatic Detection of COVID 19 Infection Using Deep Learning Models from X-Ray Images. In IOP Conference Series: Materials Science and Engineering (Vol. 1099, No. 1, p. 012050). IOP Publishing.
Yadav, S., & Shukla, S. (2016, February). Analysis of k-fold cross-validation over hold-out validation on colossal datasets for quality classification. In 2016 IEEE 6th International conference on advanced computing (IACC) (pp. 78-83). IEEE.
Yang, Y., Yang, R., Pan, L., Ma, J., Zhu, Y., Diao, T., & Zhang, L. (2020). A lightweight deep learning algorithm for inspection of laser welding defects on safety vent of power battery. Computers in Industry, 123, 103306.
Yu, C., Qi, X., Ma, H., He, X., Wang, C., & Zhao, Y. (2020). LLR: Learning learning rates by LSTM for training neural networks. Neurocomputing, 394, 41-50.
Zou, Z., Shi, Z., Guo, Y., & Ye, J. (2019). Object detection in 20 years: A survey. arXiv preprint arXiv:1905.05055.