Industrial Defect Detection Using Modern Computer Vision Techniques

  • Authors

    • Dr. Idris Mohammed Department of Physics, Khartoum Science University, Sudan. Author

    DOI:

    https://doi.org/10.67228/30713498/IJADSMC-2023PII0Q6X

    Published 08-05-2023

  • Industrial Inspection, Defect Detection, Computer Vision, Deep Learning, Convolutional Neural Networks, Quality Control, Automation

    Issue

    Vol. 6 No. 2 (2023)

    Section

    Articles

    How to Cite

    [1]
    I. Mohammed, “Industrial Defect Detection Using Modern Computer Vision Techniques”, IJADSMC, vol. 6, no. 2, pp. 01–15, Aug. 2023, doi: 10.67228/30713498/IJADSMC-2023PII0Q6X.

  • Abstract

    The manufacturing process of the industry requires very high quality standards to guarantee reliability and suitability of the product, level of customer satisfaction and economical efficiency of the product. The fact is that even small defects may cause some heavy loss of money, risks, and damaged reputation. Conventional defect inspection systems are often based on manual inspection and classical machine vision methods, which is time consuming, subjective and prone to errors and it is hard to scale into large scale production set ups today. The recent fast development of computer vision technology using the methods of deep learning and artificial intelligence has also changed the nature of automated defect detection systems dramatically. The current computer vision methods including convolutional neural networks (CNNs), transformer networks, and mixed learning systems have shown outstanding results when detecting surface defects, structural deformities and functional defects in many industrial settings. The paper is an in-depth research on the industrial defect detection with the help of the modern computer vision methodology, including the theoretical bases, the current approach methods, and the practical aspect of their implementation. The suggested framework incorporates the advanced image acquisition, preprocessing and deep features extraction, and intelligent classification algorithms to obtain a high level of detection and strong stability in adverse industrial environments. Extensive trials indicate the efficiency of the contemporary vision-based solutions to decrease false positives, accelerate the speed of inspection, and maintain the quality control. The outcomes verify that defect detection systems based on deep learning are far more effective than conventional inspection and can be utilized in practice in the industrial environment on a real-time basis. The paper ends with the determination of the current limitations and research areas to explore in the future in the field of scalable and explainable defect detectors in smart factories.

  • References

    [1] Xie, X. (2008). A review of recent advances in surface defect detection using texture analysis techniques. Electronic Letters on Computer Vision and Image Analysis, 7(3), 1–22.

    [2] Tsai, D. M., & Hsieh, C. Y. (1999). Automated surface inspection using wavelet reconstruction. Pattern Recognition, 32(8), 1317–1330.

    [3] Kumar, A., & Pang, G. K. H. (2002). Defect detection in textured materials using Gabor filters. IEEE Transactions on Industry Applications, 38(2), 425–440.

    [4] Zhang, Y., & Chen, Y. (2013). Defect detection using morphological image processing. Journal of Manufacturing Systems, 32(2), 221–230.

    [5] Vapnik, V. (1998). Statistical learning theory. New York: Wiley.

    [6] Cortes, C., & Vapnik, V. (1995). Support-vector networks. Machine Learning, 20(3), 273–297.

    [7] Cover, T., & Hart, P. (1967). Nearest neighbor pattern classification. IEEE Transactions on Information Theory, 13(1), 21–27.

    [8] Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5–32.

    [9] LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278–2324.

    [10] Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). ImageNet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems (NIPS), 25, 1097–1105.

    [11] Simonyan, K., & Zisserman, A. (2015). Very deep convolutional networks for large-scale image recognition. International Conference on Learning Representations (ICLR).

    [12] He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 770–778.

    [13] Tan, M., & Le, Q. (2019). EfficientNet: Rethinking model scaling for convolutional neural networks. International Conference on Machine Learning (ICML).

    [14] Dosovitskiy, A., et al. (2021). An image is worth 16x16 words: Transformers for image recognition at scale. International Conference on Learning Representations (ICLR).

    [15] Bergmann, P., Fauser, M., Sattlegger, D., & Steger, C. (2019). MVTec AD—A comprehensive real-world dataset for unsupervised anomaly detection. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

  • Downloads