製造產業長期以來，就是支持全球經濟的重要經濟活動之一。當前，有許多類型的製造商在傳統或現代環境下運行。利用傳統環境的製造商傾向於更多地使用人力資源來執行其操作，例如移動，檢查或生產物品。此外，現代的機器人比傳統的機器人更傾向於使用機器人或機器，並且許多操作是自動完成的，無需人工幫助。因此，它有助於提高製造企業的運營效率以及收入。
儘管一些新建的工廠使用智能數位設備，但傳統的工廠仍然具有手動開關，照明燈，數字顯示器，甚至指針式儀表板。因此，此類公司無法在互聯網上傳輸數位化訊息或資訊。同樣，由於他們需要工人檢查某些類型設備的狀態或數據值，因此可能浪費人力、時間和成本。因此，為了克服這些問題，可以實現自動化系統。自動化系統可以幫助傳統製造工廠獲得更高的運營效率並獲得更高的收入。因此，它們在進入工業4.0時代時可以提高競爭力。
基於上述陳述，我們提出了幾種系統來支持基於人工智能（AI）和深度學習的製造環境，以升級環境本身並提昇運營和製造收益的有效性。該系統包括自動圖像識別，產品缺陷檢測，火災和/或煙霧檢測以及水檢測。所有系統都是自動化系統，由AI自動完成和智慧化操作。
核心技術提出了一種用於多尺度缺陷圖像檢測的改進模型。我們修改了SSD 512的演算法部分流程，因為它可以檢測高維圖像。所提出的系統針對不同尺寸的圖像進行模型調整，並使其調整以適合自動檢測的模型，可以達成快速訓練模型、高精準度偵測的目標，適合在製造產業應用在工業製造產線上使用，進而發展智慧工廠適用的AI平台。

Manufacturing is one of the principal parts of business which has been supporting the world’s economy for a long time. Currently, there are many types of manufacturers running under traditional or modern environments. Manufactures which utilize traditional environments tend to use human resources more to execute its operations, such as moving, checking, or producing items. Besides, modern ones tend to use robots or machines more than the traditional ones and many operations are done automatically without the help of humans. Thus it helps to increase the effectiveness of a manufacture’s operations as well as the income.
Though some newly built factories, use smart digital devices, traditional ones still have manual switches, lights, numerical displays, even pointer-type dashboard. So, such companies can not transmit digitized signal or values on the internet. Also, since they need workers to check status or data values of certain types of equipment, it may waste human resources, time, as well as cost. Therefore, to overcome these issues, an automated system can be implemented. Automated systems can help traditional manufacturing plants to get greater operational efficiency and make a higher income. Consequently, their competitiveness can be upgraded as they go towards the era of Industry 4.0. Based on the above statements, we propose several systems to support manufacturing environments based on artificial intelligence (AI) and deep learning in order to upgrade the environments themselves and escalate the effectiveness of the operations and manufacturing’s gain. The systems include automated image recognition, product’s defect detection, fire and/or smoke detection, and water detection. All of the systems are automated systems, done and operated by AI.
This work proposes an improved model for multi-scale defect image detection. We modified the algorithm of SSD 512 since it can detect higher-dimensional images. The proposed system makes model adjustments for images of different sizes and adjusts it fit the model suitable for automated detection.

[1] J. Redmon, S. Divvala, R. Girshick, and A. Farhadi, "You Only Look Once: Unified, Real-Time Object Detection", in IEEE conference on computer vision and pattern recognition, Las Vegas, NV, USA, 2016, pp. 779-788.

[2] C. Y, Cao. J. C. Zheng, Y. Q. Huang, J. Liu, and C. F. Yang, "Investigation of a Promoted You Only Look Once Algorithm and Its Application in Traffic Flow Monitoring", Appl. Sci. Vol. 9, No. 17, 2019, pp. 1-14.

[3] W. Liu, D. Anguelov, D. Erhan, C. Szegedy, S. Reed, C. Y. Fu, and A. C. Berg, "SSD: Single Shot Multibox Detector", in European conference on computer vision, Amsterdam, 2016, pp. 21-37.

[4] F. Yang, H. Chen, J. Li, F. Li, L. Wang, and X. Yan, "Single Shot Multibox Detector With Kalman Filter for Online Pedestrian Detection in Video", IEEE Access, 2019, pp. 15478-15488.

[5] R. Girshick, J. Donahue, T. Darrel, and J. Malik, "Rich feature hierarchies for accurate object detection and semantic segmentation", in IEEE conference on computer vision and pattern recognition, Columbus, Ohio, 2014, pp. 580-587.

[6] R. Girshick, "Fast R-CNN", in IEEE International Conference on Computer Vision, Santiago, Chile, 2015, pp. 1440-1448.

[7] S. Ren, K. He, R. Girshick, and J. Sun, "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks", in Proceedings of the 28th International Conference on Neural Information Processing Systems, Montreal, Canada, 2015, pp. 91-99.

[8] K. He, G. Gkioxari, P. Doll´ar, and R. Girshick, "Mask R-CNN", in IEEE international conference on computer vision, Venice, Italy, 2018, pp. 2961-2969.

[9] R. Zhang, C. Cheng, X. Zhao, and X. Li, "Multiscale Mask R-CNN-Based Lung Tumor Detection Using PET Imaging", Artificial Intelligence in Molecular Imaging Clinics, Vol. 18, 2019, pp. 1-8.

[10] J. Min, Y. Kim, S. Lee, T. W. Jang, I. Kim, J. Song, "The Fourth Industrial Revolution and Its Impact on Occupational Health and Safety, Worker's Compensation and Labor Conditions", 2019.

[11] J. Hull, "The second industrial revolution and the staples frontier in Canada: rethinking knowledge and history", Scientia Canadensis, Canada, 1994, Vol.18, No. 1, pp. 22-37.

[12] A. Toffler, "The third wave", Bantam books, New York, 1980, Vol. 484.

[13] J. Bloem, M. V. Doorn, S. Duivestein, D. Excoffier, R. Maas, E. V. Ommeren, "The fourth industrial revolution: things to tighten the link between IT and OT contents", Groningen Sogeti VINT, 2014.

[14] A. Mikołajczyk and M. Grochowski, "Data augmentation for improving deep learning in image classfication problem", International Interdisciplinary PhD Workshop, Swinoujście, 2018, pp. 117-122.

[15] D. Li, D. Yu, "Deep Learning: Methods and Applications, Foundations and Trends in Signal Processing", Now Publishers, 2014.

[16] Y. Bengio, "Learning deep architectures for AI, Foundations and Trends in Machine Learning", Vol. 2, 2009, pp. 1-127.

[17] Y. Bengio, A. Courville, and P. Vincent, "Representation learning: A review and new perspectives, Pattern Analysis and Machine Intelligence", IEEE Transactions, Vol. 35, 2019, pp. 1798-1828.

[18] J. Schmidhuber, "Deep learning in neural networks: An overview, Neural Networks", Vol. 61, 2015, pp. 85-117.

[19] Y. LeCun, Y. Bengio, and G. Hinton, "Deep Learning", Nature, Vol. 521, 2015, pp. 436-444.

[20] I. Arel, D. C. Rose, and T. P. Karnowski, "Deep machine learning-a new frontier in artificial intelligence research [research frontier]", Computation Intelligence Magazine, IEEE, Vol. 5, 2010, pp. 13-18.

[21] J. Kuruvilla, D. Sukumaran, A. Sankar, and S. P. Joy, "A review on image processing and image segmentation", International Conference on Data Mining and Advanced Computing, Ernakulam, 2016, pp. 198-203.

[22] T. R. Rao, "Metal casting: Principles and Practice", New Age International, 2007.

[23] K. I., "The future of manufacturing: 2020 and beyond", Industry Week Special Research Report, p. 12, 2016.

[24] R. Rajkolhe and J. Khan, "Defects, causes and their remedies in casting process: A review", International Journal of Research in Advent Technology, Vol. 2, No. 3, pp. 375–383, 2014.

[25] X. Li, S. K. Tso, X.-P. Guan, and Q. Huang, "Improving automatic detection of defects in castings by applying wavelet technique", IEEE Transactions on Industrial Electronics, Vol. 53, No. 6, pp. 1927–1934, 2006.

[26] J. Xin, C, Huang, "Fire Risk Assessment of Residential Buildings Based on Fire Statistics from China", Fire Technology, Vol. 50, pp. 1147-1161, 2014.

[27] Fire Service Bureau, Ministry of Public Security, China fire services, China Personnel Press, Beijing, 2012 (in Chinese).

[28] A. K. Soe, X. Zhang, "Fire detection by static image characteristics of the light blue flame using roi-based color intensity composition detection algorithm", ICIC Express Letters, Vol. 5, No. 12, December, 2011, pp. 4479-4486.

[29] V. Vipin, "Image Processing Based Forest Fire Detection", International Journal of Emerging Technology and Advanced Engineering, Vol. 2, February, 2012, pp. 87-95.

[30] S. Ham, B. Ko, J. Nam, "Fire-flame detection based on fuzzy finite automation", Pattern Recognition (ICPR), 2010 20th International Conference, 2010, pp. 3919-3922.

[31] Dimitropoulos, K. Tsalakanidou, Filareti, Grammalidis, Nikos, "Flame detection for video-based early fire warning systems and 3D visualization of fire propagation", Proceedings of the IASTED International Conference on Computer Graphics and Imaging (CGIM), 2012, pp. 209-216.

[32] X. Trong, Tung, Kim, Jong-Myon, "Fire flame detection in video sequences using multi-stage pattern recognition techniques", Engineering Applications of Artificial Intelligence, October, 2012, Vol. 25, No. 7, pp. 1365-1372.

[33] J. Zhou, X. Du, "Image recognition technology in fire detection", Fire Science and Technology, Vol. 26, No. 4, 2007, pp. 417-420.

[34] X. Zhang, F. Xu, Z. Song, Z. Mei, "Video flame detection algorithm based on multi-feature fusion technique", Proceedings of the 2012 24th Chinese Control and Decision Conference (CCDC), 2012, pp. 4291-4294.

[35] Y. Y. Yan, B. G. Shao, Y. W. Hong, B. G. Zhi, "Contour Extraction of Flame for Fire Detection", Advanced Materials Research, Manufacturing Science, and Technology, Vol 383-390, 2012, pp. 1106-1110.
[36] Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun, “Faster R-CNN: Towards Real-Time ObjectDetection with Region Proposal Networks”, 2016
[37] Kaiming He, Georgia Gkioxari, Piotr Dollár, Ross Girshick, “Mask R-CNN” , 2018