本研究提出了一種深度學習方法，使用跨機器數據概念漂移進行CNC工具機中的多狀態鐵屑狀態識別。本研究的目的是考慮在進行跨多機台數據資料學習的過程，會因為不同資料集之間的影像差異，包括影像色調、背景光源、錄影的角度與品質差異等因素，造成在訓練資料時的準確度無法提升。為了解決這個問題，將這些新的數據包括進來是一個直接的思路。在本研究中，為了獲得更好的學習效果，考慮了針對這些新數據的概念漂移方法。實驗的結果得出在加上2%調整後的影像資料準確率可提升至93.94%，比較初始模型(initial model)的結果提高了11%。同時也去調整訓練的次數(epoch)與學習率(learning rate)對訓練結果的影響，結果顯示增加訓練的次數無法顯著的提升準確率，但調整學習率的大小將會大幅地影響最終的學習結果。本研究的另一項任務是鐵屑量的狀態分類，透過將原始影像資料從RGB型態轉灰階影像，並進行二值化後，依照不同的灰度值大小對鐵屑量進行分類，並透過前一項任務所訓練的模型，將分類的結果印在每一幀的影像上，目的在於未來進行CNC工具機水柱清潔路徑的規劃能依據機台上遍布的鐵屑多寡，進行路徑最佳化。

In this study, a deep learning approach for considering multi-state metal chips status recognition in CNC machine tools with the use of concept drift for cross machines data is proposed. The purpose of this study is to consider the process of cross-machine data learning, in which the recognition accuracy may be degraded due to the image differences between different datasets, including the image color tone, the background light source, the recording angle and the quality differences. To include those new data become a straightforward idea to resolve this problem. In this study, in order to have better learning perform, concept drift approaches are considered for those new data. The experimental results show that the accuracy of the image data with the 2% adjustment can be increased to 93.94%, which is 11% higher than the result of the initial model. we also adjust the training epoch and learning rate and the results show that increasing the number of training cannot significantly improve the accuracy rate, but adjusting the size of the learning rate will significantly affect the final learning results. Another task of this study is the state classification of metal chips volume rather than only having chips or not. The classification results can be used in the planning of the cleaning path in the future based on the amount of metal chips spread on the machine.

[1] T. Kalsoom, N. Ramzan and S. Ahmed, "Societal impact of IoT-lead smart factory in the context of industry 4.0," in 2020 International Conference on UK-China Emerging Technologies (UCET), Glasgow, UK, pp. 1-5, 2020, doi: 10.1109/UCET51115.2020.9205484.
[2] M. K. Habib and C. Chimsom, "Industry 4.0: sustainability and design principles," in 2019 20th International Conference on Research and Education in Mechatronics (REM), Wels, Austria, pp. 1-8, 2019, doi: 10.1109/REM.2019.8744120.
[3] A. Adetunla and N. Madonsela, "Harnessing industry 4.0 technology to improve productivity in manufacturing industries: a case of CNC based welding factory," in 2021 IEEE 12th International Conference on Mechanical and Intelligent Manufacturing Technologies (ICMIMT), Cape Town, South Africa, pp. 180-185, 2021, doi: 10.1109/ICMIMT52186.2021.9476203.
[4] D. Arora, M. Garg and M. Gupta, "Diving deep in deep convolutional neural network," in 2020 2nd International Conference on Advances in Computing, Communication Control and Networking (ICACCCN), Greater Noida, India, pp. 749-751, 2020, doi: 10.1109/ICACCCN51052.2020.9362907.
[5] K. He, X. Zhang, S. Ren and J. Sun, "Deep residual learning for image recognition," in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, pp. 770-778, 2016.
[6] J. Lu, A. Liu, F. Dong, F. Gu, J. Gama and G. Zhang, "Learning under concept drift: a review," IEEE Transactions on Knowledge and Data Engineering, vol. 31, no. 12, pp. 2346-2363, Dec.2019, doi: 10.1109/TKDE.2018.2876857.
[7] Y. Chen, Z. Chai, Y. Cheng and H. Rangwala, "Asynchronous federated learning for sensor data with concept drift," in 2021 IEEE International Conference on Big Data (Big Data), Orlando, FL, USA, pp. 4822-4831, 2021 doi: 10.1109/BigData52589.2021.9671924.
[8] N. Lu, G. Zhang, and J. Lu, “Concept drift detection via competence models,” Artificial Intelligence, vol. 209, pp. 11–28, 2014.
[9] S. Disabato and M. Roveri, "Learning convolutional neural networks in presence of concept drift," in 2019 International Joint Conference on Neural Networks (IJCNN), Budapest, Hungary, pp. 1-8, 2019, doi: 10.1109/IJCNN.2019.8851731.
[10] G. Ditzler, M. Roveri, C. Alippi and R. Polikar, "Learning in nonstationary environments: a survey," IEEE Computational Intelligence Magazine, vol. 10, no. 4, pp. 12-25, Nov. 2015, doi: 10.1109/MCI.2015.2471196.
[11] J. Gama, I. Žliobaitė, A. Bifet, M. Pechenizkiy and A. Bouchachia, "A survey on concept drift adaptation", ACM Computing Surveys (CSUR), vol. 46, no. 4, pp. 44, 2014.
[12] J. Konečny, ℓ H. B. McMahan, F. X. Yu, P. Richtárik, A. T. Suresh and D. Bacon, "Federated learning: strategies for improving communication efficiency", arXiv preprint arXiv:1610.05492, 2016.
[13] 李紀萱, "基於視覺應用於CNC工具機之鐵屑定位與安全警示系統", 碩士論文, 國立臺灣科技大學電機工程系, 台北市, 2023.
[14] M. Chen, X. Dai, Z. Lv, F. Sun and J. Yuan, "Terrain classification in field environment with grid-based SURF feature," in Proceedings of the 33rd Chinese Control Conference, Nanjing, China, pp. 8253-8258, 2014, doi: 10.1109/ChiCC.2014.6896383.
[15] L. Li, X. An and Q. Gu, "Research on grid-based abnormal behavior identification method of ships," in 2021 IEEE 2nd International Conference on Information Technology, Big Data and Artificial Intelligence (ICIBA), Chongqing, China, pp. 430-434, 2021, doi: 10.1109/ICIBA52610.2021.9688328.
[16] D. Gorgan et al., "Grid based data processing tools and applications for black sea catchment basin," in Proceedings of the 6th IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems, Prague, Czech Republic, pp. 223-228, 2011, doi: 10.1109/IDAACS.2011.6072745.
[17] S. Choi and S. Han, "Fast image retrieval with grid-based keypoint detector and binary descriptor," in 2014 International Conference on Information and Communication Technology Convergence (ICTC), Busan, Korea (South), pp. 679-680, 2014, doi: 10.1109/ICTC.2014.6983253.
[18] X. Wang and Z. Hu, "Grid-based pavement crack analysis using deep learning," in 2017 4th International Conference on Transportation Information and Safety (ICTIS), Banff, AB, Canada, pp. 917-924, 2017, doi: 10.1109/ICTIS.2017.8047878.
[19] Y. LeCun, Y. Bengio, and G. Hinton, "Deep learning," nature, vol. 521, no. 7553, pp. 436-444, 2015.
[20] D. Han, Q. Liu, and W. Fan, "A new image classification method using CNN transfer learning and web data augmentation," Expert Systems with Applications, vol. 95, pp. 43-56, 2018.
[21] X. Chi, S. Huang and J. Li, "Handwriting recognition based on resnet-18," in 2021 2nd International Conference on Big Data & Artificial Intelligence & Software Engineering (ICBASE), Zhuhai, China, pp. 456-459, 2021, doi: 10.1109/ICBASE53849.2021.00091.
[22] X. Han and R. Jin, "A small sample image recognition method based on ResNet and transfer learning," in 2020 5th International Conference on Computational Intelligence and Applications (ICCIA), Beijing, China, pp. 76-81, 2020 doi: 10.1109/ICCIA49625.2020.00022.
[23] H. -C. Shin et al., "Deep convolutional neural networks for computer-aided detection: CNN architectures, Dataset Characteristics and Transfer Learning," IEEE Transactions on Medical Imaging, vol. 35, no. 5, pp. 1285-1298, May.2016, doi: 10.1109/TMI.2016.2528162.
[24] D. P. Kingma and J. Ba, "Adam: a method for stochastic optimization," arXiv preprint arXiv:1412.6980, 2014.
[25] Institute of Electrical and Electronics Engineers (IEEE), "IEEE approved draft guide for architectural framework and application of federated machine learning," IEEE P3652.1/D6.1, pp.1-70, Sept. 2020.
[26] P. K. Mallapragada, R. Jin, A. K. Jain and Y. Liu, "Semiboost: boosting for semi-supervised learning," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 31, no. 11, pp. 2000-2014, Nov. 2009, doi: 10.1109/TPAMI.2008.235.
[27] N. Sharma and M. Dutta, "Recommender system designed using an ensemble approach to semi-supervised learning," in 2022 International Conference on Innovative Computing, Intelligent Communication and Smart Electrical Systems (ICSES), Chennai, India, pp. 1-5, 2022, doi: 10.1109/ICSES55317.2022.9914228.
[28] C. A. da S. Barreto, A. M. de P. Canuto, J. C. Xavier, A. C. Gorgônio, D. F. A. Lima and R. R. F. da Costa, "Two novel approaches for automatic labelling in semi-supervised methods," in 2020 International Joint Conference on Neural Networks (IJCNN), Glasgow, UK, pp. 1-8, 2020, doi: 10.1109/IJCNN48605.2020.9207079.
[29] G. Cai, Y. Wang, L. He and M. Zhou, "Unsupervised domain adaptation with adversarial residual transform networks," IEEE Transactions on Neural Networks and Learning Systems, vol. 31, no. 8, pp. 3073-3086, Aug. 2020, doi: 10.1109/TNNLS.2019.2935384.
[30] P. Rosin, "Thresholding for change detection," in Sixth International Conference on Computer Vision (IEEE Cat. No.98CH36271), Bombay, India, pp. 274-279, 1998, doi: 10.1109/ICCV.1998.710730.