臺灣博碩士論文加值系統

Construction industry is definitely a crucial industry in Taiwan due to its economic status. However, according statistics, the number of occupational fatalities of construction industry is the highest compared to the other industries. Among all occupational injuries, a research mentioned that head injuries are prong to be serious and often cause fatality. Even though the governments of diverse countries are dedicated to establishing related regulations, workers may still take off hard hats as it is hard to carry, inconvenient, or uncomfortably sultry, thus increasing the risks of occupational hazards. Accordingly, how to efficiently monitor the hard hat wearing becomes the crucial issue which this research would study. However, safety inspection practices still rely heavily on inspectors' manual monitoring and reporting on construction site at present based on the user interview, and it would be a huge burden of manpower to check whether workers wear hard hat or not all the times just for safety assurance. Therefore, the research aims to exploit deep learning method to develop a non-hard-hat use (NHU) automatic detection system for not only increasing the convenience of construction safety management but also reducing the manpower and time consumed. The research would utilize the data collected from internet and real construction site to train the deep learning network, MRCNN, and combined with different backbones and data augmentation strategies for inspecting whether it would influence Precision, Recall and Accuracy, improving prediction performance. The results indicated that data augmentation strategies do really influence the model performance. Finally, the best model would be selected and connected to LINE App for informing the hard hat wearing violation to on-site managers. the needs of sufficient accuracy, notification of the status of workers not wearing hard hats, and evidences saved for imposing fines.

Construction industry is definitely a crucial industry in Taiwan due to its economic status. However, according statistics, the number of occupational fatalities of construction industry is the highest compared to the other industries. Among all occupational injuries, a research mentioned that head injuries are prong to be serious and often cause fatality. Even though the governments of diverse countries are dedicated to establishing related regulations, workers may still take off hard hats as it is hard to carry, inconvenient, or uncomfortably sultry, thus increasing the risks of occupational hazards. Accordingly, how to efficiently monitor the hard hat wearing becomes the crucial issue which this research would study. However, safety inspection practices still rely heavily on inspectors' manual monitoring and reporting on construction site at present based on the user interview, and it would be a huge burden of manpower to check whether workers wear hard hat or not all the times just for safety assurance. Therefore, the research aims to exploit deep learning method to develop a non-hard-hat use (NHU) automatic detection system for not only increasing the convenience of construction safety management but also reducing the manpower and time consumed. The research would utilize the data collected from internet and real construction site to train the deep learning network, MRCNN, and combined with different backbones and data augmentation strategies for inspecting whether it would influence Precision, Recall and Accuracy, improving prediction accuracy. The results indicated that data augmentation strategies do really influence the model performance. Finally, the best model would be selected and connected to LINE App for informing the hard hat wearing violation to on-site managers. the needs of sufficient accuracy, notification of the status of workers not wearing hard hats, and evidences saved for imposing fines.

由於其經濟地位(GDP)，建築業在台灣絕對是至關重要的行業。可是，根據台灣職業安全衛生管理局的2017年年報統計，建築行業中的職業死亡人數卻是所有行業中最高的。在所有職業傷害中，一項研究提到頭部受傷可能很嚴重，並經常導致死亡。即使各國政府致力於製定相關法規，但由於難以攜帶，不便的悶熱等不舒服，工人可能仍會脫掉安全帽，從而增加了職業災害的風險，因此，如何有效地監測安全帽的佩戴成為該研究將要研究的關鍵問題。
根據訪談，已知目前的安全檢查實踐仍然重度依賴檢查員在施工現場的手動監測和報告，但為了安全保證而不停檢查工人是否始終戴安全帽將會是巨大的人力負擔，因此，本研究旨在探索深度學習方法，以開發一種未配戴安全帽自動檢測系統，從而增加了施工安全管理的便利性，並減少了人力和時間。該研究將利用網路和實際台灣施工現場收集的數據來訓練深度學習網絡MRCNN，並結合不同的骨幹和數據增強策略來檢查其是否會影響Precision和Recall以及Accuracy，從而提高預測準確性。結果顯示，數據增強策略是會影響的。最後，將選擇最佳模型並將其連接到LINE App，以將安全帽違規事項通知建築工地現場管理者，達到準確預測、通知、證據保存等使用者需求。

Abdul Hamid, A. R., Wan Yusuf, W. Z. and Singh, B. (2003, 2003). Hazards at Construction Sites,. Paper presented at the Proceeding of the 5th Asia-Pacific Strutural Engineering and Construction Conference.
Ahn, Y. S., Bena, J. F., & Bailer, A. J. (2004). Comparison of unintentional fatal occupational injuries in the Republic of Korea and the United States. Injury prevention : journal of the International Society for Child and Adolescent Injury Prevention, 10(4), 199-205. doi:10.1136/ip.2003.004895
Albawi, S., Mohammed, T. A., & Al-Zawi, S. (2017). Understanding of a convolutional neural network. Paper presented at the 2017 International Conference on Engineering and Technology (ICET).
Attwell, D., & Laughlin, S. B. (2001). An energy budget for signaling in the grey matter of the brain. Journal of Cerebral Blood Flow & Metabolism, 21(10), 1133-1145.
Awwad, R., El Souki, O., & Jabbour, M. (2016). Construction safety practices and challenges in a Middle Eastern developing country. Safety science, 83, 1-11.
Barro-Torres, S., Fernández-Caramés, T. M., Pérez-Iglesias, H. J., & Escudero, C. J. (2012). Real-time personal protective equipment monitoring system. Computer Communications, 36(1), 42-50.
BoL. (2013). US Bureau of Labor Statistics. Retrieved from https://www.bls.gov/opub/mlr/2013/
Chiocha, C., Smallwood, J., & Emuze, F. (2011). Health and safety in the Malawian construction industry. Acta Structilia, 18(1), 68-80.
Conchie, S. M., Moon, S., & Duncan, M. (2013). Supervisors’ engagement in safety leadership: Factors that help and hinder. Safety science, 51(1), 109-117.
Dong, S., He, Q., Li, H., & Yin, Q. (2015). Automated PPE misuse identification and assessment for safety performance enhancement. In ICCREM 2015 (pp. 204-214).
Du, S., Shehata, M., & Badawy, W. (2011). Hard hat detection in video sequences based on face features, motion and color information. Paper presented at the 2011 3rd International Conference on Computer Research and Development.
Evanoff, B., Dale, A. M., Zeringue, A., Fuchs, M., Gaal, J., Lipscomb, H. J., & Kaskutas, V. (2016). Results of a fall prevention educational intervention for residential construction. Safety science, 89, 301-307.
Farooqui, R. U., Arif, F., & Rafeeqi, S. (2008). Safety performance in construction industry of Pakistan. Paper presented at the First International Conference on Construction In Developing Countries, Karachi, Pakistan.
Girshick, R. (2015). Fast r-cnn. Paper presented at the Proceedings of the IEEE international conference on computer vision.
Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2014). Rich feature hierarchies for accurate object detection and semantic segmentation. Paper presented at the Proceedings of the IEEE conference on computer vision and pattern recognition.
Gittleman, J. L., Gardner, P. C., Haile, E., Sampson, J. M., Cigularov, K. P., Ermann, E. D., . . . Chen, P. Y. (2010). [Case Study] CityCenter and Cosmopolitan Construction Projects, Las Vegas, Nevada: Lessons learned from the use of multiple sources and mixed methods in a safety needs assessment. Journal of Safety Research, 41(3), 263-281. doi:https://doi.org/10.1016/j.jsr.2010.04.004
Green, L., & Tominack, G. (2012). Real-time proactive safety in construction. Power, 156(1), 62-65.
Grel, T. (2017). region of interest pooling explained. Retrieved from https://deepsense.ai/region-of-interest-pooling-explained/
Han, J., & Moraga, C. (1995). The influence of the sigmoid function parameters on the speed of backpropagation learning. Paper presented at the International Workshop on Artificial Neural Networks.
He, K., Gkioxari, G., Dollár, P., & Girshick, R. (2017). Mask r-cnn. Paper presented at the Proceedings of the IEEE international conference on computer vision.
He, K., Zhang, X., Ren, S., & Sun, J. (2015). Spatial pyramid pooling in deep convolutional networks for visual recognition. IEEE transactions on pattern analysis and machine intelligence, 37(9), 1904-1916.
He, K., Zhang, X., Ren, S., & Sun, J. (2016, 27-30 June 2016). Deep Residual Learning for Image Recognition. Paper presented at the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).
Hu, K., Rahmandad, H., Smith‐Jackson, T., & Winchester, W. (2011). Factors influencing the risk of falls in the construction industry: a review of the evidence. Construction Management and Economics, 29(4), 397-416. doi:10.1080/01446193.2011.558104
Jaderberg, M., Simonyan, K., & Zisserman, A. (2015). Spatial transformer networks. Paper presented at the Advances in neural information processing systems.
Jeong, B. Y. (1998). Occupational deaths and injuries in the construction industry. Applied Ergonomics, 29(5), 355-360. doi:https://doi.org/10.1016/S0003-6870(97)00077-X
Kaskutas, V., Dale, A. M., Nolan, J., Patterson, D., Lipscomb, H. J., & Evanoff, B. (2009). Fall hazard control observed on residential construction sites. American journal of industrial medicine, 52(6), 491-499.
Kelm, A., Laußat, L., Meins-Becker, A., Platz, D., Khazaee, M. J., Costin, A. M., . . . Teizer, J. (2013). Mobile passive Radio Frequency Identification (RFID) portal for automated and rapid control of Personal Protective Equipment (PPE) on construction sites. Automation in Construction, 36, 38-52.
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Paper presented at the Advances in neural information processing systems.
L. Torrey, J. S. (2009). Handbook of Research on Machine Learning Applications. IGI Global.
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. nature, 521(7553), 436-444.
Li, H., Li, X., Luo, X., & Siebert, J. (2017). Investigation of the causality patterns of non-helmet use behavior of construction workers. Automation in Construction, 80, 95-103.
Lin, Y.-H., Chen, C.-Y., & Wang, T.-W. (2011). Fatal occupational falls in the Taiwan construction industry. Journal of the Chinese Institute of Industrial Engineers, 28(8), 586-596.
Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.-Y., & Berg, A. C. (2016). Ssd: Single shot multibox detector. Paper presented at the European conference on computer vision.
Mahalingam, A., & Levitt, R. E. (2007). Safety issues on global projects. Journal of construction engineering and management, 133(7), 506-516.
Naticchia, B., Vaccarini, M., & Carbonari, A. (2013). A monitoring system for real-time interference control on large construction sites. Automation in Construction, 29, 148-160.
Nwankpa, C., Ijomah, W., Gachagan, A., & Marshall, S. (2018). Activation functions: Comparison of trends in practice and research for deep learning. arXiv preprint arXiv:1811.03378.
OSHA. (2006). Module 13—Personal protective equipment. Retrieved from https://www.osha.gov/SLTC/personalprotectiveequipment/hazards_solutions.html.
Park, M.-W., Elsafty, N., & Zhu, Z. (2015). Hardhat-wearing detection for enhancing on-site safety of construction workers. Journal of construction engineering and management, 141(9), 04015024.
Pendlebury, M. C., Brace, C. L. and Gibb, A. G. F. (2006). Construction Health: Site Hazards and Risks. The Future of Sustainable Construction.
Rameezdeen, R., Pathirage, C. and Weerasooriya, S. (2003). Study of Construction Accidents in Sri Lanka. Built-Environment-Sri Lanka,, Vol. 04.
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. Paper presented at the Proceedings of the IEEE conference on computer vision and pattern recognition.
Ren, S., He, K., Girshick, R., & Sun, J. (2015). Faster r-cnn: Towards real-time object detection with region proposal networks. Paper presented at the Advances in neural information processing systems.
Sardana, N. (2018). Instance Segmentation. Retrieved from https://tjmachinelearning.com/lectures/1718/instance/instance.pdf
Seo, J., Han, S., Lee, S., & Kim, H. (2015). Computer vision techniques for construction safety and health monitoring. Advanced Engineering Informatics, 29(2), 239-251.
Sherratt, F., Farrell, P., & Noble, R. (2013). UK construction site safety: discourses of enforcement and engagement. Construction Management and Economics, 31(6), 623-635. doi:10.1080/01446193.2012.747689
Shrestha, K., Shrestha, P. P., Bajracharya, D., & Yfantis, E. A. (2015). Hard-hat detection for construction safety visualization. Journal of Construction Engineering, 2015(1), 1-8.
Skibniewski, M. J. (2015). Research trends in information technology applications in construction safety engineering and management. Frontiers of engineering management, 1(3), 246-259.
Smith-Jackson, T., Artis, S., Hung, Y.-H., Kim, H. N., Hughes, C., Kleiner, B., & Nolden, A. (2011). Safety critical incidents among small construction contractors: A prospective case study.
Teo Ai Lin, E., Theo, H., & Feng, Y. (2008). Construction health and safety performance in developing and developed countries: A parallel study in South Africa and Singapore. Paper presented at the Evolution of and Directions in Construction Safety and Health: Proceedings of CIB W99 14th Rinker International Conference, 9-11 March, 2008, Gainesville, Florida.
Tuma, M. A., Acerra, J. R., El-Menyar, A., Al-Thani, H., Al-Hassani, A., Recicar, J. F., . . . Maull, K. I. (2013). Epidemiology of workplace-related fall from height and cost of trauma care in Qatar. International journal of critical illness and injury science, 3(1), 3.
Vitharana, V., De Silva, G., & De Silva, S. (2015). Health hazards, risk and safety practices in construction sites–a review study. Engineer: Journal of the Institution of Engineers, Sri Lanka, 48(3).
Wells, J. (1986). The construction industry in developing countries : alternative strategies for development. London: Croom Helm.
Zhang, H., Yan, X., Li, H., Jin, R., & Fu, H. (2019). Real-time alarming, monitoring, and locating for non-hard-hat use in construction. Journal of construction engineering and management, 145(3), 04019006.
Zhao, D., Lucas, J., & Thabet, W. (2009, 13-16 Dec. 2009). Using virtual environments to support electrical safety awareness in construction. Paper presented at the Proceedings of the 2009 Winter Simulation Conference (WSC).
中華民國勞動部. (2020). 勞動統計月報 Retrieved from https://www.mol.gov.tw/statistics/2452/2453/%E5%8B%9E%E5%8B%95%E7%B5%B1%E8%A8%88%E6%9C%88%E5%A0%B1/
國家發展委員會. (2019). 重要統計資料手冊. Retrieved from 國家發展委員會: https://www.ndc.gov.tw/Content_List.aspx?n=507E4787819DDCE6
陳韋成. (2015). 工地安全監控系統. 國立臺灣大學, Available from Airiti AiritiLibrary database. (2015年)
黃倩翠. (2011). 勞工安全衛生教育訓練現況分析與複訓機制. (碩士). 國立交通大學, 新竹市. Retrieved from https://hdl.handle.net/11296/bqx83e