橋梁檢測工作通常需要檢測人員拍攝數百至上千張圖片，耗費大量時間進行檢視。為提高檢測效率並減少遺漏，本研究開發了「自動化裂縫影像雲端辨識系統」以及應用程式「auto predictor」，實現混凝土橋梁的裂縫自動辨識，或通過本研究設計的劣化偵測網站上傳劣化圖片進行混凝土橋梁的裂縫辨識。此雲端平台結合「橋梁BIM雲端管理系統」，工程人員可依據結構設計圖建立BIM橋梁模型，並儲存至雲端。在劣化檢測過程中，檢測人員能拍照記錄裂縫，並整合模型與裂縫的資訊及描述。在訓練裂縫影像辨識模型方面，本研究使用台灣地區長期橋梁檢測所拍攝的劣化圖片，這些圖片涵蓋多種實際環境條件，並挑選出具有裂縫特徵的圖片。透過有效的劣化標註策略，考量複合性劣化的標註方式，搭配YOLOv4與YOLOv7演算及本研究所建議的參數進行比較，最終獲得良好的模型供系統使用。研究結果顯示，在自動化裂縫影像雲端辨識系統中，隨機挑選的測試集劣化圖片，包括橋梁檢測及短梁剪力實驗後的試體，其裂縫皆可成功辨識。在「自動化裂縫影像雲端辨識系統」與「橋梁BIM雲端管理系統」的深度結合中，不僅能自動辨識構件裂縫，還能生成裂縫位置圖表及相關資訊。這一創新整合為決策者提供直觀的視覺化數據，展現出卓越的應用潛力與未來發展前景。採用YOLOv7演算法的模型達到87.64%的平均準確率(mAP)，顯著提升橋梁檢測效率，展現良好應用潛力。

Bridge inspection work typically requires inspectors to take hundreds to thousands of pictures, consuming a significant amount of time for review. To improve inspection efficiency and reduce omissions, this study developed an "Automated Crack Image Cloud Detection System" and the application "Auto Predictor," which enables automatic identification of concrete bridge cracks or allows users to upload deteriorated images for concrete bridge detection through a deterioration detection website designed in this study. This cloud platform integrates with the "Bridge BIM Cloud Management System," allowing engineers to create BIM bridge models based on structural design drawings and store them in the cloud. During the deterioration inspection process, inspectors can take photos to document cracks and integrate information and descriptions of the model and cracks. In training the crack image detection model, this study used deteriorated images captured during long-term bridge inspections in Taiwan, which encompass various real-world environmental conditions, selecting images that exhibit crack features. Through effective deterioration labeling strategies, considering the labeling methods for complex deterioration, and comparing calculations using YOLOv4 and YOLOv7 with the parameters recommended in this study, a robust model was ultimately obtained for system use. The research results show that in the Automated Crack Image Cloud Detection System, randomly selected test set deterioration images, including those from bridge inspections and specimens after short beam shear experiments, were all successfully identified for cracks. In the deep integration of the "Automated Crack Image Cloud Detection System" and the "Bridge BIM Cloud Management System," it is not only possible to automatically identify component cracks but also to generate crack location charts and related information. This innovative integration provides decision-makers with intuitive visual data, demonstrating exceptional application potential and future development prospects. The YOLOv7-based model achieved a mean Average Precision (mAP) of 87.64%, significantly improving bridge inspection efficiency and demonstrating exceptional application potential.

英文參考文獻
AlexeyAB. (2021). Open source darknet YOLOv4. Retrieved 2021, from GitHub: https://github.com/AlexeyAB/darknet
ASCII. (n.d.). Retrieved 2022, from ASCII Table: https://www.asciitable.com
Azhar, S., Nadeem, A., Mok, J. Y., & Leung, B. H. (2008). Building Information Modeling (BIM): A new paradigm for visual interactive modeling and simulation for construction projects. Proceedings of the First International Conference on Construction in Developing Countries, 1, pp. 435-446. Retrieved from https://www.academia.edu/download/35698780/045.pdf
Berners-Lee, T., Fielding, R., & Frystyk, H. (1996, May). Hypertext transfer protocol--HTTP/1.0，RFC 1945. Retrieved 2022, from https://doi.org/10.17487/RFC1945
Beskopylny, A. N., Shcherban', E. M., Stel'makh, S. A., Mailyan, L. R., Meskhi, B., Razveeva, I., . . . Onore, G. (2023). Detecting Cracks in Aerated Concrete Samples Using a Convolutional Neural Network. Applied Sciences, 13(3), p. 1904. Retrieved from https://doi.org/10.3390/app13031904
Bochkovskiy, A., Wang, C.-Y., & Liao, H.-Y. M. (2020). YOLOv4: Optimal speed and accuracy of object detection. arXiv preprint. Retrieved from https://doi.org/10.48550/arXiv.2004.10934
Burns, B., Beda, J., & Hightower, K. (2019). Kubernetes: Up and Running: Dive into the Future of Infrastructure. Sebastopol, CA, USA: O'Reilly Media.
Byun, N., Han, W. S., Kwon, Y. W., & Kang, Y. J. (2021, May). Development of BIM-based bridge maintenance system considering maintenance data schema and information system. Sustainability, 13(9), p. 4858. Retrieved from https://doi.org/10.3390/su13094858
Carion, N., Massa, F., Synnaeve, G., Usunier, N., Kirillov, A., & Zagoruyko, S. (2020). End-to-End Object Detection with Transformers. European Conference on Computer Vision (ECCV). Retrieved from https://doi.org/10.48550/arXiv.2005.12872
Chang, C.-Y., Chen, L.-C., & Ma, C.-C. (2017). Three-dimensional measurement of dynamic full-field displacement by stereo DIC using one high-speed camera. 15th Asia Pacific Conference for Non-Destructive Testing. 23(3). Singapore: e-Journal of Nondestructive Testing. Retrieved from https://www.ndt.net/?id=22148
Cheng, J., & Ma, Z. (2019). A Review of Building Information Modeling in Bridge Management: Current Research and Future Directions. Advances in Civil Engineering, 2019, pp. 1-12. Retrieved from https://doi.org/10.1155/2019/7681836
Cremona, C. (2003). Application of the French assessment approach to bridge maintenance strategies. Structural Engineering International, 13(3), pp. 195-200.
Dang, J., Chun, P.-J., Mizumoto, T., Liu, J., & Fujishima, T. (2021). Multi-type bridge damage detection method based on YOLO. Japan Society of Civil Engineers. Intelligence, Informatics and Infrastructure, 2, pp. 447-456. Retrieved from https://doi.org/10.11532/jsceiii.2.J2_447
Davila Delgado, J. M., Butler, L. J., Gibbons, N., Brilakis, I., Elshafie, M. Z., & Middleton, C. (2017). Management of structural monitoring data of bridges using BIM. Bridge Engineering, 170(3), pp. 204-218. Retrieved from https://doi.org/10.1680/jbren.16.00013
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., . . . Houlsby, N. (2021). An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale. International Conference on Learning Representations(ICLR). Retrieved from https://doi.org/10.48550/arXiv.2010.11929
Fielding, R. e. (1999, Jun.). Hypertext transfer protocol--HTTP/1.1，RFC 2616. Retrieved 2022, from https://www.frameip.com/rfc-2616-hypertext-transfer-protocol-http1-1/
Golabi, K., & Shepard, R. (1997). Pontis: A system for maintenance optimization and improvement of US bridge networks. Interfaces, 27(1), pp. 71-88. Retrieved from https://doi.org/10.1287/inte.27.1.71
Gonzalez, J., & Lichtenstein, M. (2020). Integrating BIM and Traditional Inspection Methods for Improved Bridge Management. International Journal of Civil Engineering, 18(6), pp. 1002-1014. Retrieved from https://doi.org/10.1007/s40940-020-00106-1
Google. (n.d.). Google Colaboratory. Retrieved 2021, from https://colab.research.google.com/notebooks/intro.ipynb
Guo, C., Lv, X., Zhang, Y., & Zhang, M. (2021, November). Improved YOLOv4-tiny network for real-time electronic component detection. Scientific Reports (Sci. Rep.), 11(1), p. 22744. Retrieved from https://doi.org/10.1038/s41598-021-02225-y
Hawk, H., & Small, E. P. (1998). The BRIDGIT bridge management system. Structural Engineering International, 8(4), pp. 309-314. Retrieved from https://doi.org/10.2749/101686698780488712
Hsieh, C.-Y. (2018). Integration of image processing, computer vision, and artificial intelligence to identify concrete surface cracks. M.S. thesis, Dept. Civil Eng., National Taiwan Univ. Taipei, Taiwan.
Hsu, S.-H., Chang, T.-W., & Chang, C.-M. (2021). Concrete surface crack segmentation based on deep learning. Proceedings of the European Workshop on Structural Health Monitoring, 128, pp. 24-34. Retrieved from https://link.springer.com/chapter/10.1007/978-3-030-64908-1_3
Hypertext Transfer Protocol -- HTTP/1.1. (1999). Retrieved 2022, from World Wide Web Consortium(W3C): https://www.w3.org/Protocols/rfc2616/rfc2616.html
Kassem, M., Kelly, G., Dawood, N., Serginson, M., & Lockley, S. (2015). BIM in facilities management applications: a case study of a large university complex. Built Environment Project and Asset Management, 5(3), pp. 261-277. Retrieved from https://doi.org/10.1108/BEPAM-02-2014-0011.
Khan, S., Naseer, M., Hayat, M., Zamir, S. W., Khan, F. S., & Shah, M. (2022, September). Transformers in Vision: A Survey. ACM Computing Surveys, 54(10), pp. 1-41. Retrieved from https://doi.org/10.1145/3505244
Krishnamurthy, B., Mogul, J. C., & Kristol, D. M. (1999, May). Key differences between HTTP/1.0 and HTTP/1.1. Computer Networks, 31(11-16), pp. 1737-1751. Retrieved from https://doi.org/10.1016/S1389-1286(99)00008-0
Kruachottikul, P., Cooharojananone, N., Phanomchoeng, G., Muangsiri, W., & Silapachote, P. (2021). Deep learning-based visual defect-inspection system for reinforced concrete bridge substructure: A case of Thailand's department of highways. Journal of Civil Structural Health Monitoring, 11, pp. 949-965. Retrieved from https://doi.org/10.1007/s13349-021-00490-z
Lauridsen, J., & Lassen, B. (1999). The Danish bridge management system DANBRO. In Bridge Management 4: Inspection, Maintenance, Assessment and Repair, (pp. 61-70).
LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998, November). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), pp. 2278-2324. Retrieved from https://ieeexplore.ieee.org/document/726791
Li, R., Yu, J., Li, F., Yang, R., Wang, Y., & Peng, Z. (2023). Automatic bridge crack detection using unmanned aerial vehicle and Faster R-CNN. Construction and Building Materials, 362, p. 129659. Retrieved from https://doi.org/10.1016/j.conbuildmat.2022.129659
Lin, P. T., Yao, Y.-T., Chen, Y.-H., Lin, S. S., Chang, C.-Y., Liu, K.-Y., . . . Lu, L.-H. (2019). Stereovision-based automatic crack detection for 3D bridge inspection. 2nd World Congress on Condition Monitoring (WCCM), (p. 150). Singapore.
Lin, Y. B., Lai, J. S., Chang, K. C., & Li, L. S. (2006). Flood scour monitoring system using fiber Bragg grating sensors. Smart Materials and Structures, 15(6), pp. 1950-1959. Retrieved from https://dx.doi.org/10.1088/0964-1726/15/6/051
Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., . . . Guo, B. (2021). Swin Transformer: Hierarchical Vision Transformer using Shifted Windows. International Conference on Computer Vision (ICCV). Retrieved from https://doi.org/10.48550/arXiv.2103.14030
Lu, W.-H., Lin, S.-P., Lin, P. T., & Wu, Y.-W. (2018, September). Development and applications of artificial intelligence image recognition system. 19th Conference on Nondestructive Testing Technology (CNDT 2018), (p. 67). Taipei, Taiwan.
Mandirola, M., Casarotti, C., Peloso, S., Lanese, I., Brunesi, E., & Senaldi, I. (2022). Use of UAS for damage inspection and assessment of bridge infrastructures. International Journal of Disaster Risk Reduction, 72, p. 102824. Retrieved from https://doi.org/10.1016/j.ijdrr.2022.102824
Miyamoto, A., & Motoshita, M. (2015). Development and practical application of a bridge management system (J-BMS) in Japan. Civil Engineering Infrastructures Journal, 48(1), pp. 189-216.
Opara, J. N., Thein, A. B., Izumi, S., Yasuhara, H., & Chun, P.-J. (2021). Defect detection on asphalt pavement by deep learning. GEOMATE Journal, 21(83), pp. 87-94. Retrieved from https://doi.org/10.21660/2021.83.6153
Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., . . . Duchesnay, É. (2011). Scikit-learn: machine learning in Python. Journal of Machine Learning Research (J. Mach. Learn. Res.), 12, pp. 2825-2830. Retrieved from https://doi.org/10.48550/arXiv.1201.0490
Rezaie, A., Achanta, R., Godio, M., & Beyer, K. (2020). Comparison of crack segmentation using digital image correlation measurements and deep learning. Construction and Building Materials, 261, p. 120474. Retrieved from https://doi.org/10.1016/j.conbuildmat.2020.120474
Tay, Y., Dehghani, M., Bahri, D., & Metzler, D. (2022). Efficient Transformers: A Survey. ACM Computing Surveys, 55(6), pp. 1-28. Retrieved from https://doi.org/10.1145/3530811
Thompson, P. D., Small, E. P., Johnson, M., & Marshall, A. R. (1998). The Pontis bridge management system. Structural Engineering International, 8(4), pp. 303-308. Retrieved from https://doi.org/10.2749/101686698780488758
Tong, Z., Gao, J., & Zhang, H. (2018). Innovative method for recognizing subgrade defects based on a convolutional neural network. Construction and Building Materials, 169, pp. 69-82. Retrieved from https://doi.org/10.1016/j.conbuildmat.2018.02.081
tzutalin. (2015). LabelImg. Retrieved 2021, from GitHub: https://github.com/tzutalin/labelImg
Wang, C.-Y., Bochkovskiy, A., & Liao, H.-Y. M. (2023). YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), (pp. 7464-7475).
WongKinYiu. (2022). Open source YOLOv7. Retrieved 2023, from GitHub: https://github.com/WongKinYiu/yolov7
Woodward, R. J., Cullington, D. W., Daly, A. F., Vassie, P. R., Haardt, P., Kashner, R., & Astudillo, R. (2001). Bridge management in Europe (BRIME): deliverable D14 final report. European Commission. European Commission.
Yau, N. J., & Liao, H. K. (2007, Mar.). Development of bridge management systems in Taiwan. Journal of the Chinese Institute of Engineers, 30(2), pp. 307-315.
Zhang, Q., Barri, K., Babanajad, S. K., & Alavi, A. H. (2021). Real-time detection of cracks on concrete bridge decks using deep learning in the frequency domain. Engineering, 7(12), pp. 1786-1796. Retrieved from https://doi.org/10.1016/j.eng.2020.07.026
Zhang, Y., Zuo, Z., Xu, X., Wu, J., Zhu, J., Zhang, H., . . . Tian, Y. (2022, 12 1). Road damage detection using UAV images based on multi-level attention mechanism. Automation in Construction, 144, p. 104613. Retrieved from https://doi.org/10.1016/j.autcon.2022.104613
Zhou, Y., Zhang, L., & Wang, X. (2021, Aug.). Integration of Building Information Modeling (BIM) and Machine Learning for Bridge Maintenance: A Case Study. Journal of Civil Engineering and Management, 27(8), pp. 621-634. Retrieved from https://doi.org/10.3846/jcem.2021.13819
Zhou, Z. (2022). An intelligent bridge management and maintenance model using BIM technology. Mobile Information Systems, 2022(1), p. 7130546. Retrieved from https://doi.org/10.1155/2022/7130546
Zhu, M. (2004). Recall, precision and average precision(Tech. Rep. 2004-09). Waterloo, ON, Canada: Department of Statistics and Actuarial Science, University of Waterloo. Retrieved from https://web.archive.org/web/20110504130953/http://sas.uwaterloo.ca/stats_navigation/techreports/04WorkingPapers/2004-09.pdf
中文參考文獻
交通部. (2018). 公路橋梁檢測及補強規範. 交通部頒布.
交通部運輸研究所. (2021). 公路橋梁檢測人員培訓教材(初稿). 交通部運輸研究所.
交通部運輸研究所. (2025年6月26日). 全國橋梁分類統計. 擷取自 全國橋梁統計資訊網: https://bss.iot.gov.tw/bss/country/statistics_all/list/
宋裕祺, 陳俊仲, 賴明俊, 許家銓, 洪曉慧, & 劉光晏. (2014). 橋梁生命週期防災管理系統建置. 國家地震工程研究中心，報告編號：NCREE-2014-027.
詹博帆. (2019). 影像辨識與深度學習在智慧交通系統的現況與未來展望. 中華技術(123), 頁 342-353.
蔡宜真. (2024). 以無人機建立即時自動化橋梁裂縫影像辨識系統. 碩士論文，國立臺灣大學. 臺北市. 擷取自 https://doi.org/10.6342/NTU202402259
謝孝勇, 劉光晏, 陳世海, & 林育賢. (2025). 應用YOLO於橋梁表面裂縫辨識與快篩之研究. 中國土木水利工程學刊, 37(1), 頁 75-88. 擷取自 https://doi.org/10.6652/JoCICHE.202503_37(1).0007
蘇振維, 張舜淵, 楊幼文, 黃俊豪, 江明益, 姚乃嘉, . . . 廖艾貞. (2018). 第二代台灣地區橋梁管理資訊系統建置規劃(三). 交通部運輸研究所.