研究生: |
謝喬丹 Richard Jordan Citra |
---|---|
論文名稱: |
Image-Based Preliminary Emergency Assessment and Evaluation of Damaged Buildings After Earthquake – A Taiwan Case Study Image-Based Preliminary Emergency Assessment and Evaluation of Damaged Buildings After Earthquake – A Taiwan Case Study |
指導教授: |
鄭明淵
Min-Yuan Cheng |
口試委員: |
鄭明淵
Min-Yuan Cheng 陳鴻銘 Hung-Ming Chen 高明秀 Minh-Tu Cao Doddy Prayogo Doddy Prayogo |
學位類別: |
碩士 Master |
系所名稱: |
工程學院 - 營建工程系 Department of Civil and Construction Engineering |
論文出版年: | 2021 |
畢業學年度: | 109 |
語文別: | 英文 |
論文頁數: | 74 |
中文關鍵詞: | Structural Damage Recognition 、Deep Learning 、Convolutional Neural Network 、EfficientNet 、Transfer Learning |
外文關鍵詞: | Structural Damage Recognition, Deep Learning, Convolutional Neural Network, EfficientNet, Transfer Learning |
相關次數: | 點閱:230 下載:0 |
分享至: |
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
Earthquake has become one of the major natural disasters that caused a lot of losses, such as economic losses and human casualties as the main losses. Death tolls in earthquakes arise from three main causes: structural collapses, non-structural causes, and follow-on disasters. The most responsible reason for death tolls in an earthquake is structural collapses. To prevent another death toll after an earthquake occurred, a rapid preliminary assessment of a building is needed to ensure whether the building is safe enough to be occupied or not after the disaster. Nowadays, structural damage recognition is done manually which consists of in-site inspection and/or images inspection. However, manual experts evaluation has some drawbacks which are time-consuming, introduces uncertainty, subjective opinion into the assessment, and also greatly relies on the experiences of the experts. Meanwhile, applications of deep learning (DL) in computer vision have increased significantly in recent years. Convolutional Neural Network (CNN) has been the most popular model of Deep Learning in computer vision. This network is capable of learning amounts of mid-to high‐level image representations, which means CNN has a powerful generalization ability to classify an image. Furthermore, this research proposed TL-EfficientNet which incorporated EfficientNet and Transfer Learning method together to develop the image-recognition classifier model to classify the images of a structural member into several levels of damage based on Taiwan’s code for emergency assessment and evaluation of a building post-disaster using a pre-trained EfficientNet from ImageNet dataset. From result findings and several analyses conducted in further discussions of this research, the proposed model has shown promising results in terms of accuracy and F1 score. Hence, this model conclusively demonstrated a great capability in structural damage recognition using CNN model performance.
Earthquake has become one of the major natural disasters that caused a lot of losses, such as economic losses and human casualties as the main losses. Death tolls in earthquakes arise from three main causes: structural collapses, non-structural causes, and follow-on disasters. The most responsible reason for death tolls in an earthquake is structural collapses. To prevent another death toll after an earthquake occurred, a rapid preliminary assessment of a building is needed to ensure whether the building is safe enough to be occupied or not after the disaster. Nowadays, structural damage recognition is done manually which consists of in-site inspection and/or images inspection. However, manual experts evaluation has some drawbacks which are time-consuming, introduces uncertainty, subjective opinion into the assessment, and also greatly relies on the experiences of the experts. Meanwhile, applications of deep learning (DL) in computer vision have increased significantly in recent years. Convolutional Neural Network (CNN) has been the most popular model of Deep Learning in computer vision. This network is capable of learning amounts of mid-to high‐level image representations, which means CNN has a powerful generalization ability to classify an image. Furthermore, this research proposed TL-EfficientNet which incorporated EfficientNet and Transfer Learning method together to develop the image-recognition classifier model to classify the images of a structural member into several levels of damage based on Taiwan’s code for emergency assessment and evaluation of a building post-disaster using a pre-trained EfficientNet from ImageNet dataset. From result findings and several analyses conducted in further discussions of this research, the proposed model has shown promising results in terms of accuracy and F1 score. Hence, this model conclusively demonstrated a great capability in structural damage recognition using CNN model performance.
Cha, Y.-J., Choi, W., Suh, G., Mahmoudkhani, S., & Büyüköztürk, O. (2018, 2018/09/01). Autonomous Structural Visual Inspection Using Region-Based Deep Learning for Detecting Multiple Damage Types. Computer-Aided Civil and Infrastructure Engineering, 33(9), 731-747. https://doi.org/10.1111/mice.12334
Chen, C.-C. (2014). Research on Revision of Emergency Evaluation Technology and System for Dangerous Buildings after Earthquake (震災後危險建築物緊急評估技術及制度之修訂研究). M. O. I. Construction and Planning Agency.
Coburn, A. W., Spence, R. J., & Pomonis, A. (1992). Factors determining human casualty levels in earthquakes: mortality prediction in building collapse. Proceedings of the tenth world conference on earthquake engineering,
Construction and Planning Agency, M. O. I. (2010). Emergency assessment of dangerous buildings after disaster Teaching material for staff workshop (災害後危險建築物緊急評估作業人員講習會教材).
Deng, J., Dong, W., Socher, R., Li, L., Kai, L., & Li, F.-F. (2009, 20-25 June 2009). ImageNet: A large-scale hierarchical image database. 2009 IEEE Conference on Computer Vision and Pattern Recognition,
Elnashai, A. S., & Di Sarno, L. (2008). Fundamentals of earthquake engineering. Wiley Online Library.
Gao, Y., & Mosalam, K. M. (2018, 2018/09/01). Deep Transfer Learning for Image-Based Structural Damage Recognition. Computer-Aided Civil and Infrastructure Engineering, 33(9), 748-768. https://doi.org/10.1111/mice.12363
Gao, Y., & Mosalam, K. M. (2019). PEER Hub ImageNet (Φ-Net): A Large-Scale Multi-Attribute Benchmark Dataset of Structural Images. PEER Report No.2019/07.
Goodfellow, I., Bengio, Y., Courville, A., & Bengio, Y. (2016). Deep learning (Vol. 1). MIT press Cambridge.
Hu, J., Shen, L., & Sun, G. (2018). Squeeze-and-excitation networks. Proceedings of the IEEE conference on computer vision and pattern recognition,
Huang, Y., Cheng, Y., Chen, D., Lee, H., Ngiam, J., Le, Q. V., & Chen, Z. (2018). GPipe: Efficient Training of Giant Neural Networks using Pipeline Parallelism.
Jovel, R. J., & Mudahar, M. (2010). Damage, Loss, and Needs Assessment Guidance Notes: Volume 3. Estimation of Post-Disaster Needs for Recovery and Reconstruction.
Kandel, I., & Castelli, M. (2020, 05/01). The effect of batch size on the generalizability of the convolutional neural networks on a histopathology dataset. ICT Express, 6. https://doi.org/10.1016/j.icte.2020.04.010
Kingma, D. P., & Ba, J. (2014). Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980.
Lin, C., & Sun, H. (2000). In witness of Chi-Chi Earthquake—The Causes of Damage and Countermeasures. McGraw-Hill, Taipei, Taiwan.
Mostafa, M., Hogan, L., Stephens, M., Olsen, M., & Elwood, K. (2020). EVALUATING STATE-OF-THE-ART METHODOLOGIES FOR POST- EARTHQUAKE STRUCTURAL RECONNAISSANCE.
Ng, A. Y.-T. (2021). Machine Learning - Coursera MOOC. Stanford Online. https://www.coursera.org/learn/machine-learning
Ng, S. M., Angelier, J., & Chang, C.-P. (2009). Earthquake cycle in Western Taiwan: Insights from historical seismicity. Geophysical Journal International, 178(2), 753-774. https://doi.org/10.1111/j.1365-246X.2009.04164.x
Pan, X., & Yang, T. Y. (2020, 2020/05/01). Postdisaster image-based damage detection and repair cost estimation of reinforced concrete buildings using dual convolutional neural networks [https://doi.org/10.1111/mice.12549]. Computer-Aided Civil and Infrastructure Engineering, 35(5), 495-510. https://doi.org/https://doi.org/10.1111/mice.12549
Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., & Chen, L.-C. (2018). Mobilenetv2: Inverted residuals and linear bottlenecks. Proceedings of the IEEE conference on computer vision and pattern recognition,
Shao, L., Zhu, F., Li, X. J. I. t. o. n. n., & systems, l. (2014). Transfer learning for visual categorization: A survey. 26(5), 1019-1034.
Szegedy, C., Wei, L., Yangqing, J., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015, 7-12 June 2015). Going deeper with convolutions. 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),
Tan, M., Chen, B., Pang, R., Vasudevan, V., Sandler, M., Howard, A., & Le, Q. V. (2019). Mnasnet: Platform-aware neural architecture search for mobile. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition,
Tan, M., & Le, Q. V. J. a. p. a. (2019). Efficientnet: Rethinking model scaling for convolutional neural networks.
Vetrivel, A., Gerke, M., Kerle, N., Nex, F., Vosselman, G. J. I. j. o. p., & sensing, r. (2018). Disaster damage detection through synergistic use of deep learning and 3D point cloud features derived from very high resolution oblique aerial images, and multiple-kernel-learning. 140, 45-59.
Yeum, C. M., Dyke, S. J., & Ramirez, J. (2018, 2018/01/15/). Visual data classification in post-event building reconnaissance. Engineering Structures, 155, 16-24. https://doi.org/https://doi.org/10.1016/j.engstruct.2017.10.057
Zagoruyko, S., & Komodakis, N. J. a. p. a. (2016). Wide residual networks.
Zeiler, M. D., & Fergus, R. (2014). Visualizing and understanding convolutional networks. European conference on computer vision,
Zhang, A., Wang, K. C. P., Li, B., Yang, E., Dai, X., Peng, Y., Fei, Y., Liu, Y., Li, J. Q., & Chen, C. (2017, 2017/10/01). Automated Pixel-Level Pavement Crack Detection on 3D Asphalt Surfaces Using a Deep-Learning Network. Computer-Aided Civil and Infrastructure Engineering, 32(10), 805-819. https://doi.org/10.1111/mice.12297
Zoph, B., Vasudevan, V., Shlens, J., & Le, Q. V. (2018). Learning transferable architectures for scalable image recognition. Proceedings of the IEEE conference on computer vision and pattern recognition,