Estimation Method for Roof‐damaged Buildings from Aero-Photo Images During Earthquakes Using Deep Learning

Information Systems Frontiers - Tập 25 - Trang 351-363 - 2021
Shono Fujita1, Michinori Hatayama2
1Graduate School of Informatics, Kyoto University, Kyoto, Japan
2Disaster Prevention Research Institute, Kyoto University, Uji, Japan

Tóm tắt

Issuing a disaster certificate, which is used to decide the contents of a victim’s support, requires accuracy and rapidity. However, in Japan at large, issuing of damage certificates has taken a long time in past earthquake disasters. Hence, the government needs a more efficient mechanism for issuing damage certificates. This study developed an estimation system of roof-damaged buildings to obtain an overview of earthquake damage based on aero-photo images using deep learning. To provide speedy estimation, this system utilized the trimming algorithm, which automatically generates roof image data using the location information of building polygons on GIS (Geographic Information System). Consequently, the proposed system can estimate, if a house is covered with a blue sheet with 97.57 % accuracy and also detect whether a house is damaged, with 93.51 % accuracy. It would therefore be worth considering the development of an image recognition model and a method of collecting aero-photo data to operate this system during a real earthquake.

Tài liệu tham khảo

Cabinet Office in Japan. (2013). Operation standard of damage certification for houses in disaster, pp. 1–3 (in Japanese). Available at: http://www.bousai.go.jp/taisaku/pdf/shishinall.pdf. Accessed 18 Feb 2020.

Cabinet Office in Japan. (2017). Operation manual of damage certification work for houses in disaster pp. 2, (in Japanese) Available at: http://www.bousai.go.jp/taisaku/pdf/saigai_tebiki_full.pdf. Accessed 18 Feb 2020.

Cabinet Office in Japan. (2018). Summary of revision in March 2018. (in Japanese) Available at: http://www.bousai.go.jp/taisaku/pdf/h3003kaitei.pdf. Accessed 18 Feb 2020.

Crisis Mappers Japan. (2020). Crisis Mappers Japan. (in Japanese) Available at: http://crisismappers.jp/about.html. Accessed 1 Apr 2021.

Curran, K., Crumlish, J., & Fisher, G. (2012). OpenStreetMap. International Journal of Interactive Communication Systems and Technologies, 2(1), 69–78.

Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) ImageNet: A large-scale hierarchical image database. In Proc. CVPR, pp. 2–9.

DRONE BIRD. (2020). DRONE BIRD. (in Japanese) Available at: http://dronebird.org. Accessed 1 Apr 2021.

Fujiu, M., Ohara, M., & Meguro, K. (2013). Study on operation test of remote building damage assessment systemduring large scale earthquake disaster, Institute of Industrial Science, University of Tokyo, Bulletin of ERS, No. 46.

Geospatial Information Authority of Japan. (2020). About aero-photo images. (in Japanese) Available at: https://www.gsi.go.jp/gazochosa/gazochosa41006.html. Accessed 1 Apr 2021.

Girres, J.-F., & Touya, G. (2010). Quality assessment of the French OpenStreetMap dataset. Transactions in GIS, 14(4), 435–459.

He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 770–778).

Hezaveh, M. M., Kanan, C., & Salvaggio, C. (2017). Roof damage assessment using deep learning, 2017 IEEE Applied Imagery Pattern Recognition Workshop (AIPR).

Horie, K., Shigekawa, K., Maki, N., Tanaka, S., & Hayashi, H. (2005). Application of Damage Assessment Training System (DATS) to Ojiya City following the 2004 Niigata-ken Chuetsu Earthquake -through a disaster response support activity for issuing victim certificate- (in Japanese). Journal of Social Safety Science, 7, 123–132.

Hu, J., Shen, L., & Sun, G. (2017). Squeeze-and-excitation networks. arXiv preprint arXiv:1709.01507.

IMAGENET Large Scale Visual Recognition Challenge. (2012). ImageNet large scale visual recognition competition 2012 (ILSVRC2012). Available at: http://image-net.org/challenges/LSVRC/2012/results.html. Accessed 1 Apr 2021.

Inoue, M., Suetomi, I., Fukuoka, J., Onishi, S., Numada, M., & Meguro, K. (2018). Development of estimation formula of disaster response-work volume based on the Kumamoto earthquake (in Japanese). Monthly Journal of the Institute of Industrial Science, University of Tokyo, 70(4), 289–297

Ji, M., Liu, L., Du, R., & Buchroithner, M. F. (2019). A comparative study of texture and convolutional neural network features for detecting collapsed buildings after earthquakes using pre- and post-event satellite imagery. Remote Sensing, 11, 1202.

Kamagatani, Y., & Matsuoka, M. (2017). Damaged building recognition of the 2016 Kumamoto earthquakes using deep learning with aerial photographs (in Japanese). Tono Research Institute of Earthquake Science “the report of Disaster Prevention Research Committee 2017”. pp. 49–57.

Kashani, A. G., & Graettinger, A. J. (2015). Cluster-Based roof covering damage detection in ground-basedlidar data. Automation in Construction, 58, 19–27.

Krizhevsky, A., Sutskever, I., Hinton, G. E. (2012). ImageNet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems 25 (NIPS 2012).

Matsuoka, Y., Fujiu, M., Takayama, J., Nakayama, S., Suda, S., & Sakaguchi, H. (2017). Development of new hazard map after a disaster using unmanned aerial vehicle (in Japanese), 54th Research Presentation of Infrastructure Planning and Management.

Ministry of Internal Affairs and Communications. (2018). Study about the issue of damage certification in large scale earthquakes-focus on the 2016 Kumamoto Earthquake-. (in Japanese) Available at: https://www.soumu.go.jp/main_content/000528758.pdf. Accessed 1 Apr 2021.

Murakami, S., Hayashi, H., Tamura, K., Maki, N., Higashida, M., Horie, K., Hamamoto, R., & Komatsu, R. (2012). Analysis for effective operation of victim’s certificate -in case of 2012 Kyoto-fu Nanbu flood disaster (in Japanese). Journal of Social Safety Science, 23, 1–10.

National Network for Emergency Mapping. (2020). National Network for Emergency Mapping. (in Japanese) Available at: https://www.n2em.jp. Accessed 18 Feb 2020.

Nex, F., Duarte, D., Tonolo, F. G., & Kerle, N. (2019). Structural building damage detection with deep learning: assessment of a state-of-the-Art CNN in operational conditions. Remote Sensing, 11, 2765.

Ogawa, N., & Yamazaki, F. (2000). Photo-interpretation of building damage due to earthquakes using aerialphotographs, Proceedings of the 12th World Conference on Earthquake Engineering.

Open Aerial Map. (2020). Open Aerial Map. Available at: https://openaerialmap.org. Accessed 1 Apr 2021.

Prechelt, L. (1998). Automatic early stopping using cross validation: quantifying the criteria. Neural Networks, 11(4), 761–767.

Sharan, C., Woolley, C., Vandermersch, P., Cohen, J., & Tran, J. (2014). cuDNN: efficient primitives for deep learning, arXiv preprint, arXiv:1410.0759.

Shimizu, S., Komaru, Y., Wakaura, M., Tokizane, Y., Nakamura, H., & Fujiwara, H. (2019). Fragility function of wooden house considering the roofing type (in Japanese). Journal of social safety science, 34 , 63–73.

Suppasri, A., Mas, E., Charvet, I., Gunasekera, R., Imai, K., Fukutani, Y., Abe, Y., & Imamura, F. (2013). Building damage characteristics based on surveyed data and fragility curves of the 2011 Great East Japan tsunami. Natural Hazards, 66(2), 319–341.

Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2015)Going deeper with convolutions. In CVPR.

Zeiler, M. D., & Fergus, R. (2014). Visualizing and understanding convolutional networks. ECCV. 818–833. https://doi.org/10.1007/978-3-319-10590-1_53.

Thông tin
Thông tin xuất bản

Information Systems Frontiers

Tập 25

351-363

Thông tin tác giả