Journal of Korea Water Resources Association (한국수자원학회논문집)

Korea Water Resources Association (한국수자원학회)
권호 표지
DOI QR코드

DOI QR Code

Applying deep learning based super-resolution technique for high-resolution urban flood analysis

고해상도 도시 침수 해석을 위한 딥러닝 기반 초해상화 기술 적용

  • Choi, Hyeonjin (Department of Civil Engineering, Kumoh National Institute of Technology) ;
  • Lee, Songhee (Department of Civil Engineering, Kumoh National Institute of Technology) ;
  • Woo, Hyuna (Department of Civil Engineering, Kumoh National Institute of Technology) ;
  • Kim, Minyoung (Department of Civil Engineering, Kumoh National Institute of Technology) ;
  • Noh, Seong Jin (Department of Civil Engineering, Kumoh National Institute of Technology)
  • 최현진 (금오공과대학교 토목공학과) ;
  • 이송희 (금오공과대학교 토목공학과) ;
  • 우현아 (금오공과대학교 토목공학과) ;
  • 김민영 (금오공과대학교 토목공학과) ;
  • 노성진 (금오공과대학교 토목공학과)
  • Received : 2023.08.18
  • Accepted : 2023.10.05
  • Published : 2023.10.31
Download PDF
⟨ Previous Next ⟩

Abstract

As climate change and urbanization are causing unprecedented natural disasters in urban areas, it is crucial to have urban flood predictions with high fidelity and accuracy. However, conventional physically- and deep learning-based urban flood modeling methods have limitations that require a lot of computer resources or data for high-resolution flooding analysis. In this study, we propose and implement a method for improving the spatial resolution of urban flood analysis using a deep learning based super-resolution technique. The proposed approach converts low-resolution flood maps by physically based modeling into the high-resolution using a super-resolution deep learning model trained by high-resolution modeling data. When applied to two cases of retrospective flood analysis at part of City of Portland, Oregon, U.S., the results of the 4-m resolution physical simulation were successfully converted into 1-m resolution flood maps through super-resolution. High structural similarity between the super-solution image and the high-resolution original was found. The results show promising image quality loss within an acceptable limit of 22.80 dB (PSNR) and 0.73 (SSIM). The proposed super-resolution method can provide efficient model training with a limited number of flood scenarios, significantly reducing data acquisition efforts and computational costs.

기후변화와 도시화의 영향으로 인해 자연재해의 발생빈도와 규모가 증가하고 있다. 특히 도시 침수는 발생 시간이 짧고 막대한 인명 및 경제적 손실을 초래할 수 있기 때문에 신속하고 정확도 높은 예측 정보 생산이 중요하다. 하지만, 기존 물리과정 및 인공지능 기반 기법은 고해상도 침수 해석을 위해 많은 전산 자원이나 데이터가 요구되는 한계가 있다. 본 연구에서는 딥러닝 기반 초해상화(Super-Resolution) 기법을 통한 고해상도 도시 침수 해석 방법을 제안하고 적용성을 평가한다. 제안된 방법은 고해상도 물리 모형의 결과로 훈련된 초해상화 딥러닝 모형을 이용하여 저해상도 침수 해석 이미지를 고해상도로 변환한다. 미국 포틀랜드 도심지의 두 가지 침수 사례에 대해 적용, 4 m 공간해상도 물리 모의 결과를 1 m 급 고해상도 침수 해석 정보로 초해상화 하였으며, 초해상화 이미지와 고해상도 원본 간 높은 구조적 유사성이 확인되었다. 성능 지표로 평가한 결과, 전체 검증 대상 이미지에 대한 평균 PSNR 22.77 dB, SSIM 0.77로 우수하여, 초해상화 기법의 도시 침수 해석 적용성이 검증되었다. 제안된 방법은 적은 양의 침수 시나리오만으로도 효율적인 딥러닝 모형 훈련이 가능하고, 물리 모형의 정보를 최대한 활용할 수 있기 때문에, 고해상도 도시 침수 정보 생산에 효과적으로 사용될 수 있을 것으로 기대된다.

Keywords

Acknowledgement

이 성과는 정부(과학기술정보통신부)의 재원으로 한국연구재단(No.2022R1A4A5028840)의 지원과 환경부의 재원으로 한국환경산업기술원(RS-2023-00218973)의 지원을 받아 수행된 연구임.

References

  1. Ballard, T., and Erinjippurath, G. (2020). "FireSRnet: Geoscience-driven super-resolution of future fire risk from climate change." arXiv Preprint, arXiv:2011.12353.
  2. Ballard, T., and Erinjippurath, G. (2022). "Contrastive learning for climate model bias correction and super-resolution." arXiv Preprint, arXiv:2211.07555.
  3. Bashir, S.M.A., Wang, Y., Khan, M., and Niu, Y. (2021). "A comprehensive review of deep learning-based single image super-resolution." PeerJ Computer Science, Vol. 7, e621.
  4. Caballero, J., Ledig, C., Aitken, A., Acosta, A., Totz, J., Wang, Z., and Shi, W. (2017). "Real-time video super-resolution with spatio-temporal networks and motion compensation." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, U.S., pp. 4778-4787.
  5. Cheng, J., Kuang, Q., Shen, C., Liu, J., Tan, X., and Liu, W. (2020). "ResLap: Generating high-resolution climate prediction through image super-resolution." IEEE Access, Vol. 8, pp. 39623-39634. https://doi.org/10.1109/ACCESS.2020.2974785
  6. Cooley, A., and Chang, H. (2017). "Precipitation intensity trend detection using hourly and daily observations in Portland, Oregon." Climate, Vol. 5, No. 1, 10.
  7. Dai, T., Cai, J., Zhang, Y., Xia, S.-T., and Zhang, L. (2019). "Second-order attention network for single image super-resolution." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Long Beach, USA, pp. 1106511074.
  8. Dong, C., Loy, C.C., He, K., and Tang, X. (2014). "Learning a deep convolutional network for image super-resolution." Computer Vision - ECCV 2014: 13th European Conference, Zurich, Switzerland, pp. 184-199.
  9. Galar, M., Sesma, R., Ayala, C., Albizua, L., and Aranda, C. (2020). "Super-resolution of Sentinel-2 images using convolutional neural networks and real ground truth data." Remote Sensing, MDPI, Vol. 12, No. 18, 2941.
  10. Guidolin, M., Chen, A.S., Ghimire, B., Keedwell, E.C., Djordjevic, S., and Savic, D.A. (2016). "A weighted cellular automata 2D inundation model for rapid flood analysis." Environmental Modelling & Software, Vol. 84, pp. 378-394.
  11. Guo, Z., Leitao, J.P., Simoes, N.E., and Moosavi, V. (2021). "Data-driven flood emulation: Speeding up urban flood predictions by deep convolutional neural networks." Journal of Flood Risk Management, Vol. 14, No. 1, e12684.
  12. He, J., Zhang, L., Xiao, T., Wang, H., and Luo, H. (2023). "Deep learning enables super-resolution hydrodynamic flooding process modeling under spatiotemporally varying rainstorms." Water Research, Vol. 239, 120057.
  13. Hwang, S.-G., and Lee, J.-H. (2023). "Super-resolution technique of underwater image based on lightweight convolutional neural network for marine accident cause analysis." Journal of Korean Institute of Intelligent Systems, Vol. 33, No. 2, pp. 127-132. https://doi.org/10.5391/JKIIS.2023.33.2.127
  14. Jia, Y., Ge, Y., Chen, Y., Li, S., Heuvelink, G.B.M., and Ling, F. (2019). "Super-resolution land cover mapping based on the convolutional neural network." Remote Sensing, MDPI, Vol. 11, No. 15, 1815.
  15. Kim, J., Lee, J.K., and Lee, K.M. (2016). "Deeply-recursive convolutional network for image super-resolution." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, U.S., pp. 1637-1645.
  16. Kwon, O.-S. (2023). "Vehicle detection algorithm using super resolution based on deep residual dense block for remote sensing images." Journal of Broadcast Engineering, Vol. 28, No. 1, pp. 124-131. https://doi.org/10.5909/JBE.2023.28.1.124
  17. Ledig, C., Theis, L., Huszar, F., Caballero, J., Cunningham, A., Acosta, A., Aitken, A., Tejani, A., Totz, J., Wang, Z., and Shi, W. (2017). "Photo-realistic single image super-resolution using a generative adversarial network." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, U.S., pp. 4681-4690.
  18. Lee, S., Nakagawa, H., Kawaike, K., and Zhang, H. (2016). "Urban inundation simulation considering road network and building configurations." Journal of Flood Risk Management, Vol. 9, No. 3, pp. 224-233. https://doi.org/10.1111/jfr3.12165
  19. Lim, B., Son, S., Kim, H., Nah, S., and Lee, K.M. (2017). "Enhanced deep residual networks for single image super-resolution." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, Honolulu, HI, U.S., pp. 136-144.
  20. Lombana, L., and Martinez-Grana, A. (2022). "A flood mapping method for land use management in small-size water bodies: Validation of spectral indexes and a machine learning technique." Agronomy, MDPI, Vol. 12, No. 6, 1280.
  21. Mei, Y., Fan, Y., Zhou, Y., Huang, L., Huang, T.S., and Shi, H. (2020). "Image super-resolution with cross-scale non-local attention and exhaustive self-exemplars mining." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Seattle, WA, U.S., pp. 5690-5699.
  22. Moreno-Rodenas, A.M., Bellos, V., Langeveld, J.G., and Clemens, F.H.L.R. (2018). "A dynamic emulator for physically based flow simulators under varying rainfall and parametric conditions." Water Research, Vol. 142, pp. 512-527. https://doi.org/10.1016/j.watres.2018.06.011
  23. Noh, S.J., Lee, J.-H., Lee, S., and Seo, D.-J. (2019). "Retrospective dynamic inundation mapping of hurricane harvey flooding in the Houston Metropolitan Area using high-resolution modeling and high-performance computing." Water, Vol. 11, No. 3, 597.
  24. Saharia, C., Ho, J., Chan, W., Salimans, T., Fleet, D.J., and Norouzi, M. (2021). "Image super-resolution via iterative refinement." IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 45, No. 4, pp. 4713-4726.
  25. Sara, U., Akter, M., and Uddin, M.S. (2019). "Image quality assessment through FSIM, SSIM, MSE and PSNR - A comparative study." Journal of Computer and Communications, Vol. 7, No. 3, pp. 8-18. https://doi.org/10.4236/jcc.2019.73002
  26. Szegedy, C., Ioffe, S., Vanhoucke, V., and Alemi, A. (2016). "Inception-v4, Inception-ResNet and the impact of residual connections on learning." Proceedings of the AAAI Conference on Artificial Intelligence, Phoenix, AZ, U.S., Vol. 31, No. 1, pp. 4278-4284.
  27. Wang, X., Yi, J., Guo, J., Song, Y., Lyu, J., Xu, J., Yan, W., Zhao, J., Cai, Q., and Min, H. (2022). "A review of image super-resolution approaches based on deep learning and applications in remote sensing." Remote Sensing, MDPI, Vol. 14, No. 21, 5423.
  28. Wang, Y., Chen, A.S., Fu, G., Djordjevic, S., Zhang, C., and Savic, D.A. (2018). "An integrated framework for high-resolution urban flood modelling considering multiple information sources and urban features." Environmental Modelling & Software, Vol. 107, pp. 85-95. https://doi.org/10.1016/j.envsoft.2018.06.010
  29. Wang, Z., Bovik, A., Sheikh, H., and Simoncelli, E. (2004). "Image quality assessment: From error visibility to structural similarity." IEEE Transactions on Image Processing, Vol. 13, No. 4, pp. 600-612. https://doi.org/10.1109/TIP.2003.819861
  30. Zha, L., Yang, Y., Lai, Z., Zhang, Z., and Wen, J. (2021). "A lightweight dense connected approach with attention on single image super-resolution." Electronics, MDPI, Vol. 10, No. 11, 1234.