All Issue

2025 Vol.58, Issue 6

Research Article

Foundation model for waterbody segmentation based on ground camera images

수체 영역 분할을 위한 지상 카메라 이미지 기반 파운데이션 모델 활용 연구

Jinyong Kima
,
Sohee Kimb
,
Donghwi Jungc*
김 진용a
,
김 소희b
,
정 동휘c*
aResearcher, Future and Fusion Lab of Architectural, Civil and Environmental Engineering, Korea University, Seoul, Korea
bIntegrated Master & Ph.D. Student, Department of Civil, Environmental and Architectural Engineering, Korea University, Seoul, Korea
cAssociate Professor, School of Civil, Environmental and Architectural Engineering, Korea University, Seoul, Korea
a고려대학교 미래건설환경융합연구소 연구원
b고려대학교 건축사회환경공학과 석·박사통합과정
c고려대학교 건축사회환경공학부 부교수
*Corresponding Author
30 June 2025. pp. 449-457
PDF XML Citation
Abstract

With the increasing flash floods due to climate change, real-time flood detection methods have gained significant attention. For real-time flood detection, ground-based camera images are primarily used. Unlike satellite images, deep learning-based waterbody segmentation techniques are applied to handle the high variability of waterbody segmentation. To address the challenge of constructing a large volume of ground truth images, Segment Anything Model 2(SAM 2), based on the foundation model and the concept of prompts, has been proposed. Applying this model to flood detection is expected to enable high-accuracy waterbody segmentation with relatively small training datasets through fine-tuning. Therefore, this study proposes an advanced waterbody segmentation model based on SAM 2 and compares their performance with the existing model (V-FloodNet). The improved model, which selectively applies prompts, achieved a 10% improvement in performance compared to the existing model, demonstrating higher accuracy and greater efficiency in the segmentation process.

Keywords
Foundation model
Waterbody segmentation
Ground camera image

기후 변화로 인해 돌발 홍수가 증가함에 따라 실시간 홍수 탐지 방법이 주목받고 있다. 실시간 홍수 탐지 목적으로는 주로 지상 카메라 이미지가 사용되며, 위성 이미지와는 달리 변동성이 큰 수체 영역 분할을 위해 딥러닝 기반 수체 영역 분할 기법이 적용되고 있다. 방대한 양의 참값 이미지 구축의 어려움을 해결하기 위해 파운데이션 모델과 프롬프트 개념을 기반으로 한 Segment Anything Model 2(SAM 2)가 제안되었으며, 이를 홍수 탐지에 적용하면 비교적 적은 학습 데이터로도 미세 조정을 통해 높은 정확도의 수체 영역 분할이 가능할 것으로 예상된다. 본 연구에서는 SAM 2 모델의 구조를 개선하여 수체 영역 분할 모델을 구축하고, 기존 모델(V-FloodNet)과의 성능을 비교하였다. 프롬프트를 선택적으로 적용하도록 개선된 모델은 기존 모델 대비 성능이 10% 향상되었으며, 더 높은 정확성과 분할 과정에서의 효율성을 확보한 것으로 확인되었다.

키워드
파운데이션 모델
수체 영역 분할
지상 카메라 이미지
References
1

Ayzenberg, L., Giryes, R., and Greenspan, H. (2024). "Protosam: One-shot medical image segmentation with foundational models." arXiv preprint, arXiv:2407.07042.

2

Blanch, X., Wagner, F., and Eltner, A. (2023). River water segmentation dataset (RIWA), accessed 20 December 2024, <https://doi.org/10.34740/KAGGLE/DSV/4901781>.

3

Bommasani, R., Hudson, D.A., Adeli, E., Altman, R., Arora, S., von Arx, S., Bernstein, M.S., Bohg, J., Bosselut, A., and Brunskill, E., et al. (2021). "On the opportunities and risks of foundation models." arXiv preprint, arXiv:2108.07258.

4

Bukhari, S.A.S., Shafi, I., Ahmad, J., Butt, H.T., Khurshaid, T., and Ashraf, I. (2024). "Enhancing flood monitoring and prevention using machine learning and IoT integration." Natural Hazards, Vol. 121, pp 4837-4864.

10.1007/s11069-024-06986-3
5

Fang, Z., Wang, Y., Peng, L., and Hong, H. (2021). "Predicting flood susceptibility using LSTM neural networks." Journal of Hydrology, Vol. 594, 125734.

10.1016/j.jhydrol.2020.125734
6

Intergovernmental Panel on Climate Change (IPCC) (2022). Summary for policymakers. in: climate change 2023: Synthesis report. Contribution of working group II to the Sixth assessment report of the Intergovernmental Panel on Climate Change, accessed 13 January 2025, <https://www.ipcc.ch/report/ar6/wg2/chapter/summary-for-policymakers/>

7

Jia, M., Tang, L., Chen, B.C., Cardie, C., Belongie, S., Hariharan, B., and Lim, S.N. (2022). "Visual prompt tuning." Proceedings of the European Conference on Computer Vision, Springer Nature Switzerland, Cham, pp. 709-727.

10.1007/978-3-031-19827-4_41
8

Jun, S., Jang, H., Kim, S., Lee, J.S., and Jung, D. (2024). "A review of ground camera-based computer vision techniques for flood management." Computers and Concrete, Vol. 33, No. 4, pp. 425-443.

9

Khanuja, G.S. (2019). A study of real time search in flood scenes from UAV videos using deep learning techniques. Ph.D. dissertation, Purdue University, West Lafayette, IN, U.S.

10

Kim, M., Jin, N., Lee, S., Hong, S., Kim, C., and Ko, M. (2024). "A framework for analyzing road flooding damage using Semantic Segment Anything." Proceedings of the Korea Information Science Society Conference, pp. 1040-1042.

11

Kim, S.E. (2024). "Deep learning-based real-time road flood depth monitoring." Journal of the Korean Society of Civil Engineers, Vol. 72, No. 7, pp. 40-45.

12

Kim, Y.S., Yang, S.K., Yu, K., and Kim, D.S. (2014). "Flood runoff calculation using disaster monitoring CCTV system." Journal of Environmental Science International, KENSS, Vol. 23, No. 4, pp. 571-584.

10.5322/JESI.2014.4.571
13

Kirillov, A., Mintun, E., Ravi, N., Mao, H., Rolland, C., Gustafson, L., Xiao, T., Whitehead, S., Berg, A.C., and Lo, W.Y., et al. (2023). "Segment anything." Proceedings of the IEEE/CVF International Conference on Computer Vision, IEEE, Paris, France, pp. 4015-4026.

10.1109/ICCV51070.2023.00371
14

Kwon, S.H., and Lee, S. (2024). "Deep learning to recognize water level for agriculture reservoir using CCTV imagery." Water Resources Management, Vol. 38, No. 3, pp. 1165-1180.

10.1007/s11269-023-03714-7
15

Kwon, S.H., Ha, C., and Lee, S. (2023). "A study on the application of the agricultural reservoir water level recognition model using CCTV image data." Journal of Korea Water Resources Association, Vol. 56, No. 4, pp. 245-259.

10.3741/JKWRA.2023.56.4.245
16

Lee, H.Y., Kim, K.S., and Lee, K.W. (2024). "Application of Geo- Segment Anything Model (SAM) scheme to water body segmentation: An experiment study using CAS500-1 images." Korean Journal of Remote Sensing, Vol. 40, No. 4, pp. 343-350.

17

Liang, Y., Li, X., Tsai, B., Chen, Q., and Jafari, N.(2023). "V-FloodNet: A video segmentation system for urban flood detection and quantification." Environmental Modelling & Software, Vol. 160, 105586.

10.1016/j.envsoft.2022.105586
18

Lim, S.H., Kwon, S.H., and Lee, S.Y. (2024). "Development of a water body segmentation model for estimating water levels in agricultural reservoirs." Journal of Korea Water Resources Association, Vol. 57, No. 11, pp. 935-944.

19

Lin, T.Y., Goyal, P., Girshick, R., He, K., and Dollár, P. (2017). "Focal loss for dense object detection." Proceedings of the IEEE International Conference on Computer Vision, IEEE, Venice, Italy, pp. 2980-2988.

10.1109/ICCV.2017.324
20

Minaee, S., Boykov, Y., Porikli, F., Plaza, A., Kehtarnavaz, N., and Terzopoulos, D. (2021). "Image segmentation using deep learning: A survey." IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 44, No. 7, pp. 3523-3542.

10.1109/TPAMI.2021.305996833596172
21

Moghimi, A., Welzel, M., Celik, T., and Schlurmann, T. (2024). "A comparative performance analysis of popular deep learning models and Segment Anything Model (SAM) for river water segmentation in close-range remote sensing imagery." IEEE Access, Vol. 12, pp. 52067-52085.

10.1109/ACCESS.2024.3385425
22

Muhadi, N.A., Abdullah, A.F., Bejo, S.K., Mahadi, M.R., and Mijic, A. (2020). "Image segmentation methods for flood monitoring system." Water, Vol. 12, No. 6, 1825.

10.3390/w12061825
23

Nagaiah, K., Kalaivani, K., Palamalai, R., Suresh, K., Sethuraman, V., and Karuppiah, V. (2024). "A logical remote sensing based disaster management and alert system using AI-assisted internet of things technology." Remote Sensing in Earth Systems Sciences, Vol. 7, pp. 457-471.

10.1007/s41976-024-00139-7
24

National Aeronautics and Space Administration (NASA) (n.d.). MODIS water mask, accessed 17 March 2025, <https://modis.gsfc.nasa.gov/data/dataprod/mod44w.php>.

25

Pekel, J.F., Cottam, A., Gorelick, N., and Belward, A.S. (2016). "High-resolution mapping of global surface water and its long-term changes." Nature, Vol. 540, pp. 418-422.

10.1038/nature2058427926733
26

Pi, Y., Nath, N.D., and Behzadan, A.H. (2021). "Detection and semantic segmentation of disaster damage in UAV footage." Journal of Computing in Civil Engineering, Vol. 35, No. 2, 04020063.

10.1061/(ASCE)CP.1943-5487.0000947
27

Ravi, N., Gabeur, V., Hu, Y.T., Hu, R., Ryali, C., Ma, T., Khedr, H., Rädle, R., Rolland, C., and Gustafson, L. et al. (2024). "SAM 2: Segment anything in images and videos." arXiv preprint, arXiv:2408.00714.

28

Rentschler, J., Avner, P., Marconcini, M., Su, R., Strano, E., Vousdoukas, M., and Hallegatte, S. (2023). "Global evidence of rapid urban growth in flood zones since 1985." Nature, Vol. 622, No. 7981, pp. 87-92.

10.1038/s41586-023-06468-937794266
29

Rentschler, J., Salhab, M., and Jafino, B.A. (2022). "Flood exposure and poverty in 188 countries." Nature Communications, Vol. 13, No. 1, 3527.

10.1038/s41467-022-30727-435764617PMC9240081
30

Ryali, C., Hu, Y.T., Bolya, D., Wei, C., Fan, H., Huang, P.Y., Aggarwal, V., Chowdhury, A., Poursaeed, O., and Hoffman, J., et al. (2023). "Hiera: A hierarchical vision transformer without the bells-and-whistles." Proceedings of the International Conference on Machine Learning, PMLR, Honolulu, Hawaii, U.S., pp. 29441-29454.

31

Shorten, C., and Khoshgoftaar, T.M. (2019). "A survey on image data augmentation for deep learning." Journal of Big Data, Vol. 6, No. 1, pp. 1-48.

10.1186/s40537-019-0197-0
32

Tauro, F., Olivieri, G., Petroselli, A., Porfiri, M., and Grimaldi, S. (2016). "Flow monitoring with a camera: A case study on a flood event in the Tiber river." Environmental Monitoring and Assessment, Vol. 188, pp. 1-11.

10.1007/s10661-015-5082-526812952
33

Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, L., and Polosukhin, I. (2017). "Attention is all you need." Advances in Neural Information Processing Systems, Long Beach, CA, U.S., Vol. 30, pp 5998-6008.

34

Wang, J., Liu, Z., Zhao, L., Wu, Z., Ma, C., Yu, S., Dai, H., Yang, Q., Liu, Y., and Zhang, S. et al. (2023). "Review of large vision models and visual prompt engineering." Meta-Radiology, Vol. 1, No. 3, 100047.

10.1016/j.metrad.2023.100047
35

Wu, X., Zhang, Z., Xiong, S., Zhang, W., Tang, J., Li, Z., An, B., and Li, R. (2023). "A near-real-time flood detection method based on deep learning and SAR images." Remote Sensing, Vol. 15, No. 8, 2046.

10.3390/rs15082046
36

Yan, Z., Sun, W., Zhou, R., Yuan, Z., Zhang, K., Li, Y., Liu, T., Li, Q., Li, X., and He, L. et al. (2024). "Biomedical SAM 2: Segment anything in biomedical images and videos." arXiv preprint, arXiv:2408.03286.

37

Zhu, J., Qi, Y., and Wu, J. (2024). "Medical SAM 2: Segment medical images as video via Segment Anything Model 2." arXiv preprint, arXiv:2408.00874.

Information
  • Publisher :KOREA WATER RESOURECES ASSOCIATION
  • Publisher(Ko) :한국수자원학회
  • Journal Title :Journal of Korea Water Resources Association
  • Journal Title(Ko) :한국수자원학회 논문집
  • Volume : 58
  • No :6
  • Pages :449-457
  • Received Date : 2025-01-13
  • Revised Date : 2025-03-17
  • Accepted Date : 2025-04-01
  • DOI :https://doi.org/10.3741/JKWRA.2025.58.6.449
Journal Informaiton Journal of Korea Water Resources Association Journal of Korea Water Resources Association
Online Submission
Archives
  • scopus
  • KCI-수자원
  • NRF
  • KOFST
  • KISTI Current Status
  • KISTI Cited-by
  • crosscheck
  • orcid
  • open access
  • ccl
Journal Informaiton Journal Informaiton - close