Development of Short-Term Rainfall Prediction Model Based on BiLSTM and Its Application to CCTV-Integrated Flood Monitoring System

Sangheon Kim; Sangsu Park; Il-Soo Park

doi:10.21729/ksds.2026.19.1.23

All Issue

2026 Vol.19, Issue 1 Preview Page Next Page

Original Article

Development of Short-Term Rainfall Prediction Model Based on BiLSTM and Its Application to CCTV-Integrated Flood Monitoring System BiLSTM 기반 단기 강우예측 모델의 개발 및 CCTV 통합형 침수 모니터링 시스템 적용

31 March 2026. pp. 23-34

PDF XML

Abstract

This study developed a short-term rainfall prediction model based on BiLSTM neural networks and applied it to a CCTV-integrated flood monitoring system. The model receives 24-hour time-series meteorological data and predicts rainfall one hour ahead using a 3-layer BiLSTM architecture with weighted MSE loss function. Performance evaluation was conducted using data from 18 ASOS stations nationwide, selected based on climate zone classification, during May to October 2024. The model achieved average MAE of 0.394 mm, RMSE of 1.628 mm, and correlation coefficient of 0.576. The developed model is deployed on NVR hardware, achieving average inference time of 0.8 seconds for real-time predictions. The system issues graduated three-level alarms based on predicted rainfall intensity. Installation is planned in flood-vulnerable areas including Yongin, Jeju, and Gwangju cities.

Keywords

BiLSTM

Rainfall prediction

CCTV-integrated system

Deep learning

Flood monitoring

본 연구는 BiLSTM 신경망 기반의 단기 강우예측 모델을 개발하고 CCTV 통합형 침수 모니터링 시스템에 적용하였다. 모델은 24시간 시계열 기상 데이터를 입력받아 3층 BiLSTM 구조로 1시간 후 강수량을 예측하며, 가중 MSE 손실 함수로 예측 정확도를 높였다. 기후 지역 분류 기반으로 선정된 전국 18개 ASOS 관측소의 2024년 5–10월 데이터로 평가한 결과, 평균 MAE 0.394 mm, RMSE 1.628 mm, 상관계수 0.576을 달성하였다. NVR 하드웨어에 탑재되어 평균 0.8초의 추론 시간으로 실시간 예측을 수행하며, 3단계 알람(주의보 30 mm/h, 경보 50 mm/h, 위험 70 mm/h)을 발령한다. 용인시, 제주시, 광주시 등 침수 취약 지역에 설치 예정이다.

키워드

BiLSTM

강우예측

CCTV 통합 시스템

딥러닝

침수 모니터링

References

Graves, A. and J. Schmidhuber. (2005). Framewise Phoneme Classification with Bidirectional LSTM and other Neural Network Architectures. Neural Networks. 18(5-6): 602-610.

10.1016/j.neunet.2005.06.042

Hochreiter, S. and J. Schmidhuber. (1997). Long Short-Term Memory. Neural Computation. 9(8): 1735-1780.

10.1162/neco.1997.9.8.1735

Hong, S. Y. and J. W. Lee. (2009). Assessment of the WRF Model in Reproducing a Flash-Flood Heavy Rainfall Event over Korea. Atmospheric Research. 93(4): 818-831.

10.1016/j.atmosres.2009.03.015

Intergovernmental Panel on Climate Change. (2023). Climate Change 2023: Synthesis Report. Geneva: IPCC.

Kim, Jin-Uk, Tae-Jun Kim, Do-Hyun Kim, Jae-Kwan Shim, Young-Hwa Byun. (2024). Classification of Climate Zones and Future Projections of Regional Climate Based on High-Resolution Scenarios over South Korea. Journal of Climate Change Research. 15(6): 1233-1244.

10.15531/KSCCR.2024.15.6.1233

Kratzert, F., D. Klotz, C. Brenner, K. Schulz, and M. Herrnegger. (2018). Rainfall–Runoff Modelling Using Long Short- Term Memory (LSTM) Networks. Hydrology and Earth System Sciences. 22: 6005-6022.

10.5194/hess-22-6005-2018

Lam, R., A. Sanchez-Gonzalez, M. Willson, P. Wirnsberger, M. Fortunato, F. Alet, S. Ravuri, T. Ewalds, S. Eaton-Rosen, W. Hu, A. Merose, S. Hoyer, G. Holland, O. Vinyals, J. Stott, A. Pritzel, S. Mohamed, and P. Battaglia. (2023). Learning Skillful Medium-Range Global Weather Forecasting. Science. 382(6677): 1416-1421.

10.1126/science.adi2336

Lee, Y. and S. D. Brody. (2018). Examining the Impact of Land Use on Flood Losses in Seoul, Korea. Land Use Policy. 70: 500-509.

10.1016/j.landusepol.2017.11.019

Ministry of the Interior and Safety. (2021). 2020 Disaster Yearbook. Sejong: MOIS.

Pathak, J., S. Subramanian, P. Harrington, S. Raja, A. Chattopadhyay, M. Mardani, T. Kurth, D. Hall, Z. Li, K. Azizzadenesheli, P. Hassanzadeh, K. Kashinath, A. Anandkumar. (2022). FourCastNet: A Global Data-Driven High-Resolution Weather Model Using Adaptive Fourier Neural Operators. arXiv preprint arXiv:2202.11214.

10.48550/arXiv.2202.11214

Shi, W., J. Cao, Q. Zhang, Y. Li, and L. Xu. (2016). Edge Computing: Vision and Challenges. IEEE Internet of Things Journal. 3(5): 637-646.

10.1109/JIOT.2016.2579198

Shi, X., Z. Chen, H. Wang, D. Y. Yeung, W. K. Wong, and W. C. Woo. (2015). Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting. Advances in Neural Information Processing Systems. 28: 802-810.

Shi, X., Z. Chen, H. Wang, D. Y. Yeung, W. K. Wong, and W. C. Woo. (2017). Deep Learning for Precipitation Nowcasting: A Benchmark and a New Model. Advances in Neural Information Processing Systems. 30: 5617-5627.

Xiang, Z., J. Yan, and I. Demir. (2020). A Rainfall-Runoff Model with LSTM-Based Sequence-to-Sequence Learning. Water Resources Research. 56(1): e2019WR025326.

10.1029/2019WR025326

Korean References Translated from the English

김진욱, 김태준, 김도현, 심재관, 변영화 (2024). 고해상도 기후변화 시나리오 기반의 남한 기후 지역 분류 및 미래 전망. 한국기후변화학회지. 15(6): 1233-1244.

10.15531/KSCCR.2024.15.6.1233

행정안전부 (2021). 2020 재해연보. 세종: 행정안전부.

Information

Publisher :Korean Society of Disaster and Security
Publisher(Ko) :한국방재안전학회
Journal Title :Journal of Korean Society of Disaster and Security
Journal Title(Ko) :한국방재안전학회 논문집
Volume : 19
No :1
Pages :23-34
Received Date : 2025-12-22
Revised Date : 2026-01-23
Accepted Date : 2026-01-27
DOI :https://doi.org/10.21729/ksds.2026.19.1.23

Journal of Korean Society of Disaster and Security ISSN:2466-1147(Print) 2508-285X(Online) 한국방재안전학회 논문집

All Issue

Korean References Translated from the English