Jin et al. (2026) Video‐Based Rainfall Opportunistic Sensing in Hydrology: A Lightweight Machine Learning Approach

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Water Resources Research
• Year: 2026
• Date: 2026-07-01
• Authors: Yongcheng Jin, Zhe Zhu, Xinwei Xue, Guangtao Fu, Chi Zhang, Yu Li
• DOI: 10.1029/2025wr041953

Research Groups

• Department of Computer Science, University of Shenzhen
• Institute of Remote Sensing and Digital Earth, Chinese Academy of Sciences
• Department of Electrical Engineering, Zhejiang University

Short Summary

This paper proposes a lightweight machine-learning approach called CIEGA-NET for accurate and efficient rainfall measurement from dynamic videos. The method achieves a mean relative error of 8.2% in estimating event rainfall.

Objective

• Develop a video-based rainfall measurement system that balances accuracy and computational efficiency.

Study Configuration

• Spatial Scale: City-scale (Shenzhen, Hangzhou, Dalian)
• Temporal Scale: Real-time monitoring

Methodology and Data

• Models used: Coupled Image Enhancement and Gradual Aggregation Network (CIEGA-NET)
• Data sources: Surveillance cameras, smartphones, public synthetic data sets

Main Results

• CIEGA-NET estimates event rainfall with a mean relative error of 8.2%.
• Reduces errors by nearly 10% overall and up to 25% in heavy rainfall compared to existing video-based methods.
• Achieves a processing speed of 26 frames per second at commonly used resolutions.

Contributions

• Provides a lightweight, efficient solution for high-density rainfall monitoring networks.
• Offers improved accuracy and reduced computational requirements compared to existing video-based methods.

Funding

• National Natural Science Foundation of China (Grant No. 61872251)
• Shenzhen Science and Technology Innovation Commission (Grant No. JCYJ20190808154816491)

Citation

@article{Jin2026VideoBased,
  author = {Jin, Yongcheng and Zhu, Zhe and Xue, Xinwei and Fu, Guangtao and Zhang, Chi and Li, Yu},
  title = {Video‐Based Rainfall Opportunistic Sensing in Hydrology: A Lightweight Machine Learning Approach},
  journal = {Water Resources Research},
  year = {2026},
  doi = {10.1029/2025wr041953},
  url = {https://doi.org/10.1029/2025wr041953}
}