Yu et al. (2026) Dynamic Multi-Parameter Sensing Technology for Ecological Flows Based on the Improved DSC-YOLOv8n Model

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

Identification

• Journal: Water
• Year: 2026
• Date: 2026-01-06
• Authors: Jun Yu, Yongsheng Li, Ting Wang, Peipei Zhang, Wenlong Jiang, Lei Xing
• DOI: 10.3390/w18020146

Research Groups

Not specified in the provided text.

Short Summary

This paper proposes an improved DSC-YOLOv8n-seg model for accurate, real-time monitoring of water levels and a deep learning-based spectral principal direction recognition method for surface flow velocities, addressing challenges in ecological flow management with enhanced robustness.

Objective

• To develop an efficient and accurate real-time monitoring system for water levels and surface flow velocities to improve ecological flow management, addressing limitations of traditional methods and environmental interferences.

Study Configuration

• Spatial Scale: Riverine environments and water bodies (implied by "ecological flow management," "river surface flow velocities," "staff gauge characters").
• Temporal Scale: Real-time and dynamic monitoring.

Methodology and Data

• Models used: Improved DSC-YOLOv8n-seg model (for joint water level line and staff gauge character recognition), Deep learning-based spectral principal direction recognition method (for surface water flow velocity calculation).
• Data sources: Visual data from real-time monitoring (e.g., cameras capturing staff gauges and water surfaces).

Main Results

• The improved DSC-YOLOv8n-seg model achieved an average recognition error of ±0.012 m for water levels, with a model accuracy of 93.1%, recall rate of 94.5%, and mAP50:95 of 93.9%.
• The deep learning-based spectral principal direction recognition method calculated surface water flow velocity with a relative error of 0.005 m/s.
• Both methods demonstrated enhanced robustness in challenging conditions, including low-light and nighttime scenarios, effectively addressing environmental interference.

Contributions

• Proposes an improved DSC-YOLOv8n-seg model for highly accurate and robust joint recognition of water level lines and staff gauge characters.
• Introduces a deep learning-based spectral principal direction recognition method for stable and efficient real-time surface water flow velocity calculation.
• Provides a comprehensive, robust, and accurate solution for dynamic multi-parameter sensing (water level and flow velocity) crucial for ecological flow management, particularly effective in challenging environmental conditions.

Funding

Not specified in the provided text.

Citation

@article{Yu2026Dynamic,
  author = {Yu, Jun and Li, Yongsheng and Wang, Ting and Zhang, Peipei and Jiang, Wenlong and Xing, Lei},
  title = {Dynamic Multi-Parameter Sensing Technology for Ecological Flows Based on the Improved DSC-YOLOv8n Model},
  journal = {Water},
  year = {2026},
  doi = {10.3390/w18020146},
  url = {https://doi.org/10.3390/w18020146}
}