中圖分類號：TP273 文獻標志碼：A DOI: 10.16157/j.issn.0258-7998.256409
中文引用格式： 吳昊澤，李博，阮斌. 智能水位標尺監測系統的設計實現[J]. 電子技術應用，2025，51(8)：40-46.
英文引用格式： Wu Haoze，Li Bo，Ruan Bin. Design and implementation of intelligent water level gauge detection system[J]. Application of Electronic Technique，2025，51(8)：40-46.

Design and implementation of intelligent water level gauge detection system

Abstract： With the development of deep learning, video-based water level detection has become a research hotspot in recent years. To improve the accuracy of water level measurements in urban rivers, reservoirs, and other water bodies, this paper proposes a 5G intelligent water level gauge detection system. The system primarily involves the hardware design centered around the Starshine SSC338G and a water level detection method based on an improved YOLOv8n. The method first identifies the orientation of the water gauge, then performs grayscale and binarization processing on the gauge images. Finally, YOLOv8n is used to recognize characters and scale information on the gauge, allowing for the calculation and analysis of water level data. By incorporating the attention mechanism EMA and replacing the loss function with Focal-EIoU, the improved model achieves a 21% reduction in the number of model parameters, a 17% reduction in model size, and a 21% reduction in floating-point operations.Comparative experiments with human-eye observations demonstrate that the model's accuracy meets the requirements, thus fulfilling the design goals for an intelligent water level gauge system.

Key words : water level monitoring；intelligent recognition of water gauge；YOLO；gimbal；5G