A Water Level Estimation System Based on Image Recognition of Water Level Gauge

There are various water reservoir tanks such as septic tanks, industrial cisterns, agricultural cisterns and fire protection cisterns. Generally, they are installed outdoors and water levels vary depending on weather conditions. By monitoring water level in water reservoir tanks, the flooding can be predicted and flood damage can be mitigated. The water level is usually measured by a water level gauge. In general, installing the water level gauge is difficult due to the construction environment of the water reservoir tanks. However, this can be done by using image recognition taken by cameras installed around the water reservoir tanks. In this paper, we design and implement an estimation system for water level based on image recognition of water level gauge. For object recognition, we use YOLOv5 intelligent algorithm. The experimental results using the proposed system shows that the water level can be estimated by recognizing the numerical values and red symbols on the water level gauge through image recognition.