Image-based System for Water Detection on Ephemeral Streams

Despite the availability of systems to detect water like those based on electric resistance, monitoring ephemeral and intermittent streams is still a challenge. For this reason we developed an image-based system to detect the presence of water in streams using Machine Learning and a micro-controller. The system main components are a micro-controller and a coordination unit. The micro-controller is equipped with a 2MP color camera which supports Machine Learning and Micro Python. The coordination unit is used to control relays and transmit data via LoRa. The system is low-power and can be powered with a battery making it a standalone system for analyzing and processing images on the edge. Using machine learning, we trained a model to specifically identify the presence of water on ephemeral streams. More-over, and using relays, this system can activate a more complex and power-consuming device to measure the discharge of the stream only when it detects the presence of water. Making this monitoring system more efficient for ephemeral stream monitoring. This work has been supported by the European Union’s Horizon 2020 research and innovation program under grant agreement No 1010178