DEEPEaRth Lab develops AI-powered edge computing systems embedded on drones to extract quantitative information directly from dynamic environments. By bringing computation closer to data collection, the system transform observations into real-time information for monitoring, process understanding and informed decision-making.
Lightweight neural networks process data directly on-board drones, reducing the delay between data acquisition and information extraction.
Raw observations are transformed into spatially distributed measurements and indicators describing environmental conditions and dynamics.
AI-derived observations provide new opportunities to investigate how environmental processes evolve and interact across space and time.
RivAIr is our drone-based platform for real-time observation of river dynamics. An onboard edge-computing device integrates fine-tuned lightweight AI models into a single real-time workflow, transforming drone imagery into quantitative river observations directly during flight. RivAIr provides real-time flow observations to investigate interactions between hydraulics and river morphology.
A closer look at the system and its scientific foundations.
Our systems capture environmental processes as they evolve, providing spatially distributed observations that support monitoring, process understanding and modelling.
Beyond point measurements, observing processes in space and time.
DEEPEaRth Lab activities are supported by research institutions, national innovation centers, and technology-driven companies.