| Abstract |
Monitoring High-tide Flooding (HTF) is challenging because HTF usually spreads widely and forms localized water accumulations depending on the natural processes and infrastructure. Stationary monitoring systems and satellite imaging have their certain limitations. To date, citizen science is considered as the most promising means to monitor HTF, which provides wide and continuous coverage of the community and real-time first-hand witness of the flooding event. Here, we present a flexible Artificial Intelligence (AI)-supported citizen science platform for HTF monitoring. Flood extent is identified through standard photogrammetry algorithms and a Computer vision technique called monoplotting, and water depth can be estimated using reference objects. In this paper, monoplotting is employed to establish a correlation between photos and the corresponding digital elevation model (DEM) data, allowing to map the flood extent and water depth to the DEM map to minimize the data uncertainty and enhance the data credibility, resolution, and overall value. © 2020 ACM. |
