Blupix: Citizen science for flood depth estimation in urban roads

Abstract

Existing flood monitoring tools (such as flood gauges, overhead satellite imagery, and hydrological modeling) are susceptible to reshaped surface topography and microtopographic variations commonly seen in urban environments, ultimately yielding large estimation errors that can limit their applications particularly when real-time, high spatial resolution estimates of floodwater depth and/or movement patterns is desirable. In recent years, thanks to advancements in mobile computing and cloud storage capacities, crowdsourcing has become an invaluable source of collecting and sharing field data in disaster mitigation and response. In this research, we introduce Blupix, a mobile application with embedded computing functionality for on-demand floodwater depth estimation from geocoded photos of submerged stop signs. Blupix allows the user to capture a photo of a submerged stop sign using the camera of the mobile device, automatically obtains and stores the geographical location and azimuth angles of the mobile device when the photo is taken, provides an interactive interface in Google Street View to locate the stop sign in a flood-free view, runs a light object detection model (EfficientDet03) trained to detect stop signs and poles and measure their sizes in pre-and post-flood images, and finally, calculates the depth of floodwater as the difference in pole lengths. The app interface is designed to be easy-to-use and motivated by an operational need in flood emergencies where ordinary people and first response teams need reliable, ad-hoc flood water depth information in their surroundings with minimum access to flood monitoring data or other pre-installed infrastructure.