Physics-aware Masked Diffusion-based Flood Simulation for Urban Fisheye Disaster Detection
Computer Vision and Pattern Recognition
Summary
The authors created a system called PhysFlood that can make realistic flood images from just one picture taken with a fisheye lens, which usually distorts images. This helps because there aren’t many real flood pictures to study, and fisheye images are tricky for simulations. Their system uses a type of AI called Diffusion Models to add floods in a way that looks real and can be adjusted by changing things like water level. They tested it by asking people to rate the images, and found that the simulated floods looked believable and consistent. This work supports better flood prediction and detection using limited image data.
Authors
Sodtavilan Odonchimed, Tsogt Enkhbayar, Oyunzul Munkhtamga, Munkhjargal Gochoo
Abstract
Physical simulations that predict the behavior of urban disasters, such as climate-related flooding, play a crucial role in disaster prevention and the development of anomaly detection models. However, the severe shortage of flood data in real-world environments, combined with the inherent distortions of fisheye lens images, which are used for urban surveillance, has made high-precision simulations challenging. To address this, we propose a new physical simulation system PhysFlood that leverages Diffusion Models to synthesize realistic floods from just a single image captured by a fisheye lens. Our system not only enables simulation from a single image, but also features the ability to freely control and generate diverse flood scenarios by manipulating physically meaningful variables, such as water levels. In our evaluation experiments, we conducted a qualitative human study and demonstrated that the simulation images generated by PhysFlood exhibit both acceptable realism and robustness.