Automated flood detection from Sentinel-1 GRD time series using Bayesian analysis for change point problems

Flood monitoring using satellite imagery faces significant limitations in existing methods, particularly regarding training data requirements and distinguishing temporary inundation from permanent water bodies. This paper presents a novel approach using Bayesian analysis for change point problems (BCP) applied to Sentinel-1 SAR time series data, which automatically detects temporal discontinuities in backscatter patterns. Our method processes raw SAR data directly with minimal preprocessing and operates without requiring training data or ancillary information, offering significant advantages for operational flood mapping in diverse geographical contexts. We validate our approach using the UrbanSARFloods benchmark dataset across three distinct geographical settings (Weihui, China; Jubba, Somalia; and NovaKakhovka, Ukraine). Our BCP approach achieves F1 scores ranging from 0.41 to 0.76 (IoU: 0.25-0.61), significantly outperforming both Otsu's thresholding (F1: 0.03-0.12, IoU: 0.02-0.08) and Siamese convolutional neural network approaches (F1: 0.08-0.34, IoU: 0.05-0.24). Further analysis reveals exceptional performance in open areas with F1 scores of 0.47-0.81 (IoU: 0.31-0.68) and high recall (0.36-0.84), contrasted with substantially lower performance in urban areas, indicating a common challenge across current flood detection methods in complex environments. The proposed method's training-free nature enables immediate deployment to new regions without model retraining or adaptation, while its ability to differentiate flood inundation from permanent water bodies without ancillary data represents a significant methodological advancement in SAR-based flood detection.