This study was conducted in response to the Beijing–Tianjin–Hebei mega heavy rainfall event at the end of July 2023, and the severely affected and representative Jiangou village in Beijing was selected as the study area. A variety of methods were used to synthesize and analyze the situation and propose an adaptive response to heavy rainfall and flooding in the village. Based on multi-source remote sensing (RS) data, a comprehensive topographic and hydrological characterization was carried out, and the precipitation before and after the disaster was analyzed; the flood inundation area was extracted using the improved normalized water body index (MNDWI) and OTSU thresholding methods, and the changes of water bodies during the flooding period were quantitatively analyzed; and an improved convolutional-neural-network-based building identification and extraction model was constructed to extract the research distribution of buildings in the area. The sponge city construction (SPCC) method was improved to obtain a method that can mitigate flood risk and adapt to villages by constructing small artificial lakes and local topographic buffers to improve the water storage and drainage capacity of villages. The study shows that these methods are innovative in flood hazard analysis and mitigation but still need further improvement in data accuracy, simulation depth, and system evaluation.
Read full abstract