Effects of flooding on rice biomass in Hanoi city on the basis of deep learning application and Sentinel-1A data

Despite being Vietnam's largest city, Hanoi's economy still relies on agriculture. Recent weather events, like floods, have significantly impacted rice biomass. Mapping and monitoring rice growth using synthetic aperture radar (SAR) data and the Artificial Bee Colony—Deep Neural Network (ABC-DNN) can provide reliable data on rice production affected by floods. Sentinel-1 satellite images from January to October 2022 showed that VH polarization yielded more detailed information than VV polarization. Field data and Support Vector Machine (SVM) classification estimated rice cultivation areas at approximately 81 ha for Winter-Spring and 77 ha for Summer-Autumn crops, with over 90% accuracy. The ABC-DNN model predicted aboveground biomass (AGB) with coefficients of determination (R2) ranging from 0.722 to 0.745. The model effectively identified flood-prone areas, aiding policymakers in developing strategies to mitigate agricultural damage, particularly in lowland regions of Hanoi.