Paper Contents
Abstract
A great number of communities in Africa are threatened by flood disasters. While mapping the spatial extent of flooding is necessary for emergency response as well as for adaptation decisions, accurately mapping the extent of these floods across regions requires a significant amount of training data, usually obtained via field surveys. Field surveys can be time-consuming, costly and impractical in inaccessible terrain. This necessitates the application of automatic algorithms for flood detection. Therefore, it is important to assess the effectiveness of the current automated techniques to guarantee their precision. This study employed the RF method to delineate flood extent as a basis for the validation of automatic flood detection algorithms using Sentinel-1 data within the Google Earth Engine (GEE) platform. RF's overall accuracy was 99% while Otsu's automatic flood detection algorithms were 76%. The validation results highlight how combining machine learning techniques with SAR data might improve flood monitoring and aid in disaster management efforts.
Copyright
Copyright © 2024 Sa'ad Ibrahim. This is an open access article distributed under the Creative Commons Attribution License.
