Comparison of the Dynamic World global classification product with the SVM algorithm-based classification product in the Hanoi area

Abstract

Currently, the use of global land cover products can yield certain benefits for research and applications across various industries. Among these, products with high spatial resolution, reaching up to 10 meters, have the potential to provide significant advantages in land management and data analysis. These products encompass various sources, with notable examples being Dynamic World (DW), Esri’s Land Cover (Esri) and ESA’s World Cover (EWC)… which utilize satellite imagery and synthesize global land cover results with high resolution and near-real-time updates. This study employs artificial intelligence models and machine learning algorithms to classify land cover in the Hanoi area using Sentinel-2 imagery and compares the results with existing global products to assess their effectiveness. The results indicate that these products are relatively effective and can be applied to specific specialized fields in Vietnam, especially on the land management.