2m resolution fine classification dataset of urban wetlands in Guilin City in 2023

Urban wetlands play a crucial role in urban sustainable development. Effective extraction of information on urban wetland area, type, and distribution is fundamental to understanding the functions and services of urban ecosystems. However, inconsistencies in urban wetland classification systems and the lack of effective fine classification methods for urban wetlands constrain the accurate extraction of urban wetland information. In this study, a two-stage classification approach combining object-oriented random forest and knowledge-based hierarchical decision tree algorithms was used to construct a fine classification dataset of urban wetlands in Guilin City at 2m resolution using GF-1 and GF-6 PMS images in 2023, aiming to provide basic data on the distribution of urban wetlands in Guilin. The dataset includes eight types of wetlands, including woody wetlands, herbaceous wetlands, rivers, lakes, inland beaches, reservoirs/ponds, aquaculture ponds, and canals/channels. The overall accuracy of the validated classification results is 94.89%. This dataset can provide fundamental data to support the conservation and sustainable development of urban wetlands.