CD-UWD 1.0: Chengdu Urban Waterbody Semantic Segmentation Dataset

CD-UWD 1.0 is a Chengdu Urban Waterbody Semantic Segmentation Dataset for city-scale urban water body extraction from very high-resolution remote sensing imagery. The dataset was constructed to support precise semantic segmentation of urban water bodies in complex urban environments, including main rivers, small rivers, lakes, small water bodies, other water bodies, and non-water scenes. It can be used for urban waterbody extraction, urban hydrological analysis, ecological monitoring, flood risk assessment, and remote sensing semantic segmentation research. The dataset is based on very high-resolution RGB imagery of Chengdu and provides image samples and corresponding pixel-level semantic segmentation labels. Two sample-size versions are included: 1024-Set: image and label patches with a size of 1024 × 1024 pixels. 512-Set: image and label patches with a size of 512 × 512 pixels. The dataset contains six subsets: CDUWD-1: Main rivers.CDUWD-2: Small rivers.CDUWD-3: Lakes.CDUWD-4: Small water bodies.CDUWD-5: Other water bodies.CDUWD-6: Non-water scenes. This dataset is associated with the paper “Precise City-Scale Urban Water Body Semantic Segmentation and Open-Source Sample Set Construction Based on Very High-Resolution Remote Sensing: A Case Study in Chengdu”. Project repository: https://github.com/xicheng79/CDUWD-1.0---Chengdu-Urban-Waterbody-Semantic-Segmentation-Dataset Acknowledgements: This project was supported by the Graduate Quality Engineering Construction Funding Program of Chengdu University of Technology (2024YAL016).