HK-DOM&nDSM-Dataset

This dataset is designed for building extraction from remote sensing data and is divided into three subsets: Train, Val, and Test. Each subset contains three folders: DOM, nDSM, and label. The DOM data provide high-resolution orthophoto imagery, the nDSM data provide height-related information, and the label data provide the corresponding building masks for supervised semantic segmentation. The DOM, nDSM, and label files are matched by the same file name or sample identifier, ensuring one-to-one correspondence among optical imagery, height information, and ground-truth annotations. The training set is used for model optimization, the validation set is used for model selection and hyperparameter tuning, and the test set is reserved for final performance evaluation. This dataset can support research on building footprint extraction, multimodal remote sensing segmentation, DOM–nDSM feature fusion, and urban object extraction.