A comprehensive suite of datasets for VHR land cover mapping in Pearl River Delta, China

A comprehensive suite of datasets for VHR land cover mapping in Pearl River Delta, China Junshen Luo, Yikai Zhao, Mingyang Xuan, Jizhou Zheng, Yan Zhou*, Xiaoping Liu 1. Overview We present a comprehensive suite of datasets for Very High Resolution (VHR) land cover mapping in complex subtropical regions, specifically designed for the Pearl River Delta (PRD) of China. This collection addresses critical challenges including annotation dependency, image heterogeneity, and spectral limitations through three complementary components: an unlabeled pretraining dataset (PRD262K), a large annotated semantic segmentation dataset (PRDLC-PRO), and a high-quality point sample set. 2. Dataset Components 2.1 PRD262K: Unlabeled Pretraining Dataset Purpose: Self-supervised pretraining of remote sensing foundation modelsContent: 262,436 unlabeled VHR image patches (512×512 pixels)Spatial Resolution: 1.19 metersSource: Google Earth VHR optical imagery (2023)Coverage: Complete PRD region (170 map sheets following 1:50,000 cartographic standard)Format: .TIF (RGB, 0-255 pixel range) 2.2 PRDLC-PRO: Annotated Semantic Segmentation Dataset Purpose: Supervised training and evaluation of land cover classification modelsContent: 33,342 annotated image-label pairsSpatial Resolution: 1.19 metersSource: Google Earth VHR optical imagery (2023) and integrated from multiple public datasets (LoveDA [1], DLRSD [2], WHDLD [3], Globe230k [4], OpenEarthMap [5])Classification System: 8 classes which respectively represent 1#cropland, 2#forest, 3#grass, 4#shrubland, 5#wetland, 6#water, 7#impervious,8#bare + 0#background Dataset Construction: (1) Source Integration: Samples from 5 established datasets were selected based on climatic similarity to PRD (2) Label Harmonization: Original labels were mapped to unified 8-class system (see Table A1 in manuscript) (3) Quality Control: Manual verification and consistency checking (4) Distribution: Sequential contributions from source datasets Formats: Images: .jpg (512×512, RGB) Labels: .png (512×512, single-channel with class indices) 2.3 PRD-MR-points: Medium-Resolution Point Sample Set Purpose: Training and validation for medium-resolution land cover mapping and fusionContent: 15,000 geographically distributed sample pointsSource: Multi-source fusion of 6 land cover products: General Products: CLCD [6], Dynamic World [7], Esri 10m [8], GLC_FCS10 [9] Wetland-specific: GLWD [10], GWL_FCS30 [11]Resolution: 10 meters (resampled to consistent scale)Construction Workflow: Data Integration: All products clipped to PRD and reclassified to PRDLC-PRO system Weighted Voting: Class-specific agreement thresholds: Strict (3/4): Forest, Water, Impervious Loose (2/4): Cropland, Grass, Shrubland, Bare Weighted Fusion: Wetland (wetland products weighted 1.5×) Manual Verification: Google Earth VHR imagery reference Stratified Sampling: Balanced class distribution Format: .shp (Point shapefile with column “class” in attribute table) 3. Data Access Available Files: 1_PRD262K.zip - Unlabeled pretraining images 2_PRDLC-PRO.zip - Annotated segmentation dataset (images + labels) 3_MR_points.zip - Point sample set pretrained.txt – All pretrained image lines train.txt/val.txt/test.txt - Recommended train/val/test splits Download Links: Baidu Yun: https://pan.baidu.com/s/1NvvssWLtfr2yDjrxKyCXnw?pwd=tfd2 (access code:tdf2) Zenodo: https://doi.org/10.5281/zenodo.18301135 4. Contact For questions regarding dataset usage, please contact: Junshen Luo ([e-mail]) More information: https://doi.org/10.3390/rs18060897 License: CC BY-NC-SA 4.0Last Updated: 2026.03.15 5. Citation Please kindly cite the papers if this code is useful and helpful for your research: Luo, J., Zhao, Y., Xuan, M., Zheng, J., Zhou, Y., & Liu, X. (2026). Investigating Very-High-Resolution Land Cover Mapping in the Pearl River Delta with Remote Sensing Foundation Models and Multi-Source Data Bayesian Fusion. Remote Sensing, 18(6), 897. https://doi.org/10.3390/rs18060897 6. Reference 1. Wang, J.; Zheng, Z.; Ma, A.; Lu, X.; Zhong, Y. LoveDA: A Remote Sensing Land-Cover Dataset for Domain Adaptive Semantic Segmentation 2022. 2. Shao, Z.; Yang, K.; Zhou, W. Performance Evaluation of Single-Label and Multi-Label Remote Sensing Image Retrieval Using a Dense Labeling Dataset. Remote Sens. 2018, 10, 964, doi:10.3390/rs10060964. 3. Shao, Z.; Zhou, W.; Deng, X.; Zhang, M.; Cheng, Q. Multilabel Remote Sensing Image Retrieval Based on Fully Convolutional Network. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 2020, 13, 318–328, doi:10.1109/JSTARS.2019.2961634. 4. Shi, Q.; He, D.; Liu, Z.; Liu, X.; Xue, J. Globe230k: A Benchmark Dense-Pixel Annotation Dataset for Global Land Cover Mapping. J. Remote Sens. 2023, 3, 0078, doi:10.34133/remotesensing.0078. 5. Xia, J.; Yokoya, N.; Adriano, B.; Broni-Bediako, C. OpenEarthMap: A Benchmark Dataset for Global High-Resolution Land Cover Mapping 2022. 6. Brown, C.F.; Brumby, S.P.; Guzder-Williams, B.; Birch, T.; Hyde, S.B.; Mazzariello, J.; Czerwinski, W.; Pasquarella, V.J.; Haertel, R.; Ilyushchenko, S.; et al. Dynamic World, near Real-Time Global 10 m Land Use Land Cover Mapping. Sci. Data 2022, 9, 251, doi:10.1038/s41597-022-01307-4. 7. Yang, J.; Huang, X. The 30 m Annual Land Cover Dataset and Its Dynamics in China from 1990 to 2019. Earth Syst. Sci. Data 2021, 13, 3907–3925, doi:10.5194/essd-13-3907-2021. 8. Karra, K.; Kontgis, C.; Statman-Weil, Z.; Mazzariello, J.C.; Mathis, M.; Brumby, S.P. Global Land Use / Land Cover with Sentinel 2 and Deep Learning. In Proceedings of the 2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS; July 2021; pp. 4704–4707. 9. Zhang, X.; Liu, L.; Zhao, T.; Zhang, W.; Guan, L.; Bai, M.; Chen, X. GLC_FCS10: A Global 10-m Land-Cover Dataset with a Fine Classification System from Sentinel-1 and Sentinel-2 Time-Series Data in Google Earth Engine 2025. 10. Lehner, B.; Anand, M.; Fluet-Chouinard, E.; Tan, F.; Aires, F.; Allen, G.H.; Bousquet, P.; Canadell, J.G.; Davidson, N.; Ding, M.; et al. Mapping the World’s Inland Surface Waters: An Upgrade to the Global Lakes and Wetlands Database (GLWD V2). Earth Syst. Sci. Data 2025, 17, 2277–2329, doi:10.5194/essd-17-2277-2025. 11. Zhang, X.; Liu, L.; Zhao, T.; Chen, X.; Lin, S.; Wang, J.; Mi, J.; Liu, W. GWL_FCS30: A Global 30 m Wetland Map with a Fine Classification System Using Multi-Sourced and Time-Series Remote Sensing Imagery in 2020. Earth Syst. Sci. Data 2023, 15, 265–293, doi:10.5194/essd-15-265-2023.