30-meter Resolution Cropland Distribution Dataset of the YarlungZangbo-Lhasa-Nyangqv River Region, Tibet Autonomous Region, 1990-2023

The spatial distribution dataset of cropland in the Yarlung-Zangbo-Lhasa-Nyangqv River region from 1990 to 2023 was extracted based on Landsat 5 Level 2, Collection 2, Tier 1 and Landsat 8 Level 2, Collection 2, Tier 1 data. High-resolution images were used for the cropland sample labeling in 2023, and for generating validation point sets in 2015 and 2020. The ResUnet model was applied to extract cropland information from Landsat data. The model underwent fine-tuning to adapt to the characteristics of Landsat 5 data, ensuring consistency in accuracy across different sensors and years. Eight year’s products were finally obtained, corresponding to the distribution of cropland in the region in 1990, 1995, 2000, 2005, 2010, 2015, 2020 and 2023.

1990至2023年雅鲁藏布江-拉萨河-年楚河流域耕地空间分布数据集，基于陆地卫星（Landsat）5号Level 2、Collection 2、Tier 1与陆地卫星（Landsat）8号Level 2、Collection 2、Tier 1数据提取生成。2023年借助高分辨率影像完成耕地样本标注，并以此生成2015年与2020年的验证点集。本研究采用残差U-Net（ResUnet）模型从Landsat系列遥感数据中提取耕地信息，并对模型进行微调以适配Landsat 5数据的特征，确保不同传感器与年份间的精度一致性。最终共得到8个年份的耕地分布产品，分别对应该流域1990、1995、2000、2005、2010、2015、2020及2023年的耕地空间分布情况。