IrrMap
收藏资源简介:
IrrMap是一个大规模、全面的数据集,包含来自LandSat和Sentinel的多分辨率卫星图像,以及关键辅助数据,如作物类型、土地利用和植被指数。该数据集涵盖了美国西部多个州从2013年到2023年的1,687,899个农场和14,117,330英亩土地,为灌溉分析和确保地理空间对齐和质量控制提供了丰富多样的基础。数据集已经准备好用于机器学习,具有标准化的224×224 GeoTIFF补丁、多种输入模式、精心选择的训练-测试分割数据,以及用于无缝深度学习模型训练和灌溉制图基准测试的伴随数据加载器。数据集还附带了一个完整的管道,用于数据集生成,使研究人员能够轻松地将IrrMap扩展到新的区域进行灌溉数据收集,或经过少量修改即可用于农业和地理空间分析的其他类似应用。
IrrMap is a large-scale and comprehensive dataset containing multi-resolution satellite imagery from Landsat and Sentinel, alongside key auxiliary data including crop types, land use categories, and vegetation indices. Spanning multiple states in the western United States from 2013 to 2023, the dataset covers 1,687,899 farms and 14,117,330 acres of land, providing a rich and diverse foundation for irrigation analysis, geospatial alignment assurance, and quality control. The dataset is ready for machine learning applications, featuring standardized 224×224 GeoTIFF patches, multiple input modalities, carefully curated train-test splits, and accompanying data loaders that enable seamless deep learning model training and irrigation mapping benchmarking. Additionally, the dataset comes with a complete dataset generation pipeline, allowing researchers to easily extend IrrMap to new regions for irrigation data collection, or adapt it with minimal modifications for other similar applications in agricultural and geospatial analysis.



