GeoPlant
收藏资源简介:
GeoPlant数据集由法国国家信息与自动化研究所等机构创建,是一个用于植物物种预测的欧洲尺度数据集。该数据集包含超过500万条异质性仅存在记录和9万条标准化存在-不存在调查记录,覆盖超过1万种植物。数据集内容包括多种环境栅格(如海拔、人类足迹和土壤)、Sentinel-2 RGB和NIR卫星图像、20年气候变量时间序列和Landsat卫星时间序列。数据集的创建旨在通过集成多模态遥感数据,解决物种分布模型中的数据异质性和不平衡问题,从而提高生态监测和保护的效率。
The GeoPlant dataset, developed by institutions such as the French National Institute for Informatics and Automation (INRIA), is a Europe-scale dataset for plant species prediction. This dataset contains over 5 million heterogeneous presence-only records and 90,000 standardized presence-absence survey records, covering more than 10,000 plant species. The dataset includes multiple types of environmental grids (e.g., elevation, human footprint, and soil data), Sentinel-2 RGB and NIR satellite imagery, 20-year time series of climatic variables, and Landsat satellite time series. The dataset was created to address the issues of data heterogeneity and imbalance in species distribution models by integrating multi-modal remote sensing data, thereby improving the efficiency of ecological monitoring and conservation.




