GeoLifeCLEF 2023

GeoLifeCLEF 2023数据集由法国国家信息与自动化研究所创建，旨在通过高空间分辨率（约十米）评估和预测欧洲大陆植物物种的分布。数据集包含约五百万条仅存在记录和六千条详尽的存在-不存在调查，涵盖了大约一万种植物物种。创建过程中，研究人员从选定的GBIF数据集中提取并协调了这些异质性记录，并将其与多种环境栅格数据关联，包括10米分辨率的RGB和近红外卫星图像以及20年的气候变量和卫星点值时间序列。该数据集的应用领域包括生态模型和机器学习，旨在解决物种分布预测中的偏差问题，特别是在大规模和高空间分辨率下的预测。

The GeoLifeCLEF 2023 dataset was developed by the French National Institute for Research in Computer Science and Automation (INRIA). Its core objective is to evaluate and predict the distribution of plant species across the European continent at a high spatial resolution of approximately 10 meters. The dataset contains roughly 5 million occurrence-only records and 6,000 comprehensive presence-absence surveys, covering approximately 10,000 plant species. During the creation process, researchers extracted and harmonized these heterogeneous records from selected Global Biodiversity Information Facility (GBIF) datasets, and associated them with various environmental raster data, including 10-meter resolution RGB and near-infrared satellite imagery, as well as 20-year climate variables and satellite pixel value time series. The dataset has applications in ecological modeling and machine learning, aiming to address bias issues in species distribution prediction, especially for large-scale and high-spatial-resolution prediction tasks.