2018-2019年欧盟10国主要农作物10m种植分布数据集

本数据集涵盖了2018-2019年欧盟10个国家冬麦、春麦、玉米和油菜的种植空间分布。基于哨兵2号(Sentinel-2) 遥感影像数据，利用英国和法国可公开获取的地面样本训练随机森林模型，并将模型迁移到欧盟其它缺乏地面样本的国家进行作物分类。基于2万多个LUCAS实地调查样本进行验证，10个国家的主要作物识别总体精度从89%到97%，生产者精度从72%到98%，用户精度从73%到98%。基于NUTS-2级统计面积进行验证，平均决定系数（R2）大于0.9。

This dataset covers the spatial distribution of winter wheat, spring wheat, maize and rapeseed cultivation in 10 European Union (EU) countries during 2018–2019. Based on Sentinel-2 remote sensing imagery data, a random forest model was trained using publicly available ground samples from the United Kingdom and France, then transferred to other EU countries lacking ground samples for crop classification. Validated using over 20,000 LUCAS field survey samples, the overall accuracy of major crop identification across the 10 countries ranges from 89% to 97%, with producer’s accuracy ranging from 72% to 98% and user’s accuracy ranging from 73% to 98%, respectively. Validation based on NUTS-2 level statistical areas yielded an average coefficient of determination (R²) greater than 0.9.

搜集汇总

背景与挑战

背景概述

该数据集提供了2018-2019年欧盟10个国家主要农作物（冬麦、春麦、玉米和油菜）的10米分辨率种植分布图，基于哨兵2号遥感影像和随机森林模型生成，分类精度高（总体精度89%-97%），适用于农业遥感研究。