基于Sentinel-2的时间序列多国基准数据集(Sen4AgriNet)

在这项工作中，我们介绍了Sen4AgriNet，这是一个基于Sentinel-2的时间序列多国基准数据集，专为使用机器和深度学习的农业监测应用而定制。Sen4AgriNet数据集根据通过地块识别系统（LPIS）收集的农民申报进行注释，以协调全国范围内的标签。这些声明最近才作为开放数据提供，首次允许根据地面实况数据对卫星图像进行标记。我们将根据联合国粮食及农业组织（FAO）的指示性作物分类方案，在整个欧洲提出并标准化一种新的作物类型分类，以满足共同农业政策（CAP）的需求。具体使用方法请曾参阅https://github.com/Orion-AI-Lab/S4A?tab=readme-ov-fileIn this work, we introduce Sen4AgriNet, a Sentinel-2-based temporal multi-national benchmark dataset customized for agricultural monitoring applications using machine and deep learning. The Sen4AgriNet dataset is annotated based on farmer declarations collected via the Land Parcel Identification System (LPIS) to harmonize labels across national boundaries. These declarations have recently been made available as open data, allowing satellite imagery to be labeled with ground truth data for the first time. We propose and standardize a new crop type classification across Europe in accordance with the indicative crop classification scheme of the Food and Agriculture Organization of the United Nations (FAO), to meet the requirements of the Common Agricultural Policy (CAP). For specific usage instructions, please refer to https://github.com/Orion-AI-Lab/S4A?tab=readme-ov-file

搜集汇总

背景与挑战

背景概述

Sen4AgriNet是一个基于Sentinel-2卫星时间序列的多国基准数据集，专为农业监测的机器学习和深度学习应用设计。其标注基于农民通过土地地块识别系统(LPIS)的声明，并采用联合国粮农组织的作物分类方案进行标准化，覆盖2019-2020年全球范围，旨在支持欧洲共同农业政策的需求。