加拿大农田数据集

加拿大农田数据集是一个包含78,536个高分辨率（10米/像素，640x640米）地理参考图像的数据集，涵盖10种主要作物类别，收集自2017至2020年的四个作物生产年份及五个月（六月至十月）。每个实例包含12个光谱带、一个RGB图像以及额外的植被指数带。数据集由加拿大农业和农业食品部创建，旨在通过提供精确和连续的农田覆盖监测，推动建立强大的农业环境模型，加速对复杂农业区域的理解。此外，数据集还提供了模型和源代码，允许用户使用单个图像或一系列图像预测作物类别，适用于多时相深度学习分类。

The Canadian Agricultural Field Dataset comprises 78,536 georeferenced high-resolution images, each with a spatial resolution of 10 meters per pixel and a spatial coverage of 640×640 meters. It encompasses 10 major crop categories, and was collected across four crop production years from 2017 to 2020, as well as five consecutive months between June and October. Each data instance contains 12 spectral bands, one RGB image, and additional vegetation index bands. Developed by Agriculture and Agri-Food Canada, this dataset aims to advance the development of robust agricultural environmental models and accelerate the understanding of complex agricultural regions by providing accurate and continuous farmland cover monitoring. Furthermore, the dataset provides accompanying models and source code, allowing users to predict crop categories using either a single image or a sequence of images, making it suitable for multi-temporal deep learning classification.