Multi-temporal Crop Classification

该数据集包含了2022年从哨兵卫星收集的美国本土范围内的图像，以及来自美国农业部作物数据层的作物和土地覆盖标签。具体来说，共有3,083张图像用于训练，771张图像用于测试，每张图像包含224×224像素，跨越3个时间步骤的6个波段。该数据集的空间分辨率为中等，任务是对作物进行分类。

This dataset contains images collected over the contiguous United States in 2022 from Sentinel satellites, along with crop and land cover labels from the USDA Crop Data Layer. Specifically, there are 3,083 images for training and 771 images for testing. Each image has a resolution of 224×224 pixels and includes 6 spectral bands across 3 time steps. This dataset has a moderate spatial resolution, and its task is crop classification.