x-bigearthnet

BigEarthNet (BEN) is a large-scale remote sensing dataset designed for multi-label land cover classification. This version of the dataset consists of 22,000 paired Sentinel-1 SAR and Sentinel-2 multispectral image patches, selected from the original BEN collection. Each image pair is geo-referenced and annotated with one or more land cover classes based on the CORINE Land Cover (CLC) database. The dataset includes 12 multispectral bands from Sentinel-2 and 2 SAR bands (VV, VH) from Sentinel-1, making it suitable for multimodal learning tasks.

BigEarthNet（BEN）是一款专为多标签土地覆盖分类任务设计的大规模遥感数据集。本版本数据集从原始BEN数据集集中选取了22000对配对的Sentinel-1合成孔径雷达（Synthetic Aperture Radar，简称SAR）与Sentinel-2多光谱图像切片。每一对图像均经过地理参考，并基于CORINE土地覆盖（CLC）数据库标注了一个或多个土地覆盖类别。该数据集包含Sentinel-2的12个多光谱波段以及Sentinel-1的2个SAR波段（VV、VH），适用于多模态学习任务。