BigEarthNet Remote Sensing Image Understanding Dataset

BigEarthNet is a new large-scale multi-label Sentinel-2 benchmark archive for remote sensing image understanding research. The dataset consists of 590,326 Sentinel-2 image patches, each of which is annotated with multiple land cover categories (i.e., multiple labels) provided by the 2018 CORINE Land Cover Database (CLC 2018).

The patch details are as follows:

• 10m band pixel is 120*120
• 20m band pixel is 60*60
• 60m band pixel is 20*20