The HySpecNet-11k dataset is constructed by the Remote Sensing Image Analysis (RSiM) group at TU Berlin and the Big Data Analytics in Earth Observation group at the Berlin Institute for the Foundations of Learning and Data (BIFOLD).
HySpecNet-11k is a large-scale hyperspectral benchmark dataset made up of 11,483 nonoverlapping image patches acquired by the EnMAP satellite. Each patch is a portion of 128 × 128 pixels with 224 spectral bands and with a ground sample distance of 30 m.
To construct HySpecNet-11k, a total of 250 EnMAP tiles acquired during the routine operation phase between 2 November 2022 and 9 November 2022 were considered. The considered tiles are associated with less than 10% cloud and snow cover. The tiles were radiometrically, geometrically and atmospherically corrected (L2A water & land product). Then, the tiles were divided into nonoverlapping image patches. The cropped patches at the borders of the tiles were eliminated. As a result, more than 45 patches per tile are obtained, resulting in 11,483 patches for the full dataset.
We provide predefined splits obtained by randomly dividing HySpecNet into: i) a training set that includes 70% of the patches, ii) a validation set that includes 20% of the patches, and iii) a test set that includes 10% of the patches. Depending on the way that we used for splitting the dataset, we define two different splits: i) an easy split, where patches from the same tile can be present in different sets (patchwise splitting); and ii) a hard split, where all patches from one tile belong to the same set (tilewise splitting).
For further details about HySpecNet-11k, please see our paper:
M. H. P. Fuchs and B. Demir, "HySpecNet-11k: a Large-Scale Hyperspectral Dataset for Benchmarking Learning-Based Hyperspectral Image Compression Methods," IEEE International Geoscience and Remote Sensing Symposium, Pasadena, CA, USA, 2023, pp. 1779-1782, doi: 10.1109/IGARSS52108.2023.10283385.
If you use HySpecNet-11k in your research, please cite our paper:
M. H. P. Fuchs and B. Demir, "HySpecNet-11k: a Large-Scale Hyperspectral Dataset for Benchmarking Learning-Based Hyperspectral Image Compression Methods," IEEE International Geoscience and Remote Sensing Symposium, Pasadena, CA, USA, 2023, pp. 1779-1782, doi: 10.1109/IGARSS52108.2023.10283385.
hyspecnet-11k-*.tar.gz tarballs use the following command:
cat *.tar.gz | tar -ixzv*-DATA.npy files have to be generated by running the tif_to_npy.ipynb notebook from the HySpecNet Tools.
*-DATA.npy files the number of bands is reduced from 224 to 202 by removing bands [127 – 141] and [161 – 167] that are affected by strong water vapor absorption. The data is clipped and rescaled to the range [0 – 1] using min-max normalization. Furthermore, the data is converted to float32.