Published July 6, 2021 | Version v1
Dataset Open

Sea-Ice Data Content Representation Based on Latent Dirichlet Allocation for Belgica Bank in Greenland


Data description

File type : -.npy (python numpy file)

File content: Each file is a numpy array of size (number of 256x256 patches, 4096) indexed by id of the patch (each scene contains 6,400 patches, each patch has 4,096 micropatches of size 4x4, assigned one topic [1] per micropatch, resulting in 4,096 topics per patch). Each file has 4 months of observation. Array size is 25600 x 4096. We provide 6 files containing 24 months of observation (see the excel file for the Sentinel-1 ids) [2].

Software to open with: Python

Example code:

import numpy

Data= numpy.load(“filename_with_path”)



1. C. Karmakar, C.O. Dumitru, G. Schwarz, and M. Datcu, “Feature-Free Explainable Data Mining in SAR Images Using Latent Dirichlet Allocation”, IEEE JSTARS, vol. 14, pp. 676-689, 2021.


Files (244.7 MB)

Name Size Download all
244.7 MB Preview Download

Additional details

Related works

Is cited by
Journal article: 10.1109/JSTARS.2020.3039012 (DOI)


ExtremeEarth – From Copernicus Big Data to Extreme Earth Analytics 825258
European Commission