The ISPRS HyRANK Benchmark

The HyRANK Hyperspectral Dataset and Benchmark Framework

The HyRANK Dataset and Benchmark has been developed in the framework of the ISPRS Scientific Initiatives. The main objective was to fill the current gap regarding the limited availability of hyperspectral datasets and benchmarking frameworks for validating new classification methods against the state-of-the-art.

In the current phase the dataset contains satellite hyperspectral data from the Hyperion sensor (EO-1, USGS). In particular, the “HyRANK Hyperspectral Satellite Dataset I” contains five Hyperion hyperspectral images along with the corresponding reference/ ground truth data.

You can download the “HyRANK Hyperspectral Satellite Dataset I” here

[Reference/ Cite as:] Karantzalos, Konstantinos, Karakizi, Christina, Kandylakis, Zacharias, & Antoniou, Georgia. (2018). HyRANK Hyperspectral Satellite Dataset I (Version v001) [Data set]. Zenodo. http://doi.org/10.5281/zenodo.1222202

 

The training set contains two hyperspectral images (i.e., Dioni and Loukia) and the validation set contains three hyperspectral images (i.e., Erato, Nefeli, Kiriki). The employed nomenclature which generally follows the CORINE Land Cover (CLC) principles, contains 14 land cover classes.

The training set contains two hyperspectral images (i.e., Dioni and Loukia) along with the corresponding ground truth.

The validation set contains three hyperspectral images (i.e., Erato, Nefeli, Kiriki).

 

Therefore, start training your algorithms.

The online evaluation platform will be !! soon online !! in order to quantitative validate your classification maps.