Data Repository

The following collection of datasets has been prepared for ImageLab, the files can be loaded directly into ImageLab(1).

  Imaging Method(s) Image Size Spectral Size Download
Coffee beans mixed with stones
near infrared 319x359 156
[69 MB]
This remarkable dataset contains the hyperspectral image of a mixture of coffee beans and rocks. The spectra exhibit lots of spikes and saturated regions which form an excellent test case for algorithms detecting and correcting artifacts in hyperspectral images.
Many thanks to Viktoria Parrag for providing the data; Reference: Parrag, V; Felföldi, J; Firtha, F (2014): Roasted and green coffee beans with stones: a hyperspectral training data set for multivariate image analysis. CORVINUS University of Budapest, Department of Physics and Control, Faculty of Food Science.
mass spectrometry 40x50 8
[1.5 MB]
Concrete sample taken from the wall of an road tunnel. MS image acquired by laser ablation in combination with ICP/MS; eight specific isotopes have been measured: 13C, 25Mg, 27Al, 29Si, 35Cl, 42Ca, 49Ti, and 57Fe.
Data courtesy Andreas Limbeck, Vienna Univ. of Technology.
EDX (energy dispersive xray) + Raman 101x101 12 (EDX) + 1024 (Raman)
[61 MB]
A multisensor dataset which combines energy dispersice xray (EDX) and Raman spectra of a sample of particulate matter obtained in the salt lake district of Western Australia.
Data courtesy Johannes Ofner, Vienna University of Technology.
Reference: H. Lohninger, J. Ofner: Multisensor hyperspectral imaging as a versatile tool for image-based chemical structure determination. Spectroscopy Europe Vol. 26 No. 5 (2014), 6-10“.
Gulf of Mexico Wetlands
visible & near infrared 320x401 360
[64 MB]
This data sample is taken from SpecTIR’s extensive hyperspectral baseline environmental dataset covering multiple National Wildlife Reserves in the Gulf of Mexico during the period of May-June 2010. This specific sample is from the Lower Suwanee NWR, collected at a 2m spatial resolution and a nominal 5nm band spacing covering the spectral range of 395 nm to 2.0 µm.
Atmospheric Particulate Matter
EDX (energy dispersive xray, 12 elements + backscatter signal)
101x101 13 (EDX, incl. backscatter)
1024 (Raman)
[80 MB]
This multisensor dataset contains data fused from three imaging methods (Raman, energy-dispersive x-ray, and scanning electron microscope).
Data courtesy Johannes Ofner; Reference: J. Ofner, K.A. Kamilli, E. Eitenberger, G. Friedbacher, B. Lendl, A. Held, H. Lohninger: Chemometric analysis of multisensor hyperspectral images of precipitated atmospheric particulate matter.
Analytical Chemistry 87 (2015) pp. 9413-9420
48x48 768
[7.2 MB]
An HSI of an electronic banking card showing the transponder chip and parts of the copper coil. The data contain both the THz spectra and the pulses for time domain spectroscopy.
Data courtesy Harald Pühringer, Recendt GmbH, Linz, Austria.
mass spectrometry 120x120 4
[0.7 MB]
Petal of a Paeonia which has been watered in water containing blue ink. The hyperspectral image has been obtained by LA-ICP-MS.
Data courtesy Max Bonta of Vienna University of Technology, Inst. of Chemical Technologies and Analytics.
ToF-SIMS 512x512 52
[5.2 MB]
The combination of ToF-SIMS and HRSEM allows the investigations of hydrogen embrittlement in duplex stainless steel microstructures.
Data courtesy Wiley.
Reference: O. Sobol, G. Holzlechner, M. Holzweber, H. Lohninger, T. Boellinghaus, W.E.S. Unger: "First Use of and Multivariate Analysis and Data Fusion of ToF-SIMS and HR-SEM image data for Studying Deuterium Assisted Degradation Processes in Duplex Steels", ECASIA 2015 Proceedings, to be published in Surface and Interface Analysis, Wiley 2016

(1) Please note: the datasets published on this page have been specially signed to allow full access to the data even with the evaluation copy of ImageLab (required version: 0.99 or higher). Any change of the data or the corresponding meta information will render the digital signature invalid.