Lossy Compression of Landsat Multispectral Images

Ruslan Kozhemiakin (1) , Sergey Abramov (1) , Vladimir Lukin (1) , Blazo Djurovic , Igor Djurovic (2) , Benoit Vozel (3, 4)

1 Kharkov National University
2 UCG - University of Montenegro
3 ENSSAT - École Nationale Supérieure des Sciences Appliquées et de Technologie
4 IETR - Institut d'Électronique et des Technologies du numéRique

We consider practical aspects of lossy compression with application to multispectral images provided by Landsat sensor. Two facts are taken into account: 1) the inherent noise presence and its properties; 2) rather high degree of component correlation. These properties in different degree are used in 2D and 3D lossy compression. Comparison of the suggested approaches has been carried out for some compression ratios. It is demonstrated that 3D-compression using techniques based on discrete cosine transform (DCT) can provide some benefits but only under condition of proper grouping of sub-band images.

Keywords: noise

Domains: Computer Science [cs]; Networking and Internet Architecture [cs.NI]; Engineering Sciences [physics]; Electronics

Ruslan Kozhemiakin, Sergey Abramov, Vladimir Lukin, Blazo Djurovic, Igor Djurovic, et al.. Lossy Compression of Landsat Multispectral Images. 5th Mediterranean Conference on Embedded Computing (MECO), Jun 2016, Bar, Montenegro. pp.104--107. ⟨hal-01413287⟩