Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/16ACD52E-556A-4660-8626-9131F913A83D | ||
| Year | 2017 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Ioannis Gkouzionis, "Spectral cube reconstruction from multiplexed spatial and spectral data", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2017 https://doi.org/10.26233/heallink.tuc.69353 | ||
| Appears in Collections |
Hyper-spectral imaging is a non-destructive detection technology that integrates conventional imaging and spectroscopy to get both spatial and spectral information of a target. This technique has in recent times become a powerful tool for scientific and industrial analysis in many different fields. Its applications range from aerial agricultural surveillance, through food safety and forensics to space exploration and biomedical applications. Hyper-spectral images represent observations of a scene at many different wavelengths and most importantly associate to each pixel in the imaged scene a full spectral vector or spectral signature. The work presented in this thesis deals with a new method in acquiring and reconstructing the spectral cube in hyper-spectral imaging. The method employs an electro-optical device that acquires and stores the spectral cube in a spatio-spectral multiplexed fusion. The next step is to disentangle the multiplexed information, so that the spectral cube to contain a stack of images each one captured at a different wavelength. Two methods for the reconstruction of the spectral cube were implemented and analyzed in the present diploma dissertation. The first one is based on image interpolation, while the second one is based on the diagonal connection of the spectral bands (tomographic reconstruction). Image interpolation refers to the estimation of intensity values at unknown locations, while tomographic reconstruction relies on the resampling of the stored data with the aid of geometrical transformations. In this way, we achieve a straightforward de-multiplexing of the dataset and the reconstructed spectral cube consists of hundreds of narrow band spectral images. The above features make our approach suitable in demanding spectral imaging applications, such as microscopic images and non-destructive analysis.
Copyright © DIAS 2013
