A Background Correction Algorithm for Hyperspectral Images
- Resource Type
- Conference
- Authors
- Sadeg, Said; Cauzid, Jean; Fabre, Cecile; Song, Yingying; Brie, David; Djermoune, El-Hadi
- Source
- 2023 31st European Signal Processing Conference (EUSIPCO) European Signal Processing Conference (EUSIPCO), 2023 31st. :486-490 Sep, 2023
- Subject
- Signal Processing and Analysis
Fluctuations
Geology
Signal processing algorithms
Europe
Signal processing
Fluorescence
Surface fitting
Hyperspectral image processing
background estimation
spatial and spectral regularization
- Language
- ISSN
- 2076-1465
This paper introduces a method for background (or baseline) correction in hyperspectral images. The method is based on the optimization of a criterion incorporating a non-quadratic robust loss (data fidelity term) and both spatial and spectral regularization terms to enforce baseline smoothness. Unlike the classical approach based on a pixel-by-pixel baseline correction, the proposed algorithm exploits jointly the spatial and spectral information. The effectiveness of the the proposed algorithm is demonstrated using simulated and geological hyperspectral images.