Estimating the Number of Endmembers to Use in Spectral Unmixing of Hyperspectral Data with Collaborative Sparsity

Lucas Drumetz 1 Guillaume Tochon 1 Jocelyn Chanussot 1 Christian Jutten 2
1 GIPSA-SIGMAPHY - SIGMAPHY
GIPSA-DIS - Département Images et Signal
2 GIPSA-VIBS - VIBS
GIPSA-DIS - Département Images et Signal
Abstract : Spectral Umixing (SU) in hyperspectral remote sensing aims at recovering the signatures of the pure materials in the scene (endmem-bers) and their abundances in each pixel of the image. The usual SU chain does not take spectral variability (SV) into account, and relies on the estimation of the Intrinsic Dimensionality (ID) of the data, related to the number of endmembers (NOE) to use. However, the ID can be significantly overestimated in difficult scenarios, and sometimes does not correspond to the desired scale and application dependent NOE. Spurious endmembers are then frequently extracted and included in the model. We propose an algorithm for SU incorporating SV, using collaborative sparsity to discard the least explicative endmembers in the whole image. We compute an algorithmic regularization path for this problem to select the optimal set of endmembers using a statistical criterion. Results on simulated and real data show the interest of the approach.
Type de document :
Communication dans un congrès
Petr Tichavsky; Massoud Babaie-Zadeh; Olivier Michel; Nadège Thirion-Moreau. 13th International Conference on Latent Variable Analysis and Signal Separation (LVA/ICA 2017), Feb 2017, Grenoble, France. Springer, Latent Variable Analysis and Signal Separation, 10169 (10169), pp.381 - 391, 2017, Theoretical Computer Science and General Issues. 〈10.1007/978-3-319-53547-0_36〉
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01479420
Contributeur : Christian Jutten <>
Soumis le : mardi 28 février 2017 - 18:27:12
Dernière modification le : samedi 28 juillet 2018 - 01:14:48
Document(s) archivé(s) le : lundi 29 mai 2017 - 16:17:40

Fichier

drumetz_lva_ica2017.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Citation

Lucas Drumetz, Guillaume Tochon, Jocelyn Chanussot, Christian Jutten. Estimating the Number of Endmembers to Use in Spectral Unmixing of Hyperspectral Data with Collaborative Sparsity. Petr Tichavsky; Massoud Babaie-Zadeh; Olivier Michel; Nadège Thirion-Moreau. 13th International Conference on Latent Variable Analysis and Signal Separation (LVA/ICA 2017), Feb 2017, Grenoble, France. Springer, Latent Variable Analysis and Signal Separation, 10169 (10169), pp.381 - 391, 2017, Theoretical Computer Science and General Issues. 〈10.1007/978-3-319-53547-0_36〉. 〈hal-01479420〉

Partager

Métriques

Consultations de la notice

514

Téléchargements de fichiers

297