TY - GEN
T1 - Successive nonnegative projection algorithm for linear quadratic mixtures
AU - Kervazo, Christophe
AU - Gillis, Nicolas
AU - Dobigeon, Nicolas
N1 - Publisher Copyright:
© 2021 European Signal Processing Conference, EUSIPCO. All rights reserved.
PY - 2021/1/24
Y1 - 2021/1/24
N2 - In this work, we tackle the problem of hyperspectral unmixing by departing from the usual linear model and focusing on a linear-quadratic (LQ) one. The algorithm we propose, coined Successive Nonnegative Projection Algorithm for Linear Quadratic mixtures (SNPALQ), extends the Successive Nonnegative Projection Algorithm (SNPA), specifically designed to address the unmixing problem under a linear non-negative model and the pure-pixel assumption (a.k.a. near-separable assumption). By explicitly modeling the product terms inherent to the LQ model along the iterations of the SNPA scheme, the nonlinear contributions of the mixing are mitigated, thus improving the separation quality. The approach is shown to be relevant in realistic numerical experiments, which further highlight that SNPALQ is robust to noise.
AB - In this work, we tackle the problem of hyperspectral unmixing by departing from the usual linear model and focusing on a linear-quadratic (LQ) one. The algorithm we propose, coined Successive Nonnegative Projection Algorithm for Linear Quadratic mixtures (SNPALQ), extends the Successive Nonnegative Projection Algorithm (SNPA), specifically designed to address the unmixing problem under a linear non-negative model and the pure-pixel assumption (a.k.a. near-separable assumption). By explicitly modeling the product terms inherent to the LQ model along the iterations of the SNPA scheme, the nonlinear contributions of the mixing are mitigated, thus improving the separation quality. The approach is shown to be relevant in realistic numerical experiments, which further highlight that SNPALQ is robust to noise.
KW - Linear-Quadratic Models
KW - Non-linear Blind Source Separation
KW - Non-linear Hyperspectral Unmixing
KW - Nonnegative Matrix Factorization
KW - Separability and Pure-Pixel Assumption
U2 - 10.23919/Eusipco47968.2020.9287788
DO - 10.23919/Eusipco47968.2020.9287788
M3 - Conference contribution
AN - SCOPUS:85099276365
T3 - European Signal Processing Conference
SP - 1951
EP - 1955
BT - 28th European Signal Processing Conference, EUSIPCO 2020 - Proceedings
PB - European Signal Processing Conference, EUSIPCO
T2 - 28th European Signal Processing Conference, EUSIPCO 2020
Y2 - 18 January 2021 through 22 January 2021
ER -