Comparing nonlinear mixing models for vegetated areas using simulated and real hyperspectral data
Joint work with L. Tits, B. Somers, Y. Altmann and P. Coppin
Most of the unmixing algorithms proposed in the
literature rely on the widely acknowledged linear mixing model to
describe the observed pixels. Unfortunately, this model has been
shown to be of limited interest for specific scenes, in particular
when acquired over vegetated areas. Consequently, in the past few
years, several nonlinear mixing models have been introduced to take
nonlinear effects into account while performing spectral unmixing. These models have been proposed
empirically, however without any thorough validation. In this paper,
the authors take advantage of two sets of real and physical-based
simulated data to validate the accuracy of various nonlinear models
in vegetated areas, namely the Fan model, the Nascimento/Somers model, the generalized bilinear model and the polynomial post-nonlinear model (to have a brief overview of these models, the interested reader is invited to consult this paper). These physics-based models, and
their corresponding unmixing algorithms, are evaluated with respect
to their ability of fitting the measured spectra and of providing an
accurate estimation of the abundance coefficients, considered as the
spatial distribution of the materials in each pixel.
The preliminary results are reported in the paper published in IEEE J. Sel. Topics Appl. Earth Observations Remote Sensing in 2014:
- article
[ .pdf ].
The datasets used in this study are publicly available online:
The Matlab codes for the unmixing algorithms based on the models under test are available here:
- matlab codes
[ .zip ].
