Image Processing Reference
In-Depth Information
Fig. 9.4
R-G-B 1-3-5 of an MNF transformation (same subset as Fig.
9.1
)
9.4.2
Classification and Material Detection
In principal, the same fundamentals apply to the classification of multispectral and
hyperspectral data sets. Well known supervised and unsupervised classification tech-
niques will hence not be considered here. More recent developments, such as the use
of image segmentation and object oriented analysis techniques, are also applicable to
spectral high resolution data and described elsewhere in this volume. In this chapter,
a focus is put on those methods that are more often used with hyperspectral data or
that appear particularly advantageous when applied with hyperspectral data.
There are numerous techniques focusing on either the ability to detect absorp-
tion features in surface materials from imaging spectrometer data or on the extended
feature space of hyperspectral imagery as a whole (or MNF-transformed input).
Absorption based detection of single materials originates from geological applica-
tions, but is also useful in urban environments, where
diverse and spectrally distinct materials occur. This capa-
bility of spectrometric data is generally enhanced by
normalizing spectra via a so-called
convex-hull transfor-
mation
. A mathematically derived curve is fitted to
envelop the original spectrum (hull), utilizing local spec-
tral maxima to connect the hull segments, while leaving
absorption features as spectral gaps below the hull.
Dividing the original spectrum by the hull values results
in a baseline along 1 (or 100% of the hull) and relative
absorption features with depths between 0 and 1 (Fig.
9.5
).
These features are quantifiable in a sense that for exam-
ple the absorption depth or the full width at half maxi-
mum (FWHM) of the absorption feature can be measured
regardless of potential albedo differences in the individual
materials.
It is then possible to compare transformed spectra from imaging spectrometry
data with equally processed spectra from a spectral library. This may be done by
calculating the band-wise residuals between image and reference spectrum and
cumulating these in a root mean squared error (RMSE). A perfect match (which is
a rather theoretical assumption) should yield in zero residuals and would indicate
It is important
to choose the
appropriate
analysis technique
depending on the
questions to be
answered. This
may include
feature based
methods or
illumination
insensitive
techniques
