Hyperspectral
Research - Home
|
The hyperspectral research unit of VITO develops several
applications based on airborne imaging spectroscopy.
However, spaceborne data or data originating from other
sensors are frequently combined with airborne hyperspectral
data.
This website gives a brief overview of these research
activities and lists all recent publications. Our activities
are mainly concentrated on vegetation mapping, on coastal
and aquatic applications and on the development and
application of classification techniques:
|
A semi-empirical
method was developed to produce near-surface suspended
sediment concentration maps of the Scheldt
river. Maps of sediment concentration at a given time
are essential to advise government agencies and dredging
companies to assure access to the harbour of Antwerp.
|
| VITO uses airborne
hyperspectral images (and geomorphology maps and bathymetry
maps) to monitor coral reefs. Maps
of bottom types occurring on the coral reef structure
can be produced. These maps can assist in the identification
of critical areas of the coral reef that require intensified
protection.
|
VITO
developed a method to map sand types
along the Flemish coast based on airborne hyperspectral
images. Laser altimetry offers the opportunity to
establish accurate digital elevation maps of beaches.
By combining hyperspectral measurements and laser
altimetry, and by comparing the different acquisitions
in time, sand transport along the coast can be estimated.
Knowledge of sand transport enables policy makers
to take preventive measures (e.g. beach nourishment).
|
| VITO
also uses airborne hyperspectral images to investigate
and to map sediment composition on
difficult accessible locations. Accurate spectral
information enables deducing the grain size of the
sediment as well as the moisture content, the organic
matter content and even the main pigments among which
chlorophyll a. This information can be used to map
sediment ecotope types which are important to determine
the biological value of the sediment, but also to
verify the susceptibility to erosion.
|
VITO developed a
semi-automatic method to map vegetation.
Accurate knowledge of the spatial distribution of
different vegetation types is important for
- studying the biodiversity in nature reserves
- managing recreational parks
- spatial planning
- preventive detection of fire risk.
|
| VITO developed a
generic procedure for the classification of
hyperspectral images. Features are extracted
from the complete reflectance spectrum or features
from wavelet coefficients are used. The classification
procedure selects features using a floating search
procedure and uses linear discriminant as an optimization
function. This generic procedure was validated for
several applications like stress detection in orchards,
water quality monitoring, …. The developed procedure
also provides information on the discriminating spectral
positions and band width and the spectral resolution
required for a particular application or for the development
of new sensors.
|
|
|
|
 |
|
