×

You are using an outdated browser that does not fully support the intranda viewer.
As a result, some pages may not be displayed correctly.

We recommend you use one of the following browsers:

Full text

Title
Proceedings of the Symposium on Global and Environmental Monitoring

386
The advantages of the newer generations of
airborne linear array systems include: high
spatial resolution, good radiometric resolution and
sensitivity, flexible wavelength choice and the
ability to use narrow wavelength bands. Some
linear array systems also offer the ability to
produce stereo images from which tree and stand
height estimates should be possible.
Another technological advance is the
incorporation of high qualify inertial navigation
data with airborne linear array data to permit
more efficient and accurate geometric correction to
cartographic coordinates. Linear array data and
forest interpretation can then be input directly
into a cartographically accurate forest inventory
database on a geographic information system. In
fact, with stereo capability it may be feasible to
produce more accurate base maps using geo-
referenced imagery and digital terrain models
created from stereo linear array data.
The following sections describe in more
detail the unique characteristics, advantages and
disadvantages of airborne linear array sensor
systems.
2.1 Spatial Resolution
The spatial resolution of airborne linear
array data can be much higher then equivalent
satellite-based systems and consequently airborne
imagery can used as a viable alternative to
medium-scale aerial photography. For forestry
applications this means that stand-specific or even
tree-specific information can be obtained from
airborne imagery.
The spatial resolution of these systems is
determined by the field of view and the aircraft
altitude. The MEIS II system, for example, can
provide data at resolutions ranging from 0.25 to
10 meters. For forestry applications where
information is often required at a variety of scales,
from reconnaissance mapping to detailed forest
measurements, imagery can be acquired at spatial
resolutions appropriate for each specific
application. In addition, this flexibility to choose
spatial resolution makes airborne linear array
imagery an ideal data source for multistage or
multiphase sampling purposes.
2.2 Spectral Range
For forestry applications where
information on the vigour or condition of the
vegetation is important there is often a need to
use imagery acquired beyond the visible range in
the middle-infrared portion of the electromagnetic
spectrum. The MEIS sensor with a spectral range
from 390 - 1100 nm can provide data in the
important middle-infrared region.
The FLI and CASI systems, on the other
hand, were designed primarily for ocean
applications and consequently are only able to
provide data in the visible and near-infrared
portions of the spectrum from 430 - 805 nm.
There are plans to modify the CASI system to
accommodate different sensor heads which would
extend its spectral range farther into the infrared
region (G.A. Borstad, personal communication)
which would make these systems more suitable for
forestry applications.
2.3 Spectral Resolution
Spectral resolution (the number and width
of spectral bands that can be selected) is controlled
in one-dimensional linear array systems (e.g.
MEIS) through interchangeable optical filters and
in two-dimensional systems (e.g. FLI and CASI)
through diffraction gratings.
The MEIS sensor acquires imagery in up
to eight spectral bands. Various sets of spectral
filters are available for the MEIS system which
have been optimized for specific applications, such
as SPOT and TM simulations, stereo applications
and forestry. Of particular interest for forestry
have been the filter sets for vegetation stress
studies (incorporating passbands of 3 nm width
located on the chlorophyll red reflectance edge)
which have provided more accurate forest insect
and disease damage assessments (Kneppeck and
Ahem, 1989; Epp and Reed, 1986).
The vegetative red reflectance edge, which
occurs in the 650-800nm spectral region, has
received increased attention during the past
several years as a potential indication of vegetative
stress (McColl et aL, 1983). The ability to produce
filters which isolate spectral bands as small as
3nm in this region provides the potential to detect
very small spectral changes in the reflective red
edge which could be important for early detection
of vegetative stress resulting from insect or disease
damage.
The disadvantage of using optical filters to
obtain spectral bands is the time required to
determine the appropriate spectral range for each
filter (usually accomplished by using data from
non-imaging spectrometers) and the time and cost
involved in manufacturing the appropriate filters.
As imaging spectrometers, the FLI and
CASI systems, have the ability to image 288
bands. For the FLI system each band can be as
close as 1.3 nm and as narrow as 2.5 nm (Buxton,
1988). The CASI system can image 288 bands as
narrow as 1.8 nm (Gower et al., 1987), when
operating in its "spectral" mode. The spectral data
from these bands can be used to examine and
analyze the spectral signature of a target. The
spectral signature can be important in helping to
determine optimal spectral band width and
positioning for use in acquiring high spatial
resolution data. When acquiring data in the
"spatial mode" eight spectral bands can be