×

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
Remote sensing for resources development and environmental management
Author
Damen, M. C. J.

311
Symposium on Remote Sensing for Resources Development and Environmental Management / Enschede / August 1986
Spectral signatures of soils and terrain conditions
using lasers and spectrometers
H.Schreier
Department of Soil Science, University of British Columbia, Vancouver, B.C., Canada
ABSTRACT: A pulsed infrared laser system was flown over a number of test sites and the results showed that
high resolution data on terrain and vegetation height, and surface reflection can be obtained simultaneously.
With this combined data set it is possible to examine the effect of surface roughness on spectral reflection
measurements and this greatly facilitates terrain and vegetation assessments. To explore the development
of multispectral lasers ground based reflection measurements were made with a spectrometer using three
different sets of soil samples. The results showed that % carbon, % iron and % sand content were significantly
correlated with spectral reflection measurements at 550, 630 and 1600 nm wavelengths respectively. However,
these relationships are not universal and are only applicable on a site specific basis. Organic carbon could
only be predicted with confidence from soils originating from a field where the organic carbon content was
highly variable and greater than 2.0%. In a different sample set where organic carbon content was below 2%
the % sand variability could be predicted. Finally, a third set of samples, originating from mine tailings
having no carbon present and high iron variability, showed significant relationships with reflection at 630 nm
wavelength. These results suggest that if we are interested in quantifying soil fertility conditions the
development and use of a multispectral laser might provide a new dimension to remote sensing since it would
provide both essential reflection and surface roughness assessments at the same time.
1 INTRODUCTION 2 REMOTE SENSING WITH AN AIRBORNE LASER
Surface roughness has a significant influence on de
tailed spectral reflection measurements and until
recently this subject has been largely ignored
because of the difficulties in measuring roughness
by remote means. With the introduction of airborne
lasers it is now possible to not only measure the
terrain surface reflection but also to quantify
surface roughness or height variations. As shown by
Schreier et al (1984), Krabill et al (1984) and
Schreier et al (1985) terrain height variation can be
determined with a pulsed airborne laser reaching
vertical accuracies of better than 20 cm and a newer
model laser as indicated by Jepsky (1986) has shown
even greater accuracy. In this way tree height
measurements and tracings of the canopy of individual
trees is readily possible. The laser generates an
active energy source and laser amplitude or reflec
tion measurements can be carried out simultaneously
with the height measurements and this provides a new
dimension in detailed terrain assessments by remote
sensing. The surface roughness or height component
is of particular interest in vegetation studies which
involve biomass determinations, species identifica
tions, nutrient deficiency assessment through foliage
analysis, and assessment of toxicity or plant stress
for geochemical prospecting. Another application
where the assessment of roughness is of importance is
in the analysis of soil fertility where the surface
structure and cultivation pattern at the soil surface
has a profound influence on reflection.
Current lasers operate at single narrow wavelength
bands and in order to fully exploit this technology
multispectral laser capabilities need to be explored.
It is the aim of this paper to first document the
results of airborne laser test flights so as to
emphasize the laser capabilities at a single wave
length band. The second aim is to provide background
information for spectral properties which are essen
tial for the development, application and use of
multispectral airborne lasers. Examples from the
airborne laser test flights focus on vegetation
applications, while the quantification of soil types
for fertilizer assessments is emphasized in the
second aim.
2.1 Description of laser system
A gallium arsenide laser built by Associated Controls
and Communication Inc. (ACCI) was used in this study
The system operates at 904 nm wavelength, has a
pulse rate of 2000 pps and a peak power output of
80 watts. In order to provide optimum ground
coverage and to facilitate data verification an
inertial navigation system, a photogrammetric camera,
and an airborne data acquisition system were inter
linked with the laser. The footprint of the laser
covered an area of 50 x 50 cm on the ground.
2.2 Terrain assessment with the airborne laser
The system was tested for height and reflection
accuracy over the National Research Council photo
grammetric test site in Sudbury and the research
forest at the Petawawa National Forestry Institute.
As reported elsewhere (Schreier et al 1984) average
height accuracies between 10 and 24 cm were obtained
by the airborne laser. A wide range of vegetation
could be differentiated on the basis of reflection
measurements alone and this in spite of the limita
tion of using a single wavelength frequency.
An example of the dual capability of height and
reflection measurement is provided in Figure 1 which
shows a height and reflection profile of a mixed
forest transect at the Petawawa research station.
As indicated in Figure 1 grass, broadleaf, and
coniferous trees could be separated by reflection
measurements alone but the height profile clearly
facilitates the interpretations since some broadleaf
trees reflect the near infrared energy at a rate
similar to some of the understory vegetation. An
additional component which helps in the interpreta
tion is the fact that the laser penetrates the coni
fer tree canopy much more frequently than the broad
leaf tree canopy. This gives a more spiky tree
profile for spruce and pine trees and a more rounded
profile for maple and poplar trees which in fact
reflects the tree structure. Using a very high pulse