RESEARCH: SPECTRAL
IMAGING |
|
|
Recent developments in
spectrometer instrumentation facilitate precision measurements of
color and infra-red reflectance from plant leaves and canopies. These
measurements have tremendous potential to quantify incipient water and
nutrient stresses. In 2002 we received a 3-year, $650,000 grant,
jointly funded by the USDA and NASA, to refine the use of spectral
imaging in agriculture. |
CLICK ON THE TITLES TO VIEW ABSTRACTS:
|
|
|
-
Changing Integration Time Has
a Minimal Effect on the
Accuracy of the Apogee/StellarNet
Spectroradiometer
Nick Knighton and B. Bugbee - 2004
|
INTRODUCTION
Changing the
integration time has the potential to change the accuracy of the
Apogee/StellarNet spectroradiometer. All Apogee/StellarNet
spectroradiometers are calibrated to an NIST traceable calibration
lamp. This lamp (LI-COR Inc., model 1800-02; Lincoln, NE) has an
output of 206.2 μmol m-2 s-1, which requires a
relatively long integration time (about 500 ms) to obtain a full scale
response. Sunlight and high-wattage electric light sources often have
a much higher intensity, which require much shorter integration times
(10 to100 ms) to prevent light saturation of the detectors. The
accuracy might be affected differently by increasing or decreasing
integration times from the initial calibration time.
|
|
-
A Mixture of Barium Sulfate
and White Paint is a
Low-Cost Substitute Reflectance Standard for Spectralon
Nick Knighton and B. Bugbee - 2004
|
INTRODUCTION
Barium sulfate is
a white powder that has historically been used as a reference standard
(Weidner and Hsia, 1981). It may be a less expensive alternative to
higher priced white standards that use sintered PTFE (polytetrafluoroethylene;
Spectralon®, Labsphere, Inc., North Sutton, NH). A 500 g
bottle of BaSO4 costs $25, whereas reflectance standard made of
Spectralon® and with a 99% reflectance factor costs $385.
However, pure barium sulfate dries to a powder and easily rubs off
surfaces. We sought to increase the durability of barium sulfate by
mixing it with untinted white latex paint. Reflectance and durability
of different ratios of this BaSO4 and paint mixtures were
measured from 430 to 950 nm with an Apogee-StellarNet
spectroradiometer.
|
|
-
Spectrometer
Curve Smoothing Using
Replicate Scans and Running Averages
Nick Knighton and B. Bugbee - 2004
|
INTRODUCTION
Two techniques are commonly used to reduce noise
in spectral measurements: 1) averaging replicate scans and 2) the
use of smoothing algorithms. We examined the advantages and
disadvantages of curve smoothing by each of these techniques.
Spectrawiz, the software
package available with Apogee-StellarNet spectrometers, allows users
to reduce noise in spectra using two methods:
1) averaging up to 99
replicate scans
2) running average
smoothing algorithm called Boxcar Pixel Smoothing.
|
|
|
|
ABSTRACT Ground-based spectral
imaging devices offer an important supplement
to satellite imagery. Hand-held, ground-based sensors allow rapid,
inexpensive measurements that are not affected by the earth’s
atmosphere. They also provide a basis for high altitude spectral
indices.
We
quantified the spectral reflectance characteristics of hard red
spring wheat (Triticum aestivum cv. Westbred 936) in research
plots subjected to either nitrogen or water stress in a two year
study. Both types of stress reduced ground cover, which was
evaluated by digital photography and compared with ten spectral
reflectance indices. On plots with a similar soil background,
simple indices such as the normalized difference vegetation index,
ratio vegetation index, and difference vegetation index were equal
to or superior to more complex vegetation indices for predicting
ground cover. Yield was estimated by integrating the normalized
difference vegetation index over the growing season. The
coefficient of determination (r2) between integrated
normalized difference vegetation index and final yield was 0.86.
Unfortunately, none of these indices were able to differentiate
between the intensity of green leaf color and ground cover fraction,
and thus could not distinguish nitrogen from water stress. We
developed a reflective index that can differentiate nitrogen and
water stress over a wide range of ground cover. The index is based
on the ratio of the green and red variants of the normalized
difference vegetation index. The new index was able to distinguish
nitrogen and water stress from satellite data using wavelengths less
than 1000 nm. This index should be broadly applicable over a wide
range of plant types and environments.
|
|
-
Comparison
of the Analytical Spectral Devices
FieldSpec Pro JR and the
Apogee/StellarNet
Model Spec-PAR/NIR Spectroradiometers
Dan Dallon
- 2003
|
ABSTRACT
Spectroradiometers can be used to
identify plant stress by measuring crop reflectance.
Spectroradiometers can measure nitrogen and water stress, and
subsequent adjustments in applications can improve crop yield and
reduce over-application. Two portable spectroradiometers and their
accompanying software were compared for accuracy and ease of use
while conducting experiments to determine water and nitrogen stress
in crested wheat grass and soybean plants. The first unit was a
higher priced ($40,000) visible/NIR triple diode spectroradiometer
(ASD FieldSpec Pro JR) with a range of 350-2500 nm. It uses ASD’s
FieldSpec Pro RS3 software. The second was a low-priced
($3600) visible/NIR fixed diode spectroradiometer (Apogee/StellarNet
SPEC-PAR/NIR) with a range of 350-950 nm. It comes with
StellarNet’s SpectraWiz software. The study focused on the software
provided by the two manufacturers because the most noticeable
differences were software-based. Comparisons were made to determine
if the advantages of the ASD spectroradiometer and its software
justified their substantially higher cost. The StellarNet SpectraWiz
software was much more user-friendly and conveniently designed than
the ASD software, which appeared outdated and time consuming to
operate. However, results indicate that the ASD hardware is easier
to use and simpler to configure than the Apogee/StellarNet model and
that the FieldSpec Pro’s increased range was advantageous for water
stress analysis. However, both spectroradiometers showed similar
consistency within the 350-900 nm range.
|
|
|
|
ABSTRACT
Nitrogen stress in crops has been
measured for many years using inexpensive spectrometers with ranges
up to 1000 nm. Detectors for spectral measurements above 1000 nm
are expensive, but they can be used to analyze water stress in
plants by measuring reflectance of the major water bands at 1450 nm
and 1900 nm. Although there is a water band at 970 nm, it is
generally considered to be too small and inconsistent to use as an
indicator of water stress. This experiment aims to determine the
correlation between reflectance at 970 nm and 1450 nm. This was
done by taking single-leaf spectral measurements of excised soybean
leaves over a two hour period. This data was then analyzed using
three water stress indices and compared to the values at 1450 nm.
Reflectance values at 970 nm were highly correlated with values at
1450 nm. The data indicate that water stress can be determined
using the 970 water band.
|
|
-
Selective
Filters Increase Response of
Silicon Diode Spectroradiometers above
600 nm
Dan Dallon, G. Ritchie, and B. Bugbee - 2002
|
INTRODUCTION Reflectance of wavelengths above 600 nm is an important indicator of plant health.
Silicon diode sensors have high sensitivity to green light (500-600 nm),but low sensitivity above 700 nm.
The green radiation in sunlight saturates the detectors before significant NIR radiation is absorbed, making NIR measurements less accurate.
The NIR signal can be increased by using a filter that decreases green radiation without decreasing NIR radiation. This allows increased
instrument integration time and higher NIR absorption.
|
|
-
Wheat
Stress and Reflectance:
Ground Cover, Chlorophyll, and Nitrogen
Glen Ritchie, D. Dallon, D. Wright, V.P. Rasmussen, and Bruce Bugbee
- 2002
|
INTRODUCTION Vegetation reflectance indices can be
used to identify plant health and separate plant growth from spectral contaminants, such as the soil. Leaf
transmittance and reflectance were compared with plant canopy reflectance to determine the
usefulness of using reflectance to detect nitrogen stress in spring wheat
(Triticum aestivum cv. Westbred 936).
|
|
-
Imaging Canopies to Detect Stress Using a Portable Spectrometer
Glen Ritchie, D. Wright, V. Rasmussen,
S. Klassen and Bruce Bugbee - 2002
|
ABSTRACT We conducted field studies in Idaho and Utah
to determine canopy spectral differences between
nitrogen-stressed, water-stressed, and healthy winter wheat.
Spectrometer reflectance measurements made with a portable
spectrometer correlated with plant health and seed yield (highest
r2 = 0.68). Derivative spectra provided a method for
differentiating water-stress from nitrogen deficiency in test
plots.
|
