Agriculture Reference
In-Depth Information
strata in a large cotton field. McKinion et al. (2009) demonstrated an automated
technology for early season control of cotton insect pests, including software sys-
tems for processing RS imagery to spatially variable application maps and wireless
local area networks for automated field delivery of data. Their study indicated that,
comparing spatially variable application of insecticide for the tarnished plant bug
control with the traditional uniform application, the spatially variable method could
reduce a cost of insect control by up to 60%.
Machine-harvested cotton must be properly defoliated before harvesting in order
to achieve high fiber quality. Defoliation is conducted by applying abscission chemi-
cals that cause plants to drop their leaves and promote boll opening. Because crop
maturity varies across a field, defoliants should not be applied uniformly within the
field. Fridgen et al. (2003) used remotely sensed imagery for variable application
of cotton defoliant. Compared to conventional uniform application, VRAs reduced
chemical use by 18%, an economically and environmentally beneficial result.
According to a report by Bagwell (2004), VRA of defoliants based on NDVI could
achieve defoliant cost savings of up to 40% without significant impact on lint yield
and quality. The actual savings depends on crop in-field variability; the greater the
variability, the greater the savings. However, a 20-30% cost reduction was found to
be common.
6.2.4 S
ENSOR
-B
ASED
S
YSTEM
FOR
V
ARIABLE
R
ATE
N A
PPLICATION
Cotton must receive appropriate rates of N fertilizer for optimal yield and quality; both
underfertilization and overfertilization with N can negatively affect the desired growth
pattern of cotton plants, and thus degrade fiber quality and reduce yield (Fernandez
et al., 1996; Gerik et al., 1998). Additionally, overfertilization with N will increase
production costs while increasing the potential for negative environmental impacts
(Bakhsh et al., 2002; Potter et al., 2001). Therefore, it is desirable to have a system that
is able to apply the appropriate amount of N according to plant needs.
Nitrogen concentration in cotton plants can be detected with spectral reflectance
measurements. To use this ability in a system for on-the-go N application, a real-
time ground-based sensing system is required to determine N needs of the plants.
Heege and Thiessen (2002) used on-the-go sensing of crop canopy reflectance to
control site-specific N top dressing. Kostrzewski et al. (2003) tested the ability of
a ground-based system in measuring water and N stresses in cotton. Both systems
depended on reflectance measurements based on incident solar radiation on the crop
canopy. However, variations in solar angle and atmospheric conditions changed the
amplitude and the spectral characteristics of sunlight reaching the crop canopy and
reduced the accuracy of spectral data (Pinter, 1993). To eliminate the effect of ambi-
ent light change on sensor performance, Stone et al. (1996) and Sui et al. (1998,
2005) included artificial illumination in developing ground-based optical sensors for
measuring plant canopy reflectance. Sensors that provide their own light source are
referred to as active optical sensors.
A sensor-based VRA system involves equipment capable of diagnosing plant
growth status and applying appropriate amounts of inputs according to plant needs
on the go. A sensor-based VRA system usually consists of sensors, controllers, and
