Agriculture Reference
In-Depth Information
and chemical application will be critically important in this type of system. With
the use of real-time kinematic (RTK) GPS systems, about 2.5 cm of accuracy can
be achieved.
The real-time sensing-based approach controls the application rate based on
the current knowledge of pest stress or canopy characteristics. Real-time sensing
systems involve both contact and noncontact sensing to identify either pests that
need to be controlled or the crop and foliage/canopy that needs to be protected.
Sometimes, the sensor system also involves sensing of other indirect variables such
as soil organic matter, soil type, temperature, wind velocity, and rainfall to define
application rate through established relationships. Yet another parameter that is
important to be sensed is actual deposition of chemicals onto the target plant parts
and/or ground. GPS information is not essential in this type of system as the decision
on the rate of application is made based on local information gathered in real time.
However, accuracy of sensing the system in detecting, identifying, measuring, and/
or locating pests, desired parts of plants, or other essential parameters will be criti-
cally important in achieving accurate chemical application. To be applicable, sensors
and onboard computing systems should be capable of providing this information in
real or near-real time. To minimize the effect of delays caused by data processing,
control algorithms, and nozzle actuation, sensors generally precede the nozzles by a
sufficient distance, and a sufficiently fast onboard computing system is used.
Various types of sensors have been investigated and evaluated in the past to achieve
variable rate, targeted, and robotic pesticide application. Noncontact sensors used in
these advanced application systems involved a wide range of electromagnetic spec-
trum from visible to ultrasonic waves. Color cameras, photodetectors, laser scanners,
multispectral and hyperspectral cameras, thermal cameras, and ultrasonic sensors
are some of the examples. These sensors have been used to determine parameters
such as color, shape, and size (Bezenek, 1994; Zhang and Chaisattapagon, 1995), tex-
ture (Meyer et al., 1998), reflectance (Franz et al., 1991; Zhang and Chaisattapagon,
1995), and temperature of pests. This information is then used to categorize pest or
canopy patterns, and to identify and locate them. These sensing systems are also
capable of detecting and localizing target plant canopy or parts (e.g., fruits, foliage).
For example, in specialty crop spraying, ultrasonic and laser scanners are used to
scan and detect the presence or absence of plant canopies (e.g., Solanelles et al.,
2006). In row crops, boom-mounted sensors (e.g., color cameras) have been used to
identify weeds in real time (e.g., Steward et al., 2002). Different types of vision sen-
sors have also been investigated in the past for insect and disease detection in row and
specialty crops (Larios et al., 2008; Martin et al., 2008; Qian et al., 2004; Ridgway
et al., 2001). This sensor input is then used to control the location, direction, and rate
of chemical application. It is, however, important to note that such control for precise
chemical application is a challenging task. Researchers around the world have been
working on such technologies, but with only limited success (Sections 10.3.3-10.3.4).
Ultrasonic and laser sensors have been used in various applications as a low-
cost solution to provide rough estimation of crop canopy shape and size. Ultrasonic
sensors measure the time needed by sound to reflect back from a target to detect if
there is a target canopy within a predefined distance from the sensor. Laser sensors
operate on a similar principle, but use laser light. Wei and Salyani (2004, 2005)
