6.4.1 In Situ Data Collection
In situ data collection is a traditional data collection approach that could be traced
back to thousands of years ago when census enumerator went door by door
collecting information about each family. Nowadays, in situ is still a useful way to
collect data. In agricultural activities, scientists use transducer, such as thermometer,
anemometer, and psychrometer, or other in situ instrument, to make measurements
(Jensen 2015). Among the various transducers and other in situ instruments, the
camera is the most common instrument to collect visible-light in situ agricultural
images. Spectroradiometer is another commonly used in situ instrument when
collecting agricultural image data.
However, no matter which transducer we adopt, error is inevitable during in situ
data collection process. It is difficult to get rid of intrusive in situ data collection,
method-produced error, and measurement-device calibration error when
performing in situ data collection. To deal with this problem, some advanced
technologies are adopted in collecting in situ agricultural images. For example,
Haug and Ostermann (2014) propose a benchmark dataset for crop/weed discrimination, to minimize the error brought by an artificial measurement during the in
situ data collection; all images were collected using a camera which is mounted to
the autonomous field robot.
6.4.2 Airborne-Based Data Collection
Airborne-based data collection mission is performed by suborbital aircraft such as
airplanes, helicopters, and unmanned aerial vehicles (UAVs). The airborne digital
cameras mounted on the aircraft could be divided into small-format digital cameras
medium-format digital cameras and large-format digital cameras. The images collected by the small-format digital camera are less than 16 megapixels (MP) per band.
Images collected by the medium-format digital cameras are greater than 16 MP per
band. Large-format digital cameras such as the Leica Geosystems, Ag, and
Microsoft Ultramap, Inc., compared with small-format and medium-format digital
cameras, have very large charge-coupled device (CCD) linear or area CCD arrays
which provide the capability to generate the larger images.
As the first generation of remote sensing data collection approach, airborne-based
data collection has made a huge progress during the past few years. With the
widespread of consumer drone, airborne-based data collection is more accessible
and easier to use than ever. In many agricultural data collection tasks, airborne-based
data collection is an efficient and economic tool to capture the field-level image with
low cost.
88
C. Zhang and L. Lin
In situ data collection is a traditional data collection approach that could be traced
back to thousands of years ago when census enumerator went door by door
collecting information about each family. Nowadays, in situ is still a useful way to
collect data. In agricultural activities, scientists use transducer, such as thermometer,
anemometer, and psychrometer, or other in situ instrument, to make measurements
(Jensen 2015). Among the various transducers and other in situ instruments, the
camera is the most common instrument to collect visible-light in situ agricultural
images. Spectroradiometer is another commonly used in situ instrument when
collecting agricultural image data.
However, no matter which transducer we adopt, error is inevitable during in situ
data collection process. It is difficult to get rid of intrusive in situ data collection,
method-produced error, and measurement-device calibration error when
performing in situ data collection. To deal with this problem, some advanced
technologies are adopted in collecting in situ agricultural images. For example,
Haug and Ostermann (2014) propose a benchmark dataset for crop/weed discrimination, to minimize the error brought by an artificial measurement during the in
situ data collection; all images were collected using a camera which is mounted to
the autonomous field robot.
6.4.2 Airborne-Based Data Collection
Airborne-based data collection mission is performed by suborbital aircraft such as
airplanes, helicopters, and unmanned aerial vehicles (UAVs). The airborne digital
cameras mounted on the aircraft could be divided into small-format digital cameras
medium-format digital cameras and large-format digital cameras. The images collected by the small-format digital camera are less than 16 megapixels (MP) per band.
Images collected by the medium-format digital cameras are greater than 16 MP per
band. Large-format digital cameras such as the Leica Geosystems, Ag, and
Microsoft Ultramap, Inc., compared with small-format and medium-format digital
cameras, have very large charge-coupled device (CCD) linear or area CCD arrays
which provide the capability to generate the larger images.
As the first generation of remote sensing data collection approach, airborne-based
data collection has made a huge progress during the past few years. With the
widespread of consumer drone, airborne-based data collection is more accessible
and easier to use than ever. In many agricultural data collection tasks, airborne-based
data collection is an efficient and economic tool to capture the field-level image with
low cost.
88
C. Zhang and L. Lin