2.1 Introduction
The remote sensing technique provides an indirect approach to monitor agricultural
landscapes. Remote sensing data are widely used in crop type identification, crop
growth condition monitoring, crop water use and stress assessment, crop yield
estimation, etc. Agricultural application is one of the most successful applications
in remote sensing. Today, remote sensing data can be acquired from near-surface
using unmanned aerial vehicles (UAV) or aircraft, and from space using satellites
and space stations. The uses of different remote sensing data sources depend on
specific applications and study regions. Remote sensing data have different spatial,
temporal, and spectral resolutions that are designed for different purposes. The
spatial resolution is related to the spatial domain and determines the minimum size
of the target that can be distinguished. The temporal resolution is associated with the
time domain and determines the temporal changes that can be captured. The spectral
resolution reflects the spectral characteristics of targets and determines the types/
features that can be recognized. These are the three main characteristics of remote
sensing techniques. Other features such as multiangular observation and active lidar
system are valuable for improving surface parameterization and can provide surface
structure information.
Spatial resolution in remote sensing refers to the size of a pixel that can be
identified in an image. The spatial resolution for agricultural applications spans
from submeter to meters and kilometers. Different terminologies were used for
describing spatial resolution. Generally, a coarse resolution refers to the pixel size
of 100 m and coarser. Satellite sensors such as Advanced Very High Resolution
Radiometer (AVHRR), Moderate Resolution Imaging Spectroradiometer (MODIS),
and Visible Infrared Imaging Radiometer Suite (VIIRS) belong to coarse spatial
resolution sensors. Although the AVHRR sensor uses “very high resolution,” it is a
coarse-resolution sensor. The medium spatial resolutions are usually used for
10–100 m. Some literature also called this moderate resolution. However, since
the MODIS instrument uses “moderate resolution” for spatial resolutions of 250 m to
1 km (coarse resolution), many publications later used medium resolution for
10–100 m. This range of spatial resolution includes some well-known sensors
such as these aboard Landsat and Sentinel-2 satellites. High spatial resolution
usually means 10 m or finer, and very high spatial resolution normally refers to
the submeter resolution.
Temporal resolution refers to the frequency of observations over the same
location. For UAV or airborne remote sensing observations, observation frequency
depends on the need of the application. It could be just one or a few times during a
short period. Satellite observations have a routine revisit schedule. Coarse-resolution
sensors such as AVHRR, MODIS, and VIIRS have a wide swath width (or a large
field of view) and can acquire global images daily. Medium-resolution sensors such
as Landsat and Sentinel-2 have a relatively longer revisit cycle. For example, the
revisit cycle for Landsat-8 is 16 days, and for Sentinel-2 constellation (A and B
satellites combined) is 5 days. The coarse- and medium-spatial-resolution satellite
8
F. Gao
The remote sensing technique provides an indirect approach to monitor agricultural
landscapes. Remote sensing data are widely used in crop type identification, crop
growth condition monitoring, crop water use and stress assessment, crop yield
estimation, etc. Agricultural application is one of the most successful applications
in remote sensing. Today, remote sensing data can be acquired from near-surface
using unmanned aerial vehicles (UAV) or aircraft, and from space using satellites
and space stations. The uses of different remote sensing data sources depend on
specific applications and study regions. Remote sensing data have different spatial,
temporal, and spectral resolutions that are designed for different purposes. The
spatial resolution is related to the spatial domain and determines the minimum size
of the target that can be distinguished. The temporal resolution is associated with the
time domain and determines the temporal changes that can be captured. The spectral
resolution reflects the spectral characteristics of targets and determines the types/
features that can be recognized. These are the three main characteristics of remote
sensing techniques. Other features such as multiangular observation and active lidar
system are valuable for improving surface parameterization and can provide surface
structure information.
Spatial resolution in remote sensing refers to the size of a pixel that can be
identified in an image. The spatial resolution for agricultural applications spans
from submeter to meters and kilometers. Different terminologies were used for
describing spatial resolution. Generally, a coarse resolution refers to the pixel size
of 100 m and coarser. Satellite sensors such as Advanced Very High Resolution
Radiometer (AVHRR), Moderate Resolution Imaging Spectroradiometer (MODIS),
and Visible Infrared Imaging Radiometer Suite (VIIRS) belong to coarse spatial
resolution sensors. Although the AVHRR sensor uses “very high resolution,” it is a
coarse-resolution sensor. The medium spatial resolutions are usually used for
10–100 m. Some literature also called this moderate resolution. However, since
the MODIS instrument uses “moderate resolution” for spatial resolutions of 250 m to
1 km (coarse resolution), many publications later used medium resolution for
10–100 m. This range of spatial resolution includes some well-known sensors
such as these aboard Landsat and Sentinel-2 satellites. High spatial resolution
usually means 10 m or finer, and very high spatial resolution normally refers to
the submeter resolution.
Temporal resolution refers to the frequency of observations over the same
location. For UAV or airborne remote sensing observations, observation frequency
depends on the need of the application. It could be just one or a few times during a
short period. Satellite observations have a routine revisit schedule. Coarse-resolution
sensors such as AVHRR, MODIS, and VIIRS have a wide swath width (or a large
field of view) and can acquire global images daily. Medium-resolution sensors such
as Landsat and Sentinel-2 have a relatively longer revisit cycle. For example, the
revisit cycle for Landsat-8 is 16 days, and for Sentinel-2 constellation (A and B
satellites combined) is 5 days. The coarse- and medium-spatial-resolution satellite
8
F. Gao