et al. 2008; Atzberger 2013). Crop development stage is also estimated and reported
using the Crop Growth Monitoring System (CGMS) (Supit et al. 2012).
The Crop Watch (CropWatch) program at the Institute of Remote Sensing
Applications (IRSA) of the Chinese Academy of Sciences (CAS) extensively uses
a suite of satellite remote-sensed data to model and monitor crops worldwide
(Wu and Li 2004; Wu et al. 2010, 2014). The crop condition indices include
vegetation health index (VHI) and vegetation condition index (VCI) that are derived
from remote sensing. Four geospatial levels of crop condition are assessed which are
global monitoring and reporting units, regional major production zones, 31 major
national report, and subnational reports of 9 large counties (Wu et al. 2015).
10.3.4 Limitations and Perspectives
The limitations of the remote sensing approach for crop status monitoring are as
follows.
1. Noncrop-specific spatial resolution of high-temporal remote-sensed data: The
high temporal resolution is crucial in monitoring crop status (Basso et al.
2013). The problem becomes even more serious when small household farms
are investigated (Fermont and Benson 2011). This leaves out many of the veryhigh-spatial-resolution satellite sensors since they have a revisit frequency of
more than 5 days or even weeks. One week during the growing season makes a lot
of difference on crops, while continuous monitoring of crops requires shorter
revisits with high-quality data. Most of the commonly used satellite sensors, like
AVHRR and MODIS, have moderate or even coarse spatial resolution where the
footprint of each pixel is a mixture of many ground features. The mixture makes it
much harder to get crop-specific status over time.
2. The significant effect of cloud and haze on crop condition indices: Optical remote
sensing is the most commonly used technology in monitoring crop conditions.
Cloud covers make it unusable for crop condition. Cloud-free images are hard to
get in certain agricultural areas (King et al. 1995; Eberhardt et al. 2016). Often
cloud masking is applied, while high temporal resolution would help in making
up the missing parts. The surrounding areas of cloud mask are often affected by
light clouds or haze which significantly reduced the vegetation indices (Eberhardt
et al. 2016). This also poses a serious negative effect on crop condition which
leads to a false reading.
3. Saturation of crop condition indices: Vegetation indices are often used as indicators to the health of crops. However, they are suffering saturation problem
when the crop coverage and leaf density are over certain thresholds (Haboudane
2004). The saturated vegetation indices lead to their low correlation to crop yield
or condition (Zarco-Tejada et al. 2005). This made it difficult to relate satellite
information to quantitative crop yield estimates at different scales (King et al.
1995).
194
E. G. Yu and Z. Yang
using the Crop Growth Monitoring System (CGMS) (Supit et al. 2012).
The Crop Watch (CropWatch) program at the Institute of Remote Sensing
Applications (IRSA) of the Chinese Academy of Sciences (CAS) extensively uses
a suite of satellite remote-sensed data to model and monitor crops worldwide
(Wu and Li 2004; Wu et al. 2010, 2014). The crop condition indices include
vegetation health index (VHI) and vegetation condition index (VCI) that are derived
from remote sensing. Four geospatial levels of crop condition are assessed which are
global monitoring and reporting units, regional major production zones, 31 major
national report, and subnational reports of 9 large counties (Wu et al. 2015).
10.3.4 Limitations and Perspectives
The limitations of the remote sensing approach for crop status monitoring are as
follows.
1. Noncrop-specific spatial resolution of high-temporal remote-sensed data: The
high temporal resolution is crucial in monitoring crop status (Basso et al.
2013). The problem becomes even more serious when small household farms
are investigated (Fermont and Benson 2011). This leaves out many of the veryhigh-spatial-resolution satellite sensors since they have a revisit frequency of
more than 5 days or even weeks. One week during the growing season makes a lot
of difference on crops, while continuous monitoring of crops requires shorter
revisits with high-quality data. Most of the commonly used satellite sensors, like
AVHRR and MODIS, have moderate or even coarse spatial resolution where the
footprint of each pixel is a mixture of many ground features. The mixture makes it
much harder to get crop-specific status over time.
2. The significant effect of cloud and haze on crop condition indices: Optical remote
sensing is the most commonly used technology in monitoring crop conditions.
Cloud covers make it unusable for crop condition. Cloud-free images are hard to
get in certain agricultural areas (King et al. 1995; Eberhardt et al. 2016). Often
cloud masking is applied, while high temporal resolution would help in making
up the missing parts. The surrounding areas of cloud mask are often affected by
light clouds or haze which significantly reduced the vegetation indices (Eberhardt
et al. 2016). This also poses a serious negative effect on crop condition which
leads to a false reading.
3. Saturation of crop condition indices: Vegetation indices are often used as indicators to the health of crops. However, they are suffering saturation problem
when the crop coverage and leaf density are over certain thresholds (Haboudane
2004). The saturated vegetation indices lead to their low correlation to crop yield
or condition (Zarco-Tejada et al. 2005). This made it difficult to relate satellite
information to quantitative crop yield estimates at different scales (King et al.
1995).
194
E. G. Yu and Z. Yang