both topics of crop pattern mapping and crop status monitoring are covered, a
summary conclusion is briefed along with major future trends.
10.2 Crop Pattern Mapping
10.2.1 Statistical Approach
The crop pattern at different administrative levels is often estimated using statistical
sampling (Basso et al. 2013). The statistic approach is still the official, operational
approach by most statistic agencies (Bosecker 1988; Abreu et al. 2010, 2011; Irwin
et al. 2014). In NASS of the USDA, crop acreage is determined mainly by the June
Agricultural Survey (JAS) which is based on stratified sampling frameworks over
the national agricultural states and areas (Bosecker 1988; Good and Irwin 2006;
Lopiano et al. 2011). The European survey of crop acreage uses a classical statistical
scheme based on area frame sampling and ground visits to obtain the main estimation variables in the MARS (Monitoring Agriculture with Remote Sensing) project
(Gallego 1999; Pradhan 2001). The acreage pattern of India is based on a national
survey using a sampling framework (Parihar and Oza 2006). The crop acreage
estimation of China is also mainly based on a stratified sampling framework
(Yang et al. 2007; Wu and Li 2012).
10.2.2 Remote Sensing Approach
The general remote sensing approach includes the following major steps as shown in
Fig. 10.1(Lu and Weng 2007):
1. Selection of remotely sensed observations: Three major aspects are to be considered during the selection of proper remotely sensed data, i.e., spatial, temporal,
and radiometric resolution. Table 10.1 lists some selected sensors and their
resolution characteristics. Besides these intrinsic characteristics of remote sensing, the external factors need to be weighed in during the selection of observations. These may include costs and feasibility. The results are valuable only if
they can be produced and available in the right time frame. In general, higher
resolution would lead to higher accuracy if cost and time are not issues. However,
in reality, these may be the very constraints that the researcher and practitioner
have to work with. The higher resolution also means higher cost and more time
needed to process the data, while these resources may be limited. Higher resolution may open up possibilities for achieving improved classification, while
current technologies may not be able to take advantage of the added features
due to the computing speed or other similar constraints. It is also possible to lead
10 Crop Pattern and Status Monitoring
177
summary conclusion is briefed along with major future trends.
10.2 Crop Pattern Mapping
10.2.1 Statistical Approach
The crop pattern at different administrative levels is often estimated using statistical
sampling (Basso et al. 2013). The statistic approach is still the official, operational
approach by most statistic agencies (Bosecker 1988; Abreu et al. 2010, 2011; Irwin
et al. 2014). In NASS of the USDA, crop acreage is determined mainly by the June
Agricultural Survey (JAS) which is based on stratified sampling frameworks over
the national agricultural states and areas (Bosecker 1988; Good and Irwin 2006;
Lopiano et al. 2011). The European survey of crop acreage uses a classical statistical
scheme based on area frame sampling and ground visits to obtain the main estimation variables in the MARS (Monitoring Agriculture with Remote Sensing) project
(Gallego 1999; Pradhan 2001). The acreage pattern of India is based on a national
survey using a sampling framework (Parihar and Oza 2006). The crop acreage
estimation of China is also mainly based on a stratified sampling framework
(Yang et al. 2007; Wu and Li 2012).
10.2.2 Remote Sensing Approach
The general remote sensing approach includes the following major steps as shown in
Fig. 10.1(Lu and Weng 2007):
1. Selection of remotely sensed observations: Three major aspects are to be considered during the selection of proper remotely sensed data, i.e., spatial, temporal,
and radiometric resolution. Table 10.1 lists some selected sensors and their
resolution characteristics. Besides these intrinsic characteristics of remote sensing, the external factors need to be weighed in during the selection of observations. These may include costs and feasibility. The results are valuable only if
they can be produced and available in the right time frame. In general, higher
resolution would lead to higher accuracy if cost and time are not issues. However,
in reality, these may be the very constraints that the researcher and practitioner
have to work with. The higher resolution also means higher cost and more time
needed to process the data, while these resources may be limited. Higher resolution may open up possibilities for achieving improved classification, while
current technologies may not be able to take advantage of the added features
due to the computing speed or other similar constraints. It is also possible to lead
10 Crop Pattern and Status Monitoring
177