LiDAR provide information beyond spectral signatures. The structures of the
crop ecosystem may be revealed to remote sensor as an effective signature.
4. Advancement of machine learning: Deep learning is found its great success in
image identification and pattern recognition. These newly improved classification
methods will help in improving the extraction and classification of cropland from
very-high-resolution remote sensing data.
10.3 Crop Status Monitoring
10.3.1 Statistical Approach
The statistical ground survey on a sampling framework is the traditional, operational
approach in estimating crop yield and monitoring crop status in many countries
(Hanuschak 2013). The yield forecast in the United States is based on two surveys:
one is the subjective survey, the Agricultural Yield Survey (AYS), from selected
farmers monthly, and the other is the objective plot measurement survey, the
Objective Yield (OY) survey, under a sampling framework for major crops (i.e.,
wheat, corn, soybeans, cotton, and potatoes) (Hale et al. 1999; Good and Irwin 2006;
NASS 2012; Irwin et al. 2014; Good and Irwin 2016). Yield estimates in India are
done using a survey of crop cutting experiments (CCE) under a stratified multistage
random sampling framework (Parihar and Oza 2006).
10.3.2 Remote Sensing Approach
Remote sensing can be an efficient technology to get quick and updated crop
condition throughout the growing season. The most widely used approach is to
use satellite sensors with high temporal resolution and derive certain conditionsensitive indices to evaluate the status against normal values over multiple years
(Yu et al. 2012a). Figure 10.4 shows the typical workflow for monitoring crop status
using time series of remotely sensed observations.
The following are the typical steps for crop status monitoring as shown in
Fig. 10.4:
1. Selection of remotely sensed data: The monitoring of crop status requires frequently revisited observations during the crop growing season. Unlike the crop
mapping which needs high-spatial-resolution multispectral observations, the crop
status monitoring highlights a specific requirement on the temporal resolution
among the three resolution properties of satellite remote sensors. Table 10.3 lists
selected satellite sensors that have been used in crop status monitoring.
10 Crop Pattern and Status Monitoring
185
crop ecosystem may be revealed to remote sensor as an effective signature.
4. Advancement of machine learning: Deep learning is found its great success in
image identification and pattern recognition. These newly improved classification
methods will help in improving the extraction and classification of cropland from
very-high-resolution remote sensing data.
10.3 Crop Status Monitoring
10.3.1 Statistical Approach
The statistical ground survey on a sampling framework is the traditional, operational
approach in estimating crop yield and monitoring crop status in many countries
(Hanuschak 2013). The yield forecast in the United States is based on two surveys:
one is the subjective survey, the Agricultural Yield Survey (AYS), from selected
farmers monthly, and the other is the objective plot measurement survey, the
Objective Yield (OY) survey, under a sampling framework for major crops (i.e.,
wheat, corn, soybeans, cotton, and potatoes) (Hale et al. 1999; Good and Irwin 2006;
NASS 2012; Irwin et al. 2014; Good and Irwin 2016). Yield estimates in India are
done using a survey of crop cutting experiments (CCE) under a stratified multistage
random sampling framework (Parihar and Oza 2006).
10.3.2 Remote Sensing Approach
Remote sensing can be an efficient technology to get quick and updated crop
condition throughout the growing season. The most widely used approach is to
use satellite sensors with high temporal resolution and derive certain conditionsensitive indices to evaluate the status against normal values over multiple years
(Yu et al. 2012a). Figure 10.4 shows the typical workflow for monitoring crop status
using time series of remotely sensed observations.
The following are the typical steps for crop status monitoring as shown in
Fig. 10.4:
1. Selection of remotely sensed data: The monitoring of crop status requires frequently revisited observations during the crop growing season. Unlike the crop
mapping which needs high-spatial-resolution multispectral observations, the crop
status monitoring highlights a specific requirement on the temporal resolution
among the three resolution properties of satellite remote sensors. Table 10.3 lists
selected satellite sensors that have been used in crop status monitoring.
10 Crop Pattern and Status Monitoring
185