382
E. Taira
Further, NIR instrument operations such as calibration transfer, database management, and measurement settings require considerable time and effort. To solve these
problems, 16 NIR instruments connected through a networked system have been
introduced in each sugarcane factory for operations such as updating calibration
models and performing database maintenance.
As these NIR instruments were distributed across distances of 1000 km over
these islands, an effective monitoring and control system was necessary. A computer
network was the best solution for this problem. The network’s control center was
located at the sugar association on Okinawa Main Island, where monitoring and
control operations were performed. Each local system was connected to the control
center through the Internet. The network system worked very well and was easy
to operate. The network system operator could conduct daily checks and address
any minor problems experienced by the slave instruments. Further, sugar content
measurement results collected from all the slave systems could be monitored daily.
Such network systems are advantageous when a calibration manager needs to
update the calibration model on all instruments. Once a sample has been measured
using all instruments, the user can correct biases arising due to slight instrument
differences by using a “repeatability file”. Taira et al. showed the calibration result
for Pol in cane (sugar index for payment) with and without a repeatability file [2].
These calibration models showed lower pooled standard error (P-SE) and pooled
bias than the no repeatability file models, with the first-derivative standard normal
variate (1DSNV) pre-treatment showing the lowest root mean square errors of prediction (RMSEPs). The network system can be used to apply calibration models to all
instruments in all regions. Furthermore, this system can be used to estimate nutrient
compositions for supporting fertilization operations (Fig. 17.1).
17.2 Assisting Smart Agriculture in Sugarcane Production
In Japan, growers’ sugarcane prices are based on the sugarcane quality. Therefore,
growers try to maintain the unit yield and preserve and/or increase the sugar content.
Quality data, such as sugar content and yield, for all sugarcane fields are automatically
collected by the quality payment system. Thus, this system functions as a big data
collection system.
Watanabe et al. investigated the nutrients present in sugarcane juice to identify
the key factors affecting sugarcane quality [3]. Juice analysis over a 3-year period
showed that potassium (K
+ ) and chloride (Cl
− ) were the most abundant cation and
anion in the juice, respectively, and that both negatively correlated with the sucrose
concentration. Further, K
+ and Cl
− concentrations varied significantly depending on
the production area. Traditionally, most growers could not obtain information about
the soil and plant chemistry on their farm, and for decades, they simply applied
fertilizers in an unscientific way. Even for sugarcane production, fertilizers, such as
the potassium fertilizer KCl, were applied without confirming the soil condition and
chemical compositions. When typhoons strike these islands, many farms suffer soil
E. Taira
Further, NIR instrument operations such as calibration transfer, database management, and measurement settings require considerable time and effort. To solve these
problems, 16 NIR instruments connected through a networked system have been
introduced in each sugarcane factory for operations such as updating calibration
models and performing database maintenance.
As these NIR instruments were distributed across distances of 1000 km over
these islands, an effective monitoring and control system was necessary. A computer
network was the best solution for this problem. The network’s control center was
located at the sugar association on Okinawa Main Island, where monitoring and
control operations were performed. Each local system was connected to the control
center through the Internet. The network system worked very well and was easy
to operate. The network system operator could conduct daily checks and address
any minor problems experienced by the slave instruments. Further, sugar content
measurement results collected from all the slave systems could be monitored daily.
Such network systems are advantageous when a calibration manager needs to
update the calibration model on all instruments. Once a sample has been measured
using all instruments, the user can correct biases arising due to slight instrument
differences by using a “repeatability file”. Taira et al. showed the calibration result
for Pol in cane (sugar index for payment) with and without a repeatability file [2].
These calibration models showed lower pooled standard error (P-SE) and pooled
bias than the no repeatability file models, with the first-derivative standard normal
variate (1DSNV) pre-treatment showing the lowest root mean square errors of prediction (RMSEPs). The network system can be used to apply calibration models to all
instruments in all regions. Furthermore, this system can be used to estimate nutrient
compositions for supporting fertilization operations (Fig. 17.1).
17.2 Assisting Smart Agriculture in Sugarcane Production
In Japan, growers’ sugarcane prices are based on the sugarcane quality. Therefore,
growers try to maintain the unit yield and preserve and/or increase the sugar content.
Quality data, such as sugar content and yield, for all sugarcane fields are automatically
collected by the quality payment system. Thus, this system functions as a big data
collection system.
Watanabe et al. investigated the nutrients present in sugarcane juice to identify
the key factors affecting sugarcane quality [3]. Juice analysis over a 3-year period
showed that potassium (K
+ ) and chloride (Cl
− ) were the most abundant cation and
anion in the juice, respectively, and that both negatively correlated with the sucrose
concentration. Further, K
+ and Cl
− concentrations varied significantly depending on
the production area. Traditionally, most growers could not obtain information about
the soil and plant chemistry on their farm, and for decades, they simply applied
fertilizers in an unscientific way. Even for sugarcane production, fertilizers, such as
the potassium fertilizer KCl, were applied without confirming the soil condition and
chemical compositions. When typhoons strike these islands, many farms suffer soil