AGRISERVER is a sensor network with software for agricultural use. It has been operating for more than 6 years in a vineyard, a chestnut orchard, an apple orchard, and a green-house in Obuse, Nagano Prefecture, Japan. It provides data on temperature, humidity, and solar intensity at each farm, as well as crop images, through the Internet. It has supplied large amounts of data for us in multiple applications in agriculture-related industries. To distribute the AGRISERVER data to the public, we opened a new portal site, "Know-Live". We propose several applications for the effective use of the data, including cultivation and management, food-chain traceability, agribusiness advertising, education, communication between farmers and consumers, and agritourism. Discussions on the applications confirmed that AGRISERVER can be a powerful tool, especially in an agriculture- based society such as that in Obuse.

View full abstract

"Know-Live" is a Web-based farm information system that shows observations from field monitoring devices and encourages users to submit comments (subjective information) on weather and image data. Using a questionnaire survey, we interviewed 110 people, including businesspeople, farmers, and householders, to investigate the effectiveness of Know-Live. The results show that people recognized the system as being useful in improving agricultural technology and providing information on the safety and security of crops.

View full abstract

We developed a high-definition image comparison system to extract information about crop growing conditions. Physiological and environmental data are important for increasing yields. However, it is burdensome for farmers to collect such data every day. Against this backdrop, we developed a monitoring system to automatically collect high-definition crop images, which can then be viewed on a specialized Web-based image viewer and a synchronous viewpoint image viewer. Users can easily observe detailed images over the Internet and find differences among them by using the two viewers. Moreover, they can add comments on an image for other users to view. The results of the observation of apples and grapes suggested that the proposed system provide an efficient way to extract information for identifying the growth stages of crops.

View full abstract

Environmental information on agricultural fields is necessary for farmers to improve crop yields and quality. It has recently become possible to obtain such information at field sites by using networked sensors. However, only a few long-term cases have been stably and continuously observed because of the effort required to manage these sensors in the field. We investigated data collection rates and operational experience to understand data losses and errors in the installation and operation of the Agriserver sensing network. Data were collected through the coordinated actions of a local agent installed in the local network and a remote agent that accessed Agriserver data via the Internet. We designed and implemented a smart algorithm to perform these operations. On the basis of the operational experiences and failures of Agriserver over 6 years, we compiled a useful and practical maintenance manual divided into three sections: Daily Check, Troubleshooting, and Frequently Asked Questions. The manual should help users improve operations and simplify maintenance of field sensors and to use Agriserver more efficiently.

View full abstract

Measuring humidity in agricultural fields is important in forecasting disease and pest infestations, but humidity sensors used under field conditions for long periods are prone to malfunctions because of dust and high humidity. To solve this problem, we developed a way to estimate humidity using a Field Server with an internal humidity sensor to reduce degradation over time. The sensor is protected from dust and high humidity because the outside air is drawn into the Field Server through a filter and warmed. The humidity outside is estimated from the inner humidity, inner temperature, and outer temperature. Simulations showed that the accuracy of the estimated humidity depends in large part on the accuracy of the inner humidity sensor and the temperature sensors. In field experiments, outside humidity estimates were highly accurate; the inner humidity sensor readings averaged 10%RH lower than the actual outdoor humidity. In a long-term field experiment, a humidity sensor installed inside of a Field Server was in good condition after six months and the estimated humidity had similar accuracy rates at the beginning and end of the experiment, whereas a humidity sensor installed in a traditional position over the same time period had deteriorated.

View full abstract

Field Servers can collect real-time sensor data, control devices such as solenoids for irrigation, and test devices such as newly developed sensors in open fields. Therefore, they should be multifunctional, but they also must portable and safe to operate. In creating Field Servers, cost and performance must be balanced, especially in the early stages of development, so we proposed the creation of an "Open-FS (Open Field Server)" from open-source hardware. We also proposed the use of a "Sensor Cloud" as a low-cost method to collect, share, and view sensor data, and developed a system that could be used on publicly available cloud sources such as Twitter. A viewer for the Sensor Cloud was also developed in HTML5. The performance of the system was evaluated through long-term experimental operations at several sites in India and Japan. The Open-FS and the Sensor Cloud worked sufficiently well as we expected.

View full abstract

Real-time weather data were used in the simulation to promote renewable energy utilisation in the smart grid approach. The case study was conducted for the 11 cities in the Tochigi prefecture to show the suitability of renewable energy and to determine the regions that might use it. Solar power had the highest capability of producing electricity, followed by biomass, hydropower, and wind power. The ratio of each renewable energy generation depends on the characteristics of each city's geographical and climatic conditions. The amount of surplus electricity for one year was not considered in the yearly and monthly estimations. In the hourly estimation, 3,363 MWh/year of surplus electricity was available for use in city shortfalls using smart-grid approaches. The maximum surplus electricity was found to be 4,235 MWh/year through a daily estimation; in contrast, 9,075 MWh/year was observed as the surplus in hourly estimations for Nasu city of the Tochigi prefecture. We have found that surplus electricity was higher at Nasu due to the geographic and climatic conditions available for wind power generation. A similar approach to renewable energy potential can be considered for rural areas in Japan for additional sources of energy besides fossil fuel and nuclear plants.

View full abstract

Agricultural information processing systems using GIS (geographical information system)-compatible paddy field maps have been proposed as a way to increase efficiency for regional farmers. But producing paddy field maps is time consuming, because the maps are drawn manually. We modified a raster-to-vector conversion method for a personal computer (PC) to automatically convert a binarized cadastral map in raster format to a paddy field map in vector format. The method, which uses a pixel tracing algorithm, was developed in the programming language C# with the use of 9,757- × 13,084-pixel binarized cadastral maps. It cannot efficiently convert an entire image, but efficiency is improved by dividing the image into several smaller ones and separately processing each. The method performs multi-threaded raster-to-vector conversion of each image, which increases the processing speed, and then assembles the vector images to produce the entire paddy field map. A raster-to-vector conversion of an entire image takes about 6 min on a PC, or 1/8 to 1/20 of the time required to manually create a paddy field map, depending on the area. The method should facilitate the development of GIS-based agricultural programs.

View full abstract