In the agriculture sector, one of the issues for farmers is unstable income due to fluctuations in yields and produce prices, among other factors. One of the main factors affecting yields other than external environmental factors such as natural disasters is disease. The most effective way of preventing disease is to apply preventative agrochemical treatment around the time of infection. However, since disease is not visible to the eye until it has actually taken hold, it has been difficult to ascertain the optimal timing to applying the treatment. Moreover, to manage the dosage and timing of the agrochemical, it is important to recognize the signs of disease. In greenhouse cultivation, along with the basic parameters of temperature and humidity, disease occurrence is also influenced by the quantity of sunlight, leaf wetness, the growing environment and the outside weather. Plantect ™ uses AI to analysis these factors that affect crop development in order to predict disease.
Cloud-based data analysis using AI achieves disease prediction accuracy of 92 percent
Plantect ™ is a service comprising hardware for measuring the environment inside a greenhouse and software for predicting disease occurrence based on the measurement data. The hardware includes sensors for measuring temperature, humidity, sunlight, and carbon dioxide. These are installed inside a greenhouse and the measurement data is sent to a cloud. Users can access the data in the cloud via a web-based app using various devices, such as smartphones or PCs. This enables them to check the inside environment in the greenhouse in real time, or browse past data, at any time and from any location.
In addition to the monitoring function, Plantect ™ also has a function for predicting the occurrence of disease. Using the data sent to the cloud by the monitoring function, Bosch's proprietary algorithm analyses elements related to disease occurrence, such as leaf wetness, links the analysis with weather forecasts, and displays plant disease infection risk notifications on the app. In 2017, Bosch established a new research center, where it has brought together all of the internal organizations that had been working on AI research. The goal of the center is to bolster the company's expert knowledge of AI. Bosch plans to invest 300 million euros through to 2021 on expanding research and development in this field. For Plantect ™, Bosch developed a disease prediction algorithm at its new center, using data from over 100 greenhouses and its expertise in AI technology. Plantect ™ predicts disease based on Bosch's proprietary algorithm and the monitoring data from each greenhouse. This is a different approach from the wide-area warnings that have been used to date, enabling each user to receive customized disease predictions. In testing on past data, the system has recorded a prediction accuracy of 92 percent.
Focus on user experience making it easy for small- and medium-scale farmers to implement
Wireless: Plantect ™ uses a long-range wireless communication protocol (LoRa) due to low energy consumption. Furthermore, since it is battery-powered, initial installation requires no wiring such as electrical power sockets or communication cables. It can be easily installed in any position within the greenhouse. It uses readily available alkali batteries, which can last for up to one year.
Reasonable operation cost: The Plantect ™ fee plan is designed to make it easy for users to introduce the service. It is available for a monthly usage charge, with no initial cost. Service usage charges have been set for the basic monitoring function and disease prediction function separately.
Easy-to-use user interface: The system has an integrated design that is easy to understand. Users can enlarge the data that they wish to monitor on the display. If they wish to obtain more detailed information, they can check it simply by tapping the display. The intuitive operation enables even those who are not familiar with computers and smartphones to check the environment in the greenhouse easily.
Using the IoT to expand business domains
Currently, the Plantect ™ disease prediction function is limited to greenhouse grown tomatoes; but Bosch plans to expand its application to strawberries, cucumbers, ornamentals and flowers, and other agricultural products going forward. The company is also planning sales in other countries with highly promising markets for greenhouse cultivation. Regarding Plantect ™, Dr. Wolz says: "Bosch is a global supplier of sensors. In addition to our strength in hardware, recently we have greatly accelerated our investment in IoT, such as middleware and cloud platform operations to support IoT solutions. Plantect ™ is an example of how our investment in IoT has resulted in expansion into new business areas. I am very proud that it was created by a Japanese Bosch start-up."
|Product name||Plantect ™|
|Start of order receiving||From June 8th, 2017|
|Start of shipping||From August, 2017 onward|
|Monthly usage fees||Monitoring function: 4,980 yen / month|
Disease prediction function: 3,350 yen / month
|Product website URL||http://www.bosch.co.jp/plantect|
|Inquiries||Inquiries can be made via the following telephone number or website.|
|Announcement of innovative service for smart agriculture|
FUJI Project, Project Leader