Elevated hydroponic strawberry culture has advantages in saving labor over conventional ground-based soil culture. As the growing beds stand off the ground, inputs and outputs of matter can be measured accurately. I developed a low-cost nutrient management system using a ubiquitous environment control system platform with open-source Arduino hardware for a total cost of about JPY 60,000 per set. The system precisely and continuously measures the volume and electric conductivity of the input and output nutrient solutions. Visualization of water and nutrient uptake by strawberries improves production through precise control of the hydroponic system. Results of practical tests over 2 years improved the management of strawberry culture. The system has been made open source as a smart agricultural tool to support small- to medium-scale greenhouses that grow strawberries in Japan.

‘Yamada Nishiki’ brewer’s rice is grown in a specific part of Hyogo Prefecture. In 2016, we surveyed members of three agriculture cooperatives (called JAs) in Hyogo Prefecture (JA Hyogo-Rokko, JA Minori, and JA Hyogo-Mirai) by questionnaire on their recognition of the benefit of field surveys for growth diagnosis of ‘Yamada Nishiki’ and their preference for doing those surveys. Statistical analysis distinguished two groups: independent farmers, who prefer to diagnose rice plant growth in their fields themselves (20%), and client farmers, who heavily depend on the field surveys by the three JAs (>50%). The independent farmers are willing to develop the necessary skills for such field surveys, such as with a cloud-based analytical system, which uses photos of rice plants taken by farmers on their smartphones, as a decision support system for effective harvest diagnosis and fertilizer usage.

Artificial pollination is a heavy-duty task in pear production. Our research aims at developing a method to detect flowers to be pollinated from camera images in an application of robot vision. Images are taken of flower clusters from a certain distance for determining their position within the tree and of flowers taken from close up for artificial pollination. All images are analyzed by the Faster R-CNN (Regions with Convolutional Neural Networks) model. Branch area information is extracted by a multi-scale filter from the distant view images. The Faster R-CNN model trained on the pear flower image data with the branch area information outperformed the model without that information. When the Intersection over Union is 0.5, the average precision was 0.747 for the distant view images and 0.939 for the close-up images.

Solar sterilization is a physical pest control technique applied by covering the field surface with a polyethylene film in summer and maintaining a soil temperature of ≥40°C for a certain time. If unsettled weather or typhoons occur, farmers need to determine whether the soil has had enough treatment before they decide on their subsequent field management. Thus, prediction of integrated soil temperature during the solar sterilization period is very important for judging the sterilization effect. We developed a method for predicting this from NARO Agro-Meteorological Grid Square Data. By using soil temperatures of the previous 2 weeks, we successfully predicted the future integrated soil temperature. When the temperature trends in July (during the soil temperature survey period) and August (during the prediction period) were similar, the prediction error was <2 days after 9 days. We developed a prototype application for smartphones that incorporates this method to set a prediction period and a target integrated soil temperature for effective solar sterilization, and to display the date when the target temperature should be achieved and the fraction still to achieve.