Kyomaru・en, an agricultural corporation in Hamamatsu City, Shizuoka Prefecture, is working with Itochu Techno Solutions (CTC; Minato-ku, Tokyo) to develop an artificial harvest forecasting system. We are running an intelligence (AI) model demonstration test. Using an artificial intelligence camera installed in the house, we determine the amount of growth of vegetables and predict the date of harvest, along with information about the environment in the house. It is intended to be useful for planned production and shipment. Hydroponics grows leeks, mini varieties “Hime Chingensai” and “Hime Mitsuba” on a plot of 1-3 hectares. In the demo test, the camera measures the leaf elongation of “Hime Mitsuba” and combines the solar radiation, temperature and humidity, and water temperature data obtained by the ambient sensor with the AIC, creating a predictive model. In addition, the actual measurement confirms whether it can be assembled according to the model. It will start on May 6, 2023 and will last until May 5.
Efficiency in distribution
In Kyomaru-en, the date of sowing, date of planting, number of plants, etc. are recorded in a personal computer table for planned production and dispatch. Based on previous experience, we have predicted the expected harvest date. If there is a difference in growth than planned, you can change the concentration of the hydroponic nutrient solution or open a skylight or curtain 2 or 3 times a week. It is also used to adjust plant growth by closing the plant. Since there were 3 items, it took a lot of time and labor to measure, so we came up with a CTC system that can take measurements with a camera and reduce the burden of actually measuring. I saw it. Kyomaru-en manager Hisashi Kawaguchi said, “We measure and confirm the predicted values of the implemented system every day. It’s almost right,” he says. involve the personnel needed for the harvest to estimate the volume and quantity of shipments. “We can provide a destination for distribution.”
A source: https://news.yahoo.co.jp