[4Yin2] Interactive session 2

[4Yin2-05] Tomato yield prediction in greenhouse horticulture based on the average days to harvest

〇Yusei Yoshida¹, Yosuke Kobayashi¹, Kazuhiko Sato¹, Tatsuro Horie², Shinya Watanabe¹ (1.Muroran Institute of Technology, 2.AIR WATER INC.)

Keywords:Facility cultivation, Short-term yield forecast, Plant factory, Tomato, Machine Learning

In this study, we propose a new approach for the short-term yield prediction of tomatoes in greenhouse horticulture.
The main feature of our approach is to use the average number of days to harvest for achieving more higher accuracy forecasting. The short-term yield prediction is very important for the farm manager because this prediction accuracy is directory connected to farm profit.
However, high accuracy short-term prediction is so difficult because tomato growth is affected by a variety of factors and varies widely from individual to individual, even though the environment of the greenhouse can be controlled to some extent, such as temperature and humidity. In this study, two methodologies for predicting the harvest were implemented and compared these accuracies. One is the" direct prediction of harvest date" approach and the other is the" prediction of deviation based on the average number of days to harvest" approach.
In conclusion, there was no significant difference in the results of the current prediction between the methods.