In conventional plant factories, the judgment of whether a seedling is good or bad depends on the subjective judgment of the operator, and only low-precision sorting is possible. This technology uses two AI models, one that estimates the height, width, and weight of spinach seedlings from their images, and the other that estimates the seedlings that will grow well based on these estimated values, to determine whether the seedlings are good or bad, including the possibility of growth after transplanting. The AI model can properly select normal seedlings at the time of transplanting, thereby increasing the yield and productivity of the plant factory.
Farm Ship Co., Ltd. and Pye Material Design Co., Ltd. will verify, under a contract of NEDO, the accuracy of this technology and accumulate cultivation data, including the weight of transplanted seedlings at the time of harvest, with the aim of achieving advanced judgment and automation of transplanting work. Furthermore, in combination with the demand and weight forecasting technology and growth control technology developed as part of this project, they aim to realize a highly accurate supply and demand adjustment system that will lead to food loss reduction.