Supporting Table Grape Berry Thinning with Deep Neural Network and Augmented Reality Technologies

• DOI: 10.2139/ssrn.4110968
• Publication Date: 2022-05-28T12:13:08Z
• Authors: Prawit Buayai; Kabin Yok-In; Daisuke Inoue; Hiromitsu Nishizaki; Koji Makino; Xiaoyang Mao
• Keywords: Table grape; Thinning; Table (database)

Berry thinning is a crucial process in table grape cultivation. This because not only are visual features, such as bunch compactness, form, and berry size, important factors affecting market value, giving sufficient space for each to grow also largely affects the quality, sugar concentration, of final product. requires professional skills, it usually accomplished by experienced farmers. Furthermore, appropriate period limited two weeks; hence, tasks have become bottleneck increasing yield products. paper addresses aforementioned issue proposing system empowering naïve farmers begin without in-person coaching from expert The proposed employs deep neural network model learn knowledge required identifying berries be removed, uses augmented reality technology present instructions based on this through smart glasses. was validated throughout entire growing season real field Yamanashi prefecture, Japan. It confirmed that unskilled can execute immediately training. they familiar with quickly. products who used an 8.18% higher average quality score than those skilled