Development of weight prediction 2D image technology using the surface shape characteristics of strawberry cultivars

Hyeonchae Yoo1   Jongguk Lim1   Giyoung Kim1   Moon Sung Kim2   Jungsook Kang1   Youngwook Seo1   Ah-yeong Lee1   Byoung-Kwan Cho3   Soon-Jung Hong4   Changyeun Mo5,*   

1National Institute of Agricultural Sciences, Rural Development Administration, Jeonju 54875, Korea
2USDA-ARS Environmental Microbial and Food Safety Laboratory, Henry A. Wallace Beltsville Agricultural Research Center, Beltsville, MD 20705, USA
3Department of Biosystems Machinery Engineering, College of Agricultural and Life Science, Chungnam National University, Daejeon 34134, Korea
4Korea National College of Agriculture and Fisheries, Jeonju 54874, Korea
5Department of Biosystems Engineering, College of Agricultural and Life Science, Kangwon National University, Chuncheon 24314, Korea


The commercial value of strawberries is affected by various factors such as their shape, size and color. Among them, size determined by weight is one of the main factors determining the quality grade of strawberries. In this study, image technology was developed to predict the weight of strawberries using the shape characteristics of strawberry cultivars. For realtime weight measurements of strawberries in transport, an image measurement system was developed for weight prediction with a charge coupled device (CCD) color camera and a conveyor belt. A strawberry weight prediction algorithm was developed for three cultivars, Maehyang, Sulhyang, and Ssanta, using the number of pixels in the pulp portion that measured the strawberry weight. The discrimination accuracy (R2) of the weight prediction models of the Maeyang, Sulhyang and Santa cultivars was 0.9531, 0.951 and 0.9432, respectively. The discriminative accuracy (R2) and measurement error (RMSE) of the integrated weight prediction model of the three cultivars were 0.958 and 1.454 g, respectively. These results show that the 2D imaging technology considering the shape characteristics of strawberries has the potential to predict the weight of strawberries.

Figures & Tables

Fig. 1. Strawberry picture of three cultivars