A Feasibility Study for a Smartphone-Based Vibratory Bowl System for Peanut Sorting

Abstract

HighlightsConventional peanut sorting and harvest day prediction were time-consuming and inefficient.A smartphone-based vibratory bowl system was fabricated for fast peanut sorting.The system could transport and sort 200 peanuts in less than 5 minutes.The accuracy of peanut counting was greater than 87%, and the accuracy of color detection was greater than 90%.Abstract. Peanut is an important crop in the southern region of the United States. The southeastern states of Georgia, Florida, and Alabama account for over two-thirds of peanut production. Determining an optimum harvest maturity for peanuts is critical to the industry because it directly impacts the yield and quality of the peanuts. The conventional way of determining peanut maturity is a visual inspection method that could take a long time to generate unreliable and inconsistent results due to human errors. In this study, a new method of determining peanut harvest time was proposed. The black peanuts, which is a class of ready-to-pick one, were selected through an automated process. The machine consists of two processes. The transport process used a mechanism of bowl vibration to align the peanuts into a line before entering the second process-the automated sorter, which would separate the peanuts according to their maturity level. Base on the transport analysis, the vibrational bowl was able to transport 200 peanuts to the automated sorter for less than 5 minutes. As for the sorting process, the experimental results showed the accuracy of peanut counting was more than 87%. The average F1 score for peanut color sorting was greater than 90%. All the findings suggested it is feasible to use a smartphone-based vibratory bowl system for fast peanut counting and color sorting. Keywords: Dynamic color detection, Image processing, Peanut maturity, Vibrational bowl.