Studies on Sorting Systems for Fruits and Vegetables

(Part I) Color Grading of Apple by Neural Network Systems

Abstract

Uniform color grading of apple is still a serious problem because of judgment errors caused by the seasonal fluctuation in grading criterion, the difference of criterion in producing area and the fatigue of grading workers.
In this study, a method for color grading of apple is examined by neural network systems used image processing instead of human eyes. As the results, the grade judgment ratio is more than 90% at the last recognition error e=0.01 with the neural network systems used in this study. It was proved that the neural network systems has the reappearance of color grading of apple.