Studies on Automatic Sorting of Agricultural Products (II)

Application of Fuzzy Set Theory to Grading

Abstract

In the present grading of agricultural products, the shape and peel color are major factors, and they are estimated on the status of appearance, generally, the Fuzzy set theory was applied to calculate the discriminant functions for grading agricultural products.
According to the property factors such as shape index, colored degree and colored uneveness that were reported in the previous papers, the satsuma oranges were attemped to resort into two groups A and B. The results were obtained as follows:
The values of membership functions (estimated functions) of shape index S*, colored degree C_m, and colored uneveness C_v(1) that belong to A group increased (from 0 to 1) as S* increased, C_m and C_v(i) decreased.
After resorting by this method, the shape flakiness of satsuma oranges belong to A group was greater (ratio of diameter and height was smaller), and the peel color was more even and progressive than those belong to B group. Additionally, the standard deviations of all factors (shape index, colored degree, colored uneveness) of A group became smaller as compared with the results that were obtained from manual sorting at the fruits sorting plant, It shows that this method is available for grading agricultural products.