Sorting and Grading of Cherries on the Basis of Ripeness, Size and Defects by Using Image Processing Techniques

ABSTRACT: Sorting of different fruits according to their ripeness, size, color, appearance, and also proper separation of impurities and appropriate supply to markets of great importance in the fruit trade. In this research, Image Processing Techniques were used in order to analyze the ripeness of cherry and its sorting. The samples were picked from Mashhad black cherry varieties gardens during four stages and with interval of 5 days. Using a CCD camera and in controlled environment, 250 photographs of the samples were taken. For analyzing the fruit ripeness, simultaneously with imaging of the samples, the percentage of their Total Soluble Solids (TSS)(index for ripeness of corps) was measured by Refractor meter. By data analysis and review of images in different color spaces, RGB space was detected as the best space to study fruit ripeness. Analyzing the correlation of different functions between color component sand the index for cherry ripeness showed that the red color component has the highest correlation with TSS(R2=0.9). Also the results showed that the algorithm accuracy in sorting cherries according to their ripeness was 92 %. The results showed that the average accuracy of algorithms on sorting defective cherries was more than 90%, among which the highest accuracy was related to worm-eaten cherries, and the least accuracy was related to squashed cherries. The results for analysis of algorithms in different conditions showed that the maximum accuracy of the algorithm for sorting the cherries is related to sorting due to size (accuracy of 96%).