BAYESIAN APPROACH AND IMAGE PROCESSING TECHNIQUE FOR MANGO CLASSIFICATION

Abstract

This paper presents Bayesian approach for mango classification based on digital image processing. We design algorithm to classify mango into two groups: green mango and ripe mango. Color features are extracted from RGB mango image and red intensity of the mango image using statistic calculation. The Bayesian is applied to classify the color-based feature of the mango image. We evaluate the performance of the Bayesian approach against another technique; the support vector machine (SVM). Both methods were implemented and tested with 100 mango images. The experimental results show the superiority of the proposed method, with an accuracy of 88%, as compared to SVM-based method, with an accuracy of 83%.