Identification and classification of export quality carabao mangoes using image processing

Abstract

The Carabao mango is the most prevalent and most exported mango variety in the Philippines due to its exotic taste and sweetness, which puts the nation on the global map. As practiced, the quality of mango is assessed by its physical look and weight. Currently, the evaluation of mango is done through manual checking. The utilization of scientific strategy for quality evaluation of mango in this study is done through image processing, which is a more efficient, non-destructive, and cost-effective grading method. Classified sample carabao mangoes from a mango export company were analyzed and become the data sets of the device then undergo image processing procedure through the Support Vector Machine (SVM) algorithm. Carabao Mangoes in the study are classified to be Export Quality, Reject Quality, and Unknown. In this paper, the proposed methodology is divided into three parts, namely: (i) identifying the color of the mangoes through RGB color recognition, (ii) grading of mango based on its weight, (iii) determining the size of the mango by its length and width. The functionality test and statistical analysis revealed 90 percent overall accuracy of the device. © 2019 ACM.