A STUDY ON NOVEL FRAMEWORK FOR DETECTING CHEMICALLY RIPENED MANGO FRUITS

Abstract

Computer vision systems are now commonly utilised for identifying, classifying, and grading various types of fruits. For mango fruit classification and maturity detection, existing research focuses on characteristics such as fruit size, colour, shape, and texture. Though much work is put into identifying, maturing, and detecting defects in mango using colour cues, there are few attempts made in the direction of identifying artificially ripened mango fruits. In contrast, just a few attempts have been made to delve deeper into the colour characteristics of the fruits. To address these challenges, this study provides a novel framework based on MPEG-7 colour descriptors for detecting fake ripening of mango fruit. The proposed scheme consists of three stages: first, image pre-processing, which includes masking, filtering, segmenting, and cropping; second, dominant colour extraction using dominant colour descriptors, which is then mapped with the help of clustering to identify the artificial ripening of mango fruit; and finally, dominant colour extraction using dominant colour descriptors, which is finally mapped with the help of clustering to identify the artificial ripening of mango fruit. Experiments were conducted on two separate datasets including four different varieties of mangoes, demonstrating the suggested method's robustness and efficiency in comparison to existing approaches.