Optimization of Sorting Processes of Partially Rotten Mango for Food Processing with Image Processing

Abstract

This research aims to develop a method for improving the process of separating partially rotten fruit of ripe mango to the process of creating new creative Mango Stir-Fry. The partial rotten approximation process was based on image processing using Gaussian Mixture Model, HSV color model, dilation, erosion, Otsu’s binarization thresholding, and image color proportions techniques for mobile application development. The results showed that the techniques presented in this study could sort and approximate partially rotten mangoes with an accuracy of 99.06%. Moreover, the developed mobile application can approximate the net weight of ripe mango pulp and the net weight of Mango Stir-Fry to be obtained. These approximation errors were 0.68% and 1.91% for the net weight of net ripe mango pulp and Mango Stir-Fry, respectively. Therefore, this work helps the cooking process to classify partially rotten ripe mangoes and estimate the volume of processed products more conveniently.