Determining ‘Carabao’ Mango Ripeness Stages Using Three Image Processing Algorithms
Abstract
Harvested mangoes are commonly classified or sorted manually. This method is tedious, time consuming, inaccurate and prone to errors. Human inspection is also subjective and factors like visual stress and tiredness may arise that can result in the inconsistencies in judgment. The use of a chroma meter is reliable but the equipment is expensive. This study explored the use of three digital image processing algorithms to determine harvested ‘Carabao’ mango ripeness stages. Canny edge detection, Sobel edge detection, and Laplacian of Gaussian detection algorithms were used to extract a mango image from its original image. The mean red, green, and blue (RGB) values of the detected images were converted to L*a*b* color values which were used to identify the ripeness level of the mango image based on the standard generated from gathered data of ‘Carabao’ mangoes. The standard generated was also based on the mango peel color index scale from University of the Philippines Los Baños Postharvest Horticulture Training and Research Center (PHTRC). The algorithms’ performance had an overall accuracy of 80.5% for canny edge detection algorithm and L*a*b* color extraction using neural networks; 63.88% for Sobel edge detection algorithm and L*a*b* color extraction using rgb2lab function in MATLAB software; and 17.33% Laplacian of Gaussian detection and L*a*b* color extraction using OpenCV. Overall, the implementation of Canny edge detection algorithm for image processing and L*a*b* color extraction using neural networks performed best among the algorithms used in classifying ‘Carabao’ mango ripeness stages. To improve the performances of the algorithms, it is recommended to improve the quality of the sample images by controlling the light, exposure, and camera to be used, matching it with more chroma meter sample points on the ‘Carabao’ mango to attain a better color average of the sample mango.
Keywords: ‘Carabao’ mango · image processing · L*a*b* color values · neural networks
Copyright (c) 2019 Miguel Carlos S Guillermo, Deborah S Naciongayo, Mary Grace Galela, Czaryl P Dioquino, Emma Ruth V Bayogan, Cinmayii G Manliguez
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
In the spirit of open access and free distribution of knowledge, the Banwa journals do not charge authors for the publication of their articles. All costs for journal production and online hosting and archiving are borne by the publisher. In return for cost-free publication of their articles, the authors give the university the nonexclusive right to publish articles in electronic and print versions of the journal and to include the article in future reprinting of the issue in whole or in part. Authors may make use of their articles published in the journal, whether in its original or altered form, provided that acknowledgment is given to the journal for publishing them first.