Evaluating the ripening stages of Musa acuminata × balbisiana (Saba) using 2D-image analysis
ILYSSA MARIE D. MOSO, SEAN RAFAEL L. PUDA, PIA O. TIMAAN, and MARIA DEANNA B. JOLITO
Philippine Science High School Western Visayas Campus – Department of Science and Technology (DOST-PSHSWVC), Brgy. Bito-on, Jaro, Iloilo City 5000, Philippines
Abstract
Saba banana is one of the Philippines’ main export products but has a fast ripening process which affects its quality and marketability. Most instruments utilized in measuring banana quality are destructive. Current image analysis studies aim to address this; however, few have used 2D-image analysis in assessing the ripeness of the banana. This paper presents the evaluation of the ripening stage of Saba banana using 2D-image analysis as an alternative, non-destructive method. The 120 banana images were pre-processed by obtaining the mask to remove the background, retaining only the region of interest. Then, further converted into HSV and Grayscale. Ninety of the processed images were utilized as training data sets and the rest as testing data sets for the Support Vector Machine (SVM) where the overall correctness is 70%. The researchers recommend that future studies increase the texture features of the GLCM Analysis to improve the performance of the SVM.
Keywords: banana, image analysis, Python, ripening stage, SVM