Banana Ripeness Classification: A Comparative Approach with SVM, CNN, and KNN Integrating VGG16

This study presents a novel method for classifying the maturity of bananas by fusing handcrafted characteristics with deep learning based on VGG16. The accuracy of banana characterization is improved by combining data extracted by VGG16 with additional features as contrast, coarseness, directional features, ripeness factor, hue, and HSV values. By addressing the inherent difficulties in manual maturity assessment, this hybrid methodology provides the banana industry with a more precise and effective classification system. To evaluate the efficacy of the Support Vector Machine (SVM), k-Nearest Neighbours (KNN), and Convolutional Neural Network (CNN) models in classifying bananas using the combined feature set, the study uses a comparative analysis. The findings highlight the transformational potential of deep learning in agriculture, particularly in automating visual assessments for enhanced banana maturity determination, and have significant implications for customer happiness and supply chain efficiency. The findings show how machine learning may be applied in the real world to enhance agricultural practices and regulate fruit quality better.