Design of Machine Learning Solutions to Post-Harvest Classification of Vegetal Species

This paper presents a machine learning approach to automatically classifying post-harvest vegetal species. Color images of species were applied convolutional neural networks (CNNs) and support vector (SVM) classifiers. We focused on okra as the target classified it into two quality types. However, our could also be other The solution consists several components, each design process its combinations are essential for classification quality. Therefore, we carefully investigated their effects accuracy. Through experimental evaluation, confirmed following: (1) in color space selection, HLG (hue, lightness, green) HSL saturation, lightness) species; (2) suitable preprocessing techniques required owing complexity data noise load; (3) diversity extension image by mixing different datasets obtained under conditions is quite effective reducing overfitting possibility. results this study will assist AI practitioners development classifications based learning.