In recent days, deep learning has been considered as the state-of-the-art computer vision technique for image classification task. The introduction of Convolutional Neural Network (CNN) made the feature engineering task simple. The classification of various stages of maturity of a fruit is a challenging task using machine learning techniques as it is hard to differentiate the visual feature of the fruits at different maturity stages. In this proposed work, four different ripeness stage of banana were classified using proposed CNN model and compared with the state-of-the-art CNN model using transfer learning. Classification using CNN model requires a huge number of training images to achieve better classification result. The proposed CNN model was trained and tested with both original and augmented images. The CNN model was trained with overall validation accuracy of 96.14%.

近年来，深度学习已被视为用于图像分类任务的最新计算机视觉技术。卷积神经网络（CNN）的引入使特征工程任务变得简单。使用机器学习技术对水果的各个成熟阶段进行分类是一项艰巨的任务，因为很难区分处于不同成熟阶段的水果的视觉特征。在这项拟议的工作中，使用拟议的CNN模型对香蕉的四个不同成熟阶段进行了分类，并与使用转移学习的最新CNN模型进行了比较。使用CNN模型进行分类需要大量的训练图像，以获得更好的分类效果。所提出的CNN模型已通过原始图像和增强图像进行了训练和测试。