Classification of Fruit Ripeness Grades using a Convolutional Neural Network and Data Augmentation

Abstract

Currently the classification processes of the degree of maturity of fruits require the use of complex systems, which, most of the times, are not within the reach of small farmers or consumers who do not have knowledge of the characteristics that a fruit must have in order to be catalogued as immature, mature or rotten. For this reason, a tool that can be accessed by anyone, was designed and implemented through a mobile application that served as an interface. This article describes the use of a convolutional neural network for the classification of the degree of maturity of the following fruits: red apple, green apple, banana, orange and strawberry. First, two sets of images were constructed. Secondly, the data argumentation technique was performed and then the training of the convolutional neuronal network was performed using the dataset images as input. In order to know the performance of the different models generated, the following metrics were used: precision, accuracy, recall, log loss, and f1 score. The best average precision obtained was 96.34%.

Keywords

• data augmentation
• fruit ripenning
• maturity
• deep learning
• convolutional neural networks.