Online Grading of Fruits using Deep Learning Models and Computer Vision

Fruit quality evaluation is an important task in many industrial applications especially, in processing units. In Industries separating bad quality fruits manually is expensive and time consuming. The classification and grading of fruits when done manually is not precise. This work presents an efficient methodology for fruit classification using deep learning. Most existing works done to address the fruits classification based on quality focus on a single variety of fruit and a more general system with good accuracy is not available. In this paper, a Convolutional Neural Network based quality evaluation system for multiple fruits is presented. The proposed work is evaluated by comparing with state-of-the-art works based on two datasets and it achieves an accuracy of 99.12% and 97.67% for those. To make the proposed work available to public a web application has been created and the classification model is integrated to that.