AUTOMATIC FRUIT CLASSIFICATION USING DEEP LEARNING FOR INDUSTRIAL APPLICATIONS

Abstract

Fruit classification is an important task for many industrial applications. A fruit classification system may be used to help a supermarket cashier identify the fruit species and prices. It may also be used to help people decide whether specific fruit species are meeting their dietary requirements. In this paper, we propose an efficient framework for fruit classification using deep learning. More specifically, the framework is based on two different deep learning architectures. The first is a proposed light model of six convolutional neural network layers, while the second is a fine-tuned visual geometry group-16 pre-trained deep learning model. Two color-image datasets, one of which is publicly available, are used to evaluate the proposed framework. The first dataset (dataset 1) consists of clear fruit images, while the second dataset (dataset 2) contains fruit images, that are challenging to classify. Classification accuracies of 99.49% and 99.75% were achieved on dataset 1 for the first and second models, respectively. On dataset 2, the first and second models obtained accuracies of 85.43% and 96.75%, respectively. 

Existing System

Fuzzy model, fitness-scaled chaotic artificial bee colony (FSCABC) optimization technique.

Proposed System

This system presents an efficient deep learning based framework for visual fruit classification. More specifically, our framework is based on the convolutional neural networks (CNNs), which are multi-layered feed-forward neural networks that are able to learn task-specific invariant features in a hierarchical manner. Deep learning models have intensively utilized during the last few years for automatic feature engineering. These models showed robust ability in feature representation. They became a common solution to deal with the rapid growth of heterogeneous big data. These models achieved excellent accuracy in different approaches such as image classification, object recognition, and speech recognition. For this reason, we were encouraged to employ deep learning for fruit classification. The proposed framework avoids the problems of other shallow learning-based approaches. Moreover, the need for huge annotated data to fit deep models was solved via the application of the transfer learning and augmentation principles. The presented framework investigates two different deep learning architectures. It is evaluated on two distinct datasets of color fruit images. 

Architecture

BLOCK DIAGRAM