A Symbolic Representation and Classification of Fruits

Abstract

In the real word scenario, automation of digital system plays a vital role, especially in the field of agriculture which needs a good automated system for classification of different fruits as it consumes customer’s time and is also very useful to the farmers. In this paper, we propose a system to classify different fruit classes using symbolic representation and classifier. Firstly, texture, color and shape features are extracted, and then natural clustering is applied on the feature fusion matrix. By making use of interval of mean and standard deviation, intra class variation is captured. For experimentation, 1200 images of 10 fruit classes is collected and totally 12000 samples are used. Further, symbolic classifier is used to obtain the confusion matrix and comparative study is made to show the robustness of symbolic representation and classifier with existing methods, SVM and KNN classifiers.