NON-DESTRUCTIVE CLASSIFICATION OF FRUITS BASED ON COLOR BY USING MACHINE LEARNING TECHNIQUES

Inclusion of computer vision and image processing technique in agriculture field providing a user friendly environment in quality testing and grading of fruit before placing it in market. Automated quality testing and grading of fruits influenced with the extracted quality parameters of fruit image and number of dataset used in training phase of machine. Color, size, texture and surface defect are the basic parameters in quality measure of fruits in agriculture field. This paper focused on classification accuracy based on extracted color and geometric features of fruit mango (magniferia Indica) and used sample size in training phase of machine learning algorithm. In this study maturity of mango is predicted with extracted color features by using a combination of RGB, HSI, HSV color model and classification is done using  Naivebayes and BPNN machine learning algorithms. This complete method passes from three phases 1) image pre-processing 2) feature extraction and 3) classification. The experimental results illustrate the usefulness of these measures by providing the prior information in classification. BPNN results are satisfactory in both cases of used features like i) color features ii) color and geometric features.