AUTOMATIC CLASSIFICATION OF NORMAL AND ABNORMAL ORANGES USING IMAGING TECHNIQUES

Abstract

The objective of this study is to classify normal and abnormal of orange using Image Processing and Artificial Neural Network. Machine vision and image processing techniques have been found increasingly useful in the vegetable industry, especially for applications in quality inspection and evaluation of food grains, fruits, vegetables, processed foods, and defect sorting. Orange fruit is one of the highly demanded vegetables for daily consumption of the world population. In this paper, a new automated orange sorting technique is presented. It sorts orange fruit based on color, shape, and texture defects. We use image processing techniques for segmentation of oranges and then based on our algorithm categorize them according to the type of defect. The defects considered are anthracnose, canker and blue mold. We created an image data set containing images of 60 orange samples. Color, shape and textural features extracted from 300 images. For each scenario, 5 images were captured. ANN was used to classify and identify orange fruits by using extracted features. For the classification, 12 color, 5 morphological, and 12 textural features, a total of 29 predominant features were extracted from images of four classes of orange fruits. We used a three-classification setup design by using color, shape and textural data set were used as input of a set of the network. From these data sets 70 %( 210), 15% (45), and 15% (45) were used in the network for training, validation, and testing respectively. The accuracy of classification using color, shape and texture features was perfectly classified. The performance network was best for all classifications.