Fruits Sorting with Instance Based Image Processing

Abstract

In the agricultural industry which is among the most prominent sectors on a world scale, image processing for efficiency and quality has also become an important domain. Automation systems are working with image processing to increase the efficiency of the classification of fruits. Image processing is a method developed for digital image formatting and performing some operations for obtaining or extracting useful information from a specific image. There are several methods and programs for processing images. In this study, the results obtained by the classification of apple, pear and peach fruits which are similar to each other, but which differ in terms of roughness and color, by using image processing techniques are presented. After the data sets of these fruits are formed, color and roughness are determined by image processing techniques, and success rates are compared by using Instance Based (IB) and Bayesian Network classifiers in the analysis phase. Finally, the performances of the fruit classifications were investigated using Receiver Operating Character Curve (ROC).