Classification of Iraqi Anber Rice by Using Image Processing and KNN Algorithm

Abstract

Image classification takes a large area in computer vision in term of quality or type or data sharing and so on Iraqi Anber Rice in they need this kind of work, where few in the field of computer science that deal with the types of Iraqi Anber rice, and because of the Anber Rice are grown and produced in Iraq only, and because of the importance of rice around the world and especially in Iraq. In this paper a proposed system distinguishes between the classes of Iraqi Anber Rice that Grown in different parts of Iraq, and have their own specifications for each class by using moment invariant and KNN algorithm. Iraqi Anber Rice that is more than Fiftieth class Cultivated and irrigated in different parts of Iraq, and because of the different methods of agriculture and irrigation, they differ in their characteristics and qualities and taste, and the image shape of grains differs from one class to another one. All grain enters the image processing stage to prepare the image to the next stage. A feature extraction stage to extract seven moments for each grain and then began the classification process using an algorithm 1-nearest neighbor (KNN) and it calculates the Euclidean distance between test image and training images. After the implementation of the proposed system the result was good, where its compared test image (one image) with the training image(100 image). The success rate of the classification(83%) and after applying the confusion matrix to calculate the recall and precision. The value f recall is (84.0585) and the value of precision is (82.6358), these results for the use of nine classes of Iraqi Anber Rice.