فردین اخلاقیان طاب

This paper describes a classifier fusion method for identification of rice variety as bulk samples using run length matrix features. Forty four features from run length matrix were extracted from bulk samples of different rice varieties. To find the best features, imperialist competition based feature selection was used. Finally, seventeen features were selected as the superior ones. The overall classification accuracy, found by classifier fusion technique, was achieved as %93.38.