LAUSR

BACKGROUND In this study, artificial intelligence models that identifies sunn pest damaged wheat grains (SDG) and healthy wheat grains (HWG) are presented. Svevo durum wheat cultivated from Konya area are used for the process. 150 HWG and 150 SDG are used for classification. Thanks to the constructed imaging setup, the photos of 300 wheat grains are obtained. 17 various visual features of each wheat are extracted by means of image processing techniques and these features evaluated in 3 different groups of dimension, texture and pattern as visual parameters. Artificial bee colony (ABC) optimization based Artificial Neural Network (ANN) and Extreme Learning Machines (ELM) algorithms are firstly implemented to classify the damaged wheat grains. RESULTS The correlation-based feature selection (CFS) technique is also utilized to find the most effective features among the 17 features. In the classification process by using selected 5 features, the mean absolute error values for ABC based ANN and ELM are calculated as 0.00174 and 0.00433, respectively. The proposed technique is integrated into the Graphical User Interface software and an effective detection system is constructed for practice use. CONCLUSION The results indicate that, thanks to the modified ANN algorithm and implemented CFS algorithm, the detection accuracy of the damaged wheat grains is considerably increased. This article is protected by copyright. All rights reserved.