FECT

Abstract: Classifying rice grains plays a crucial role in quality assessment and sorting processes within the agricultural and food industries. This study presents a comprehensive approach to rice grain classification by integrating image processing and machine learning techniques. The objective is to develop an efficient and accurate system for automating the classification of rice grains based on their intrinsic characteristics. Image processing methods are employed to extract relevant features from digital images of rice grains. These extracted features serve as informative inputs for machine learning algorithms, facilitating the classification of rice grains into empty paddy rice grain (aat), ripe paddy rice grain (tasak), and under-ripe paddy rice grain (matak) classes. The support Vector Machines (SVM) and the K-Nearest Neighbor (KNN) algorithm are explored for their effectiveness in handling this task. The experimental result showed that the SVM algorithm achieved 97.17% accuracy compared to the KNN algorithm which achieved only 76.16% accuracy.

Keywords: Rice Grain, Image Processing, Machine Learning, Classification, SVM, KNN.