Feature extraction for head and broken rice detection using image processing technique

• Main Authors: Bibi Hanibah, Siti Sharifah, Bejo, Siti Khairunniza, Wan Ismail, Wan Ishak, Wayayok, Aimrun
• Format: Conference or Workshop Item
• Language: English
• Published: Faculty of Engineering, Universiti Putra Malaysia 2012
• Online Access: http://psasir.upm.edu.my/id/eprint/33581/1/33581.pdf; http://psasir.upm.edu.my/id/eprint/33581/; http://cafei.upm.edu.my/download.php?filename=/TechnicalPapers/CAFEi2012_47.pdf

Rice (Oryza Sativa) is the most important staple food for a large part of human population, especially in Southeast Asia such as Malaysia and Indonesia. Rice has been graded based on three main components namely: grain composition, milling quality and defective parts. Rice grading is important to ensure only edible rice reaches the consumer standard. It also protects consumers from price manipulation. In this paper, a new approach of image processing technique has been developed to extract rice features. The features used were area, perimeter, minor axis length and major axis length of the rice. The rice images were first segmented automatically from its background by using Otsu’s method. Morphological operation was later being applied to the segmented image in order to eliminate unwanted region(s). Results from the experiment have shown that area gave more consistent results of head and broken rice detection compared to the other features. It is due to the difference in surface coverage area of the rice. Meanwhile, minor axis length gave the worst results due to same value for both broken and head rice. The method give the overall accuracy of 98% when tested using 600 samples of rice image taken from six different percentage of broken.