Classification of Rice Grain Varieties Based on Morphological Features Using Image Processing and Machine Learning Techniques
Patrick Neil M. Santiago, Rodelio P. Pascual, Marites M. Cumbe, Joel A. Macaso, Filwyn P. Villanueva, Isaac Angelo M. Dioses
Abstract
Rice is one of the world's most important cereal grain crops. Nowadays, most Filipinos are expected to finish their dinner with rice. Rice is the most popular food item, both domestically and internationally, due to its nutritional benefits. The greatest area is devoted to palay, the nation's main crop and in 2019, 4.65 million hectares were harvested, with 336 thousand hectares of rice, Nueva Ecija was the highest on the list. The province has many still farms and leads the world in palay production in 2019. Manually identifying palay grains can be laborious and skill-intensive, and accuracy may be hampered by conditions like color blindness or poor vision. In this study, using image processing and MATLAB, the researchers successfully extracted different features from palay grains such as major and minor axis, perimeter, area, and geometric mean diameter as parameters in identifying palay. Using the KNN algorithm we successfully created a model that can classify variety of palay with 92.5% accuracy which is high. In this way researchers can set the standards of rice grain variety ensuring the quality of rice exportation in the future.