Klasifikasi Mutu Beras Putih Menggunakan Faster Region Convolutional Neural Network

Abstract

Rice is a food derived from rice plants (Oryza sativa) produced after going through the milling process. In Indonesia, rice is the main staple food of the people, especially the people of Western Indonesia. Most people in Indonesia consume white rice. White rice has a level of quality with different prices and benefits for each quality. The provisions of SNI 6128 in 2015 can classify the quality of rice based on the number of head grains, broken grains, groats, red grains and damaged or yellow grains in a set of rice samples. In addition, the determination of rice quality is influenced by the color and texture of rice. This research produces a sistem that can detect the quality of white rice by looking at the color, texture and shape of whole rice. The quality or quality of white rice is then divided into three levels, namely quality A, quality B, and quality C. Quality A in this study is a type of premium rice, quality B is a type of medium rice and quality C is a type of low quality rice. The data used in this study amounted to 450 data which were then divided into 360 training data and validation data and 90 testing data. After testing, this research resulted in an accuracy of 95.55%.