Rice Grain Classification using Image Processing & Machine Learning Techniques
Biren Arora, Nimisha Bhagat, L.R. Saritha, Sonali Arcot
Abstract
Rice Grain Classification becomes very important as there are multiple rice grain types available in the market today. Classifying rice grains as per rice types manually is not feasible nor efficient. Classification can be a really tedious task when it comes to doing it manually instead of automatically. This would consume a lot of efforts as well as a lot of time would be wasted. There is a need for an intelligent and smart system which can overcome this difficulty by automating this process. It should be able to identify and classify individual rice grains according to the respective type automatically. The collection of data set should be the primary process. This includes extraction of various parameters of individual rice grains like major axis, minor axis, eccentricity, length, breadth, just to name a few. The system will utilize this information to train the computer. Each rice grain or image would be allocated to its respective class. Classes used in this project are surti kolam, idli rice, long grain basmati and boiled rice. Any rice sample that has been encountered in the system will be first classified and then will be segregated into its respective class. This would keep the entire system organized and segregated. Managing and keeping a track of different rice types is important and its proper classification in an industrial environment becomes crucial. Automating the system would encourage the industry to have future scope for its implementation according to the changes required as per the industry requirements.