An Efficient Rice Variety Identification Scheme Using Shape, Harlick & Color Feature Extraction and Multiclass SVM

Abstract

Rice is primary food harvests that each and every one person eats in throughout the globe, particularly in Asian nation. It is mostly classified in relation to its texture, color, grain shape etc. In this work, machine vision system is used for rice classification in order to distinguish rice varieties by using some special features like color, harlick and shape. Initially, real rice images are taken from camera for variety of rice such as Basmati rice, IR 18, Ponni, Ponni Leader and Ration rice. These images are taken as input image. Then special preprocessing schemes are introduced like Image thresholding, image enhancement, sharpening and filtering are used to analyze the rice variety. After that feature extraction processes are carried out for both training and testing images. Finally, the multiclass support vector machine (M-SVM) is incorporated to identify the rice variety based on matching between the feature values of training and testing images. These rice classification results such as accuracy and complexity are compared with all other existing classification processes.

Computer Vision and Machine Learning Analysis of Commercial Rice Grains: A Potential Digital Approach for Consumer Perception Studies

“…Emerging technologies such as computer vision (CV) and machine learning (ML) techniques have been applied to classify images of rice varieties [20], whole and broken rice grains [21], chalky rice [22,23], and discolored rice [24]. This technique requires the image acquisition of rice samples and computer vision algorithms to pre-process, analyze, and extract valuable information from the images to develop the classification models.…”