Identification of Rice Quality Through Pattern Classification Using Computer Vision Image Processing

Muhammad Tariq

doi:10.7176/ceis/11-2-01

Abstract

Rice is the source of Pakistan agriculture industry and food. For agriculture, industry and the oldest sector in the world use rice for different purpose. There are many challenges in the particular sector such as their analysis. This analysis mostly often related to its texture, color, shape, grain etc. In this study, Vision system used to check the quality of rice using some texture features such as color, shape and characteristics. In this study Computer Vision Image Processing tool applied on three different types of rice. Using this tool we apply pattern classification using nearest neighbor and K-nearest neighbor algorithm. Using these algorithms we get results of three varieties of rice Bastmati, Jasmine and White rice. In both algorithms white rice result show best from Basmati rice and Jasmine rice. White rice result is 93.75 % which is best as quality wise. Other tool also available like as MATLAB, Mazda etc for future more best prediction. Keywords: RST-Invariant features, Histogram features, Texture features, Nearest Neighbor algorithm, K-nearest neighbor algorithm DOI: 10.7176/CEIS/11-2-01 Publication date: February 29 th 2020

Highlights

Rice grows in over one hundred countries with an annual harvest of 158 million hectares, producing more than 700 million tons a year (470 million tons of rice)
The biggest increase in sugar production was due to higher yields, which was about 1.74% per year compared to the growth rate of 0.49% for the harvested areas
Rice production increased by 51.1 kg / ha per year, this growth rate has been reduced in percentage and overall efficiency [1]

Summary

Introduction

Rice grows in over one hundred countries with an annual harvest of 158 million hectares, producing more than 700 million tons a year (470 million tons of rice). When we look at recent studies of cereal foods using optical devices and optical processing methods, it can be seen that www.iiste.org products will be tested on many physical properties, such as the color, style, quality and size. The object of this research to analysis of different rice varieties through their images and identify best quality of rice. Real study was created to identify eight different types of rice through their own combination and use machine varieties to develop a grape variety verification system.

Results

Conclusion