Classification of rice varieties using Near-Infra red Spectroscopy

Abstract

Rice is consumed in many different forms (brown, milled and parboiled) and cultivated in different size varieties (short, medium and long grain). Many of the traditional methods of analysis for determining the physical, chemical and mechanical properties to ensure the quality of rice are time consuming, destructive, require expensive harmful reagents. The desire is to replace the traditional methods to find its quality with rapid, non-destructive, non invasive methods. All cereal grains contain starch (soluble carbohydrate) as the principal component. Starch makes up about 90% of the dry matter content of milled rice. The objective of this study is to classify rice samples based on the carbohydrate content by using the Near-Infrared Spectroscopy (NIRs). NIR spectra were taken on every 250 gm of rice in the range of 1100nm to 2200nm. All the spectral data were processed statistically and resulting, the rice samples were classified using Principle Component Analysis (PCA).