A machine vision based approach towards identification of adulterant in turmeric powder

Abstract

Turmeric quality mainly depends on Curcumin which not only imparts yellow color of turmeric but also the principal Curcuminod of turmeric. Chemicals with yellow colors e.g. Metanil yellow are often mixed to turmeric powder for achieving the attractive yellow color without much change in taste. Consumption of adulterant can cause health hazards. The detection of unwanted mixing of adulterant with food is vital but difficult to achieve manually. The paper presents a machine vision based approach for detection of adulterant with turmeric powder. The frequency domain analysis of color projection features along with principal component analysis is being performed in this paper for identification between adulterant mixed and unmixed verities of turmeric powder samples. Here a class separability measure is used to find the separation index to validate the class separation objectively. The experimental results show that the presented method may be considered as a potential tool.