Possibility of image processing technique in inspection of impurities and undesired chickpea seeds

Possibility of image processing technique in inspection of impurities and undesired chickpea seeds

Chickpea is one of main agricultural products in both points of production and consumption. Existence of impurities and seeds with different ripeness levels and colors indicates the necessity of chickpea visual sorting. The goal of the present research is applying image processing to identify impurities and undesirable chickpeas in desirable seeds. Unsuitable chickpea were greenish, brownish, split, and wrinkles seeds and impurities were stones and stalks. The image of different samples were captured and various color, texture and shape features were extracted after image preprocessing. Some benefit features were classified based on support vector machine and linear discriminant analysis methods. The results showed high ability of image processing technique for identifying of the different investigated samples. The ability of linear discriminant analysis with accuracy of 85.19% was higher than that of support vector machine with accuracy of 76.66% in classification of the samples.