Journal of Applied Computer Science & Mathematics (Jan 2009)

Effect of Foreign Bodies on Recognition and Classification of Bulk Food Grains Image Samples

  • Dayanand G. Savakar,
  • Basavaraj S. Anami

Journal volume & issue
Vol. 3, no. 6
pp. 77 – 83

Abstract

Read online

This paper presents an effect of Foreign Bodies (FB)on accuracies of recognition and classification of bulk foodgrain image samples using a Neural Network Approach. Anymatter other than major food grains is considered as a foreignbody in this work, such as stones, soil lumps, plant leaves, piecesof stems, weed, other types of grains etc. The amount of foreignbodies decides the quality of the food grains and hence it isnecessary to determine the amount of foreign body present infood grains automatically to help farmers in sowing andmarketing. Different food grains like, Green gram, Groundnut,Jowar, Rice and Wheat are considered in the study. The colorand texture features are presented to the neural network fortraining and later of the unknown grain types mixed withforeign bodies. The combination of both color and texturefeatures is employed in the work. The study reveals that thepresence of even 10 percent of foreign bodies within food grainimage samples reduces its recognition and classificationaccuracies as low as 60%. When the foreign body percentage isgreater than 50, it becomes difficult to recognize and classifyfood grain image samples.

Keywords