Detection of contaminated hazelnuts by multispectral imaging

Agricultural products (hazelnuts, peanuts, figs, corn etc.) can be effected by aflatoxin producing molds on the growing, processing and storage stages. In this study, a method based on multispectral imaging is developed to separate the contaminated hazelnut kernels from the healthy ones. The multispectral images of the hazelnuts of contaminated and uncontaminated classes are analyzed and the bands that give the best statistical difference are determined. It is observed that the reflectance images at 460 to 500 nm are the most discriminative bands for aflatoxin contamination.