Multispectral laser imaging for advanced food analysis

A hardware-software apparatus for food inspection capable of realizing multispectral NIR laser imaging at four different wavelengths is herein discussed. The system was designed to operate in a through transmission configuration to detect the presence of unwanted foreign bodies inside samples, whether packed or unpacked. A modified Lock-In technique was employed to counterbalance the significant signal intensity attenuation due to transmission across the sample and to extract the multispectral information more efficiently. The NIR laser wavelengths used to acquire the multispectral images can be varied to deal with different materials and to focus on specific aspects. In the present work the wavelengths were selected after a preliminary analysis to enhance the image contrast between foreign bodies and food in the sample, thus identifying the location and nature of the defects. Experimental results obtained from several specimens, with and without packaging, are presented and the multispectral image processing as well as the achievable spatial resolution of the system are discussed. (C) 2016 Elsevier B.V. All rights reserved.