Image-Based Screening for the Identification of Bright Greenish Yellow Fluorescence on Pistachio Nuts and Cashews

Abstract

The development of screening methodologies for a rapid identification of crops contaminated with aflatoxin is of great interest to agro-food industry. The objective of this work was to develop an image algorithm able to identify bright greenish yellow fluorescence (BGYF) on pistachio nuts and cashews. Previous researchers established that the presence of BGYF indicates that there is a high probability of aflatoxin contamination. Since BGYF is not a definitive indicator of aflatoxin contamination, samples emitting fluorescence should be removed and tested for aflatoxins by chemical means. This study, conducted in a static way, is an important step towards the development of a new more accurate and automatic aflatoxin screening method based on a vision system. In this work, a total of 352 samples of pistachio nuts and cashews were evaluated, half of which came from lots contaminated with aflatoxin. Two images in the 410–600 nm optical range were acquired for each sample. Imaging algorithms were developed to identify samples with fluorescent stains caused by BGYF. According to the image analysis results, nut samples were classified into two groups: fluorescent stains (FS) and non-fluorescent stains. Both BGYF and non-fluorescent samples were analyzed for aflatoxin. The laboratory analysis results showed a high correlation with the camera classification: pistachios and cashews placed in the FS group by the vision system contained 92 % and 82 % of the total number of nuts contaminated with aflatoxin, respectively. Moreover, a discriminant analysis of reflectance data was carried out in order to select the optimal optical range to detect BGYF, both in pistachio nuts, i.e., 480 and 520 nm, and in cashews, i.e., 440 and 600 nm.