Aspergillus colonization on peanuts is a growing concern, as it results in reduced crop yields or livestock productivity due to consumption of contaminated feed, and most importantly, they are famous as causative agents of opportunistic infestation in man. In order to defeat and evaluate peanuts suffering from Aspergillus infection, it is essential to establish proper technique and track the quality of peanuts at both farm and market levels. This study focuses on usage of Fourier transform mid-infrared spectroscopy (FTIR) with a non-invasive reflectance apparatus, photoacoustic spectroscopy (PAS) to identify and separate infected peanuts based on spectral characteristics of infrared radiation on peanuts. Classes were defined as “clean” representing no-aflatoxin/no-mold in peanuts, “moldy” representing peanuts with non-aflatoxigenic Aspergillus strains, and lastly, “toxic” representing peanuts infected with aflatoxigenic strains of Aspergillus spp. Classes were analyzed using discriminant analysis algorithm, and distance values were calculated in Mahalanobis distance units. The spectral ranges between 3600 and 2750, 1800 and 1480, and 1200 and 900 cm−1 were assigned as the key bands, and corresponding vibration modes and intensities were labeled. All healthy, clean peanuts (15 healthy/69 total peanut pods) were successfully separated from moldy ones. Separation was further detailed to distinguish peanuts infected with aflatoxin producing strains of Aspergillus flavus and Aspergillus parasiticus, and as a result, 87 % of samples were separated correctly as only moldy or toxic streams. Performance test was conducted with 36 different samples including toxic, moldy, and clean samples; 80.6 % correct classification was achieved.
Bibliographical notePublisher Copyright:
© 2015, Springer Science+Business Media New York.
Copyright 2017 Elsevier B.V., All rights reserved.
- Aspergillus species
- Discriminant analysis