Exploiting multispectral imaging for non-invasive contamination assessment and mapping of meat samples

Panagiotis Tsakanikas; Dimitris Pavlidis; Efstathios Panagou; George-John Nychas

doi:10.1016/j.talanta.2016.09.019

Exploiting multispectral imaging for non-invasive contamination assessment and mapping of meat samples

Talanta. 2016 Dec 1:161:606-614. doi: 10.1016/j.talanta.2016.09.019. Epub 2016 Sep 6.

Authors

Panagiotis Tsakanikas¹, Dimitris Pavlidis², Efstathios Panagou³, George-John Nychas⁴

Affiliations

¹ Laboratory of Microbiology and Biotechnology of Foods, Department of Food Science and Human Nutrition, School of Food, Biotechnology and Development, Agricultural University of Athens, Athens, Greece. Electronic address: p.tsakanikas@aua.gr.
² Laboratory of Microbiology and Biotechnology of Foods, Department of Food Science and Human Nutrition, School of Food, Biotechnology and Development, Agricultural University of Athens, Athens, Greece. Electronic address: d.pavlidis@aua.gr.
³ Laboratory of Microbiology and Biotechnology of Foods, Department of Food Science and Human Nutrition, School of Food, Biotechnology and Development, Agricultural University of Athens, Athens, Greece. Electronic address: stathispanagou@aua.gr.
⁴ Laboratory of Microbiology and Biotechnology of Foods, Department of Food Science and Human Nutrition, School of Food, Biotechnology and Development, Agricultural University of Athens, Athens, Greece. Electronic address: gjn@aua.gr.

PMID: 27769454
DOI: 10.1016/j.talanta.2016.09.019

Abstract

Recently, imaging and machine vision are gaining attention to food stakeholders since these are considered to be the emerging tools for food safety and quality assessment throughout the whole food chain. Herein, multispectral imaging, a surface chemistry sensor type, has been evaluated in terms of monitoring aerobically packaged beef filet spoilage at different storage temperatures (2, 8, and 15°C) and storage time. Spectral data acquired from the surface of meat samples (with/without background flora; +BF/-BF respectively) along with microbiological analysis. Qualitative analysis was employed for the discrimination of meat samples in two microbiological quality classes based on the values of total viable counts (TVC<2log₁₀CFU/g and TVC>2log₁₀CFU/g). Furthermore, a Support Vector Regression model was developed to provide quantitative estimations of microbial counts during storage. Results exhibit good performance with overall correct classification rate for the two quality classes ranging from 89.2% to 80.8% for model validation. The calculated regression results to an R-square of 0.98.

Keywords: Machine learning; Multi-spectral sensor; Non-invasive; Spoilage of meat; Surface chemistry.

MeSH terms

Bacteria
Colony Count, Microbial
Food Contamination / analysis*
Food Microbiology
Meat / microbiology*
Spectrum Analysis
Support Vector Machine