Application of hyperspectral imaging on meat and meat products

Citation

Prieto, N., Osika, E., Aalhus, J.L., Lopez-Campos, O., Juarez, M., Pawluczyk, O. (2018). Application of hyperspectral imaging on meat and meat products, 13 http://dx.doi.org/10.1079/PAVSNNR201813042

Plain language summary

Consumers' demand for quality and healthfulness has led to a higher need for quality assurance in meat production, which has increased interest in hyperspectral imaging (HSI). By merging both spatial and spectral features in one single system, HSI has the advantage of being effective for the rapid, environmentally friendly and non-invasive quality assessment, without further sample preparation. This review highlights recent publications using HSI technology for quantitative and qualitative purposes in the meat sector, where discrepancies between published studies and potential sources of variability are discussed. Many of the studies surveyed in this review show the potential of HSI systems to become a rapid and objective assay for quality assessment and classification/detection purposes of meat and meat products. However, some limitations of this technology, mainly associated with intrinsic high dimensionality of data, have also been indicated. Several methods to overcome the disadvantages, such as implementing fast and more efficient algorithms or selection of the optimal wavelengths for analysis through new dimension reduction methods, are discussed and their performance evaluated. Using selected studies, the methods leading to more efficient prediction results are illustrated. The process of improvement starting from full spectral range through the selection of optimal wavelengths for each attribute separately to finally systems in which one set of optimal wavelengths is used for prediction of different attributes or in which one set of wavelengths is used for the determination of the same attribute of different meat species is shown. The HSI systems have undergone significant changes and advances in lighting systems, spectral data acquisition devices, detectors and algorithms in the last few years. Efforts to develop dedicated instrument for particular applications (specifically tenderness prediction) are discussed. Some issues remain, however, unresolved as more improvement in terms of speed and processing is needed before the HSI technology can be effectively implemented in commercial meat processing plants for the non-destructive and rapid quality assessment and sorting in real-time.

Abstract

© CAB International 2018Consumers’ demand for quality and healthfulness has led to a higher need for quality assurance in meat production, which has increased interest in hyperspectral imaging (HSI). By merging both spatial and spectral features in one single system, HSI has the advantage of being effective for the rapid, environmentally friendly and non-invasive quality assessment, without further sample preparation. This review highlights recent publications using HSI technology for quantitative and qualitative purposes in the meat sector, where discrepancies between published studies and potential sources of variability are discussed. Many of the studies surveyed in this review show the potential of HSI systems to become a rapid and objective assay for quality assessment and classification/detection purposes of meat and meat products. However, some limitations of this technology, mainly associated with intrinsic high dimensionality of data, have also been indicated. Several methods to overcome the disadvantages, such as implementing fast and more efficient algorithms or selection of the optimal wavelengths for analysis through new dimension reduction methods, are discussed and their performance evaluated. Using selected studies, the methods leading to more efficient prediction results are illustrated. The process of improvement starting from full spectral range through the selection of optimal wavelengths for each attribute separately to finally systems in which one set of optimal wavelengths is used for prediction of different attributes or in which one set of wavelengths is used for the determination of the same attribute of different meat species is shown. The HSI systems have undergone significant changes and advances in lighting systems, spectral data acquisition devices, detectors and algorithms in the last few years. Efforts to develop dedicated instrument for particular applications (specifically tenderness prediction) are discussed. Some issues remain, however, unresolved as more improvement in terms of speed and processing is needed before the HSI technology can be effectively implemented in commercial meat processing plants for the non-destructive and rapid quality assessment and sorting in real-time.