Journal of Spectral Imaging,   Volume 1   Article ID a1   (2010)

Peer reviewed Paper

Use of spectral pre-processing methods to compensate for the presence of packaging film in visible–near infrared hyperspectral images of food products

  • A. A. Gowen
  • G. Downey
  • C. Esquerre
  • C. P. O'Donnell
Biosystems Engineering, School of Agriculture, Food Science and Veterinary Medicine, University College Dublin, Belfield, Dublin 4, Ireland

 Search for papers by this author
Biosystems Engineering, School of Agriculture, Food Science and Veterinary Medicine, University College Dublin, Belfield, Dublin 4, Ireland

 Search for papers by this author
Biosystems Engineering, School of Agriculture, Food Science and Veterinary Medicine, University College Dublin, Belfield, Dublin 4, Ireland

 https://orcid.org/0000-0003-1729-6074
 Search for papers by this author
 Corresponding Author
Biosystems Engineering, School of Agriculture, Food Science and Veterinary Medicine, University College Dublin, Belfield, Dublin 4, Ireland
[email protected]
 https://orcid.org/0000-0002-9494-2204
 Search for papers by this author

The presence of polymeric packaging film in images of food products may modify spectra obtained in hyperspectral imaging (HSI) experiments, leading to undesirable image artefacts which may impede image classification. Some pre-processing of the image is typically required to reduce the presence of such artefacts. The objective of this research was to investigate the use of spectral pre-processing techniques to compensate for the presence of packaging film in hyperspectral images obtained in the visible–near infrared wavelength range (445–945 nm), with application in food quality assessment. A selection of commonly used pre-processing methods, used individually and in combination, were applied to hyperspectral images of flat homogeneous samples, imaged in the presence and absence of different packaging films (polyvinyl chloride and polyethylene terephthalate). Effects of the selected pre-treatments on variation due to the film’s presence were examined in principal components score space. The results show that the combination of first derivative Savitzky–Golay followed by standard normal variate transformation was useful in reducing variations in spectral response caused by the presence of packaging film. Compared to other methods examined, this combination has the benefits of being computationally fast and not requiring a priori knowledge about the sample or film used.

Keywords: hyperspectral, imaging, pre-processing, packaging, film, polymer, food, test

Metrics

Downloads:

1,754

Full-Text HTML Views:

212

Abstract Views:

1,330