Quantitative and non-destructive evaluation of ground beef based on multi-spectral imaging
Conference Paper
-
- Overview
-
- Research
-
- Identity
-
- Additional Document Info
-
- View All
-
Overview
abstract
-
The quality of meat-based products is usually tested by subjective and analytical methods. There are laboratory tests which can accurately estimate the content of a sample. Yet, they imply the sample destruction. In addition, they are time consuming and not suitable for industrial applications. Spectral unmixing is wide popular in remote sensing and biomedical applications for a quantitative analysis of an image. In this study, we apply an optical characterization of a ground-beef sample by blind linear unmixing. We prepare samples of ground beef with fixed fat/protein content. The samples are employed to evaluate the characterization provided by linear unmixing of multi-spectral data. We use an eight-band multi-spectral camera and halogen lamps as illumination source. A constrained quadratic optimization algorithm is employed to estimate end-members and their abundances in the sample. The linear unmixing was applied to estimate four end-members and their abundances in the ground beef samples. These abundances match the visual characteristics of the sample such as positions with high concentration of fat. © 2020 IEEE.
publication date
funding provided via
published in
Research
keywords
-
Multispectral imaging; Nondestructive evaluation; Spectral unmixing Constrained optimization; Medical applications; Nondestructive examination; Quadratic programming; Remote sensing; Spectroscopy; Biomedical applications; Constrained quadratic optimization; Illumination sources; Multi-spectral cameras; Multi-spectral data; Multispectral imaging; Non destructive evaluation; Optical characterization; Beef
Identity
Digital Object Identifier (DOI)
Additional Document Info
start page
end page
volume