We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
PREDICTION OF BEEF FRESHNESS USING A HYPERSPECTRAL SCATTERING IMAGING TECHNIQUE.
- Authors
Ma Shibang; Xue Dangqin; Wang Xu; Xu Yang
- Abstract
A rapid and non-destructive method based on hyperspectral scattering technique for determination of beef freshness (i.e.,TVB-N) was studied in this study. Hyperspectral images of 60 beef samples stored at 4oC for 1-18 days were acquired using a VIS/NIR hyperspectral imaging system in the wavelength range of 400-1100 nm, and the scattering profiles of every wavelength were fitted to a Lorentzian distribution function to give three parameters a (asymptotic value), b (peak value) and c (full width at b/2). The single parameters (a, b or c) or their combined parameters (a+b/c or (b-a)/c) were used to develop least squaresupport vector machine (LS-SVM) models for prediction of beef freshness. The performance of a LS-SVM model developed using (b-a)/c was the best among all the tested models, showing Rc=0.91, Rp=0.86, SEC=5.83 mg/100 g and SEP=5.21 mg/100 g. A genetic algorithm was used to optimize the parameter (ba)/ c for developing a GA-LS-SVM model, which performed best showing Rc=0.97, Rp=0.96, SEC=3.38 mg/100 g and SEP=3.85 mg/100 g. This study provided a new non-destructive method based on a hyperspectral imaging technique combined with a genetic algorithm for rapid prediction of beef freshness.
- Subjects
BEEF research; HYPERSPECTRAL imaging systems; BEEF spoilage
- Publication
INMATEH - Agricultural Engineering, 2016, Vol 50, Issue 3, p55
- ISSN
2068-4215
- Publication type
Article