Prediction Of Beef Freshness Using A Hyperspectral Scattering Imaging Technique

Ma Shibang,Xue Dangqin,Wang Xu,Xu Yang

2016-01-01

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 4 degrees C 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 square-support 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 R-c=0.91, R-p=0.86, SEC=5.83 mg/100 g and SEP=5.21 mg/100 g. A genetic algorithm was used to optimize the parameter (b-a)/c for developing a GA-LS-SVM model, which performed best showing R-c=0.97, R-p=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.