Author/Authors :
Rّnsholdt، نويسنده , , Bent and McLean، نويسنده , , Ewen، نويسنده ,
Abstract :
Multivariate calibration of VIS spectra (396–705 nm) was examined as a method for quantifying carotenoid content in fish flesh. Rainbow trout (1.6–4.2 kg; 2+ and 3+ year) fillet samples (2 mm thick) varying in pigment (0.02–21.11 mg astaxanthin eq kg−1) were examined by reflectance spectrophotometry (396–705 nm) using black (air, Rb) and white (methacrylate plate, Rw) backgrounds. Homogenized samples of lower dorsal fillets were measured in a similar manner. R∞(corresponding to an infinitely thick sample) and K/S spectra were calculated using Kubelka–Munk theory. Results illustrated that multivariate was superior to univariate calibration, with the root mean square errors of prediction being 2.34, 1.41, 1.78 and 2.34 (2 mm, mean of triplicates), 1.82, 1.94, 1.74, 1.76 (homogenized samples) for Rb, Rw,R∞ and K/S respectively, compared to 3.52, 3.53, 3.67 and 2.89 in the univariate case (units: mg astaxanthin eq kg−1).
Keywords :
Colour , carotenoids , Kubelka–Munk , rainbow trout , Multivariate calibration , Astaxanthin , Reflectance spectrophotometry