Nondestructive detection of chilled mutton pH value using hyperspectral imaging technique
[Objective]To detect the pH value of chilled mutton quickly and non-destructively.[Methods]Visible/near-infrared hyperspectral(400～1 000 nm)imaging technology was used to collect scattering images on the surface of chilled mutton,and the reflectance spectral curve of the region of interest of the sample was extracted to obtain the original spectral data.Four single methods of savitzky-golay(SG),multiplicative scatter correction(MSC),standard normal variant transformation(SNV)and first-order derivative(FD)and three combination methods of MSC-SG,SNV-SG and FD-SG were used to preprocess the original spectral data.Linear regression model:partial least squares regression(PLSR)and nonlinear regression model:Random Forest(RF),support vector regression(SVR),extreme gradient boosting(XGB)were used to construct the prediction model of pH value of chilled mutton based on full wavelength.[Results]FD-SG was the optimal pretreatment method,and the XGB model was the optimal model.The R2C and R2P were 0.930 1 and 0.830 0,and the RMSEC and RMSEP were 0.052 0 and 0.079 2.The XGB model was used to calculate the pH value of each pixel in the chilled mutton sample,and a pseudo-color image was established to show the spatial distribution of the pH value of the chilled mutton sample more intuitively.[Conclusion]Visible/near-infrared hyperspectral imaging technology can realize non-destructive detection of pH value of chilled mutton.
国家科技期刊平台
NSTL
万方数据
