Nondestructive Detection of Moisture Content in Walnut Kernel by Near-Infrared Diffuse Reflectance Spectroscopy

The rapid and accurate detection of the moisture content is great significance to quality evaluation oil extraction process walnut kernel. Near-infrared (NIR) spectroscopy an ideal method for measuring in In this study, a regression model kernel was developed based on NIR diffuse reflectance using chemometric methods. different spectral pretreatment methods were adopted preprocess original data. whole spectra band divided into 5 subbands, 10 15 20 subbands screen specific wavelengths relevant content. PLS (partial least square regression), MLR (multivariate linear PCR (principle component SVR (support vector regression) used establish relationship between data measurement values comparison, optimized modeling conditions determined as follows: wavelength 1349–1490 nm, SNV-FD (standard normal variate transformation first derivative) preprocessing method, algorithm. Under these conditions, correlation coefficient (R2) root mean error prediction (RMSEP) 0.9865 0.0017, respectively. results study provided feasible To improve performance applicability model, it necessary continuously expand size sample set.