| Abstract:The chemical composition and basic density of four species of poplar which were widely used as raw material in pulping were determined by using traditional methods and the near-infrared (NIR) spectra of the samples were also collected. Partial least squares (PLS) method and cross-validation were used to confirm the best principal component numbers and build the calibration models for holocellulose, lignin, benzene ethanol extractive and basic density of poplar samples. The independent verification of the calibration models showed the coefficients of determination (R2val) were 0.9050, 0.9098, 0.9112, 0.9165, respectively. The root mean square errors of prediction (RMSEP) were 0.40%, 0.42%, 0.19%, 0.0050 g/cm3, respectively. The relative percent deviations (RPD) were 3.24, 3.33, 3.36 and 3.46, respectively. And the absolute deviations (AD) were -0.49%～0.77%, -0.66%～0.63%, -0.28%～0.33%, -0.0094 g/cm3～0.0068 g/cm3, respectively. The root mean square error of prediction and the absolute deviation basically met the error requirement and the four calibration models could realize the rapid determination of the properties of poplar wood used in paper industry. |
