Fast Discrimination of Bamboo Species Using VIS/NIR Spectroscopy

摘要

The potential of visibleear-infrared (Vis/NIR) spectroscopy to discriminate different bamboo species was investigated. Vis/NIR spectra were collected on three bamboo species, Bashania fargesii, Fargesia qinlingensis, and Phyllostachys glauca, in the wavelength range of 350-2500 nm. The range of 425-2400 nm was chosen for the spectra modeling. Multiplicative signal correction, standard normal variate with detrending, and 1st and 2nd derivatives were used to preprocess the raw spectral data, and the results were compared. Soft independent modeling of class analogy (SIMCA) and partial least squares discriminant analysis (PLS-DA) methods were applied for building discriminant models. The recognition ratio of 30 samples in the validation set was 100% by both SIMCA and PLSDA models. These results indicate that Vis/NIR spectroscopy may provide a fast and nondestructive technique to discriminate different bamboo species in the field.