Classification of bee pollen using visible near infrared spectroscopy

Ping Lin,Yong He,Yongming Chen
2009-01-01
Abstract:The potential of visible and near infrared reflectance spectroscopy (Vis/NIRS) was investigated to nondestructively classify bee pollen varieties. Noise Reduction (NR), standard normal variate transformation (SNV) and Multiplicative Scatter Correction (MSC) were applied for preprocessing of original spectral data. A new method was developed by the combination of Genetic Algorithm (GA) with partial least squares (PLS) and BPNN for spectra analysis and modeling, and an excellent prediction performance was demonstrated. Total 15 characteristic wavelengths were abstracted by GA, 7 principal components (PCs) were obtained by PLS as the best number of principal components, and finally 30 unknown bee pollen samples were successfully predicted based on 7 PCs as the input of BP neural network. The accuracy of prediction of different bee pollen varieties was 100%. Vis/NIRS technique combined with chemometrics was promising for classifying the varieties of bee pollen fast and nondestructively. This new combination of GA with PLS and BPNN might be applied for species recognition based on NIR spectroscopy for other materials.
What problem does this paper attempt to address?