Skip to Main Content
This paper evaluated the feasibility of using visible and near infrared (Vis-NIR) spectroscopy for the non-invasive discrimination of Spirulina platensis. The reflectance spectra of Spirulina platensis culture medium were measured. Full-spectrum obtained an 86.69% correct answer rate (CAR) for the discrimination of twelve species. Several spectral wavelength variable selection algorithms were operated. A hybrid variable selection algorithm of uninformation variable elimination (UVE) and successive projections algorithm (SPA) obtained the best result of 87.22% of CAR for the prediction set than other methods, including SPA, UVE and UVE-genetic algorithms(GA)-partial least squares(PLS). These results show the possibility for the non-invasive discrimination of Spirulina platensis using Vis-NIR spectroscopy, and UVE-SPA is a good spectral variable selection algorithm. It is necessary to operate UVE before SPA, which can both reduce the calculation time and increase the model' s performance.