Study On Reference Wavelength Based On Diffusion Reflectance

The spectrum is easily affected by uncontrollable interference such as physiological noise during the process of non-invasive blood glucose sensing by near infrared spectroscopy. It is a challenge solution that taking the signal at the reference wavelength where the diffuse reflectance light intensity is insensitive to different glucose concentration as an internal reference to correct spectra. In this paper, the existent characteristic of reference wavelength is studied and so is the spectra correction method.First of all, the reference wavelength of glucose is calculated due to Boltzmann diffuse transmission formula in 10% and 20% intralipid respectively. Then the Monte Carlo method is used to prove the theoretical result. Furthermore, the in vitro experiments is taken for the validation of the reference wavelength. All the results show that the three results are basically the same, and the reference wavelengths exists at 1567nm and 1556nm in this two media respectively. In this paper, the influence of optical coefficient, temperature and source-detector separation on the reference wavelength is also studied. The results show that the reference wavelength will move towards the direction of long wavelength if the absorption coefficient or the source-detector distance increases, and to the opposite direction if the scattering coefficient increases, but not change at different temperatures.Besides, the interferential factors in non-invasive blood glucose sensing are analyzed. A spectrum correction method based on the reference wavelength is deducted to improve the glucose measurement precision by simulation and in vitro experiment. A serious of diffuse reflectance are got at difference glucose concentration in 20% intralipid. The calibration models of glucose are built by the primal spectrum, the spectra after deducting background and the spectra corrected by reference wavelength respectively. The results show that the measurement precisions of glucose concentration are improved by 56.51% and 33.84% after using this newly method.