Study On Selection And Application Of Reference Measuring Method On Non-invasive Blood Glucose Sening By Near-infrared Spectroscopy

Nowadays,a key problem in non-invasive blood glucose sensing by near-infrared(NIR)is that the influence of the background changes during measurements,which restricts the effective extraction of specific glucose information.For in vitro analysis,background variations are usually corrected by the spectra of a standard reference sample which has similar optical properties with the analyte interested.Currently,little systematic research and investigation has been conducted on the effectiveness of this background correction method based on reference subtraction,as well as the selection principle of various reference measuring methods.Aiming at this situation,two common background correction methods respectively based on(1)single-beam and time-sharing measurement;(2)double-beam and synchronization measurement have been studied systemically in this thesis.Firstly,this thesis reviews the development of non-invasive blood glucose sensing in recent years,and addresses a significant barrier to the clinical application of noninvasive blood glucose sensing by NIR——the influence from the complicated background variations.As the main sources of background variations in the measurement system,the influence of drift of light intensity and system hardware noise on the specificity of useful signal,limitation detection concentration of the analyte as well as the prediction accuracy of the multivariate calibration model are analyzed quantitatively.Results show that the impact of light drift is more pronounced.Secondly,this thesis carries a comprehensive study of single-beam and doublebeam reference measuring methods in both diffuse transmission mode and diffuse reflectance mode:(1)taking measurement system as the example,the structure formula of the actual signal obtained by reference measurement is deduced firstly;(2)comparative analysis of the correction effect on drift of light intensity of two reference methods are conducted;(3)aiming at different background variations sources,this thesis first provides the theoretical formulas of limitation concentration in two methods;(4)for different instruments with various precision,a quantitative selection formula of optimal reference measuring method is presented first in this thesis;(5)at last,the necessary precision of the measurement system in order to achieve a certain detection concentration target is summarized,hoped provides some bases and instruction for the design of optical system and the application of floating reference method which is based on synchronization measurement mode.Thirdly,a multi-day experiments is carried out on aqueous glucose solution to explore the long-term effectiveness of the optimal reference measuring methods:(1)with a custom-developed NIR spectrometer based on AOTF splitting,a comprehensive evaluation on the performance of it is conducted firstly.Then,based on the evaluation results,the optimal reference measuring method is determined according to the selection formula mentioned above,that is,the double-beam measurement.(2)Next,two experimental systems with double-beam design that applicable to diffuse transmission measurement and diffuse reflectance measurement respectively,are built.Several matching optical probes which could realize the synchronous transmitting and receiving of NIR light is designed.(3)At last,experiments are conducted in pure absorption medium to validate the correction theory,the selection formula of optimal reference measuring method as well as the long-term effectiveness of the optimal reference measuring method.Finally,a preliminary verification of double-beam mearing method is studied in strong scattering medium.Building on this foundation,a method combined doublebeam measurement and NAS processing is proposed for a better extraction of specific signal information while eliminating the inherent difference between the reference beam and sample beam.