Simulation Of Photon Migration In Multi-component Media And Some New Methods Of Spectral Measurement Data Processing

Noninvasive detection of specific component content of multi-component media by spectroscopy is a research hotspot in the detection field. The light transmission in the tissue and blood composition or tissue components detection are the foreland in the area of biomedical information detection. As the model light propagation in the tissue needs to be improved and because of the influences of the weak signal, measurement conditions and individual differences, applying the near infrared spectroscopy to the clinical utility has great difficulty. The modified model to simulate the photon migration in multi-component media and new methods in the related spectral measurement data processing proposed in this paper is very helpful to develop the technology.The research achievement about light transmission in tissue and related methods of spectral measurement data processing were reviewed at the beginning of this paper. And the problems which need to be solved in biomedical optics have been illustrated: Such as the accurate construction of photon migration in multi-component media; the overlapping of the tissue absorption spectrum; the influence of the measurement environment; the model building of spectral analysis, etc. Aiming at solving these problems, the main work and innovations described in the paper are:1. According to the Monte Carlo model and its simulation model of light propagation in tissue, as its limitations, a modified model has been put forward to simulate the photon migration in multi-component media. The practical simulation, experimental verification and its application have been finished, too. This model can be used for the accurate simulation of the blood component measurement.2. In order to improve the accuracy of spectral measurement, the experience model decomposition(EMD) has been applied to the spectral measurement data preprocessing. The detection principle of eliminating individual differences in blood components measurement by dynamic spectra(DS) and the basic theory of signal filtering by EMD have been described. Introducing EMD into the spectral data pretreatment processing has been proved to improve the signal-to-noise ratio of the spectral data. The feasibility and the validity of EMD in the processing of spectrum measurement has been illustrated in the noninvasive measurement of hemoglobin concentration by DS, which could be a new method for spectral data processing and the related production design.3. The application of taking analysis of variance(ANOVA) into the spectral measurement data processing, has been used to determine the effectiveness of spectral data and its calibration model establishing. The basic concepts about ANOVA and the principle that applying ANOVA to the validation of the spectral data have been also introduced. By the evaluation index, F-value and P-value, of ANOVA method,the simulation models of spectral data under different noise levels have been analyzed. As an example, the principle that applying ANOVA to the oxygen measurement by double wavelengths and its internal mechanism have been explained. The results show that ANOVA could determine the noise level and the effectiveness of the analysis methods in spectral data processing, which can improve the reliability of composition quantitative analysis.