Study On The Detect Theories And Signal Acquisition For Dynamic Spectrum

The various chemical components present in human body carry important information about health status. Such chemical information also serves as an important indicator to a number of clinical diagnostics and therapeutic effects. Near-infrared spectroscopy receives global attention as a principal non-invasive diagnostic means. But there is still a long way to go for clinic application, because the near infrared (NIR) spectrum technology is disturbed by many factors, and the precision of the final result is restricted. The difference detect method is employed to the non-invasive blood component concentration sensing in this paper, trying to eliminate the influence of the skin and other tissues. Dynamic Spectrum (DS) which obtain the difference spectrum of artery blood from the pulse wave of blood is focused on in this paper.Firstly, the transmittance difference spectrum is studied. Results show that when the basic thickness of tissues and the additional thickness of the interested tissue are known, the difference spectrum is a function of the optical parameters of the tissues, and the latter is relative with the component concentration of the tissues. Three kinds of application of difference spectrum in the non-invasive sensor are proposed in this paper including the DS.Based on the definition of DS, three kinds of dynamic spectroscopy system are described. The phase shift error of the dynamic spectrum among different pulse wave is discussed. The detect frequency, data intercept position and the data length are studied to perfect the DS data processing.A new method to detect outliers of DS is proposed. This method can distinguish the base line wave of the pulse waves and the abnormity change of signal pulse wave, and is fit for the DS abstract.The influence of contact force is studied by experiment. Result shows that the contact force between the photoplethysmographic sensor and the measurement site influenced the photoplethysmographic signal recorded on the fingertip from young healthy subjects, but it has little influence on the rate between the normalized DS.Based on the Monte Carlo simulation, the influence of scatter in the DS detect is researched. The simulation model of signal layer blood and multilayer tissue of finger tip are founded. The normal thickness of blood layer in the models is searched by Monte Carlo simulation based on the statistical result of pulse signal. Results of the simulations show that the absolute magnitudes and relative differences in path-length factors is a function of wavelength, and the signal of DS is relative with the basic thickness of the stable tissues. The first one can be minified evidently by using the subsection equivalent path-length, and the second one indicates that the DS must be normalized before it is used.The dynamic spectrum method proposed a new idea to the NIR non-invasive detection. It has great value in this field. Theory and experimental results disclosed by this dissertation laid solid basis for the successful application of this method.