Study On Noninvasive Detection Of Hemoglobin Based On Dynamic Spectrum Method

Hemoglobin detection is a routine examination item in hospital,which has important reference significance in the diagnosis of diseases.The noninvasive detection of hemoglobin advances the efficiency and convenience of the detection of hemoglobin while changing the detection method,creating conditions for the portable and miniaturization of the detection equipment.The dynamic spectrum method is based on the photoelectric volume pulse wave recording method,which can extract the absorbance information of pulsating arterial blood,and effectively overcome the influence of individual differences of the subject and measuring conditions on blood composition detection.In this paper,the noninvasive detection technology of hemoglobin based on dynamic spectrum method is studied.Through the design and research of each module of the detection system,the miniaturization and high precision of the detection instrument are realized.The main tasks of this project are as follows:1.The multi-channel acquisition method is proposed:In view of the disadvantages of the existing spectroscopic technique,such as complex technology and large equipment volume,the eight-channel photoelectric pulse wave signal acquisition method is proposed.Based on the fact that the body’s blood components remain basically stable for a short period of time,the photoelectric pulse signal of eight fingers of two hands is collected synchronously.The correlation between low frequency physiological signals reflected by different fingertip of the same person was verified by experiments.2.Design of hardware circuit:LD lasers with different central wavelength were selected in eight channels as the source of non-invasive detection of hemoglobin.Each channel has independent signal amplification circuit,high-pass filtering circuit,low-pass filtering circuit,voltage lifting circuit,control circuit and A/D conversion circuit.The design of hardware circuit is mainly based on dynamic spectrum,and the DC and AC components of photoelectric pulse wave are obtained.First,the DC component is obtained from the operational amplifying circuit,and then the AC component of the signal is obtained by using the high-pass and low-pass filters.By eliminating background noise and interference factors through the design of related circuits,the signal-to-noise ratio of the photoelectric pulse wave is improved.The control circuit and A/D conversion circuit are designed based on STM32.A chip is used to realize the functions of circuit control and analog to digital conversion,making the equipment simple and efficient.CAN bus is used to communicate between the upper computer and the lower computer,which improves the transmission rate and ensures the real-time of the equipment.3.System software design:the software design is mainly about A/D conversion of photoelectric pulse signal,CAN bus communication,circuit control,digital signal filtering and dynamic spectrum frequency domain extraction.The lower computer takes STM32 as the core processor,and realizes functions such as A/D conversion,circuit control and CAN bus communication by calling STM32’s library function to carry out software programming.The upper computer USES Labview to restore pulse wave signal,extract and display dynamic spectrum.4.Study of modeling methods:clinical trials were conducted in the metabolic disease hospital of Tianjin medical university,and dynamic spectral data and blood test reports were collected.BP neural network was used to establish the correction model between dynamic spectral data and hemoglobin concentration,and the prediction set was predicted.The experimental results show that the dynamic spectrum extracted from photoelectric pulse wave signal is collected synchronously through eight channels,and the concentration of human hemoglobin can be predicted by using BP neural network modeling..