Design And Implementation Of A Noninvasive Detection System For Oxygen Saturation

In modern society,blood oxygen detection technology has increasingly become one of the research hotspots in the fields of control,sensing and medical treatment.For the pulse oximeter sold in the market at present,there will be a big error in the measurement result of blood oxygen under the condition of weak perfusion,moreover,it is difficult to accurately detect the signal feature points under weak perfusion by the traditional maximum method and threshold comparison method.In order to avoid the distortion of the signal and maintain the integrity of the signal,the feature points of the signal can be accurately detected under the condition of weak perfusion.In view of the above problems,the main research work of this paper includes:1.Principle and method of noninvasive blood oxygen detection based on Lambert-Beer law and near infrared spectrum theory are introduced,the formula for calculating blood oxygen saturation is derived from the theory,STM32F103VET6 as the core processor chip and blood oxygen signal acquisition simulation front-end AFE4490 combined hardware circuit design,reduce the complexity of circuit design,compared with traditional circuits,differential transmission design is adopted for blood oxygen signal input and shielding protection is added,which can effectively avoid electromagnetic interference and improve signal stability.2.According to the original pulse wave signals collected,In view of the possible signal data loss and distortion in the filtering process,in this paper,a combination of wavelet transform and FFT is proposed to preprocess the waveform.And then reconstruct the signal,the signal was denoised by using db2,sym8 and coif3 wavelet bases.Under normal perfusion,the SNR increased from 26.521390 to 29.501140,the RMS decreased from 0.025152 to 0.018415,and the RMS increased from 6.845247 to 7.331652,and the RMS decreased from 0.023766 to 0.022472,to some extent,noise such as power frequency interference and baseline drift are removed from the original signal waveform.Good results have been obtained.3.Pulse wave data obtained after preprocessing are detected for peak feature points.Due to the insufficiency of the peak feature point detection accuracy in the case of weak perfusion when the signal interference is more serious in the maximum value method and the traditional threshold comparison method,this paper proposes a peak detection algorithm that can quickly and accurately detect the crest and trough values even with a little noise interference.Compared with the traditional maximum value method and threshold comparison method for peak detection,the algorithm has been further improved,the problem of false wave peak being mistaken for effective wave peak is effectively avoided4.A large number of red and infrared pulse wave signal data of various blood oxygen states were collected on Fluke Index 2L by the designed measurement system.Blood oxygen measurement data set was constructed.Preprocessing and feature point detection are carried out according to different data collected.The red and infrared ac and dc values of bleeding oxygen pulse wave signals were obtained,respectively.The experiment determined the R value under different blood oxygen state,in terms of determination of coefficient of blood oxygen measurement formula,use least square method,Finally,the calibration curve of bleeding oxygen saturation is obtained.The oxygen saturation of hemorrhage was obtained according to the calibration curve.5.Verify the performance of the designed system.The reliability,stability and accuracy of the system are verified by experiments.Meet the design requirements.