Research On RPPG Signal Detection Method

According to statistics from relevant departments,the proportion of cardiovascular disease in urban and rural residents in China has continued to grow rapidly since 2015.Therefore,the technology of conveniently and quickly detecting cardiovascular information in daily life has become the forefront of modern medical diagnostic technology research.Although the current mature contact measurement method has high accuracy,it is easy to cause discomfort to the patient,and it is relatively expensive,and it is difficult to use in the daily physiological information detection.Therefore,the non-contact measurement method based on imaging equipment has great research significance and application prospects,In the current rPPG(remote PhotoPlethysmoGraphy,PPG)signal detection,the change in blood volume squeezes the capillary wall and connective tissue to cause it to deform,so that during the signal collection process,it is not just the change in blood that absorbs light There is also a change in light caused by the deformation of connective tissue.This situation causes a mechanism problem caused by motion artifacts.We must eliminate it by designing a noise reduction algorithm.This is the focus of our research.Based on the principle of remote photoelectric volume pulse wave,this paper has realized the research of remote PPG signal detection method.Face detection and acquisition is completed through a remote imaging device,and the original PPG signal is extracted from it.A VMD-ICA(Variational Mode Decomposition-Independent Component Analysis,VMD-ICA)joint noise reduction algorithm is applied to the noise reduction of the PPG signal.The superiority of the algorithm is discussed through data comparison and discussed.The problem of remote PPG signal detection under different light source conditions and different imaging equipment is addressed.The main research contents of this article are as follows:(1)According to the existing skin optical model of the remote PPG signal,perform the characteristic analysis.Based on this model,the signal generation principle and signal composition of the PPG signal are analyzed.Based on this,the characteristics of the PPG signal are studied,and the processing method of the PPG signal is analyzed based on various unique characteristics of the PPG signal.It provides a basis for the research and selection of PPG signal processing methods.(2)Take a face video and detect the original PPG signal from it.Based on the comparison of the characteristics of the PPG signal,the Adaboost algorithm was selected to realize the detection and positioning of the face ROI region in matlab.The original PPG signal detection was completed by RGB three-channel pixel space averaging of the face ROI region image,and analyzed The previous blind source separation methods deal with the inaccuracy of the optical characteristics of PPG signals.(3)The VMD-ICA joint noise reduction algorithm is applied to noise reduction and motion artifact removal of PPG signals.The algorithm principle and flow of the VMD-ICA joint noise reduction algorithm are analyzed,the significance and concepts of PPG signal decomposition are explained,and the flowchart and specific steps of the algorithm are explained.The remote PPG signal characteristics and algorithm principle are verified The rationality,feasibility and superiority of the algorithm,the original G-channel PPG signal is used as the input signal to implement the algorithm,and the periodic PPG signal with good periodicity and stability is obtained.(4)Comparison of detection results of remote PPG signal detection methods.The blind source separation method,wavelet threshold denoising method and VMD-ICA algorithm are analyzed in theory and signal characteristics,and the heart rate estimation of PPG signal is used as the evaluation standard of PPG signal.The instrument is a standard heart rate value,and the heart rate estimation method and error calculation method are selected to complete the comparison of the experimental results.First,the superiority of the algorithm proposed in this paper is verified in the comparison of the four signal noise reduction and motion artifact removal algorithm results,and then the heart rate estimation results under different light conditions and different imaging devices are studied,and the lighting conditions are discussed.The influence of imaging and imaging equipment on remote PPG signal detection and the existence of some possibilities.Finally,the stability,practicability and shortcomings of this detection method are discussed through experiments.Based on the previous work,this paper makes a certain research on the remote PPG signal detection method,analyzes the optical characteristics of the remote PPG signal,proposes a new method of signal noise reduction and motion artifact removal,and the detection results produced by different conditions in signal detection are studied and discussed in order to provide some reference and suggestions for the development of non-contact physiological information detection methods in the future.