Research Of Non-contact Heart Rate Detection System Based On Video

Heart rate refers to the number of heart beats per minute,and it can directly and effectively judge whether a person is healthy or not.Studies have shown that cardiovascular disease has gradually become the primary hazard to human health.With the improvement of social science and technology,as well as the gradual enhancement of people’s medical and health awareness,there is an increasing demand for non-contact medical devices.Especially,the global outbreak of COVID-19 infectious diseases since 2019 has made people pay more attention to non-contact medical and health devices.Non-contact medical testing is an effective way to avoid human contact and prevent the spread of disease at its source.Therefore,it has important research and practical significance for non-contact heart rate detection methods.This topic is based on image photoplethysmography(IPPG)to achieve non-contact heart rate measurement.The volume in the blood vessel of the human body will produce a similar trend of change due to the periodic heartbeat.By extracting this change and processing it,the human body’s heart rate information can be obtained.The main work of this thesis is as follows:(1)Explore the theoretical basis of IPPG technology,and analyze its system composition and working principle.This paper compares and analyzes the commonly used face detection algorithms,selects the AdaBoost algorithm to detect the face parts in the video data,and carries on the detailed analysis to the detection experimental results,and preliminarily collates the experimental data.In the selection stage of ROI,the disadvantages of the commonly used ROI based on human eyes are analyzed,and the ROI coordinates are calculated by returning the detected face information,so as to realize the automatic extraction of ROI position information.(2)According to different commonly used color Spaces,the advantages and disadvantages of each color space are compared and analyzed,the RGB color space is determined and adopted,and the channel R,G and B of ROI is separated.In order to extract the human physiological characteristic information from the channel signals,in this paper,the Fast ICA algorithm is used for blind signal separation to solve the problem that it is difficult to distinguish the human physiological characteristic information from the noise in each channel source signal,and determines the independent component that is most relevant to the channel source signal through calculation and analysis.The independent components were further filtered by using a bandpass filter,and finally the spectrum was obtained by using the Fast Fourier Transform(FFT)to calculate the measured heart rate.(3)The heart rate detection system was divided into several modules,the logic and implementation of each module were introduced in detail,and several subjects were tested.The test results showed that the relative error rate between the experimental group and the control group was lower than 4.800%,with high stability.Pearson correlation analysis and Bland-Altman analysis were used to verify the strong correlation and consistency between the experimental group and the control group.