| As important parameters of human vital signs,heart rate and blood oxygen saturation have important reference value in personal daily health monitoring.Imaging Photoplethysmography(IPPG)is a non-contact physiological signal detection technology developed in recent years,which uses optical sensors to obtain video images of human skin and further analyses and processes them to detect human physiological parameters.The IPPG technique has become one of the emerging research hot spots in the field of instrumentation and medicine because of its non-contact measurement,low cost and easy operation.However,the signals obtained through IPPG technology are relatively weak and susceptible to factors such as light changes and motion interference,and there are still some key issues to be solved on how to measure physiological signals more accurately,more comprehensively and more quickly.Therefore,this paper investigates the detection method of human physiological parameters based on video information,focusing on issues such as resistance to motion interference,accurate acquisition of IPPG signals adapted to light changes,and the detection of physiological parameters such as heart rate and blood oxygen saturation.The specific research content and results are as follows.Firstly,focusing on the problem of weak IPPG signals,Kalman filter is used to track the face region and draw the face object boundary box.The face region is extracted by 68 key points detection of the face,and the heart rate curves extracted from the forehead,cheeks,near the nose and the full-face skin region are compared to identify the full-face skin part as the region of interest.The Euler video amplification algorithm is then used to color enhance the video image,thus achieving signal amplification.Secondly,in order to address the impact of complex lighting conditions,the dark video is enhanced by using the improved MSRCR algorithm of fusion bilateral filtering and the prior theory that the reflection components of normal brightness image and low illumination image are consistent.While enhancing the color and contrast of the dark image,the edge information and details of the image are preserved to improve the image quality.Thirdly,to address the problems of high-frequency noise and baseline drift in the pulse wave signal,a low-pass-band trap-zero phase shift filter is used to synthetically filter out EMG interference and industrial frequency interference and correct baseline drift.For the effect of motion artifacts,an improved Log-NLMS algorithm is proposed to reduce the error of the IPPG signal,which reduces the error of measurement results by 0.9% and improves the running time by 0.01 s compared with the Log-NLMS algorithm.Finally,heart rate and blood oxygen saturation detection based on video information was designed.Firstly,heart rate measurement were designed and judged by applying Bland-Altman analysis,and the average absolute error of the experiments was obtained as1.78%.Then a variety of complex situations were designed to investigate the stability of the experiments in terms of measurement distance,measurement time,different lighting conditions and measurement status,respectively,and it was found that the smallest error in the results was obtained after a measurement distance of 0.5m and a measurement time of 20 s for the result display.Finally,blood oxygen saturation was tested and the mean absolute error of the experiment was obtained as 0.97%,while different lighting conditions and measurement states were discussed.The results show that this experiment has good usability under different conditions. |
PDF Full Text Request