Non-contact Heart Rate Measurement Using PPG And Color Videos

Heart rate is an important physiological parameter, so the long-term monitoringof heart rate is of great significance to the human health care. This project is based onthe ’observation’ in traditional Chinese medicine, and combines with the principle ofPPG technology. That collects color video of human face using camera bynon-contact manner, to detect the variation of light reflectance or transmission causedby the different light absorption effect during blood and tissue, and describe thechange of blood volume during cardiac cycles. Secondly, it used some modernmethods of signal extraction and processing to get the pulse wave, and compute thevalue of heart rate. The continuous non-contact and noninvasive measurement of heartrate is completed, and this method has many advantages like low-cost, easy to operate,stable performance, and good repeatability. It will be important for the developmentof intelligent health care in the future internet of things, and for the astronauts’physiological parameters monitoring.The main work of this paper is as follows:(1) It summarized existing heart rate monitoring methods, including contact andnon-contact monitoring methods, as well as the principle and application of PPGtechnology. And we proposed a non-contact heart rate monitoring method bycombining with the advantages of the PPG technology and the image-basednoncontact heart rate monitoring method, and described the key technologies such asface detection, blind source separation and spectrum analysis in detail.(2) We designed a non-contact heart rate measurement system, which is based onordinary webcam, i.e. the color video of human body surface (face) is obtained by anordinary camera, which is converted into a digital signal by using the method ofdigital image processing and analyzed by blind source separation (ICA), and thenPPG signal is extracted. Subsequently, heart rate is calculated by selecting, filteringand spectrum analysis of the PPG signal.(3) We used the method of face detection to extract the face zone, and blindsource separation to eliminate robustly the noise induced by motion artifacts, for thepurpose of eliminating the interference of background noise and pseudo-dynamicnoise to this measurement system and improving the accuracy. (4) In order to solve the problem of the disorder in blind source separation, weconducted correlation analysis of the independent source signals with the originalgreen channel signal respectively, and selected the most relevant signal as the originalsignal of heart rate. And we develop the software system called E-eyes.(5) Design the control experiment for this system, where we conductedBland-Altman analysis of the data by this system and the normal heart rate monitoringequipment (OMRON sphygmomanometer), the result shows that the system in thispaper meets the need of measurement of physiological parameters.