Research On Pulse Wave Extraction Method Based On Wrist Radial Artery

Pulse wave is an important physiological health index,which reflects the cardiovascular health of human body.It provides a basis for doctors to diagnose diseases.The traditional contact pulse wave detection method needs to be close to the patient’s body,which will bring a lot of inconvenience to the detection process.Photoplethysmography is a new and practical non-contact pulse wave detection technology,which can extract the color change or vibration information of human body parts by using ordinary camera.The technology has the advantages of low cost and convenient use.However,when collecting relevant video,it is vulnerable to the illumination interference in the environment and the human body’s involuntary motion interference,so the robustness of the method is poor.In order to solve the above problems and improve the accuracy of non-contact pulse wave detection,the vibration information of the radial artery is extracted from the human wrist radial artery(1)Aiming at the problem that the detection results are easily interfered by the ambient light,a heart rate extraction method based on Euler motion video amplification is proposed.Firstly,the differential pyramid is used to improve the Euler motion video amplification,so that it can better reduce the wrist irrelevant noise interference,and the method is used to visualize the small vibration of the wrist skin caused by the radial artery of the wrist,so as to reduce the interference Then the sub region block containing the most pulse wave information is selected through the optimal ROI region,and the pure pulse wave is extracted by combining with the principal component analysis algorithm for denoising,and finally the heart rate is calculated through the power spectrum.(2)Aiming at the problems of camera jitter and human body’s involuntary motion,a non-contact heart rate detection method based on vibration phase signal decomposition is proposed.The core idea is to use complex controllable pyramid to extract pulse characteristics,and design high signal-to-noise ratio selection module and robust principal component analysis module to suppress the noise,so as to finally realize the heart rate detection of pulse signal anti noise.Firstly,the video of the wrist radial artery is collected,and the direction perpendicular to the arm contour is selected as the direction of the complex controllable pyramid.Then,the phase of each pixel is decomposed,and the phase of the first frame of each pixel is selected as the standard.The phase difference signal is obtained by subtracting the phase of the corresponding pixel of each subsequent frame.The region with the highest ROI signal-to-noise ratio is selected as the region to be processed Finally,the robust principal component analysis algorithm is used to extract the pulse wave signal,and the power spectrum estimation method is used to calculate the heart rate.In order to test the performance of the algorithm,the contact pulse wave detection equipment is used as the true value,and a variety of mainstream methods are used to carry out pulse wave extraction experiments in different scenes.The accuracy and robustness of the algorithm are proved through the experiments,which proves that the method in this paper is superior to the current mainstream methods,and the performance of the algorithm is verified.