Detection Technology And System Design Of Local Blood Flow Distribution Based On RPPG

Some important physiological information functions of human body are often closely related to its blood flow distribution,and the changes of characteristic parameters can directly reflect the health and emotional status.At present,the detection of blood flow distribution is usually based on the principles of doppler,ultrasound,thermal imaging,etc.,which has a high cost and requires contact with body parts,long-term monitoring has a certain constraint on the human body.Based on the principle of remote photoplethysmography(r PPG),this paper proposed a non-contact blood flow distribution detection technology.Using the fast digital phase-locking and independent component analysis and other methods to realize non-contact detection of blood distribution and visualization analysis,which improves the system’s anti-interference ability and system robustness,and realize the long time continuous non-contact monitoring blood flow distribution,and has wide application prospect.The main work of this paper are as follows:(1)Based on the principle of r PPG,the system platform was built and the accuracy of test results was verified.20 sets of facial color video were collected,pulse wave was extracted and heart rate was calculated by fast digital phase-lock demodulation,color space conversion,independent analysis and other algorithms.The error of the results met the error requirements of medical device standard YY 1079-2008 traditional Chinese medicine oximeter.Bland-altman analysis was carried out between the detection results and the results collected by the contact device,and 95% of the data points fell within the consistency interval and with good consistency.(2)Based on fast digital lock-in technology,the signal to noise ratio and anti-interference ability of the system are improved.The simulation results show that the algorithm is effective through the phase-locked demodulation.The improvement of SNR by fast digital phase lock algorithm was evaluated by experiments of black and white polyformaldehyde board(simulated human skin).(3)Quantitative evaluation of the anti-interference ability of each heart rate extraction algorithm.Sample rate error,gaussian noise,baseline drift and motion interference of the same order as the real environment were simulated to quantitatively evaluate the influence of different interference factors on the extraction algorithm of each heart rate.(4)Verification of the the accuracy of blood distribution detection.The experimental paradigm was designed to stimulate the emotions of the subjects and change the distribution of facial blood flow by means of video.Based on the r PPG blood distribution detection system designed in this paper,facial video was collected,and the Support Vector Machine(SVM)grid was used to search for the optimal kernel function and penalty factor,and 42 data sets were modeled to achieve the dichotomy of blood distribution.The accuracy of the model was 86.7%,which proved the accuracy of the system.This study shows that the non-contact blood flow distribution detection system based on the principle of r PPG has high signal-to-noise ratio and anti-interference ability of environmental light,and it can realize long time continuous monitoring of blood flow distribution,cast off the yoke of the equipment,with many advantages such as non-contact,fast,convenient,easy to operate.It has an extremely important application value in disease screening,fatigue detection,emotional calculation and other fields.