Research Of Binocular Integrated Identity Verification And Non-contact Heart Rate Detection Based On Binocular

With the improvement of people's life level,more and more people pay more attention to their health,and rely on intelligent medical products for daily protection and detection.Heart rate is one of the most important indicators to reflect human health.How to measure heart rate conveniently,quickly and accurately is a research hotspot.The traditional heart rate measurement method has many shortcomings,such as cumbersome process,expensive instruments,direct skin contact and so on.In order to solve these problems,this paper studies a heart rate measurement method based on imaging photoplethysmography,and combines with living recognition and face recognition technology,which can quickly and accurately realize identity verification and heart rate measurement without contact,More intelligent to complete the integration of the detection process,to achieve the cloud measurement data sharing,and according to the false face detection to overcome the ippg technology using false face can also measure the heart rate value of the drawbacks,finally completed the integration of authentication and non-contact heart rate measurement system on Android platform.The main work of this paper is as follows(1)Face authentication based on binocular and false face detection based on infrared thermal imaging.Before the heart rate measurement,it is necessary to realize the identification and false face detection of personnel,so as to realize the rapid and automatic identification and upload of measured heart rate data,and overcome the false detection of heart rate value using false face.This paper studies the mtcnn algorithm,according to the algorithm to extract the position information of the face and the main key points in the image,using the face temperature information obtained by infrared thermal imaging to determine whether the average temperature of the face is in line with the normal threshold to screen the false face.Then,the standard face image is obtained by preprocessing the portrait image,and the LBP feature vector of the standard face image is extracted.The traditional SVM classification algorithm is improved,and the improved PSVM classification algorithm based on FCNN is proposed to realize face recognition.(2)Non contact heart rate detection based on ippg technology.In this paper,the principle of imaging photoplethysmography is studied.Using the face image video,after digital filtering,the RGB three channel observation signal is extracted from the region of interest,and the pulse wave signal is separated by blind source.Finally,the heart rate value is calculated by the pulse wave spectrum.(3)Heart rate correction based on light balance and linear regression model.The change,fluctuation and uneven illumination of light will affect the final result of heart rate calculation.In this paper,an improved two-dimensional gamma function is proposed to achieve light balance and eliminate the influence of uneven light.The linear regression model is used to modify the heart rate value measured by imaging photoplethysmography to improve the accuracy of heart rate measurement.(4)System implementation based on Android platform.In this paper,Android platform is selected as the development platform of the system to realize the functions of face identification information confirmation,face false face detection and non-contact heart rate detection.The hardware implementation of the system is simple and easy to operate.Finally,the integration of identity verification and non-contact heart rate measurement system is tested.The experimental results show that the system can quickly complete the identification and false face detection,and the error between the real face heart rate value and the actual heart rate value is only about 4%,which has a good application prospect.