基于面部图像的心率测量系统

吕扬; 刘杰

doi:10.3788/OPE.20152313.0652

您当前的位置：

首页 >

文章列表页 >

基于面部图像的心率测量系统

更新时间：2020-08-13

- 基于面部图像的心率测量系统
- Heart rate acquisition system on facial image
- 光学精密工程 2015年23卷第10z期页码：651-659
- 作者机构：
  
  吉林大学仪器科学与电气工程学院,吉林长春,130026
- 作者简介：
- 基金信息：
  
  吉林省科技发展计划项目重大科技专项(No.20116014)()
- DOI：10.3788/OPE.20152313.0652
  中图分类号： TH773;TP391
- 收稿日期：2015-06-05，
  
  修回日期：2015-06-21，
  
  纸质出版日期：2015-11-14
- 稿件说明：
移动端阅览
吕扬, 刘杰,. 基于面部图像的心率测量系统[J]. 光学精密工程, 2015,23(10z): 651-659

L&#220, Yang, LIU Jie. Heart rate acquisition system on facial image[J]. Editorial Office of Optics and Precision Engineering, 2015,23(10z): 651-659
吕扬, 刘杰,. 基于面部图像的心率测量系统[J]. 光学精密工程, 2015,23(10z): 651-659 DOI： 10.3788/OPE.20152313.0652.

L&#220, Yang, LIU Jie. Heart rate acquisition system on facial image[J]. Editorial Office of Optics and Precision Engineering, 2015,23(10z): 651-659 DOI： 10.3788/OPE.20152313.0652.

摘要

研究了非接触式心率采集方法

设计了基于脸部图像处理的心率测量系统。利用该系统分两步完成了对人脸部分区域图像的视频采集和处理。第一步

对单幅图像进行处理。首先建立一个非彩色模型对原始图像进行颜色校正

去除光线带来的干扰;然后

利用直方图和脸部特征定位样本区域;在空间域和时间域上对图像进行增强

将窄带灰度输入值映射到宽带输出值

并将照射分量与反射分量分离

进一步去除光线照射不均带来的干扰;最后求出单幅图像的反射均值。第二步

采用小波变换和Hilbert算法对所有图像均值构成的曲线进行处理

消除波形的基线漂移

并使R波更容易被检测到

最后得到心率值。多条件下的实验结果表明:光线充足的情况下

测得的心率误差不大于8%;光线较差的情况下

心率误差不大于20%;基本上实现了通过采集人脸区域光线变化计算心率的功能要求。由于摄像头采集速率的限制

在满足信号完整性的情况下

系统的实时性为每5 s更新一次。

Abstract

The extent of face absorption light changing along with the heart beat rate was verified

and a heart rate acquisition system based on the facial image processing was designed. Human face images were sampled in a video form by the system and they were processed in two steps. The first step was the processing of a single image. Firstly

an achromatic color model was built to correct the color of the original image and to remove the interference from light brings. Then

the sample area was positioned using a histogram and facial features

and the image was enhanced in space and time domains. Furthermore

narrowband gray input values were mapped to the output value broadband and to separate the illumination component and the reflection component

so the interference caused by uneven light irradiation could be further erased. Finally

the mean reflection of single image was calculated. In the second step

the wavelet transform and Hilbert algorithm were used to process the mean curve consisting of all the images and to eliminate the baseline drift wave

making the R wave more likely to be detected and to get the heart rate finally. Experimental results under many conditions show that the heart rate error is no more than 8% under a good light condition

and that is no more than 20% at a poor light condition. Basically

the system achieves the functional requirement of the heart rate calculation by sampling the light of some face region changes. Due to limitations of the camera sampling rate

the real-time performance of the system is updating every 5 s in the case of signal integrity ensured.

关键词

Keywords

references

TAKANO C, OHTA Y. Heart rate measurement based on a time-lapse image[J]. Medical Engineering & Physics, 2007,29(8):853-857.

张志祥. 间断动力系统的随机扰动及其在守恒律方程中的应用[D].北京:北京大学, 1998,50-55. ZHANG ZH X. Random Perturbations of Intermittent Dynamical Systems and its Application in Conservation Laws Equations[D]. Beijing:Peking University, 1998,50-55.(in Chinese)

杨福生,吕扬生. 生物医学信号的处理和识别[M]. 北京:中国中医药出版社, 2001. YANG F SH, LÜ Y SH. Processing and Recognition of Biomedical Signal[M]. Beijing:Chinese Medical Press,2001.(in Chinese)

王乐新,赵志敏,姚红兵,等. 血液的红外吸收光谱分析及应用研究[J]. 光谱学与光谱分析, 2002, 22(6):980-982. WANG L X,ZHAO ZH M, YAO H B, et al.. Infrared absorption spectrum analysis and its application to blood[J].Spectroscopy and Spectral Analysis. 2002, 22(06):980-982.(in Chinese)

GONZALEZ R C, WOODS R E. Digital Image Processing[M]. 2nd Edition. RUAN QQ Transl. Beijing:Electronic Industry Press,2003.

BAILEY D G. FPGA-based Embedded Image Processing System Design[M].YUAN K,HE WB, XIAO H Transl. Beijing:Electronic Industry Press,2013.

赵东方, 杨明, 邓世涛. 几种人脸检测算法的对比研究[J]. 内将科技,2011, 10:95-99. ZHAO D F, YANG M, DENG SH T. Comparative study of several human face detection algorithm[J]. Neijiang Technology, 2011, 10:95-99.(in Chinese)

OPPENHEIM A V, WILLSKY A S. Signals & Systerms[M]. 2nd Edition, Beijing:Electronic Industry Press, 2011.

陈春晓, 刘建业. 基于小波变换的心电信号去噪处理[J]. 生物医学工程学杂志, 2004, 21(1):89-92. CHEN CH X, LIU J Y. An approach based on wavelet transform to removal of noises[J]. Journal of Biomedical Engineering, 2004, 21(1):89-92.(in Chinese)

GE ZH X, SHA W. Wavelet Analysis Theory and MATLAB R2007 Realization[M]. Beijing:Electronic Industry Press, 2007.

BURRUS C S, GOPINATH R A. Introduction to Wavelets and Wavelet Transforms[M]. Beijing:Machinery Industry Press. 2007

PENG Y Y. Wavelet Analysis in a One-dimensional Signal Denoising[D]. Beijing:Beijing University of Posts andTelecommunications, 2010.

王国恩, 李静, 厉春生. 小波变换在信号去噪方面的应用研究[J]. 舰船电子工程, 2008, 28(11):80-83. WANG G E, LI J, LI CH SH. Research on the application of wavelet transform in terms of denoising of signal[J]. Ship Electronic Engineering, 2008, 28(11):80-83.(in Chinese)

胡俊文, 周国荣. 小波分析在振动信号去噪中的应用[J]. 机械工程与自动化, 2010(1):128-130. HU J W, ZHOU G R. Application of Wavelet analysis in vibration signal denoising[J]. Mechanical Engineering and Automation, 2010(1):128-130.(in Chinese)

曹亚君, 陈树宁. 一种基于小波分析的阈值去噪算法[J]. 无线互联科技, 2014(9):132-133. CAO Y J, CHEN SH N. An analysis based on wavelet threshold denoising algorithm[J]. Wireless Internet Technology,2014(9):132-133.(in Chinese)

浏览量

249

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于双目结构光周视扫描的3D件三维重构与位姿估计研究

重叠分数与匹配矩阵驱动的部分重叠点云配准网络设计

基于Snell窗及偏振模式的水下载体航姿测量

改进多尺度结构化融合的红外与可见光图像融合

非成对训练样本条件下的红外图像生成