An Image-Based Non-Contact Pulse Rate and Body Temperature Measurement System

碩士 === 國立臺灣科技大學 === 電子工程系 === 107 === Vital Signs are signs that appear when maintaining the body's basic physiological functions. These measurements can be used to assess an individual's physical health, and in which the pulse rate and body temperature are the indicators most commonly use...

Full description

Bibliographic Details
Main Authors: Ming-Hung Lu, 呂明紘
Other Authors: Yuan-Hsiang Lin
Format: Others
Language:zh-TW
Published: 2019
Online Access:http://ndltd.ncl.edu.tw/handle/827u73
Description
Summary:碩士 === 國立臺灣科技大學 === 電子工程系 === 107 === Vital Signs are signs that appear when maintaining the body's basic physiological functions. These measurements can be used to assess an individual's physical health, and in which the pulse rate and body temperature are the indicators most commonly used by doctors to monitor a patient's physical condition. Therefore, when visiting a doctor, patients are often asked to measure the pulse rate and body temperature first and the doctor will infer the conditions and prescribe according to the patient’s physiological values and the described symptoms. Today hospitals usually adopt contact instruments for measuring the pulse rate and body temperature. Yet the patients’ activities are limited during the process of measuring, and there is a risk of contact infection by using contact instruments in the hospital. To solve the above problems, this paper establishes an image-based contactless pulse rate and body temperature measurement system. This system can instantly perform contactless measurements by using a normal RGB camera and a far-infrared photo thermal image camera, which allows the users to move freely while measuring the pulse rate and body temperature. Thus not only can the risk of contact infection be avoided but the patients are given bigger activity space. However, according to previous research, contactless pulse rate signals are easily influenced by moving artifacts. This paper proposed an algorithm to reduce moving noise. Compared to previous research, the algorithm can measure the pulse rate more precisely in a state of severe shaking. According to the experimental results, the mean absolute error (MAE) and root mean square error (RMSE) are respectively 4.42 BPM and 6.21 BPM in a state where the user is shaking vigorously. In body temperature measurement, the paper proposed a new algorithm on human face temperature measurement that can automatically detect and track human face positions with a thermal image camera. In this way, the user can move freely and use the system more comfortably. According to the final experimental results, the MAE and RMSE of the measured body temperature were 0.375 °C and 0.439 °C, respectively.