非接触式面部穴位温度检测系统的设计

The Design of Non-Contact Facial Temperature Detection System

摘要: 针对普通测温仪以检测整个面部的温度为主，继而引起测温准确度受限的问题，提出通过改进的Dlib人脸检测算法对面部穴位进行测温。本文采用KTVC3100SY红外无感快速人体测温筛查热像仪为实验平台，借助其可见光摄像头与红外摄像头采集的信息进行穴位温度检测。针对测温仪的红外图像与可见光图像无法建立联系，提出通过借助UDP协议传输实现坐标点和红外数据的对应得到温度值。通过在VS环境下进行python编程，对人脸进行测温实验。结果表明：采用改进的Dlib人脸检测算法进行穴位温度提高了测温的精确度，弥补了整体面部测温受自然环境影响的不足；并不是所有的穴位都具有这种优势，选择正确穴位提高穴位测温精确度。

Abstract: In view of the problem that the normal thermometer mainly detects the temperature of the whole face, and then causes the problem of limited temperature measurement accuracy, it is proposed to measure the temperature of the face acuity by the improved Dlib face detection algorithm. In this paper, the KTVC3100SY infrared nonsensitive fast human body temperature screening thermal imager was used as the experimental platform, and the temperature of acupoint was detected with the information collected by its visible light camera and infrared camera. In view of the failure to establish contact between infrared image and visible image of the thermometer, it is proposed that the coordinate point and infrared data can be obtained by using the UDP protocol transmission to obtain the temperature value. The face temperature was measured by python programming in the VS environment. The results show that the acuity temperature using the improved Dlib face detection algorithm improves the accuracy of temperature measurement and makes up for the lack of in-fluence of the overall facial temperature measurement by the natural environment. But not all acupoints have this advantage. Choosing the right acupoints can improve the accuracy of temperature measurement at acupoints.