无线视频监控系统人脸跟踪功能的实现  

Realization of Face Tracking Function in Wireless Video Surveillance System

作　　者：周岳斌 杨沫 曹煜晖 ZHOU Yuebin;YANG Mo;CAO Yuhui(School of Mechanical Engineering,Hubei University of Arts and Science,Xiangyang 441053,China)

机构地区：[1]湖北文理学院机械工程学院,湖北襄阳441053

出　　处：《机械与电子》2021年第6期35-38,43,共5页Machinery & Electronics

基　　金：湖北文理学院机械工程省级重点学科开放基金项目(XK2020009)。

摘　　要：为实现视频监控的人脸跟踪功能,结合计算机视觉与舵机云台控制技术,设计了一种基于嵌入式处理器的无线视频监控系统。利用MicroPython计算机视觉库对采集图像进行滤波处理,以增强图像效果。采用改进卷积神经网络模型进行机器学习和人脸识别,将处理后的监控图像上传至服务器,用户可通过查看浏览器进行实时监控。测试结果表明,当有人脸出现在视频监控范围内时,系统可驱动舵机云台进行有效的人脸识别与追踪。In order to realize the face tracking function of video surveillance,a wireless video surveillance system based on embedded processor is designed by combining computer vision and steering engine pantilt control technology.The captured image is filtered using the MicroPython computer vision library to enhance the image effect.The improved convolutional neural network model is used for machine learning and face recognition.The processed monitoring image is uploaded to the server,and the user can monitor in real time by viewing the browser.The test results show that when a human face appears in the video surveillance range,the system can drive the steering engine pantilt to perform face recognition and tracking effectively.

关 键 词：监控系统 机器学习 卷积神经网络 人脸跟踪 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]