基于OpenCV的双摄像头活体人脸判定及实时跟踪系统

The Dual Cameras-Vivo-Face Recognition and Real-time Tracking System based on OpenCV

活体检测及人脸跟踪系统的构建已成为现代计算机视觉及OpenCV研究的热点。提出一种基于OpenCV的双摄像头活体人脸判定实时人脸跟踪系统。一种新的双摄像头活体人脸判定的方法,利用基于HAAR特征人脸检测算法,并判断距离,编写Windows程序调用串口API和下位单片机STC12C5410通信,采用CH341A芯片将RS232串口数据加载到USB协议数据栈中实现USB传输,单片机控制FUTABA舵机,获取上位机人脸检测坐标信息,实时调整舵机转动角度,使人脸始终处于图像中间290～310阈值范围内,达到人脸跟踪。该系统结构精简,可作为模块用于门禁安保系统及构建其他机器人研究平台。

Vivo detection and face tracking system has become a hotspot of modern computer vision and OpenCV research.. The dual cameras-vivo-face recognition and real-time tracking system based on OpenCV is presented. A new method of dual cameras-vivo-face recognition, account of the HAAR face detection algorithm which can judge distance, is applied. Serial API and microcontroller STC12C5410 can achieve the call communication by writing program of windows, adding the RS232 serial data to the USB protocol stack with CH341A chip to implement USB data transmission. The microprocessor controls the FUTABA Servo to obtain the coordinate information of face detection from the PC, with real-time adjustment of the servo rotating angle to guarantee the face is always in the middle of the image within a certain threshold. This lean system can be used as a module in access security systems and the construction of robots research platform.