基于双目计算机视觉的自适应识别算法及其监控应用

An Adaptive Recognition Algorithm Based on Two-eye Computer Vision and Its Application in Monitoring

双目计算机视觉是利用仿生学原理,通过标定后的双摄像头来得到同步曝光图像,然后计算获取的2维图像像素点的第3维深度信息。为了对不同环境场景进行监控提出了一种新的基于双目计算机视觉的自适应识别算法。该算法首先利用像素点的深度信息对场景进行识别判断,然后采用统计的方法为场景建模,并通过时间滤波克服光照渐变,以及通过深度算法特性克服光照突变。与单摄像头监控系统相比,利用该算法实现的视频监控原型系统,可应用于更多场合,并利用深度信息设置报警级别,来降低误检率。

Two-eye computer vision algorithm uses the principle of Bionics. It gets the depth information of each pixel by two demarcated cameras. A new adaptive recognition algorithm based on two-eyes computer vision is presented in this paper. This algorithm makes use of the depth information of each pixel. Then it adopts the statistics method to model the scene and overcomes the gradual light change using time filter. It can also overcome the sudden light change with the characteristic of the depth algorithm. The video monitor antitype system using our algorithm can be set at more places than one camera system. It can set the alarm level according to the depth. Information, which can reduce the mistake ratio.