基于视频的航站楼旅客人体特征辨识

为更好地实现航站楼智能监控,在分别分析航站楼不同功能区监控视频图像特征及特征提取效果之后,选择人头纹理特征和路径特征作为在各分区普遍适用且识别与跟踪效果良好的一组识别特征。在混合高斯背景模型前景检测算法基础上,引入基于YCbCr颜色空间阴影去除法实现阴影去除,提高前景检测精度;并基于此,分别利用基于GLCM的算法与光流法实现人头纹理特征与路径特征的提取,提高航站楼人员识别率。

Identification of Body Characteristics of Passengers Based on Video

In order to better realize the intelligent monitoring of the terminal building, after analyzing the characteristics of the video image and the feature extraction function of the different functional areas of the terminal station, select the head texture feature and the path feature, which are widely used in the terminal district and has the better recognition and tracking effect as identification features. Based on the hybrid Gaussian background model foreground detection algorithm, the shadow removal algorithm based on YCbCr color space is introduced to improve the foreground detection accuracy. And based on this, the GLCM-based algorithm and the optical flow method are used to realize the extraction of the head texture features and path characteristics, and improve the recognition rate of the terminal.