基于云台摄像机的快速移动人群的检测与跟踪

为了实现校园的安全监控,提出了基于云台摄像机的快速移动人群的检测与跟踪算法。介绍了云台摄像机用于校园安全监控的基本构成。为了更好地实现云台摄像头对于移动人群的检测与跟踪,建立了基于现实的精确摄像机模型,提出了摄像机自旋转角度约束的摄像机模型。通过最新的核相关滤波器跟踪算法(KCF)实现对运动着的人群检测与跟踪。运用Matlab仿真实验比较该方法和卡尔曼滤波跟踪算法,选择最优方法和相关滤波器跟踪算法来实现检测与跟踪要求。结果表明:相比较于传统的卡尔曼滤波跟踪算法,KFC算法的跟踪精度优于传统方法,精度多数情况下能达到90%以上,高于卡尔曼滤波跟踪算法的60%,检测与跟踪效果达到要求。

Detection and tracking of fast moving camera based on the PTZ camera

In order to realize the campus safety monitoring, we put forward the detection and tracking of fast moving crowd algorithm based on the PTZ camera. Camera for basic construction of campus safety monitoring is introduced. In order to achieve PTZ camera for better detecting and tracking, the mobile population was established based on the reality of precision of the camera model, the camera rotation angle constraint, the camera model. Detection and tracking of moving people are implemented by using the latest kernel correlation filter tracking algorithm (KCF). The Matlab simulation experiment is used to compare the proposed method and Calman filter tracking algorithm, and the optimal method is selected to detect and track the kernel correlation filter tracking algorithm. The results show that the method is feasible, compared with the traditional Kalman filter tracking algorithm, the tracking accuracy of KFC algorithm is superior to the traditional methods. The accuracy in most cases can reach more than 90%, higher than 60% of the Kalman filter tracking algorithm. Testing and track performance meet the requirements.