结合背景空间信息的相关滤波目标跟踪算法研究

目标检测跟踪算法在智能监控和人机交互中有着广泛的应用,而复杂场景下的跟踪技术研究在计算机视觉领域中具有重要的理论意义和商业价值。为解决复杂场景(光照变化、尺度变化、遮挡等)中由于主客观因素变化所导致的目标漂移问题,采集目标区域相邻的背景图像块来获得更多的背景特征,并将背景图像块添加到目标函数中来实现对目标图像块的限制。将结合了背景空间信息的方法集成到相关滤波器框架上,在现有公开数据集上进行实验。实验结果表明,在一些复杂场景下的跟踪效果得到了改善和提高,能够在不影响帧率的情况下,有效提高目标跟踪的成功率和准确性,优于其他相关滤波跟踪器。

CORREI ATION FILTERING OBJECT TRACKING ALGORITHM COMBINED WITH BACKGROUND SPACE INFORMATION

Object detection and tracking algorithm has a wide range of applications in intelligent monitoring and human-computer interaction.The research of tracking technology in complicated scenes has important theoretical significance and commercial value in the field of computer vision.In order to solve the problem of object drift caused by the change of subjective and objective factors in complicated scenes(illumination change,scale change,occlusion,etc.),the background image patches adjacent to the target area are collected to obtain more background features,and the back ground image patches are added to the target function to realize the limitation of the object image patches.We integrated the method combining the background space information into the correlation filter framework,and experimented on the existing open data set.The experimental results show that the tracking effect in some complicated scenes has been improved,which can effectively improve the success rate and accuracy of object tracking without affecting the frame rate.And it is superior to other correlation filter trackers.