基于SIFT动态背景下的视频目标跟踪方法

Study of video object tracking method in dynamic background based on SIFT

提出了一种基于尺度不变性特征变换(SIFT)的视频目标跟踪方法.该方法通过提取前后两帧的SIFT特征,利用全图SIFT特征进行匹配,通过匹配的点对计算两帧的几何变换关系,再利用变换矩阵确定下一帧目标位置,实现了连续状态下的自动跟踪.实验结果表明此方法在遮挡、噪声、旋转、光照、缩放、小视角变化等方面有较好的鲁棒性.

This paper proposes a method for object tracking in video based on scale invariant feature transform（SIFT）.The method takes the SIFT feature from two successive frames,matches the feature of the whole image,calculates the geometry transform matrix with the match points,and then locates the target in the next frame with the matrix.Such a process realizes automatic tracking in the video.The experimental results demonstrate that the proposed method has better robustness in shelter,noise,rotation,illumination change,scale change and small viewpoint change.