摘要
图像跟踪算法是成像跟踪系统的关键技术之一。如何提高计算速度,满足实时要求,始终是该技术一个值得研究的课题。该文介绍了选择目标跟踪常用的跟踪算法,提出并详细阐述了一种逐步逼近的相关跟踪算法,该算法假设目标图像在帧间位移是一有限值,从目标图像中选定一点,精心设计X、Y两特征向量,逐次逼近目标的指定点,该算法最终收敛于指定点。仿真结果表明,该算法不仅能显著加快跟踪速度,而且有较大的实用性。
Image tracking is a key technique in TV Tracking System . It is always a hot subject to reduce operation time for meeting the requirement of real time. After presenting the principle of general image tracking in brief ,The paper introduces a way of approaching gradually correlation algorithm in detail . This algorithm is based on the hypothesis that the moving displacement of an object image between frames is limited . We can choose a proper point of the object and design special X and Y eigenvector related with it to approach gradually.It can converge in the end. The result of simulation shows that this algorithm can increase remarkably the tracking speed and has practicality.
出处
《计算机仿真》
CSCD
2006年第2期93-95,共3页
Computer Simulation
关键词
图像处理
相关跟踪
算法
Image processing
Correlation tracking
Algorithm
