高级检索

    一种基于特征提取的相关跟踪算法

    An Correlation Tracking Algorithm Based on Characteristics Extraction

    • 摘要: 介绍了一种能稳定快速跟踪复杂背景下目标的算法,该算法在传统相关跟踪算法的基础上进行改进。当目标进入红外(电视)摄像机视场时,视频信号中包含有目标信息和背景信息,信号处理器先将此信号进行数字化处理,形成具有一定灰度等级的数字化图像阵列,然后采用边缘检测、阈值分割等算法对包含有目标信息的图像进行边缘处理,提取出具有特征的目标点,从而将目标从背景中分离出来,再采用相关算法对目标进行跟踪。试验证明,对于传统算法无法稳定跟踪的目标,改进后的算法能够稳定跟踪。

       

      Abstract: A sort of stabilized and quick tracking algorithm used in the complex background is introduced.The algorithm is improved based on the conventional correlation tracking algorithm.The video signal includes target information and background information when the target enters the IR(TV) camera field of view.Firstly,the signal can be digitally processed by the signal processor and the digital image array with a certain gray level is formed.Secondly,the edge processing of image containing target information is performed by using the algorithm of edge detection and threshold segmentation.The target points with characteristics are extracted to separate the target from the background.The targets can be tracked by using correlation algorithm.The experiments prove that stabilized tracking can be realized by the improved algorithm for the targets that cannot be stably tracked by the conventional algorithm.

       

    /

    返回文章
    返回