摘要： ：针对目标跟踪过程中常见的碰撞遮挡问题，提出一种视频图像序列的多目标快速跟踪方法。该方法基于卡尔曼算法预测目标位置，采用颜色匹配算法进行前后帧目标的快速匹配，建立马尔可夫链模型并结合目标置信度，判断目标所处的状态，以消除碰撞遮挡现象产生的干扰，实现准确快速的目标跟踪。实验结果表明，该方法能在快速多目标跟踪过程中有效解决目标碰撞、遮挡等问题。

关键词: 多目标跟踪, 马尔可夫模型, 置信度, 状态分析, 颜色匹配

Abstract: An algorithm for fast tracking multi-targets in image sequences of video is proposed to solve the problem of collision and cover in targets tracking. Kalman algorithm is applied to predict the position of the targets, and color match algorithm is used to determine the fast match of the targets between two frames in this algorithm, and the states of the targets are determined by building the Markov model and analyzing the confidences of the targets to avoid the interference of collision and cover, so that the targets can be tracked accurately. Experimental results indicate that this algorithm can be used to solve the problems of collision and cover in multi-target fast track.

Key words: multi-target track, Markov model, confidence, state analysis, color matching

中图分类号: 

• TP391