一种改进的运动目标跟踪与轨迹记录算法

作者:

赵雷; 陈万忠; 韩双双;

吉林大学通信工程学院; 吉林长春;

Author(s):

ZHAO Lei; CHEN Wan-zhong; HAN Shuang-shuang（College of Communication Engineering; Jilin University; Changchun 130025; China）;

关键词:

运动轨迹记录跟踪;  主动轮廓算法;  三帧差分算法;  Hausdorff匹配算法

Keywords:

recording and tracking of moving objects;  Snake algorithm;  three-frame differencing;  Hausdorff matching

分类号:

TP391.41

DOI:

CNKI:SUN:ZNXT.0.2008-02-014

文献标志码:

A

摘要:

针对目前运动目标的跟踪与记录方法占用存储空间较大的缺点,提出了一种减少存储空间的记录算法,即先用三帧差分算法和Snake算法相结合检出运动物体的轮廓,再利用Hausdorff算法对提出的轮廓进行匹配,并将匹配后的轮廓和运动轨迹以文本文件存储,大大降低了运动目标轨迹记录存储容量.实际运用表明,改进后的记录存储空间相当于通常视频文件的万分之一.该算法适于长时间记录运动目标轨迹.

Abstract:

At present,data from tracking moving targets consumes too much memory.An improved recording algorithm is proposed that reduces memory requirements.First,initial points are obtained by using a three-frame difference algorithm.Then the Snake algorithm is used to extend the initial points to determine the object’s contour.Finally,the Hausdorff distance measurement matches contours to standard figures and the results are saved in a text file.The improved algorithm uses approximately 0.01% of the usual memory re...