首页 | 本学科首页   官方微博 | 高级检索  
     

基于均值漂移与卡尔曼预测相结合的视频运动目标跟踪算法
引用本文:姜忠民,赵建民,朱信忠,徐慧英.基于均值漂移与卡尔曼预测相结合的视频运动目标跟踪算法[J].计算机时代,2010(2):22-24.
作者姓名:姜忠民  赵建民  朱信忠  徐慧英
作者单位:浙江师范大学数理与信息工程学院,浙江,金华,321004
基金项目:国家自然科学基金项目(60772071);;浙江省科技计划项目(2008C14063)
摘    要:传统的均值漂移算法是基于颜色直方图的迭代跟踪算法,在跟踪目标出现尺度变化的情况下,其跟踪结果往往出现偏差,甚至跟踪失败。鉴于此,提出了一种基于空间边缘方向直方图的均值漂移跟踪算法,使用空间分布和纹理信息作为匹配信息,将卡尔曼预测器融于均值漂移跟踪算法。实验表明,该方法能在尺度缩放等复杂背景下对目标进行准确有效的跟踪。

关 键 词:空间边缘方向直方图  均值漂移算法  卡尔曼预测器  视频目标跟踪

Video Moving Target Tracking Algorithm Based on Combination of Mean Shift and Kalman Predictor
JIANG Zhong-min,ZHAO Jian-min,ZHU Xin-zhong,XU Hui-ying.Video Moving Target Tracking Algorithm Based on Combination of Mean Shift and Kalman Predictor[J].Computer Era,2010(2):22-24.
Authors:JIANG Zhong-min  ZHAO Jian-min  ZHU Xin-zhong  XU Hui-ying
Affiliation:College of Mathematics Physics and Information Engineering/a>;Zhejiang Normal University/a>;Jinhua/a>;Zhejiang 321004/a>;China
Abstract:The traditional mean shift algorithm is the iteration tracking algorithm based on color histogram, its tracking often appears deviation or fails under the condition of scale variations occurring in tracked object. So a mean shift tracking algorithm based on spatial edge orientation histograms is presented, which uses space distribution and texture information as matching information and puts Kalman predictor in mean shift tracking algorithm. The experiments show that the algorithm can track target accuratel...
Keywords:spatial edge orientation histograms  mean shift algorithm  Kalman predictor  video target tracking  
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号