基于SIFT特征和粒子滤波的目标跟踪方法  

Tacking Object Based on SIFT Features and Particle Filter

作　　者：牛长锋[1] 陈登峰[2] 刘玉树[1] 

机构地区：[1]北京理工大学计算机学院,北京100081 [2]西安建筑科技大学信息与控制工程学院,陕西西安710055

出　　处：《机器人》2010年第2期241-247,共7页Robot

基　　金：国家部委预研基金资助项目(11415133)

摘　　要：现有的基于外观的目标跟踪算法,在光照变化和遮挡的情况下,不能准确跟踪目标.针对这个问题,考虑到尺度不变特征(SIFT特征)对于光照变换、尺度变换以及仿射变换的不变性,提出了一种利用SIFT特征建立目标模型,结合粒子滤波实现目标跟踪的新方法.在跟踪过程中,根据目标模型和候选目标中SIFT特征点在时间窗内的匹配情况,自适应更新目标模型的特征点,使模型能够适应目标外观变化.仿真结果证明了方法在不同环境下的健壮性.Existing methods based on appearance models cannot track targets correctly when illumination varies or occlusion occurs. To solve the problem, considering SIFT （scale-invariant feature transform） feature invariabilities for illumination, scale and affine, a new method is proposed in which target model is constructed by SIFT feature and particle filter is used to track object. In tracking process, the target model is updated automatically according to the matching result between target model and candidate targets in time window. As a result, the target model can adapt well to appearance variation. Simulation results show that the proposed method is robust in various scenes.

关 键 词：目标跟踪 粒子滤波 尺度不变特征 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]