机构地区: 中山大学新华学院
出 处: 《计算机仿真》 2011年第6期343-346,共4页
摘 要: 研究道路交通车辆目标实时跟踪问题,要求跟踪检测要实时有效,数据准确。针对传统的粒子滤波(M ean-Sh ift)跟踪算法计算量很大,难以实现道路交通检测中的车辆实时追踪等数据。为解决上述问题,提出了一种改进的M ean-Sh ift目标跟踪算法。首先算法引入了一个预测矢量,用来预测目标在下一帧可能出现的位置,在跟踪时算法从预测的位置开始迭代,直至收敛于目标真实位置。最后给出了一段道路交通视频仿真。仿真结果表明,所提出的算法从很大程度上提高了原有m ean-sh ift算法的效率,有利于道路交通实时检测跟踪,为设计提供了参考依据。 Mean Shift is now widely used in the field of target tracking,real-time detection of road vehicles tracking multiple targets,for the traditional Mean Shift tracking algorithm is large and difficult road test of the multi-objective real-time tracking.A modified Mean Shift Based Tracking Algorithm is proposed.Algorithm introduces a prediction vector to predict the possible target in the next frame position in the tracking algorithm for prediction of the position from the beginning iterations,until convergence to the target true position.Finally,the simulation experiment results show that the proposed algorithm is improved compared with the original mean-shift algorithm and is efficient and conducive to real-time tracking for road traffic.
领 域: [自动化与计算机技术] [自动化与计算机技术]