基于改进CamShift的多场景接力跟踪  

Multi-scene Relay Track Based on Improving CamShift

作　　者：曲巨宝[1] 

机构地区：[1]武夷学院数学与计算机系,福建武夷山354300

出　　处：《重庆师范大学学报（自然科学版）》2010年第6期69-72,82,共5页Journal of Chongqing Normal University:Natural Science

基　　金：福建省教育厅科技项目(No.JA09240);武夷学院智能计算网格科研项目(No.2009)

摘　　要：为解决多场景多摄像机下目标接力跟踪问题,减少因场景、光照、视角、遮挡等变化引起的目标丢失,构建了基于CamShift算法的多颜色模型空间,并用多目标规划最优求解法获得最优组合,以减少场景转换给目标追踪带来的干扰;并利用IIR滤波器、加速度位移方程、二阶矩阵等对目标位置、速度、旋转角度等进行预测。大量实验表明,该方法能够在无人值守的情况下,对多个摄像机视频录像中的目标进行自动搜索和跟踪,跟踪效率较高,鲁棒性较强。In order to solve the problem of target relay tracking under the multi-scene and multi-cameras,and to reduce target loss which caused by the changing of on scene,illumination,angle of view,mask and so on,a multi-color model space that based on the CamShift algorithm is introduced in this paper.The most superior combination by using the optimal solution rule of multi-objective programming to reduce the disturbance towards the target tracing caused by the scene transformation.The paper also used the IIR filter,the acceleration displacement equation,the second-order matrix etc.to make a forecast to the target location,the speed,the degrees rotation and so on.The massive experiments indicated that,in the unattended situation,this method can carry on the auto search and tracking to many video recording targets in many camera,and it also can get high track efficiency and strong robustness.

关 键 词：跟踪 多场景 视频 倾角 CAMSHIFT 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]