摘要
研究运动物体目标跟踪精确度问题,由于存在遮挡和多光源的噪声影响检测精度,而且运动目标的跟踪是在连续的图像帧间创建位置、速度、形状等存在匹配问题。传统的目标跟踪算法由于目标的动态移动速度大,而容易导致跟踪丢失目标。为了解决上述问题,提出了一种改进的基于自适应均值移动(Cam Shift)目标跟踪新算法。主要难点技术问题是提取了多运动目标视频图像,进行了背景分离。算法是一种颜色跟踪算法,根据多次迭代的计算结果,自适应调整图像,实现对运动目标的实时跟踪。仿真结果表明,提出的改进目标跟踪算法的跟踪精度和滤波效果有了较大提高,同时具有较强的鲁棒性能。
The precision of moving object tracking problem.Moving target tracking is the image frame to create a continuous position,velocity,shape and other relevant characteristics of the corresponding matching,target tracking algorithm for the traditional goal of the dynamic moving speed as large and easily lead to loss of target issues such as defect tracking,in order to solve the above problem,an improved adaptive mean shift based（Cam Shift） Target Tracking Algorithm.The algorithm is a color-based tracking algorithm,based on multiple iterations of the calculation results,adjusts the image sequence adaptive search window size and position,resulting in the current image in the center of the target to achieve real-time tracking of moving targets.Simulation results show that the proposed target tracking algorithm to improve the tracking accuracy and filtering effect has been greatly improved,and has strong robustness.
出处
《计算机仿真》
CSCD
北大核心
2012年第4期290-292,396,共4页
Computer Simulation
关键词
目标跟踪
自适应均值移动
核函数
迭代
Target tracking
Adaptive mean shift
Kernel function
Iteration
