基于KPKM算法的动目标检测与跟踪  

Moves Object Detection and Track Based on KPKM Algorithm

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：《武夷学院学报》2010年第5期48-53,共6页Journal of Wuyi University

基　　金：福建省教育厅科技项目(JA09240);武夷学院智能计算网格科研团队(2009)

摘　　要：针对均值偏移算法在跟踪目标发生形变和遮挡时丢失问题,提出了一种自适应目标检测、核函数带宽可变、Kalman滤波预测和重心轨迹跟踪的改进均值偏移算法(KPKM)。该算法利用目标检测中得到的外接矩形和重心作为均值偏移算法的初值,用改进的Kalman滤波器预测目标运动趋势,使本算法能沿着梯度方向快速收敛到目标中心。实验和仿真结果表明,该方法实现了在复杂场景下,对运动目标的精确检测和准确跟踪。Mean shift algorithm for tracking targets when the loss occurred deformation and occlusion problems, an adaptive target detection, variable kernel bandwidth, Kalman filter prediction and improved focus tracking of the mean shift algorithm （KPKM）. The target detection algorithm is used to get the bounding rectangle and the center of gravity as the initial mean shift algorithm, a modified Kalman Filter target motion trend prediction, so the algorithm can quickly converge along the gradient direction to the target center. Experimental and simulation results show that this method in complex scenarios, the precision of the moving target detection and accurate tracking.

关 键 词：图像 跟踪 自适应 MEANSHIFT 检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]