运动图像目标跟踪优化仿真

摘要

对运动图像目标进行准确跟踪,可以缩小目标的搜索范围,提高运动图像处理的效果.运动在抖动状态下,运动图像的轮廓线出现波动.传统的跟踪标记方法,主要通过提取设定的几点轮廓线轨迹点进行跟踪标记,忽略了出现大量波动时对跟踪标记的影响,导致运动图像目标跟踪效率低、误差大的问题.提出基于概率松弛的抖动状态下运动图像目标跟踪优化方法.利用高斯平滑滤波器对抖动状态下的运动图像进行预处理,融合统计背景模型对抖动状态下的运动目标进行检查,标记出目标区域运动图像轮廓线矩形的两个对角顶点,将该顶点当作跟踪的特征,利用概率松弛理论计算出前景抖动图像轮廓线最优标记,识别出当前帧中的多个抖动状态下运动图像目标区域,有效地完成了对抖动状态下运动图像跟踪标记.仿真结果表明,所提方法在抖动状态下可以精确地标记出图像中前景目标,大幅度提升了目标跟踪的稳定性.%In this paper,we propose an optimization method for target tracking of moving image under shaking state based on probabilistic relaxation.Firstly,a Gaussian-smoothing filter was used to carry out pretreatment for moving image under shaking state,and the model of statistical background was integrated to examine the moving target under shaking state.Then,two diagonal vertexes of contour line rectangle of moving image in object region were marked out and the vertexes were used as tracking feature to work out optimal mark of contour line of prospect shaking image.Moreover,the object area of moving image under multiple shaking states in current frame was recognized.The conclusion can be drawn from simulation results that the method mentioned above can mark the prospect target in image under shaking state accurately and improve the stability of target tracking.