一种基于改进快速差分进化算法的运动目标跟踪方法

A Moving Target Tracking Method Based on Improved Fast Differential Evolution Algorithm

计算机视觉用于检测和追踪图像序列中的特定目标,近年来受到了广泛的关注。跟踪过程可以被描述为一个动态优化问题,在每个图像中识别目标的最佳位置。差分进化算法(DE)具有简单、并行计算和自适应搜索全局优化等优点,被认为是一种很有前途的有效目标跟踪算法。为了更好地适应目标跟踪,在DE中进行了一些改进。

The tracking process can be described as a dynamic optimization problem that identifies the best location of the target in each image.Differential evolution algorithm(DE)has the advantages of simple,parallel computing and adaptive search global optimization,and is considered as a promising and effective target tracking algorithm.In this paper,some improvements is made in DE in order to better adapt to target tracking.