摘要
为兼顾目标跟踪的准确度和目标跟踪时图像的处理速度,文章提出一种基于运动粒子的粒子群目标跟踪算法。该算法首先对目标跟踪区域提取HSV特征得到目标特征向量;然后以高斯分布的形式撒下n个粒子构成粒子群,通过梯度收敛算法可以快速准确地搜索到最佳的目标位置,并以此位置作为跟踪点,进行下一帧跟踪。实验结果表明,在精确度方面文章提出的算法是CSK算法的1.52倍、MS算法的1.57倍,在速度方面文章提出的算法是Struck算法的22.87倍、KCF算法的1.11倍,有效地兼顾了目标跟踪的准确度和处理速度。
In order to take into account the accuracy of target tracking and the image processing speed of target tracking,a particle swarm target tracking algorithm based on moving particles is proposed in this paper.Firstly,the HSV feature is extracted from the target tracking region to get the target feature vector,and then n particles are scattered in the form of Gaussian distribution to form a particle swarm,and the best target position can be searched quickly and accurately by the gradient convergence algorithm,and this position is used as the tracking point for the next frame tracking.The experimental results show that the accuracy of the proposed algorithm is 1.52 times that of the CSK algorithm and 1.57 times that of the MS algorithm,and the speed of the proposed algorithm is 22.87 times that of the Struck algorithm and 1.11 times that of the KCF algorithm,which effectively balances the accuracy and processing speed of target tracking.
出处
《科技创新与应用》
2021年第25期10-15,共6页
Technology Innovation and Application
关键词
机器视觉
目标跟踪
HSV特征
粒子群
梯度收敛算法
machine vision
target tracking
HSV feature
particle swarm
gradient convergence algorithm