一种结合均值漂移的KCF目标跟踪算法
|
摘要
目标跟踪在机器视觉领域中发挥着巨大的作用,但是目标跟踪算法会受到目标尺度变化等因素的影响,导致算法的跟踪精度下降。针对KCF滤波算法对目标尺度变化不敏感,无法较好实现目标尺度估计的缺点,提出一种结合均值漂移的KCF目标跟踪算法。通过对KCF滤波算法的响应峰值进行判断,引入均值偏移算法对目标位置和尺度进行修正。实验结果表明,本算法在增强了算法目标尺度自适应能力的同时,能有效地融合均值漂移算法和KCF滤波算法的优点,在多种场景下的目标跟踪精度都得到了有效提升。
Target tracking plays an important role in the field of machine vision.However,the accuracy of the target tracking will be reduced by factors such as changes in the target scale.Aiming at the shortcomings of KCF filtering algorithm which is not sensitive to target scale change,this paper proposes a KCF target tracking algorithm combined with mean shift.The experimental results show that the proposed algorithm can effectively combine the advantages of the mean shift algorithm and the KCF filter algorithm while enhancing the target size estimation ability of the algorithm.The target tracking accuracy is improved in various scenarios.
Target tracking plays an important role in the field of machine vision.However,the accuracy of the target tracking will be reduced by factors such as changes in the target scale.Aiming at the shortcomings of KCF filtering algorithm which is not sensitive to target scale change,this paper proposes a KCF target tracking algorithm combined with mean shift.The experimental results show that the proposed algorithm can effectively combine the advantages of the mean shift algorithm and the KCF filter algorithm while enhancing the target size estimation ability of the algorithm.The target tracking accuracy is improved in various scenarios.
引文
[1]张娟,毛晓波,陈铁军.运动目标跟踪算法研究综述[J].计算机应用研究,2009,26(12):4407-4410.
[2]崔丹丹,吕林,方位达.无人机遥感技术在江苏海域和海岛动态监视监测中的应用研究[J].现代测绘,2013,36(06):10-11.
[3]COMAINICIU D,MEER P.Mean shift:a robust approach toward feature space analysis[J].IEEE Trans Pattern Analysis&Machine Intelligence,2002,24(5):603-619.
[4]HENRIQUES J F,CASEIRO R,MARTINS P,et al.High-Speed Tracking with Kernelized Correlation Filters[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2015,37(3):583-596.
[5]李远状,韩彦芳,于书盼.一种核相关滤波器的多尺度目标跟踪方法[J].电子科技,2018,31(10):5-9.
[2]崔丹丹,吕林,方位达.无人机遥感技术在江苏海域和海岛动态监视监测中的应用研究[J].现代测绘,2013,36(06):10-11.
[3]COMAINICIU D,MEER P.Mean shift:a robust approach toward feature space analysis[J].IEEE Trans Pattern Analysis&Machine Intelligence,2002,24(5):603-619.
[4]HENRIQUES J F,CASEIRO R,MARTINS P,et al.High-Speed Tracking with Kernelized Correlation Filters[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2015,37(3):583-596.
[5]李远状,韩彦芳,于书盼.一种核相关滤波器的多尺度目标跟踪方法[J].电子科技,2018,31(10):5-9.