[1]田应仲,刘伊芳,李 龙.联合判别式核相关滤波视觉跟随算法[J].计算机技术与发展,2021,31(03):78-83.[doi:10. 3969 / j. issn. 1673-629X. 2021. 03. 013]
　TIAN Ying-zhong,LIU Yi-fang,LI Long.Visual Following Algorithm of Joint Discriminant Kernel Correlation Filter[J].,2021,31(03):78-83.[doi:10. 3969 / j. issn. 1673-629X. 2021. 03. 013]
点击复制

联合判别式核相关滤波视觉跟随算法()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
31
期数:
2021年03期
页码:
78-83
栏目:
图形与图像
出版日期:
2021-03-10

文章信息/Info

Title:
Visual Following Algorithm of Joint Discriminant Kernel Correlation Filter
文章编号:
1673-629X(2021)03-0078-08
作者:
田应仲1; 2; 刘伊芳1; 2; 李 龙1; 2
1. 上海大学 机电工程与自动化学院,上海 200444;
2. 上海市智能制造及机器人重点实验室,上海 200444
Author(s):
TIAN Ying-zhong1; 2; LIU Yi-fang1; 2; LI Long1; 2
1. School of Mechatronics Engineering & Automation,Shanghai University,Shanghai 200444,China;
2. Shanghai Key Laboratory of Intelligent Manufacturing and Robots,Shanghai 200444,China
关键词:
核相关滤波算法; 视觉跟踪; 特征融合; 背景感知; 干扰判别
Keywords:
kernel correlation filtering algorithm; visual tracking; feature fusion; background perception; interference discrimination
分类号:
TP242
DOI:
10. 3969 / j. issn. 1673-629X. 2021. 03. 013
摘要:
目标跟踪算法是计算机视觉领域的热门技术之一,拥有广阔的发展前景。 核相关滤波视觉跟踪算法由于循环矩阵构造正负训练样本,避免求逆的大量运算,显著提高计算速度而受到广泛关注。 但是,核相关滤波算法存在一定局限性,无法应对现实环境存在的遮挡、目标尺度变化、背景模糊等复杂多变的干扰因素。 因此提出一种改进型核相关滤波算法。 该算法不仅融合多种颜色特征提高图像处理的准确度,而且通过构建自适应尺度变化策略来应对目标尺度变化的挑战。 为了进一步区分目标和背景信息,提出联合判别式背景感知与干扰判别的策略,以充分利用目标上下文信息。 相比于传统核相关滤波算法,改进算法的精度更高,鲁棒性更强。 通过在视频数据集 OTB-50 上的实验可得,改进后的核相关滤波算法性能获得较大提升。
Abstract:
Target tracking algorithm is one of the popular technologies in the field of computer vision and has a broad development prospect. The kernel-correlated filter visual tracking algorithm has been widely concerned because the cyclic matrix constructs positive and negative training samples,which significantly improves the calculation speed. However,the kernel correlation filter algorithm has certain limitations,which cannot cope with the complex and variable interference factors such as occlusion, target scale change and background blurring in the real environment. Therefore, an improved kernel correlation filter algorithm is proposed. The improved algorithm not only combines multiple color features to improve the accuracy of image processing, but also responds to the challenge of target scale change by constructing an adaptive scale change strategy. In order to further distinguish between target and background information,a strategy of joint discriminant background perception and interference discrimination is proposed to make full use of target context inform-ation. Compared with the traditional kernel correlation filter algorithm,the improved algorithm has higher accuracy and stronger robustness. Through experiments on the OTB-50 data set,the performance of the improved kernel correlation filter algorithm has been greatly improved.

相似文献/References:

[1]王业祥,朱文球,孙文静.局部加权距离度量的双向稀疏表示目标跟踪[J].计算机技术与发展,2018,28(04):60.[doi:10.3969/ j. issn.1673-629X.2018.04.013]
　WANG Ye-xiang,ZHU Wen-qiu,SUN Wen-jing.Target Tracking of Bidirectional Sparse Representation of Local Weighted Distance Metric[J].,2018,28(03):60.[doi:10.3969/ j. issn.1673-629X.2018.04.013]
[2]梁文辉,宋 涛,叶永达.间隔贝叶斯估计和空时目标运动边界视觉跟踪[J].计算机技术与发展,2021,31(09):75.[doi:10. 3969 / j. issn. 1673-629X. 2021. 09. 013]
　LIANG Wen-hui,SONG Tao,YE Yong-da.Visual Tracking Based on Interval Bayesian Estimation and Spatial-temporal Object Motion Boundaries[J].,2021,31(03):75.[doi:10. 3969 / j. issn. 1673-629X. 2021. 09. 013]

更新日期/Last Update: 2020-03-10