甘志英.一种改进的相关滤波目标跟踪算法研究[J].唐山学院学报,2024,37(6):27-33 |
一种改进的相关滤波目标跟踪算法研究 |
Research on an Improved Correlation Filter Target Tracking Algorithm |
投稿时间:2024-08-25 |
DOI:10.16160/j.cnki.tsxyxb.2024.06.005 |
中文关键词: 滤波器 目标跟踪 算法 |
英文关键词: filter target tracking algorithm |
基金项目:河北省高等学校科学技术研究项目(QN2022186) |
|
摘要点击次数: 849 |
全文下载次数: 935 |
中文摘要: |
针对目标跟踪中的目标运动模糊、光照变化、尺度变化等多种因素影响所造成的跟踪目标丢失的问题,采用复合相关滤波器实现目标跟踪。利用位置相关滤波器估计目标位置;利用尺度相关滤波器估计目标尺度;在线更新两种滤波器以适应跟踪中的目标位置、表征变化。通过OTB2015数据集对算法进行验证,该算法准确度达到74.5%,成功率达到63%,跟踪速度达到165 fps。实验表明,该算法可对目标运动模糊、光照变化、尺度变化序列中的目标进行有效跟踪,在多个指标上优于其他多个比较算法。 |
英文摘要: |
In order to solve the problem of target loss in the tracking caused by various factors such as target motion blur, illumination change and scale change, a composite correlation filter is used to track the target, the position correlation filter being used to estimate the target position and the scale correlation filter being used to estimate the target scale; and both filters are updated online to adapt to the changes in target position and representation during tracking. The algorithm is validated with the OTB2015 dataset, with an accuracy rate of 74.5%, a success rate of 63% and a tracking speed of 165 fps. Experimental results show that the algorithm can effectively track targets in sequences with motion blur, illumination change and scale change, outperforming several comparative algorithms in many indicators. |
查看全文
查看/发表评论 下载PDF阅读器 |
关闭 |
|
|
|