| 基于SIFT特征的目标跟踪算法鲁棒性研究 |
| |
| 引用本文： | 刘宏申,张传文.基于SIFT特征的目标跟踪算法鲁棒性研究[J].网络安全技术与应用,2011(6):9-11. |
| |
| 作者姓名： | 刘宏申 张传文 |
| |
| 作者单位： | 安徽工业大学计算机学院,安徽,243032 |
| |
| 摘 要： | 本文在改进的尺度不变特征转换(Scale-invariant feature transform,SIFT)特征匹配算法的基础上,提出了一种基于特征集的多目标跟踪算法。通过设置目标特征留存优先级,更新特征集,保存近几帧的稳定特征,保证特征的稳定,提高了匹配准确度。试验结果表明该算法对目标由于旋转、形变导致的跟踪性能下降具有较好的容错性,也具有较好的鲁棒性。
|
| 关 键 词： | 多目标跟踪 尺度不变特征变换 非刚性形变 |
Robustness research on Object Tracking Algorithm Based on SIFT Feature |
| |
| Liu Hongshen,Zhang Chuanwen.Robustness research on Object Tracking Algorithm Based on SIFT Feature[J].Net Security Technologies and Application,2011(6):9-11. |
| |
| Authors: | Liu Hongshen Zhang Chuanwen |
| |
| Affiliation: | Liu Hongshen,Zhang Chuanwen School of Computer Science,Anhui University of Technology,Anhui,243032,China |
| |
| Abstract: | Based on the Improved scale invariant feature transform(SIFT),a novel motion-tracking algorithm for multi-targets utilizing the feature reserving priority of preference is proposed.Retained by setting the priority of target features,updated feature set,save the stability characteristics of the last few frames to ensure the stability of features and improve the matching accuracy.The results show that the algorithm on the target due to the rotation and deformation tracking performance degradation have good fa... |
| |
| Keywords: | Multi-target tracking Scale Invariant Feature Transform non-rigid deformation |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
