中文摘要：

针对不稳定的关键点对以SIFT（Scale Invariant Feature Transform）为目标特征的视觉跟踪算法的影响,提出基于SIFT排序的视觉跟踪算法。为实现SIFT排序,提出空域稳定因子和时域稳定因子,并由此构成重要性权重,以表征各个特征点的重要程度。在SIFT排序的基础上,各个关键点按照重要性权重的不同参与跟踪,从而实现基于SIFT排序的视觉跟踪。该算法克服了不稳定的关键点对跟踪结果的影响,从而提高跟踪的准确性和鲁棒性。

英文摘要：

Aiming at the impact of unstable key points on visual tracking algorithm which uses SIFT as its target feature,we present the rank SIFT-based visual tracking algorithm. To realise rank SIFT,we propose the spatial stability factor and the temporal stability factor and compose the importance weights with them for representing the significance of each key point. Based on rank SIFT,each key point takes part in tracking according to its own importance weight,so that the rank SIFT-based visual tracking is achieved. The algorithm overcomes the impact of unstable key points on tracking outcomes,therefore improves the accuracy and robustness of tracking.