2013 | OriginalPaper | Buchkapitel
Automatic Object Tracking in Aerial Videos via Spatial-temporal Feature Clustering
verfasst von : Xiaomin Tong, Yanning Zhang, Tao Yang, Wenguang Ma
Erschienen in: Intelligence Science and Big Data Engineering
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Automatic detecting and tracking the objects from UAV videos is very important and challenging for both tactical and security applications. We present a robust object tracking system that is able to track multiple objects robustly in UAV videos. The main characteristics of the proposed system include: (1)A novel feature clustering based multiple objects tracking framework is proposed, which performs much better than the traditional foreground-blob-tracking-based methods. (2)Optical flow features are clustered both in spatial and temporal dimension to track multiple objects robustly even in the case of multiple objects cross moving. Extensive experimental results with quantitative and qualitative analysis demonstrate the robustness and effectiveness of our algorithm.