TRACKING-BASED MOVING OBJECT DETECTION

被引：0

作者：

Shen, Hao ^{[1
]}

Li, Shuxiao ^{[1
]}

Zhang, Jinglan ^{[2
]}

Chang, Hongxing ^{[1
]}

机构：

[1] Chinese Acad Sci, Inst Automat, Beijing, Peoples R China

[2] Queensland Univ Technol, Brisbane, Qld 4001, Australia

来源：

2013 20TH IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP 2013) | 2013年

关键词：

moving object detection; aerial video; tracking; real-time video processing; detection-by-tracking;

D O I：

暂无

中图分类号：

TB8 [摄影技术];

学科分类号：

0804 ;

摘要：

We present a novel approach for multi-object detection in aerial videos based on tracking. The proposed method mainly involves four steps. Firstly, both the motion history image and the tracking trajectory are employed to extract candidate target regions. Secondly, the spatial-temporal saliency is used to detect moving objects in the candidate regions. Thirdly, the previous detected objects are tracked by mean shift in the current frame. And finally, the detection results are fused with the tracking results to get refined detection results, in turn the modified detection results are used to update the tracking models. The proposed algorithm is evaluated on VIVID aerial videos, and the results show that our approach can reliably detect moving objects even in challenging situations. Meanwhile, the proposed method can process videos in real time, without the effect of time delay.

引用

页码：3093 / 3097

页数：5