Watch Out: Embedded Video Tracking with BST for Unmanned Aerial Vehicles

被引:6
作者
Battistone, Francesco [1 ]
Petrosino, Alfredo [1 ]
Santopietro, Vincenzo [1 ]
机构
[1] Univ Naples Parthenope, Dept Sci & Technol, Naples, Italy
来源
JOURNAL OF SIGNAL PROCESSING SYSTEMS FOR SIGNAL IMAGE AND VIDEO TECHNOLOGY | 2018年 / 90卷 / 06期
关键词
Tracking; Online learning; Outlier detection; Drone; Jetson; GPU;
D O I
10.1007/s11265-017-1279-x
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
The paper presents the development of a real time tracking system, named Watch Out, that is able to efficiently run on an Nvidia Jetson board mounted on a UAV (Unmanned Aerial Vehicle). The approach to long term video tracking implemented in Watch Out is named Best Structured Tracker (BST): a set of local trackers independently tracks patches of the original target in an online learning manner, while an outlier detection procedure filters out the less meaningful ones, and a resampling procedure allows to correctly reinitialise the trackers that have been filtered out. Performance of the tracking algorithm has been verified both on VOT2016 challenge datasets and in real situations using an Nvidia Jetson board mounted on a drone. Results show that the proposed system can track almost every possible target in real time.
引用
收藏
页码:891 / 900
页数:10
相关论文
共 41 条
[1]   Deformable part-based tracking by coupled global and local correlation filters [J].
Akin, Osman ;
Erdem, Erkut ;
Erdem, Aykut ;
Mikolajczyk, Krystian .
JOURNAL OF VISUAL COMMUNICATION AND IMAGE REPRESENTATION, 2016, 38 :763-774
[2]  
[Anonymous], 2014, P 2014 INT C 3D IM L
[3]  
[Anonymous], 2016, The Visual Object Tracking VOT2016 challenge results
[4]  
[Anonymous], ARXIV12032386
[5]  
[Anonymous], FLIGHT TEST BED VISU
[6]  
[Anonymous], NVDIA DEV TECHNOLOGY
[7]  
[Anonymous], 2007, GPU gems
[8]  
[Anonymous], CLUST LOC MOT EST
[9]  
[Anonymous], 2016, ARXIV160503720
[10]  
[Anonymous], 2014, REGISTRATION RECOGNI, DOI DOI 10.1007/978-3-642-44907-9_6