Yazhe Tang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Yazhe Tang has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Feng Lin
College of Computer Science, Sichuan University, Chengdu 610065, China
2 shared publications - Zhi Gao
School of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430079, China
2 shared publications - Bharath Ramesh
National University of Singapore
1 shared publication - Fang Liao
Temasek Laboratories, National University of Singapore, Singapore 117508
1 shared publication - Fei Wen
School of Remote Sensing and Information Engineering, Wuhan University, Wuhan 430079, China
1 shared publication Hailong Qin
Temasek Laboratories, National University of Singapore, Singapore 117411
1 shared publication
Latest Publications
- 1 August 2018
- Published by Elsevier in Journal of Visual Communication and Image Representation
- Vol. 55, 253-262
Visual adaptive tracking for monocular omnidirectional camera
- 1 August 2018
- Published by Elsevier in Journal of Visual Communication and Image Representation
- Vol. 55, 253-262
Visual adaptive tracking for monocular omnidirectional camera
- 9 April 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Electronics
- Vol. 66 (1) , 616-626
Vision-Aided Multi-UAV Autonomous Flocking in GPS-Denied Environment
This paper presents a sophisticated vision-aided flocking system for unmanned aerial vehicles (UAVs), which is able to operate in GPS-denied unknown environments to realize exploring and searching missions, and also able to adopt two types of vision sensors: day and thermal cameras, to measure relative motion between UAVs in different illumination conditions without using wireless communication. In order to realize robust vision-aided flocking, an integrated framework of tracking-learning-detection on the basis of multi-feature coded correlation filter has been developed. To achieve the long term tracking, a re-detector is trained on-line to adaptively re-initialize target for global sensing. An advanced flocking strategy is developed to address the autonomous multi-UAVs cooperative flight. LiDAR-based navigation modules are developed for autonomous localization, mapping and obstacle avoidance. Flight experiments of a team of UAVs have been conducted to verify the performance of this flocking system in GPS-denied environment. The extensive experiments validate the robustness of proposed vision algorithms in challenging scenarios.
- 9 April 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Electronics
- Vol. 66 (1) , 616-626
Vision-Aided Multi-UAV Autonomous Flocking in GPS-Denied Environment
This paper presents a sophisticated vision-aided flocking system for unmanned aerial vehicles (UAVs), which is able to operate in GPS-denied unknown environments to realize exploring and searching missions, and also able to adopt two types of vision sensors: day and thermal cameras, to measure relative motion between UAVs in different illumination conditions without using wireless communication. In order to realize robust vision-aided flocking, an integrated framework of tracking-learning-detection on the basis of multi-feature coded correlation filter has been developed. To achieve the long term tracking, a re-detector is trained on-line to adaptively re-initialize target for global sensing. An advanced flocking strategy is developed to address the autonomous multi-UAVs cooperative flight. LiDAR-based navigation modules are developed for autonomous localization, mapping and obstacle avoidance. Flight experiments of a team of UAVs have been conducted to verify the performance of this flocking system in GPS-denied environment. The extensive experiments validate the robustness of proposed vision algorithms in challenging scenarios.
- 6 December 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 19 (8) , 2675-2685
Synergizing Appearance and Motion With Low Rank Representation for Vehicle Counting and Traffic Flow Analysis
Appearance and motion, which are complementary, account for a dominant proportion of visual information. We propose to synergize them using a low-rank representation framework for the estimation and analysis of traffic flow. Taking advantage of the downward-looking camera configuration, we do the processing only on the measure line, called virtual gantry, instead of dealing with the whole frame, resulting in much improved efficiency. Enforcing the low-rank constraint on the spatiotemporal image which is generated via stacking pixels on virtual gantry over time, we introduce the block-sparse robust principal component analysis algorithm, in which the motion cue is leveraged to highlight the foreground and realize vehicle detection with high accuracy. The motion flow is further exploited for size normalization to classify vehicles into lite, small, medium, and large categories. Benefiting from the low-rank representation, our method is parameter insensitive, robust to illumination changes, and requires no training. We perform extensive experiments on the 24/7 videos collected over the highways in China and Singapore, obtaining nearly 100% accuracy. Meanwhile, insightful observations on the obtained traffic information are given, which could be very valuable to the users, especially to the traffic management sectors.
- 6 December 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Transportation Systems
- Vol. 19 (8) , 2675-2685
Synergizing Appearance and Motion With Low Rank Representation for Vehicle Counting and Traffic Flow Analysis
Appearance and motion, which are complementary, account for a dominant proportion of visual information. We propose to synergize them using a low-rank representation framework for the estimation and analysis of traffic flow. Taking advantage of the downward-looking camera configuration, we do the processing only on the measure line, called virtual gantry, instead of dealing with the whole frame, resulting in much improved efficiency. Enforcing the low-rank constraint on the spatiotemporal image which is generated via stacking pixels on virtual gantry over time, we introduce the block-sparse robust principal component analysis algorithm, in which the motion cue is leveraged to highlight the foreground and realize vehicle detection with high accuracy. The motion flow is further exploited for size normalization to classify vehicles into lite, small, medium, and large categories. Benefiting from the low-rank representation, our method is parameter insensitive, robust to illumination changes, and requires no training. We perform extensive experiments on the 24/7 videos collected over the highways in China and Singapore, obtaining nearly 100% accuracy. Meanwhile, insightful observations on the obtained traffic information are given, which could be very valuable to the users, especially to the traffic management sectors.