Binocular Based Moving Target Tracking for Mobile Robot

Authors:
Yingkui Du;Baojie Fan;Jianda Han;Yandong Tang
Affiliations:
State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Science, Shenyang, China 110016 and Graduate School of the Chinese Academy of Science, Beijing, China 100039;State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Science, Shenyang, China 110016 and Graduate School of the Chinese Academy of Science, Beijing, China 100039;State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Science, Shenyang, China 110016;State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Science, Shenyang, China 110016
Venue:
ICIRA '09 Proceedings of the 2nd International Conference on Intelligent Robotics and Applications
Year:
2009

Citing 8
Cited 0

Robust model-based motion tracking through the integration of search and estimation

International Journal of Computer Vision
Efficient Region Tracking With Parametric Models of Geometry and Illumination

IEEE Transactions on Pattern Analysis and Machine Intelligence
Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography

Communications of the ACM
A Taxonomy and Evaluation of Dense Two-Frame Stereo Correspondence Algorithms

International Journal of Computer Vision
Rapid and brief communication: Efficient edge-based object tracking

Pattern Recognition
Robust tracking with motion estimation and local Kernel-based color modeling

Image and Vision Computing
Handling occlusion in optical flow algorithms for object tracking

Computers & Mathematics with Applications
Fast robust correlation

IEEE Transactions on Image Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Moving target tracking is an important application of computer vision. A binocular based method is presented for mobile robot to track target reliably under the effect of occlusion, transform and rotation in unstructured environment. Point features are extracted for representing the target and environment background under middle distortion, and then are matched and tracked through consecutive stereo frames by our improved MNCC algorithm. The point features are reconstructed and utilized to estimate the relative motion by Least-Square algorithm. Because the relative motion between the point features of target and robot is inconsistent to that of environment background and robot, the point features of environment background and the errors in feature tracking are removed by RANSAC algorithm. Experiment results validate the efficiency of our method.