Research of a Real-time Feature Point Tracking Method Based on the Combination of Improved SURF and P-KLT Algorithm

CAI Jia; HUANG Panfeng

doi:10.7527/S1000-6893.2013.0206

ACTA AERONAUTICAET ASTRONAUTICA SINICA >

2013 , Vol. 34 >Issue 5: 1204 - 1214

DOI: https://doi.org/10.7527/S1000-6893.2013.0206

Electronics and Control

Research of a Real-time Feature Point Tracking Method Based on the Combination of Improved SURF and P-KLT Algorithm

CAI Jia ,
HUANG Panfeng

Expand

1. National Key Laboratory of Aerospace Flight Dynamics, Northwestern Polytechnical University, Xi'an 710072, China;
2. Research Center for Intelligent Robotics, Northwestern Polytechnical University, Xi'an 710072, China

Received date: 2012-06-13

Revised date: 2012-08-29

Online published: 2012-09-05

Supported by

National Natural Science Foundation of China (61005062,11272256)

Fold

Abstract

In order to track moving targets of image sequences in real time, a novel feature point tracking algorithm is proposed by using a combination of improved speeded up robust features (SURF) algorithm and pyramid kanade-lucas-tomasi (P-KLT) matching algorithm. First, the target box is marked manually and improved SURF algorithm is applied to extract features which are robust and distinctive in different blocks. Then, the features are tracked stably by the hierarchically iterative matching of P-KLT algorithm and mismatched points are eliminated utilizing a statistical method. Finally, the exact location of the target is obtained with the application of extracting the target contour by the Greedy Snake algorithm and the target box is updated automatically. Furthermore, discrete feature filter and adaptive feature updating strategy are designed to improve the robustness. Simulation results show that the algorithm can adapt to objective changes in attitude and size and track stably aerial vehicles with simple structures. Time consumption per frame is 31.8 ms, which is 47.1% less than SIFT+P-KLT algorithm. Geometric center error per frame is 5.03 pixel, which is 27.2% less than GFTT+P-KLT tracking algorithm and 38.6% less than SIFT+P-KLT tracking algorithm. Contour area error per frame is 16.3%, which is 56.9% less than GFTT+P-KLT tracking algorithm and 68.4% less than SIFT+P-KLT tracking algorithm.

Key words： feature extraction; SURF algorithm; KLT algorithm; target tracking; Greedy Snake algorithm

Cite this article

CAI Jia , HUANG Panfeng . Research of a Real-time Feature Point Tracking Method Based on the Combination of Improved SURF and P-KLT Algorithm[J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2013 , 34(5) : 1204 -1214 . DOI: 10.7527/S1000-6893.2013.0206

References

[1] Zhou H Y, Yuan Y, Zhang Y, et al. Non-rigid object tracking in complex scenes. Pattern Recognition Letters, 2009, 30(2): 98-102.
[2] Sivic J, Schaffalitzky F, Zisserman A. Object level grouping for video shots. The 8th European Conference on Computer Vision (ECCV), 2004.
[3] Lowe D G. Object recognition from local scale-invariant features. International Conference on Computer Vision, 1999.
[4] Bay H, Tuytelaars T, Gool L V. SURF: speeded(up robust features. The 9th European Conference on Computer Vision, 2006.
[5] Takacs G, Chandrasekhar V, Tsai S, et al. Unified real-time tracking and recognition with rotation-invariant fast features. IEEE Computer Conference on Computer Vision and Pattern Recognition (CVPR), 2010.
[6] Calonder M, Lepetit V, Strecha C, et al. BRIEF: binary robust independent elementary features. The 11th European Conference on Computer Vision, 2010.
[7] Ethan R, Vincent R, Kurt K, et al. ORB: an efficient alternative to SIFT or SURF. 2011 International Conference on Computer Vision (ICCV), 2011.
[8] Khvedchenya E. Feature descriptor comparison report. (2011-08-19)[2012-04-20]. http://computer-vision-talks.com/.
[9] Yilmaz A, Javed O, Shah M. Object tracking: a survey. ACM Computing Surveys, 2006, 38(4): 45.
[10] Lucas B, Kanade T. An iterative image registration technique with an application to stereo vision. International Joint Conference on Artificial Intelligence, 1981.
[11] Tomasi C, Kanade T. Detection and tracking of point features. Carnegie Mellon University Technical Report CMU-CS-91-132, 1991.
[12] Mian A S. Realtime visual tracking of aircrafts. Procee-dings of the 2008 Digital Image Computing: Techniques and Applications, 2008.
[13] Ben B, Ian R. Stable multi-target tracking in real-time surveillance video. IEEE Computer Conference on Computer Vision and Pattern Recognition, 2011.
[14] Gong J, Liu F Q, Song C L, et al. Research on the moving vehicle detection algorithm based on the motion vector. Advances in Intelligent and Soft Computing, 2012, 127: 41-49.
[15] Liu Y, Wang J D, Li P. A feature point tracking method based on the combination of SIFT algorithm and KLT matching algorithm. Journal of Astronautics, 2011, 32(7): 1618-1625. (in Chinese) 刘玉, 王敬东, 李鹏. 一种基于SIFT和KLT相结合的特征点跟踪方法研究. 宇航学报, 2011, 32(7): 1618-1625.
[16] Wang Y M, Wang G J. Image local invariant features and descriptors. Beijing: National Defense Industry Press, 2010: 89-100. (in Chinese) 王永明, 王贵锦. 图像局部不变性特征与描述. 北京: 国防工业出版社, 2010: 89-100.
[17] Song L H. The study of target positioning technology based on UAV image sequence. Zhengzhou: The Institute of Surveing and Mapping, PLA Information Engineering University, 2011. (in Chinese) 宋丽华. 基于无人飞行器序列影像的定位技术研究. 郑州: 中国人民解放军信息工程大学测绘学院, 2011.
[18] Bradski G, Kaebler A. Learning OpenCV. Yu S Q, Liu R Z, translated. Beijing: Tsinghua University Press, 2009: 362-363. (in Chinese) Bradski G, Kaebler A. 学习OpenCV. 于仕琪, 刘瑞祯, 译. 北京: 清华大学出版社, 2009: 362-363.

Options

Outlines

模态框（Modal）标题

Abstract

Cite this article

References