Moving Target Classification in Road Monitoring Based on Multi-Feature Fusion

Abstract:

A method for moving target classification in road monitoring based on multi-feature fusion is presented in this paper. In this method, connected component labeling and merging combined with morphology are used to achieve the target extraction. Static features in moving target are extracted. To improve the low classification accuracy, a dynamic feature, lower thirds aspect ratio variation (also named as LTVar), is proposed and added. The recognition ratio obtains the relative increasing of 3.1% compared with the static features.

• Periodical: Applied Mechanics and Materials (Volumes 401-403)
• Pages: 1208-1211
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.401-403.1208
• Online since: September 2013
• Authors: Lin Wu, Xiao Pei Wu, Juan Xu
• Keywords: Dynamic Feature, Moving Target Classification, Static Feature, Target Extraction

References

[1] Zhang S, Chan S C, Qiu R D, et al. On the design and implementation of a high definition multi-view intelligent video surveillance system[C]/Signal Processing, Communication and Computing (ICSPCC), 2012 IEEE International Conference on. IEEE, 2012: 353-357.

DOI: 10.1109/icspcc.2012.6335739

[2] Wang X, Yuan D, Li D. Moving Target Classification Technique in Video Sequence[C]/Information Technology, Computer Engineering and Management Sciences (ICM), 2011 International Conference on. IEEE, 2011, 2: 362-365.

DOI: 10.1109/icm.2011.175

[3] Cutler R, Davis L S. Robust Real-time Periodic Motion Detection, Analysis, and Applications [J]. Pattern Analysis and Machine Intelligence, IEEE Transactions on, 2000, 22(8): 781-796.

DOI: 10.1109/34.868681

[4] Xiong Z, Chen K, Gu C, et al. An algorithm of image classification based on BP neural network[C]/Computer Science and Automation Engineering (CSAE), 2012 IEEE International Conference on. IEEE, 2012, 1: 523-526.

DOI: 10.1109/csae.2012.6272651

[5] Information on http: /imagelab. ing. unimore. it/visor.