Research of Background Modeling Algorithm Method Based on Multi-Frame Average Method in Moving Target Detection

Abstract:

In intelligent video surveillance system, there exists the defect that the background is multimodal when extracting the background image, so the extracted foreground objects are not very pure with the method of background subtraction. In this paper, we put forward an algorithm which adopts adaptive threshold to update the background. Firstly, the common background model algorithms and their defect are analysed. Then we propose the strategy of updating background template based on adaptive threshold and then give the specific algorithm steps. Finally, the algorithm is simulated and the results are analyzed.

• Periodical: Applied Mechanics and Materials (Volumes 380-384)
• Pages: 1390-1393
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.380-384.1390
• Online since: August 2013
• Authors: Xiang Hua Hou, Hong Hai Liu
• Keywords: Background Subtraction, Background Template, Object Detection

References

[1] O. Yamaguchi, K. Fukui, and K. Maeda: Pro. Automatic Face and Gesture Recognition (Nara, Japan, 1998). p.318.

[2] Z.B. Guo: Research on the Algorithms of Fast Face Detection and Feature Extraction (Ph.D. Nanjing University of Science and Technology, China 2007). p.66.

[3] M. Nishiyama, O. Yamaguchi and K. Fukui: Pro. Audio- and Video-Based Biometric Person Authentication (NewYork, USA, 2005). pp.71-78.

[4] J.W. Li,Y. H. Wang and T.N. Tan: Pro. the 5th Chinese Conference on Biometric Recognition (Guangzhou, China, 2004). p.224.

[5] L.H. Zhao: Research and Realization in Face Detection and Recognition Algorithms (Ph.D. Northeastern University, China 2006). p.32.