A Novel Video Detection System on Traffic Flow Inspection

Article Preview

Abstract:

A prototype of video detection system applied to traffic flow inspection is developed, which uses CMOS linear image sensor with high resolution 2K pixels and wide dynamic range as the core of imaging device. It combines FPGA with DSP as the core of acquisition and processing of massive image data. Moreover, a novel multiscale and hierarchical clustering algorithm for image segmentation is presented. Based on the theory of graph spectral, the algorithm can construct a new graph by analyzing the feature of an original image at different clustering scales, so that image segmentation can be accomplished easily to segment the image. The simulation results show that the row scan speed of this system can reach to 1000 lines per second, the resolution being 2048 pixels.

You might also be interested in these eBooks

Info:

Periodical:

Pages:

440-444

Citation:

Online since:

June 2012

Export:

Price:

Permissions CCC:

Permissions PLS:

Сopyright:

© 2012 Trans Tech Publications Ltd. All Rights Reserved

Share:

Citation:

[1] Jiangtao Xu, Binqiao Li, and Suying Yao: Journal of Tianjin University, Vol. 42(2009), p.149.

Google Scholar

[2] F.R.K. Chung, Spectral Graph Theory. Am. Math. Soc, (1997).

Google Scholar

[3] Shengcai Zhang, and Chuan Gong: Journal of Harbin Institute of Technology, Vol. 40(2008), p.1148.

Google Scholar

[4] CMOS1421 2048x1 Linear Image Sensor Data Sheet, http: /www. fairchildimaging. com/main/ documents/CMOS1421DataSheetRevC. pdf.

Google Scholar

[5] Jianke Li: Automation & Instrumentation, Vol. 2(2009), p.19.

Google Scholar

[6] Shi J, and Malik J.: IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol22 (2000), p.888.

Google Scholar

[7] Golub G. H, and Van Loan C. F: Matrix Computations, Johns Hopkins Univ. Press Baltimore, (1989).

Google Scholar

[8] Michael Maire, European Conference on Computer Vision (ECCV), Vol. 6312(2010), p.450.

Google Scholar

[9] Zhanwen liu, Development and Research of The Traffic Flow Detection system Based On CMOS Line Image Sensor, Chang'an University, (2009).

Google Scholar