A Scalable FPGA Vehicle Monitoring and Classification Architecture

In this paper, we propose an FPGA-based vehicle monitoring and classification architecture. The proposed architecture is scalable and allows multi-lane concurrent processing. Furthermore, the system presented utilizes much less FPGA area while achieving real-time and high recognition accuracy. By utilizing an adaptive background subtraction method, we have produced accurate segmentation for varying conditions where other methods may fall short. Our experimental results show that an accuracy of 93% is feasible for most applications while each lane processing unit occupies only 13% of a Virtex-4 FPGA's slices. Furthermore, a simple post processing architecture is proposed for further improving the accuracy of the segmentation unit. Because of its small footprint, the system is suitable for portable applications such as a distributed traffic monitoring system.

[1]  Markos Papadonikolakis,et al.  A novel FPGA-based SVM classifier , 2010, 2010 International Conference on Field-Programmable Technology.

[2]  Marek Gorgon,et al.  FPGA-based Road Traffic Videodetector , 2007, 10th Euromicro Conference on Digital System Design Architectures, Methods and Tools (DSD 2007).

[3]  Shigang Wang,et al.  Video object segmentation based on frame differences and its implementation on DSP , 2008 .

[4]  Thierry Bouwmans,et al.  Background Modeling using Mixture of Gaussians for Foreground Detection - A Survey , 2008 .

[5]  Yi Jiang,et al.  Dynamic content based vehicle tracking and traffic monitoring system , 2007, Electronic Imaging.

[6]  Andrew Hunter,et al.  A single-chip FPGA implementation of real-time adaptive background model , 2005, Proceedings. 2005 IEEE International Conference on Field-Programmable Technology, 2005..

[7]  Chung-Cheng Chiu,et al.  A Robust Object Segmentation System Using a Probability-Based Background Extraction Algorithm , 2010, IEEE Transactions on Circuits and Systems for Video Technology.

[8]  Jiang-Zhong Cao,et al.  A novel online fingerprint segmentation method based on frame-difference , 2009, 2009 International Conference on Image Analysis and Signal Processing.

[9]  Hong Huo,et al.  Traffic Video Segmentation Using Adaptive-K Gaussian Mixture Model , 2006, IWICPAS.

[10]  Borko Furht,et al.  Neural Network Approach to Background Modeling for Video Object Segmentation , 2007, IEEE Transactions on Neural Networks.

[11]  M.P.T. Juvonen,et al.  Hardware Architectures for Adaptive Background Modelling , 2007, 2007 3rd Southern Conference on Programmable Logic.

[12]  Abbes Amira,et al.  An Efficient FPGA Implementation of Gaussian Mixture Models-Based Classifier Using Distributed Arithmetic , 2006, 2006 13th IEEE International Conference on Electronics, Circuits and Systems.

[13]  W. Eric L. Grimson,et al.  Adaptive background mixture models for real-time tracking , 1999, Proceedings. 1999 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No PR00149).

[14]  Viktor Öwall,et al.  Hardware accelerator design for video segmentation with multi-modal background modelling , 2005, 2005 IEEE International Symposium on Circuits and Systems.