Mono-camera based vehicle detection using effective candidate generation

This paper proposes a novel on-road vehicle detection algorithm for Advanced Driver Assistance Systems. The proposed algorithm uses shadow and edge information of original image to generate the candidate vehicles. The proposed algorithm extracts the shadow region darker than the road, and creates an edge image using the Canny edge detector. It performs AND operation between the shadow image and the edge image. The candidate vehicles are verified by SVM classifier based on the HOG feature vector. The experimental results show that the proposed algorithm provides fast and accurate vehicle detection. And the number of candidates is reasonable for the vehicle verification.

[1]  Shengcai Liao,et al.  Learning Multi-scale Block Local Binary Patterns for Face Recognition , 2007, ICB.

[2]  Y. Tamatsu,et al.  Solid or not solid: vision for radar target validation , 2004, IEEE Intelligent Vehicles Symposium, 2004.

[3]  Alberto Broggi,et al.  Vehicle and Guard Rail Detection Using Radar and Vision Data Fusion , 2007, IEEE Transactions on Intelligent Transportation Systems.

[4]  Zehang Sun,et al.  On-road vehicle detection: a review , 2006, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[5]  Zehang Sun,et al.  A real-time precrash vehicle detection system , 2002, Sixth IEEE Workshop on Applications of Computer Vision, 2002. (WACV 2002). Proceedings..

[6]  Martin Schneider,et al.  Automotive Radar – Status and Trends , 2005 .

[7]  Bill Triggs,et al.  Histograms of oriented gradients for human detection , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[8]  Bin Dai,et al.  A Vehicle Detection Method via Symmetry in Multi-Scale Windows , 2007, 2007 2nd IEEE Conference on Industrial Electronics and Applications.

[9]  Paul A. Viola,et al.  Rapid object detection using a boosted cascade of simple features , 2001, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. CVPR 2001.