VEHICLE DETECTION AND TRAFFIC ASSESSMENT USING IMAGES

The work deals with automatic vehicle detection and classification based on image features . The real images of vehicles mainly 2 wheeler and 4 wheeler are captured using a digital camera with 10 MP. The image database includes 2 wheeler and 4 wheeler separately. The images are enhanced by resizing. Then region of interest (vehicle) is obtained to count. Mainly edge descriptors (canny) is applied to obtain vehicle contour/edges . The extracted vehicle region is cropped and subjected to feature extraction. The PGH (pair wise geometrical histogram ) and edge features are used to represent the model of vehicle type. The PGH is a powerful shape descriptor which is used for polygonal shape. It also can applied to an irregular shape. Then these features are trained with neural network (NN) . The test image is also processed in similar steps. Then the vehicle is classified as 2 or 4 wheeler based on input and trained features. The percentage of accuracy is reported. Later based on the region count and some predefined knowledge about traffic surveillance the traffic is assessed. Hence the work deals with automatic vehicle detection, classification and traffic surveillance in restricted vehicle count. The work can be enhanced by including sound and video surveillance and better feature extraction techniques.

[1]  Ying Wei,et al.  A New Approach for Vision-based Rear Vehicle Tracking , 2007, 2007 2nd IEEE Conference on Industrial Electronics and Applications.

[2]  Wayne A Sarasua,et al.  Vehicle Segmentation and Tracking in the Presence of Occlusions , 2006 .

[3]  Jitendra Malik,et al.  Towards robust automatic traffic scene analysis in real-time , 1994, Proceedings of 12th International Conference on Pattern Recognition.

[4]  Hironobu Fujiyoshi,et al.  Moving target classification and tracking from real-time video , 1998, Proceedings Fourth IEEE Workshop on Applications of Computer Vision. WACV'98 (Cat. No.98EX201).

[5]  Shohreh Kasaei,et al.  A novel vehicle tracking method with occlusion handling using longest common substring of chain-codes , 2009, 2009 14th International CSI Computer Conference.

[6]  Jinhui Lan,et al.  Vehicle detection and recognition based on a MEMS magnetic sensor , 2009, 2009 4th IEEE International Conference on Nano/Micro Engineered and Molecular Systems.

[7]  Jitendra Malik,et al.  A real-time computer vision system for measuring traffic parameters , 1997, Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[8]  Huan Shen,et al.  A Robust Vehicle Tracking Approach Using Mean Shift Procedure , 2009, 2009 Fifth International Conference on Information Assurance and Security.