*Head, Department of CSE, Government College of Engineering, India
**Assistant Professor, Department of CSE, Tagore Institute of Engineering and Technology, Deviakkurichi, India
***Director, School of Civil Engineering, Karunya University, Coimbatore, India
Online published on 12 January, 2017.
Major challenges observed in traffic in road network is the issues related to efficient vehicle tracking at different dimension at different time intervals. Certain research works conducted on vehicle traffic control concentrated on localizing and road vehicle recognition based on the position and orientation of vehicle image data at single junction point. However, when the shape and posture of vehicle varies at different junction coordination, the overall traffic model based on localization and recognition rises to abnormality. Due to this, a more global feature like exact position, location of the vehicle and angular view of the camera to extract the entire road for identifying the vehicle densities and postures is required. Aerial images extracted using the ray traced templates are used for vehicle tracking and is carried out for different traffic densities. Moreover, additional features of the vehicle are obtained clearly based on the location, position, angle and height using camera fixed on the junction board based on the improvised gradient model. Furthermore, localization is performed using the time of action of the vehicle object under consideration. The effectiveness of RTT-GM is measured using improved gradient model and time of action as it travels between the source and the destination in terms of Vehicle shape and pose recovery and localization accuracy.
Vehicle object recognition, object localization, improved gradient model, ray traced templates, road extraction
