Vehicle Detection and Tracking by Localizing Rear Lamps and License Plate [ ]


Automatic vehicle detection and tracking is an essential processing system for an intelligent transport applications. However, it is posed to great challenges such as landscape variations, vehicle speed, dimension, color, etc. This paper introduces a methodology to detect and track the rear view of vehicles from videos. The methodology adopts two different color space models to detect rear lamps and rear license plates, respectively. Kalman filter is used to estimate and track the moving object and further to aid in detecting and tracking the vehicles. The relationship between the locations of the rear lamps and license plate are used to construct the Markov model, which detects and tracks the vehicle based on the kalman filter output. The experimental investigation shows the methodology maintains minimum error on detecting and tracking the vehicles. The developed user interface is believed to be useful for applying the system comfortably.