The inclined distortion is the crucial problem that influences the procession for the vehicle licenses plate image recognition. However, as a result of the inclined distortion, the vanishing points provide valuable information concerning the camera model used to capture the image. In this paper, a robust and fast algorithm for inclined distortion correction in vehicle licenses plate images based on vanishing points is put forward. This method includes three steps: detecting vanishing points in an image, calibrating camera model coefficients and correcting inclined distortion. Experimental results show this method performs well and increases the recognition rate.
Published in:
Machine Learning and Cybernetics, 2008 International Conference on
(Volume:5
)
Date of Conference: 12-15 July 2008