Abstract:

In accordance with the problem that the inclined Vehicle License Plate (VLP) image brings adverse infection for the subsequent steps of character segmentation, a VLP image correction approach based on factor zoom which is computed by the scale of character height is proposed. Firstly, according to the features of the VLP image, the VLP image is divided into four parts, and every part is detected by Radon transform. Then four equations and lengths of frame in the VLP image are solved, and the lengths, equations and ratios of the corresponding dots of the top frame and the bottom frame are computed. Finally, every dot in the tilt VLP image is corrected according to the proportion. Experiments show that the proposed VLP image correction approach can obtain better ratio than the classic correction method.

Key words: tilt correction;VLP image;Radon transform;“tie point&rdquo, method;zoom factor