Canh Nguyen and Amar Raheja
Edge detection, license plate localization, morphology
This paper presents a novel morphological and gradient difference based approach for locating vehicle license plates in natural background images. The proposed algorithm uses a gradient difference method on grayscale license plate images (converted from color) to detect vertical edges selectively and fully fill up gaps between two close enough edges and binarize the image at the same time. A series of morphological operations are exploited to eliminate all background. The method was tested with 148 vehicle images with natural background, different illumination intensities, different kinds of license plates, and different camera distances/angles. 97.8% of the images were uniquely located the license plate and the license plates of the other images were also be located beside one or two other regions.
Important Links:
Go Back
