A novel approach for license plate location in natural images

Author

Hongyao, Deng ; Xiuli, Song

Author_Institution

Dept. of Comput. Sci., Yangtze Normal Univ., Chongqing, China

fYear

2009

fDate

4-6 Aug. 2009

Firstpage

563

Lastpage

568

Abstract

Extraction of the license plate region is challenging first step in the license plate recognition system. The paper proposed approach is not only simple but also more effective than some of the existing method earlier. A new dynamic RGB threshold formula by limiting the hue change range is educed. The license image is made grids. Then we can plot and mark them by the formula and operate them with mathematical morphology because of precision-cutting the license knowledge. More than five hundred images are tested, none of the results manifested missed. Only 1 candidate-absented plate had been got. Plates were extracted and a 99% rate was achieved.

Keywords

image colour analysis; image recognition; traffic engineering computing; candidate-absented plate; dynamic RGB threshold formula; hue change range; license image; license knowledge; license plate location; license plate recognition system; license plate region; mathematical morphology; natural images; precision cutting; Character recognition; Color; Computer science; Image processing; Image recognition; Image segmentation; Licenses; Morphology; Neural networks; Vehicle dynamics;

fLanguage

English

Publisher

ieee

Conference_Titel

Applications of Digital Information and Web Technologies, 2009. ICADIWT '09. Second International Conference on the

Conference_Location

London

Print_ISBN

978-1-4244-4456-4

Electronic_ISBN

978-1-4244-4457-1

Type

conf

DOI

10.1109/ICADIWT.2009.5273920

Filename

5273920