License plate binaryzation method based on compensation principle and center section scanning

Abstract

The invention provides a two-valued method for license plate based on compensation principle and center region scanning. Firstly, the license plate region is converted through grey level stretching to realize reinforcement of the character region on the license plate; second the grey levels of the character pixel element and the background pixel element g1 and g2 of the reinforced license plate are figured out through a histogram method; thirdly the mean and the standard deviation of the reinforced license plate image is calculated and an illumination compensation factor is also obtained according to the mean and the standard deviation; fourthly an illumination penalty function is accordingly calculated based on the illumination compensation factor; fifthly a two-valued threshold value is consequently obtained according to the illumination penalty function, the standard deviation and the mean; and finally the type of the license plate can be judged through a center region scanning method with normalization treatment. Compared with the prior global dynamic thresholding method, the method of the invention has the advantages of not only quick two-valued process, but also high robustness and realization of complete separate between background and characters.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a license plate binarization method in a complex background in the license plate recognition technology. Background technique [0002] With my country's accession to WTO and the continuous development of transportation, the promotion of Intelligent Traffic System (ITS for short) is an inevitable trend. As an important part of ITS, the rapid and accurate realization of License Plate Recognition (LPR) plays a very important role in the intelligentization of traffic management and public security punishment. Since the 1990s, people have begun to study the automatic recognition of license plates. The main way is to use computer image processing technology to analyze the license plate image, automatically extract the license plate information, and determine the license plate number. At present, the theory of license plate recognition has become more and more matu...

AI Technical Summary

Problems solved by technology

[0010] The common point of the above-mentioned several license plate binarization methods is that these methods have certain conditions and restrictions, so they are easily affected by factors such as weather, background, lighting, etc., and their robustness is not good.

Once the conditions change, their binarization accuracy will fluctuate greatly, so the performance of the entire license plate recognition system will be greatly reduced

Images