Priori analysis based iterative method for segmenting characters of licence plate

Abstract

The invention relates to a priori analysis based iterative method for segmenting characters of a licence plate. Method comprises the following steps of: firstly, preprocessing the image of the licence plate, segmenting the preprocessed image of the licence plate by utilizing the vertical projection method, calculating the parameter of each character block, and carrying out preliminary screening according to the parameters; secondly, carrying out binaryzation on the characters processed by preliminary screening by utilizing the threshold value iteration method, calculating the priori knowledgecontained by each character block according to the standard licence plate system, and obtaining the characters processed by secondary screening; and thirdly, calculating the quantity of blocks of thecharacters processed by secondary screening, calculating the width of characters of the licence plate and the mean width of all characters under the condition that set conditions are met, and determining whether ultimate characters can be obtained or not according to the two parameters, or carrying projection again when the set conditions are not met. The method reduces the requirement on positioning accuracy of the licence plate, overcomes the adverse effects due to a complex background, eliminates the common segmentation results of conglutination, breakage and the like of the other methods,and further improves the accuracy of segmentation.

Description

technical field [0001] The invention relates to the field of intelligent transportation, in particular to an iterative license plate character segmentation method based on a priori analysis. Background technique [0002] License plate recognition is an important part of modern intelligent transportation system, and character segmentation is a key link in license plate recognition. Due to the fixed format of license plates, template matching has been considered as the best character segmentation method, such as literature [1] Li Zhanbin. Research on license plate recognition technology in non-stop toll collection system [D]. Qingdao: Ocean University of China, 2005. However, the template matching method has extremely high requirements on the image quality of the car body containing the license plate, so the localization and segmentation accuracy is low. As researchers have carried out a lot of research on the method of license plate character segmentation, according to the c...

AI Technical Summary

Problems solved by technology

This type of segmentation method based on prior knowledge has a certain improvement in segmentation accuracy compared with the methods proposed in literature [1]-[3], but the requirements for the accuracy of license plate location are still high, especially when the license plate is tilted and contains parts. In the case of serious interference such as the background of the car body and uneven illumination, the segmentation effect is poor, which brings great challenges and pressure to the subsequent recognition work

Images