Shadow license plate image binarization method

Abstract

The invention discloses a shadow license plate image binarization method and belongs to the technical field of intelligent traffic. The method comprises the steps of obtaining a license plate color image subjected to license plate accurate locating; extracting a red channel of the license plate color image to serve as a license plate grayscale image; judging whether the license plate image has a rule shadow or not; if yes, finding a region in which a boundary line of dark and bright regions is located, and obtaining an image binarization result through a regional binarization method; and if not, performing image binarization through a global threshold method to obtain a binarization result. According to the method, the red channel is extracted for performing gray processing on the obtained image, a difference value of grayscale values of a foreground and a background of the image is increased, and the binarization of a shadow license plate is carried out through the regional binarization method, so that an algorithm is simple, and the "artifact" problem can be improved; and different binarization polices are used for the shadow license plate and a general license plate, so that while the binarization result of the shadow license plate is improved, the binarization effect and processing efficiency of the general license plate are not influenced.

Description

technical field [0001] The invention belongs to the technical field of intelligent transportation, in particular to a binarization method for a shadowed license plate image before the license plate characters are cut. Background technique [0002] With the continuous development of machine vision, digital image processing, pattern recognition and other technologies, Intelligent Transportation System (ITS) has gradually become the main way to realize traffic management automation. The license plate recognition technology is one of the key technologies to realize the intelligent transportation system. It realizes the automatic recognition of the license plate number through the method of image processing. Image binarization is a key link in license plate preprocessing, and the clarity of the binarized image is directly related to the accuracy of subsequent license plate character segmentation and character recognition. [0003] There are many existing image binarization metho...

AI Technical Summary

Problems solved by technology

[0005] To sum up, the main shortcomings of the current solution to the problem of uneven illumination of license plate images include: (1) The image enhancement algorithm and image binarization algorithm are more complex, and the efficiency of the method is poor; (2) Although through a certain The image enhancement algorithm can enhance the contrast between the background and the foreground, but there will still be "artifacts" on the dividing line between bright and dark areas

Images