A Character Segmentation Method Based on Coarse Locating License Plate

Abstract

The invention provides a character segmentation method based on rough positioning license plate. This method can automatically judge whether the license plate is complete, and can realize accurate character segmentation on the basis of complete license plate. Firstly, select the circumscribed rectangle of the best independent connected area in the binary license plate image that can frame independent characters without interference, and form a feature mapping relationship between its width and the width of the circumscribed rectangle of the remaining connected areas to establish a training library. The theoretical character width of the test binary license plate is learned by using the training library, and the size of the circumscribed rectangle of each connected area is adjusted to obtain the rough segmentation result. Then, the rough segmentation result is directly recognized, and the character with the highest recognition confidence is selected. By identifying the center position of the character with the highest confidence and judging its position in the license plate, to determine the center point of characters in different positions, and expand to the left and right with the actual character width, the accurate segmentation result of each character can be obtained.

Description

technical field [0001] The present invention relates to the technical field of image segmentation in the field of image processing, in particular to a character segmentation method based on rough positioning license plate. Background technique [0002] License plate recognition brings convenience to parking and checkpoints in daily life, and is also an important means in traffic control and criminal investigation. At present, most algorithms need to segment the characters of the license plate before recognizing the license plate. Therefore, character segmentation is the premise of license plate recognition, and the quality of character segmentation directly affects the difficulty and recognition rate of the next character recognition. [0003] Due to vehicle movement, low resolution of acquisition equipment, and natural weather effects, the quality of license plate images is difficult to guarantee. Moreover, the license plate positioning algorithm cannot accurately locate t...

AI Technical Summary

Problems solved by technology

[0003] Due to vehicle movement, low resolution of acquisition equipment, natural weather effects, etc., the image quality of license plate is difficult to guarantee

Moreover, the license plate positioning algorithm cannot accurately locate the license plate, and inevitably there will be problems such as incompleteness and a large number of borders.

Existing license plate character segmentation methods are particularly dependent on the license plate location and the quality of the license plate image, such as the commonly used vertical projection method, vertical segmentation method, and template matching method. poor segmentation

Images