License plate character recognition method based on deep convolutional generative adversarial network

Abstract

The invention provides a license plate character recognition method based on a deep convolutional generative adversarial network (DCGAN). The method comprises specific steps of: (1) extracting a license plate image to be recognized; (2) constructing and training a DCGAN; (3) generating a license plate image; (4) building a sample set of a character recognition network; (5) building and training alicense plate character recognition network CNN; and (6) recognizing license plate characters. The license plate character recognition method based on the DCGAN can overcome network overfitting in theprior art due to insufficiency in license plate data and few training samples, enhances data samples, improves the generalization ability and the robustness of the character recognition network, andincreases a license plate character recognition rate.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a character recognition method based on a deep convolutional generative confrontation network in deep learning. The present invention aims at extracting license plate images from pictures obtained by high-definition camera equipment installed at intersections in the traffic system, and then generating a large number of license plate images with a small number of license plate images, processing the generated license plate images as training samples, and training the license plate character recognition network, Realize license plate character recognition. Background technique [0002] With the continuous improvement of the socio-economic level and the popularization of vehicles, the ever-expanding transportation industry has a greater demand for more intelligent technologies and systems, and intelligent transportation systems have become a hot issue in social life. ...

AI Technical Summary

Problems solved by technology

However, the shortcomings of this method are that the recognition speed is slow, and the recognition rate and recognition speed are difficult to meet at the same time, and the recognition rate for some vehicles without license plates or severely damaged license plates is quite low, which cannot be well satisfied for practical applications. Real-time and accuracy requirements in

However, the shortcomings of this method are, first, the serious lack of license plate data in the deep learning-based license plate recognition system

Third, too few training samples will lead to network overfitting

Images