License plate character recognition method based on contextual information

Abstract

The invention discloses a license plate character recognition method based on contextual information, and the method comprises the steps: constructing a deep neural network model which comprises a quick extraction feature network, a contextual information network and a recognition network, and the quick extraction feature network, the contextual information network and the recognition network aresequentially connected; training the deep neural network model through the obtained license plate character training sample image set; recognizing a license plate image to be recognized through the trained deep neural network model; the character recognition result is more accurate, the distinguishing capability for similar characters is stronger, and the robustness is higher.

Description

technical field [0001] The invention relates to the technical field of license plate recognition, in particular to a method for recognizing license plate characters based on context information. Background technique [0002] License plate recognition is the core technology of intelligent transportation, which includes three parts: license plate position detection, license plate character segmentation, and license plate character recognition. Among them, the license plate character recognition is the most important part of the whole technology, the quality of the license plate character recognition engine directly determines the overall performance of the license plate recognition technology. [0003] License plate character recognition refers to identifying the true letter meaning of a single license plate character that has been accurately segmented. The commonly used methods are as follows: [0004] (1) The method based on global features. This type of features uses globa...

AI Technical Summary

Problems solved by technology

The advantage of this type of feature is that it has a strong ability to distinguish characters; its disadvantage is that it pays too much attention to the local features of characters, and often mistakenly distinguishes characters with noise interference.

[0006] (3) The method based on deep learning. In recent years, deep learning technology has been widely concerned and applied in various fields by virtue of its ability to simulate the neural network of the human brain and to perform accurate nonlinear prediction. A number of classic targets have emerged. Recognition network frameworks, such as resnet, densenet, LSTM, etc. These classic network frameworks can recognize license plate characters well through transfer learning, but the disadvantage of this type of technology is that although the deeper network has better recognition effect, the model consumes a lot of memory , the amount of calculation is very large, although the shallower network model runs fast, but the recognition accuracy is average, especially for the lack of ability to distinguish similar characters

Images