License plate classification method, license plate classification apparatus, and computer-readable storage medium

Patent metadata

Field	Value
Publication number	US-12536818-B2
Application number	US-202318388938-A
Country	US
Kind code	B2
Filing date	Nov 13, 2023
Priority date	May 11, 2021
Publication date	Jan 27, 2026
Grant date	Jan 27, 2026

Abstract

A license plate classification method, a license plate classification apparatus and a computer-readable storage medium are provided. The method includes: performing a license plate recognition process on a first license plate image to obtain a license plate recognition result; performing an encoding process on the license plate recognition result to obtain a first license plate feature; performing a feature extraction process on the first license plate image to obtain a second license plate feature; and processing the first license plate feature and the second license plate feature through a classification network to obtain a first license plate classification result. In this way, an accuracy of license plate classification is improved.

First claim

What is claimed is:

1 . A license plate classification method, comprising:
performing a license plate recognition process on a first license plate image to obtain a license plate recognition result; performing an encoding process on the license plate recognition result to obtain a first license plate feature; performing a feature extraction process on the first license plate image to obtain a second license plate feature; and processing the first license plate feature and the second license plate feature through a classification network to obtain a first license plate classification result.

2 . The license plate classification method according to claim 1 , wherein the performing an encoding process on the license plate recognition result to obtain a first license plate feature, comprises:
encoding the license plate recognition result by applying a preset encoding mode to obtain a vector sequence of a character of the license plate; and processing the vector sequence of the character of the license plate by applying the Transformer model to obtain the first license plate feature.

3 . The license plate classification method according to claim 2 , wherein the preset encoding mode is one-hot encoding, the vector sequence of the character of the license plate is a vector of N×M, wherein the N is a maximum length of characters of the license plate, and the M is a type of the character of the license plate.

4 . The license plate classification method according to claim 2 , wherein the Transformer model comprises an encoding module and a decoding module, and the processing the vector sequence of the character of the license plate by applying the Transformer model to obtain the first license plate feature, comprises:
encoding the vector sequence of the character of the license plate through the encoding module to obtain an encoded vector of the character of the license plate; and decoding the encoded vector of the character of the license plate through the decoding module to obtain the first license plate feature.

5 . The license plate classification method according to claim 1 , wherein the classification network comprises a feature fusion layer and a classification layer, and the processing the first license plate feature and the second license plate feature through a classification network to obtain a first license plate classification result, comprises:
fusing the first license plate feature with the second license plate feature through the feature fusion layer to obtain a fused license plate feature; and classifying the fused license plate feature through the classification layer to obtain the first license plate classification result.

6 . The license plate classification method according to claim 5 , wherein the classification network further comprises a first shaping network and a second shaping network, the fusing the first license plate feature with the second license plate feature through the feature fusion layer to obtain a fused license plate feature, comprises:
performing a dimensionality reduction process on the first license plate feature through the first shaping network to obtain a first license plate feature that has reduced dimensionality; and performing a dimensionality reduction process on the second license plate feature through the second shaping network to obtain a second license plate feature that has reduced dimensionality; wherein dimensionality of the first license plate feature that has reduced dimensionality is equal to dimensionality of the second license plate feature that has reduced dimensionality.

7 . The license plate classification method according to claim 1 , before the performing a license plate recognition process on a first license plate image to obtain a license plate recognition result, further comprising:
acquiring an image to be processed; and cropping the image to be processed to generate the first license plate image; wherein the first license plate image comprises a license plate.

8 . The license plate classification method according to claim 7 , wherein the cropping the image to be processed to generate the first license plate image, comprises:
obtaining, by a license plate detection model, a location in the image to be processed where the license plate is located; and cropping an image of the location in the image to be processed where the license plate is located to obtain the first license plate image.

9 . The license plate classification method according to claim 1 , wherein the performing a license plate recognition process on a first license plate image to obtain a license plate recognition result, comprises:
recognizing a character in the first license plate image through a license plate recognition network to obtain the license plate recognition result.

10 . The license plate classification method according to claim 1 , before the performing a license plate recognition process on a first license plate image to obtain a license plate recognition result, further comprising:
obtaining a classification training image; performing the license plate recognition process on the classification training image to obtain a second license plate recognition result; performing the encoding process on the second license plate recognition result to obtain a third license plate feature; performing the feature extraction process on the classification training image to obtain a fourth license plate feature; repeating the above operations to obtain classification training data, wherein the classification training data comprises a plurality of sets of training features, the training features comprise the third license plate feature and a corresponding fourth license plate feature; selecting one set from the plurality of sets of training features as current training features; processing the third license plate feature and the fourth license plate feature of the current training features through the classification network to obtain a second license plate classification result; adjusting a parameter of the classification network based on the second license plate classification result; and returning to the operation of selecting one set from the plurality of sets of training features as current training features until classification accuracy of the classification network exceeds a preset threshold.

11 . A license plate classification apparatus, comprising a non-transitory memory and a processor connected to the non-transitory memory, wherein the non-transitory memory is configured to store a computer program, and when the computer program is executed, the computer program is configured to perform operations of:
performing a license plate recognition process on a first license plate image to obtain a license plate recognition result; performing an encoding process on the license plate recognition result to obtain a first license plate feature; performing a feature extraction process on the first license plate image to obtain a second license plate feature; and processing the first license plate feature and the second license plate feature through a classification network to obtain a first license plate classification result.

12 . The license plate classification apparatus according to claim 11 , wherein, while performing the encoding process on the license plate recognition result to obtain the first license plate feature, the computer program is configured to perform operations of:
encoding the license plate recognition result by applying a preset encoding mode to obtain a vector sequence of a character of the license plate; and processing the vector sequence of the character of the licen

Assignees

• Zhejiang Dahua Technology Co

Inventors

• Xue Jiale
• Dun Jingyu
• Wang Yayun
• Zhang Wanwan
• Li Yikun

Classifications

• G06V10/82

  using neural networks · CPC title

• G06V10/806

  of extracted features · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06V20/625 Primary

  License plates · CPC title

• G06V10/7715

  Feature extraction, e.g. by transforming the feature space, e.g. multi-dimensional scaling [MDS]; Mappings, e.g. subspace methods · CPC title