CN111914842A - License plate information identification method and device, computer equipment and storage medium - Google Patents
License plate information identification method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN111914842A CN111914842A CN202010796421.5A CN202010796421A CN111914842A CN 111914842 A CN111914842 A CN 111914842A CN 202010796421 A CN202010796421 A CN 202010796421A CN 111914842 A CN111914842 A CN 111914842A
- Authority
- CN
- China
- Prior art keywords
- license plate
- plate information
- target picture
- preset
- transformer model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 239000013598 vector Substances 0.000 claims abstract description 71
- 238000012549 training Methods 0.000 claims abstract description 21
- 238000012545 processing Methods 0.000 claims abstract description 14
- 238000013528 artificial neural network Methods 0.000 claims description 14
- 238000010606 normalization Methods 0.000 claims description 12
- 230000004044 response Effects 0.000 claims description 11
- 238000012790 confirmation Methods 0.000 claims description 5
- 238000010586 diagram Methods 0.000 description 9
- 230000008569 process Effects 0.000 description 9
- 238000004590 computer program Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 3
- 238000013519 translation Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000008034 disappearance Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000000977 initiatory effect Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/63—Scene text, e.g. street names
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/625—License plates
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The application discloses a license plate information identification method, which comprises the following steps: when a target picture is detected, acquiring a preset vector corresponding to the license plate image feature in the target picture; the preset vector is used as an input vector of a Transformer model, wherein the Transformer model is obtained by training according to a plurality of groups of preset vectors and license plate information corresponding to the preset vectors; and acquiring an output value of the Transformer model as license plate information corresponding to the target picture. The application also discloses a license plate information recognition device, computer equipment and a computer readable storage medium. According to the method and the device, the license plate information of the license plate image in the target picture is recognized by utilizing the efficient data processing capacity of the Transformer model, and the efficiency of recognizing the license plate information is improved.
Description
Technical Field
The present disclosure relates to the field of deep learning, and in particular, to a license plate information recognition method, a license plate information recognition apparatus, a computer device, and a computer-readable storage medium.
Background
At present, most of license plate images are recognized by dividing and recognizing image characters based on a traditional Neural Network model (such as RNN (Recurrent Neural Network) and splicing into complete license plate information.
With the rapid development of the information age, the traditional neural network model has the defects of slow iteration and slow calculation, and the development requirement of rapidly recognizing the license plate information in the picture in the modern society is more and more difficult to meet.
The above is only for the purpose of assisting understanding of the technical solutions of the present application, and does not represent an admission that the above is prior art.
Disclosure of Invention
The present application mainly aims to provide a license plate information identification method, a license plate information identification device, a computer device, and a computer-readable storage medium, and aims to solve the problem of low license plate information identification efficiency in a picture.
In order to achieve the above object, the present application provides a license plate information recognition method, including the following steps:
when a target picture is detected, acquiring a preset vector corresponding to the license plate image feature in the target picture;
the preset vector is used as an input vector of a Transformer model, wherein the Transformer model is obtained by training according to a plurality of groups of preset vectors and license plate information corresponding to the preset vectors;
and acquiring an output value of the Transformer model as license plate information corresponding to the target picture.
Further, the step of obtaining a preset vector corresponding to the license plate image feature in the target picture includes:
recognizing license plate image characteristics in the target picture by utilizing a pre-constructed residual error network, wherein the residual error network also converts the license plate image characteristics into preset vectors and outputs the preset vectors;
and acquiring an output value of the residual error network as a preset vector corresponding to the license plate image feature.
Further, the decoder of the Transformer model comprises an encoding-decoding attention layer and a feedforward neural network, wherein the encoding-decoding attention layer is used for processing output values of an encoder of the Transformer model and/or output values of a decoder of a previous layer, and the output values of the encoding-decoding attention layer are subjected to summation and normalization processing and then transmitted to the feedforward neural network.
Further, after the step of obtaining the output value of the transform model as the license plate information corresponding to the target picture, the method further includes:
and when a confirmation response of the license plate information corresponding to the target picture is received or a cancellation response of the license plate information corresponding to the target picture is not received within a preset time, taking the license plate information corresponding to the target picture and a preset vector corresponding to the target picture as training samples of the transform model, and updating the transform model according to the training samples.
Further, after the step of obtaining the output value of the transform model as the license plate information corresponding to the target picture, the method further includes:
and sending the license plate information corresponding to the target picture to remote equipment, wherein the target picture is sent to the local terminal by the remote equipment.
Further, the license plate information identification method further includes:
when a target picture is detected, detecting whether an identification record corresponding to the target picture exists or not;
if not, executing the step of acquiring a preset vector corresponding to the license plate image feature in the target picture;
and if so, acquiring the license plate information associated in the identification record as the license plate information corresponding to the target picture.
In order to achieve the above object, the present application further provides a license plate information recognition apparatus, including:
the license plate information recognition device comprises a memory, a processor and a license plate information recognition program which is stored on the memory and can run on the processor, wherein the license plate information recognition program realizes the steps of the license plate information recognition method when being executed by the processor.
To achieve the above object, the present application also provides a computer device, comprising:
the computer equipment comprises a memory, a processor and a license plate information recognition program which is stored on the memory and can run on the processor, wherein the license plate information recognition program realizes the steps of the license plate information recognition method when being executed by the processor.
In order to achieve the above object, the present application further provides a computer-readable storage medium, where a license plate information recognition program is stored, and when the license plate information recognition program is executed by a processor, the steps of the above license plate information recognition method are implemented.
According to the license plate information identification method, the license plate information identification device, the computer equipment and the computer readable storage medium, when a target picture is detected, a preset vector corresponding to license plate image features in the target picture is obtained; the preset vector is used as an input vector of a Transformer model, wherein the Transformer model is obtained by training according to a plurality of groups of preset vectors and license plate information corresponding to the preset vectors; and acquiring an output value of the Transformer model as license plate information corresponding to the target picture. Therefore, the license plate information of the license plate image in the target picture is recognized by utilizing the efficient data processing capacity of the Transformer model, and the efficiency of recognizing the license plate information is improved.
Drawings
Fig. 1 is a schematic diagram illustrating steps of a license plate information recognition method in an embodiment of the present application;
fig. 2 is a schematic diagram illustrating another step of the license plate information recognition method according to an embodiment of the present application;
fig. 3 is a schematic diagram illustrating another step of the license plate information recognition method in an embodiment of the present application;
fig. 4 is a schematic diagram of a further step of the license plate information recognition method in an embodiment of the present application;
FIG. 5 is a diagram illustrating a transform model according to an embodiment of the present application;
FIG. 6 is a diagram of an example of a transform model encoder according to an embodiment of the present application;
FIG. 7 is a diagram of an exemplary transform model decoder according to an embodiment of the present application;
fig. 8 is a block diagram illustrating a structure of a computer device according to an embodiment of the present application.
The implementation, functional features and advantages of the objectives of the present application will be further explained with reference to the accompanying drawings.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
Referring to fig. 1, in an embodiment, the license plate information identification method includes:
and S10, when a target picture is detected, acquiring a preset vector corresponding to the license plate image feature in the target picture.
In this embodiment, the target picture obtained by the terminal is captured in real time or at regular time by the camera device having a communication connection with the terminal, or may be sent to the terminal by a remote device. The remote device includes a smart phone, a computer device, a tablet computer, an etc (electronic Toll collection) system, and the like.
Wherein, the target picture is a picture with a license plate image.
Optionally, after the terminal detects the target picture, the terminal identifies license plate image features (i.e., feature images) in the target picture, and mainly identifies graphic features in a development area where the license plate image is located.
When an image with a "V" character is recognized, the V shape of the image is not focused on the word meaning of the V itself.
The feature images (image features) mainly include color features, texture features, graphic features (shape features), and spatial relationship features of images. The graphic features have two types of representation methods, one is outline features, and the other is region features. The outline features of the image are mainly directed to the outer boundary of the object, while the area features of the image are related to the entire shape area.
Optionally, the terminal is pre-established with a mapping relationship between preset vectors corresponding to different license plate image features, and the preset vectors are suitable for input vectors of a transform model.
It should be understood that the same license plate image feature may often point to multiple preset vectors, and the preset vector corresponding to the license plate image feature may be a set of all the preset vectors pointed to by the license plate image feature.
It should be noted that the Transformer model is a deep learning model proposed by the Google team in 2017, and is currently widely applied to the field of word translation. Prior to the Transformer, most neural network-based machine translation methods relied on Recurrent Neural Networks (RNNs), which utilized cycles (i.e., the output of each step fed into the next step) to perform sequential operations (e.g., translating sentences word by word). Although RNNs are very powerful in modeling sequences, their sequence means that the network is very slow in training, since long sentences require more training steps and their loop structure increases training difficulty. Compared with the RNN-based approach, the Transformer does not need a loop, but processes the data in the sequence in parallel, training much faster than RNN.
Optionally, after obtaining the license plate image features in the target picture, the terminal queries and obtains a preset vector corresponding to the license plate image features.
Optionally, the terminal is trained in advance and a residual network model is constructed, the residual network model utilizes enough target pictures with license plate images to perform feature map training to achieve convergence, and finally a mapping relation between preset vectors corresponding to different license plate image features is established.
It should be noted that the residual network is a convolutional neural network proposed by 4 scholars from Microsoft Research, and can achieve a good effect in the application of image classification and identification. The residual error network has the characteristics of being easy to optimize, improving the accuracy rate by increasing equivalent depth, and relieving the gradient disappearance problem caused by increasing the depth in the deep neural network because the internal residual error block uses jump connection.
Alternatively, the residual network may be a ResNet18 residual network.
Optionally, after the terminal inputs the target picture into the trained residual network model, the residual network identifies the license plate image features in the picture, and after the license plate image features are obtained through residual network identification, the license plate image features can be converted into a set of preset vectors corresponding to the license plate image features and output, and then output values of the residual network can be obtained and serve as the preset vectors corresponding to the license plate image features in the target picture.
And at the moment, the terminal directly acquires the output value of the residual error network as a preset vector corresponding to the license plate image characteristics.
And step S20, taking the preset vector as an input vector of a Transformer model, wherein the Transformer model is obtained by training according to multiple groups of preset vectors and license plate information corresponding to the preset vectors.
And step S30, acquiring an output value of the Transformer model as license plate information corresponding to the target picture.
Since the conventional Transformer model is generally applied to the field of word translation, the input layer of the general Transformer model is a word embedding layer to convert text information into an input vector suitable for the Transformer model.
In this embodiment, a word embedding layer of the transform model may be removed, and the preset vector corresponding to the license plate image feature obtained in step S10 is directly used as a transform model input value; alternatively, the trained residual network layer may be used as the input layer of the transform model, so that the output value of the residual network is directly used as the input value of the transform model.
Alternatively, referring to fig. 5, the transform model 10 includes a 6-layer Encoder 20(Encoder) and a 6-layer Decoder 30(Decoder), and the residual network 40 can be used as an input layer of the transform model 10 to input the predetermined vectors corresponding to the license plate image features.
Alternatively, referring to fig. 6, each layer of encoder 20 includes a self-attention layer 21(self-attention), two layers of summing and normalization layers 22(Add & normaize), and a Feed-Forward neural network 23(Feed Forward). The initial input value of the encoder 20 will reach the attention layer 21 first, the input vectors are mapped into different subspaces, the attention vectors are calculated by performing a dot product operation on all the subspaces, and finally the attention vectors calculated by all the subspaces are spliced and mapped into the original input space to obtain the final attention vector as the output.
The output values from the attention layer 21 and the initial input values of the encoder 20 are then summed and normalized in a first layer summing and normalization layer 22, and the resulting output values are fed into a feedforward neural network 23. The output value of the feedforward neural network 23 and the output value of the attention layer 21 are subjected to summation and normalization processing in a second layer summation and normalization layer 22, and the final output value is output as the output value of the encoder 20 of the current layer.
Alternatively, referring to fig. 7, each layer of the Encoder 30 includes an encoding-decoding Attention layer 31(Encoder-Decoder attribute, alternatively referred to as Context-attribute), two summing and normalization layers 32(Add & normalization), and a Feed-Forward neural network 33(Feed Forward). The output value of the encoder 20 and the output value of the decoder 30 in the previous layer will reach the coding-decoding attention layer 31 for processing, the output value obtained by the coding-decoding attention layer 31 processing and the output value of the decoder 30 in the previous layer will be subjected to summation and normalization processing in the first layer summation and normalization layer 22, and the processed output value will enter the feedforward neural network 33. The output value of the feedforward neural network 33 and the output value of the coding-decoding attention layer 31 are subjected to summation and normalization processing in a second layer summation and normalization layer 32, and the final output value is output as the output value of the current layer decoder 30.
It should be understood that the first layer decoder need not process the output values of the previous layer decoder, only the output values of the encoder.
I.e. the encoding-decoding attention layer is used to process the output values of the encoder and/or the output values of the decoder of the previous layer.
Thus, compared with the conventional decoder, the decoder of the present embodiment removes the self-attention layer, so that the input value of the decoder directly reaches the encoding-decoding attention layer for processing, the input of each word at the time t and the consideration of the information at the previous time can be focused, and the computing capability of the decoder can be improved. Of course, this is the preferred decoder structure, even if the decoder retains the self attention layer.
The Transformer model is trained in advance according to a training sample set (sample data is large enough) formed by a plurality of groups of preset vectors and license plate information corresponding to each group of preset vectors until the model converges.
It should be understood that the license plate information corresponding to each set of preset vectors is necessarily associated with the target picture of the license plate corresponding to the set of preset vectors.
It should be noted that the initial data set (each group of data includes a license plate photo and a license plate information pair) of the training sample may be obtained from the CCPD data set, which is a large data set library for license plate recognition and collects a large number of license plate photos and corresponding information. Of course, the manner in which the initial data set is acquired is not limited to this, and other alternatives are possible.
Further, after the terminal obtains preset vectors corresponding to license plate image features in the target picture, the preset vectors are used as input values of the trained Transformer model and are input into the Transformer model, the Transformer model is processed and operated, and license plate information corresponding to the preset vectors is inquired to serve as output values, namely the output values of the Transformer model are license plate information corresponding to the target picture.
Therefore, the license plate information of the license plate image in the target picture is recognized by utilizing the efficient data processing capacity of the Transformer model, and the efficiency of recognizing the license plate information is improved.
In an embodiment, as shown in fig. 2, on the basis of the embodiment shown in fig. 1, after the step of obtaining an output value of the transform model as license plate information corresponding to the target picture, the method further includes:
step S40, when a confirmation response of the license plate information corresponding to the target picture is received or a cancellation response of the license plate information corresponding to the target picture is not received within a preset time, taking the license plate information corresponding to the target picture and a preset vector corresponding to the target picture as training samples of the Transformer model, and updating the Transformer model according to the training samples.
In this embodiment, after obtaining the license plate information of the license plate image in the target picture, the terminal associates the license plate information with the target picture and outputs the license plate information corresponding to the target picture.
Optionally, if the user feels that the license plate information is identified without error, the associated terminal can be used for initiating a confirmation response of the license plate information corresponding to the target picture to the terminal; and if the user feels that the license plate information is identified wrongly, the associated terminal can be used for sending a cancellation response of the license plate information corresponding to the target picture to the terminal.
Optionally, if the terminal does not receive a cancellation response of the license plate information corresponding to the target picture within a preset time length, the license plate information is identified without errors by default.
The associated terminal may be a computer device, a smart phone, or the like.
It should be understood that the preset time period may be set by an engineer according to actual needs, such as 30 seconds, one minute, etc.
Optionally, when the terminal receives a confirmation response of the license plate information corresponding to the target picture or when the terminal does not receive a cancellation response of the license plate information corresponding to the target picture within a preset time period, it is determined that the license plate information of the target picture is identified without error, and the license plate information corresponding to the target picture and a preset vector corresponding to the target picture are further used as training samples of the transform model, so that the transform model is updated according to the training samples.
Therefore, iterative optimization can be performed on the Transformer model, and the accuracy of the Transformer model in recognizing the license plate information is improved.
In an embodiment, as shown in fig. 3, on the basis of the embodiments of fig. 1 to fig. 2, after the step of obtaining the output value of the transform model as the license plate information corresponding to the target picture, the method further includes:
and S50, sending the license plate information corresponding to the target picture to a remote device, wherein the target picture is sent to the local terminal by the remote device.
In this embodiment, the target picture is sent to the terminal by the remote device through network communication, after the terminal receives the target picture sent by the remote device and obtains license plate information corresponding to the target picture based on the identification in steps S10-S30, the license plate information is associated with the target picture and generates a data packet, and the license plate information corresponding to the target picture is fed back to the remote device in the form of the data packet.
For example, a camera device arranged at an entrance and an exit of a parking lot can be used as a remote device, and after a vehicle picture appearing at the entrance and the exit of the parking lot is captured as a target picture, the target picture is transmitted to the terminal, so that license plate information can be identified.
Therefore, even the electronic equipment with less hardware resources can quickly identify the license plate information.
In an embodiment, as shown in fig. 4, based on the embodiments of fig. 1 to 3, the method for identifying license plate information further includes:
and step S11, when the target picture is detected, detecting whether an identification record corresponding to the target picture exists.
And S12, acquiring a preset vector corresponding to the license plate image feature in the target picture.
And step S13, acquiring the license plate information associated in the identification record as the license plate information corresponding to the target picture.
In the embodiment, after the license plate information of the license plate image in the target picture is obtained by the terminal through recognition, the license plate information is associated with the target picture to generate the recognition record, and the recognition record is kept in the database.
Optionally, when obtaining a new target picture, the terminal may first detect whether an identification record of the target picture is stored in the database.
If not, the terminal executes the step of obtaining the preset vector corresponding to the license plate image feature in the target picture, further executes the steps S20-S30, and utilizes a Transformer model to identify license plate information in the currently obtained target picture.
If so, the terminal directly acquires the identification record corresponding to the target picture, and directly takes the license plate information associated in the identification record as the license plate information corresponding to the target picture.
Therefore, by introducing a recognition record storage mechanism of the target picture, when the target picture with the license plate information detected before is recorded into the system again, the license plate information corresponding to the target picture can be quickly obtained.
Certainly, the identification record may also be associated with the license plate image feature corresponding to the target image, and when the terminal obtains the target image, the license plate image feature of the target image is extracted first, and then the identification record corresponding to the license plate image feature in the database is queried according to the license plate image feature.
In addition, the present application further provides a license plate information recognition apparatus, where the license plate information recognition apparatus includes a license plate information recognition program, and the license plate information recognition program is executed by a processor to implement the steps of the license plate information recognition method according to the above embodiments.
Referring to fig. 8, a computer device, which may be a server and whose internal structure may be as shown in fig. 8, is also provided in the embodiment of the present application. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the computer designed processor is used to provide computational and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for a license plate information recognition program. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a license plate information recognition method.
Those skilled in the art will appreciate that the architecture shown in fig. 8 is only a block diagram of some of the structures associated with the present solution and is not intended to limit the scope of the present solution as applied to computer devices.
In addition, the present application also provides a computer-readable storage medium, where the computer-readable storage medium includes a license plate information recognition program, and the license plate information recognition program, when executed by a processor, implements the steps of the license plate information recognition method according to the above embodiments. It is to be understood that the computer-readable storage medium in the present embodiment may be a volatile-readable storage medium or a non-volatile-readable storage medium.
In summary, in the license plate information identification method, the license plate information identification device, the computer device, and the storage medium provided in the embodiments of the present application, when a target picture is detected, license plate image features in the target picture are identified; acquiring a preset vector corresponding to the license plate image feature; the preset vector is used as an input vector of a Transformer model, wherein the Transformer model is obtained by training according to a plurality of groups of preset vectors and license plate information corresponding to the preset vectors; acquiring an output value of the Transformer model as license plate information corresponding to the target picture; therefore, the license plate information of the license plate image in the target picture is recognized by utilizing the efficient data processing capacity of the Transformer model, and the efficiency of recognizing the license plate information is improved.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium provided herein and used in the examples may include non-volatile and/or volatile memory. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), double-rate SDRAM (SSRSDRAM), Enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, apparatus, article, or method. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, apparatus, article, or method that includes the element.
The above description is only for the preferred embodiment of the present application and not intended to limit the scope of the present application, and all modifications of equivalent structures and equivalent processes, which are made by the contents of the specification and the drawings of the present application, or which are directly or indirectly applied to other related technical fields, are intended to be included within the scope of the present application.
Claims (9)
1. A license plate information identification method is characterized by comprising the following steps:
when a target picture is detected, acquiring a preset vector corresponding to the license plate image feature in the target picture;
the preset vector is used as an input vector of a Transformer model, wherein the Transformer model is obtained by training according to a plurality of groups of preset vectors and license plate information corresponding to the preset vectors;
and acquiring an output value of the Transformer model as license plate information corresponding to the target picture.
2. The method for recognizing license plate information of claim 1, wherein the step of obtaining the preset vector corresponding to the license plate image feature in the target picture comprises:
recognizing license plate image characteristics in the target picture by utilizing a pre-constructed residual error network, wherein the residual error network also converts the license plate image characteristics into preset vectors and outputs the preset vectors;
and acquiring an output value of the residual error network as a preset vector corresponding to the license plate image feature.
3. The license plate information recognition method of claim 1, wherein the decoder of the Transformer model comprises an encoding-decoding attention layer and a feedforward neural network, wherein the encoding-decoding attention layer is used for processing an output value of an encoder of the Transformer model and/or an output value of a decoder of a previous layer, and the output value of the encoding-decoding attention layer is subjected to summation and normalization processing and then transmitted to the feedforward neural network.
4. The method for recognizing license plate information of claim 1, wherein after the step of obtaining the output value of the Transformer model as the license plate information corresponding to the target picture, the method further comprises:
when a confirmation response of the license plate information corresponding to the target picture is received or a cancellation response of the license plate information corresponding to the target picture is not received within a preset time, taking the license plate information corresponding to the target picture and a preset vector corresponding to the target picture as training samples of the transform model, and updating the transform model according to the training samples.
5. The method for recognizing license plate information of claim 1, wherein after the step of obtaining the output value of the Transformer model as the license plate information corresponding to the target picture, the method further comprises:
and sending the license plate information corresponding to the target picture to remote equipment, wherein the target picture is sent to the local terminal by the remote equipment.
6. The method for recognizing license plate information of claim 1, further comprising:
when a target picture is detected, detecting whether an identification record corresponding to the target picture exists or not;
if not, executing the step of acquiring a preset vector corresponding to the license plate image feature in the target picture;
and if so, acquiring the license plate information associated in the identification record as the license plate information corresponding to the target picture.
7. A license plate information recognition apparatus, comprising a memory, a processor, and a license plate information recognition program stored in the memory and executable on the processor, wherein the license plate information recognition program, when executed by the processor, implements the steps of the license plate information recognition method according to any one of claims 1 to 6.
8. A computer device, characterized in that the computer device comprises a memory, a processor and a license plate information recognition program stored on the memory and executable on the processor, wherein the license plate information recognition program, when executed by the processor, implements the steps of the license plate information recognition method according to any one of claims 1 to 6.
9. A computer-readable storage medium, wherein a license plate information recognition program is stored on the computer-readable storage medium, and when executed by a processor, the steps of the license plate information recognition method according to any one of claims 1 to 6 are implemented.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010796421.5A CN111914842A (en) | 2020-08-10 | 2020-08-10 | License plate information identification method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010796421.5A CN111914842A (en) | 2020-08-10 | 2020-08-10 | License plate information identification method and device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111914842A true CN111914842A (en) | 2020-11-10 |
Family
ID=73283550
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010796421.5A Pending CN111914842A (en) | 2020-08-10 | 2020-08-10 | License plate information identification method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111914842A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112508018A (en) * | 2020-12-14 | 2021-03-16 | 北京澎思科技有限公司 | License plate recognition method and device and storage medium |
| CN113139470A (en) * | 2021-04-25 | 2021-07-20 | 安徽工业大学 | Glass identification method based on Transformer |
| CN113393445A (en) * | 2021-06-21 | 2021-09-14 | 上海交通大学医学院附属新华医院 | Breast cancer image determination method and system |
| CN113569648A (en) * | 2021-06-29 | 2021-10-29 | 深圳市捷顺科技实业股份有限公司 | Vehicle position moving recording method and device |
| CN114067286A (en) * | 2022-01-18 | 2022-02-18 | 松立控股集团股份有限公司 | High-order camera vehicle weight recognition method based on serialized deformable attention mechanism |
| CN115661807A (en) * | 2022-12-28 | 2023-01-31 | 成都西物信安智能系统有限公司 | Method for acquiring license plate information |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109492640A (en) * | 2017-09-12 | 2019-03-19 | 中国移动通信有限公司研究院 | Licence plate recognition method, device and computer readable storage medium |
| CN110349229A (en) * | 2019-07-09 | 2019-10-18 | 北京金山数字娱乐科技有限公司 | A kind of Image Description Methods and device |
| CN110399798A (en) * | 2019-06-25 | 2019-11-01 | 朱跃飞 | A kind of discrete picture file information extracting system and method based on deep learning |
| CN110472688A (en) * | 2019-08-16 | 2019-11-19 | 北京金山数字娱乐科技有限公司 | The method and device of iamge description, the training method of image description model and device |
| CN111444911A (en) * | 2019-12-13 | 2020-07-24 | 珠海大横琴科技发展有限公司 | Training method and device of license plate recognition model and license plate recognition method and device |
-
2020
- 2020-08-10 CN CN202010796421.5A patent/CN111914842A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109492640A (en) * | 2017-09-12 | 2019-03-19 | 中国移动通信有限公司研究院 | Licence plate recognition method, device and computer readable storage medium |
| CN110399798A (en) * | 2019-06-25 | 2019-11-01 | 朱跃飞 | A kind of discrete picture file information extracting system and method based on deep learning |
| CN110349229A (en) * | 2019-07-09 | 2019-10-18 | 北京金山数字娱乐科技有限公司 | A kind of Image Description Methods and device |
| CN110472688A (en) * | 2019-08-16 | 2019-11-19 | 北京金山数字娱乐科技有限公司 | The method and device of iamge description, the training method of image description model and device |
| CN111444911A (en) * | 2019-12-13 | 2020-07-24 | 珠海大横琴科技发展有限公司 | Training method and device of license plate recognition model and license plate recognition method and device |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112508018A (en) * | 2020-12-14 | 2021-03-16 | 北京澎思科技有限公司 | License plate recognition method and device and storage medium |
| CN113139470A (en) * | 2021-04-25 | 2021-07-20 | 安徽工业大学 | Glass identification method based on Transformer |
| CN113393445A (en) * | 2021-06-21 | 2021-09-14 | 上海交通大学医学院附属新华医院 | Breast cancer image determination method and system |
| CN113569648A (en) * | 2021-06-29 | 2021-10-29 | 深圳市捷顺科技实业股份有限公司 | Vehicle position moving recording method and device |
| CN114067286A (en) * | 2022-01-18 | 2022-02-18 | 松立控股集团股份有限公司 | High-order camera vehicle weight recognition method based on serialized deformable attention mechanism |
| CN115661807A (en) * | 2022-12-28 | 2023-01-31 | 成都西物信安智能系统有限公司 | Method for acquiring license plate information |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111914842A (en) | License plate information identification method and device, computer equipment and storage medium | |
| CN108304882B (en) | Image classification method and device, server, user terminal and storage medium | |
| WO2020098074A1 (en) | Face sample picture marking method and apparatus, computer device, and storage medium | |
| KR20190085098A (en) | Keyword extraction method, computer device, and storage medium | |
| CN110569341B (en) | Method and device for configuring chat robot, computer equipment and storage medium | |
| CN111651992A (en) | Named entity labeling method and device, computer equipment and storage medium | |
| CN110807491A (en) | License plate image definition model training method, definition detection method and device | |
| CN112101318A (en) | Image processing method, device, equipment and medium based on neural network model | |
| CN111832581B (en) | Lung feature recognition method and device, computer equipment and storage medium | |
| CN110598603A (en) | Face recognition model acquisition method, device, equipment and medium | |
| CN110909784B (en) | Training method and device of image recognition model and electronic equipment | |
| CN111860669B (en) | Training method and device for OCR (optical character recognition) model and computer equipment | |
| CN110543823B (en) | Pedestrian re-identification method and device based on residual error network and computer equipment | |
| CN111191032A (en) | Corpus expansion method and device, computer equipment and storage medium | |
| CN110633751A (en) | Training method of car logo classification model, car logo identification method, device and equipment | |
| CN111666931B (en) | Mixed convolution text image recognition method, device, equipment and storage medium | |
| CN114647713A (en) | Knowledge graph question-answering method, device and storage medium based on virtual confrontation | |
| CN112070086B (en) | Optimization method of text recognition system, computer equipment and storage medium | |
| CN111898544B (en) | Text image matching method, device and equipment and computer storage medium | |
| CN109101984B (en) | Image identification method and device based on convolutional neural network | |
| CN117475253A (en) | Model training method and device, electronic equipment and storage medium | |
| CN116484224A (en) | Training method, device, medium and equipment for multi-mode pre-training model | |
| CN112989256B (en) | Method and device for identifying web fingerprint in response information | |
| CN115238124A (en) | Video character retrieval method, device, equipment and storage medium | |
| CN114238715A (en) | Question-answering system based on social aid, construction method, computer equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |