CN114724128A - License plate recognition method, device, equipment and medium - Google Patents
License plate recognition method, device, equipment and medium Download PDFInfo
- Publication number
- CN114724128A CN114724128A CN202210275239.4A CN202210275239A CN114724128A CN 114724128 A CN114724128 A CN 114724128A CN 202210275239 A CN202210275239 A CN 202210275239A CN 114724128 A CN114724128 A CN 114724128A
- Authority
- CN
- China
- Prior art keywords
- license plate
- recognized
- image
- vehicle
- recognition
- 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.)
- Granted
Links
Images
Classifications
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Character Discrimination (AREA)
- Traffic Control Systems (AREA)
Abstract
The application relates to the technical field of license plate recognition, in particular to a license plate recognition method, a license plate recognition device, license plate recognition equipment and a license plate recognition medium, wherein the license plate recognition method comprises the steps of obtaining a license plate image to be recognized of a vehicle to be recognized; extracting the license plate image characteristics of the license plate image to be recognized by using a convolutional neural network; carrying out license plate state recognition on the license plate image characteristics through a license plate state recognition model to obtain the license plate state of the vehicle to be recognized, wherein the license plate state at least comprises illegal shielding, non-illegal shielding and no shielding; identifying the license plate number of the vehicle to be identified by the license plate number identification model according to the license plate number characteristics; and outputting a license plate recognition result, wherein the recognition result comprises the license plate state of the vehicle to be recognized and the license plate number of the vehicle to be recognized. The method and the device have the effects of facilitating judgment of relevant workers through the recognition results and reducing the workload of the relevant workers.
Description
Technical Field
The present application relates to the field of license plate recognition technologies, and in particular, to a license plate recognition method, apparatus, device, and medium.
Background
The license plate identification is an application of a computer video image identification technology in vehicle identification, and license plate images acquired on site are transmitted to an identification system for image processing to obtain license plate information. The license plate recognition technology is widely applied to various scenes, such as license plate recognition of a parking lot, an electronic toll collection system of a highway, traffic violation monitoring and the like.
Currently, the license plate number is the most critical identity information of a vehicle, if an illegal action occurs in vehicle driving, for example, a driver is drunk, overloaded, rushing red light, speeding and the like, a traffic management department can punish the illegal vehicle by recording the license plate number, but some drivers may intentionally shield the license plate to avoid the responsibility. In the related technology, image acquisition equipment is adopted to acquire images of vehicles in a static state or a moving state for the identification of the license plate, the license plate number of the vehicle is identified according to the acquired images of the vehicle, and related workers further judge whether behaviors of illegally shielding the license plate exist or not according to the identified license plate number, so that the judgment of the related workers through identification results is not facilitated, and the workload of the related workers is increased.
Disclosure of Invention
In order to provide a more intuitive recognition result for related workers and reduce the burden of the related workers, the application provides a license plate recognition method, a device, equipment and a medium.
In a first aspect, the present application provides a method for identifying a blocking number, which adopts the following technical scheme:
a license plate recognition method includes:
acquiring a license plate image to be recognized of a vehicle to be recognized;
extracting the license plate image characteristics of the license plate image to be recognized by using a convolutional neural network;
carrying out license plate state recognition on the license plate image characteristics through a license plate state recognition model to obtain the license plate state of the vehicle to be recognized, wherein the license plate state at least comprises illegal shielding, non-illegal shielding and no shielding;
identifying the license plate number of the vehicle to be identified by the license plate number identification model according to the license plate number characteristics;
and outputting a license plate recognition result, wherein the recognition result comprises the license plate state of the vehicle to be recognized and the license plate number of the vehicle to be recognized.
By adopting the technical scheme, after the license plate image to be recognized of the vehicle to be recognized is obtained, the license plate image characteristics of the license plate image to be recognized are extracted through the convolutional neural network, the characteristics of the license plate image to be recognized are recognized through the license plate state recognition model, and the license plate state is obtained, wherein the license plate state at least comprises illegal shielding, non-illegal shielding and no shielding, the license plate image characteristics to be recognized are recognized through the license plate number recognition model, the license plate number of the vehicle to be recognized is obtained, finally the license plate state and the license plate number of the vehicle to be recognized are output together, the recognition result is more visual, meanwhile, related workers do not need to further judge whether the behavior of illegally shielding the license plate exists through the recognized license plate number, and the work burden of the related workers is reduced.
In one possible implementation manner, the acquiring a license plate map to be recognized of a vehicle to be recognized includes:
acquiring a vehicle image to be identified;
cutting a license plate image area of the vehicle image to be recognized to obtain an initial license plate image to be recognized of the vehicle to be recognized;
sequentially carrying out image processing on the initial license plate image to be recognized to obtain the license plate image to be recognized, wherein the image processing mode at least comprises any one or more of the following modes: graying, binaryzation, expansion treatment and corrosion.
By adopting the technical scheme, the vehicle image is cut to obtain the initial to-be-recognized image only containing the license plate image, the obtained initial to-be-recognized license plate image is subjected to image processing to obtain the final to-be-recognized license plate image, and the image processing mode at least comprises any one or more of the following modes: graying, binaryzation, expansion processing and corrosion, and the license plate image to be recognized after image processing is beneficial to feature extraction.
In a possible implementation manner, the cutting the license plate image area of the vehicle image to be recognized to obtain an initial license plate image to be recognized of the vehicle to be recognized includes:
determining a license plate image region from the vehicle image to be recognized;
expanding the license plate image area according to a preset proportion to obtain an expanded license plate image;
and determining the expanded license plate image as an initial license plate image to be recognized.
By adopting the technical scheme, the edge of the license plate image area to be recognized is expanded by expanding the edge of the license plate to be recognized according to the preset proportion, the expanded license plate image is cut to obtain the initial license plate image to be recognized, and the probability of incomplete image cutting in the image cutting process is reduced.
In a possible implementation manner, the identifying the license plate number of the license plate image feature by a license plate number identification model to obtain the license plate number of the vehicle to be identified includes:
performing initial recognition of the license plate number according to the license plate image characteristics of the vehicle to be recognized to obtain a character recognition result of the license plate to be recognized, wherein the character recognition result comprises each recognition character, unrecognized characters and respective corresponding positions;
replacing an unclear type character in the unrecognized character with a first character, and replacing an occlusion type character in the unrecognized character with a second character;
and updating the character recognition result according to the substitution result to obtain the license plate number.
By adopting the technical scheme, the initial recognition of the license plate number is carried out on the license plate image characteristics of the vehicle to be recognized before the license plate number recognition is carried out, the recognition result comprises the recognition characters, the unrecognized characters and the corresponding positions, the types of the unrecognized characters are distinguished, the unclear character types are replaced by the first special symbols, the shielded character types are replaced by the second special symbols, the initial recognition result of the license plate to be recognized is updated according to the replacement result, the license plate number is obtained, and the license plate number at least comprises the recognized characters, the first special symbols and the second special symbols, so that the recognition effect of the license plate number is clearer.
In one possible implementation manner, the method further includes:
carrying out digit recognition on the license plate image characteristics to obtain clear license plate digits of the vehicle to be recognized;
and outputting the clear license plate number of the vehicle to be recognized.
By adopting the technical scheme, related workers can know the license plate information of the vehicle to be recognized more visually.
In a possible implementation manner, the performing digit recognition on the license plate image features to obtain a clear license plate digit of the vehicle to be recognized includes:
performing digit recognition on the license plate image characteristics through a license plate clear digit recognition model to obtain clear license plate digits of the vehicle to be recognized;
or determining the clear license plate digit of the vehicle to be recognized according to the license plate number of the vehicle to be recognized.
By adopting the technical scheme, the number recognition is carried out by inputting the characteristics of the license plate image into the license plate clear number model, so that the clear license plate number of the vehicle to be recognized is obtained, or the clear license plate number of the vehicle to be recognized is determined by counting the license plate number of the vehicle to be recognized, so that the related workers can know the license plate information more visually.
In one possible implementation manner, the method further includes:
acquiring a license plate training sample, wherein the license plate training sample comprises a plurality of license plate training images and standard labels corresponding to the license plate training images;
acquiring a license plate state recognition model to be trained and a license plate number recognition model to be trained;
extracting the feature of each license plate training image by using the convolutional neural network;
and training the license plate state recognition model to be trained and the license plate number recognition model to be trained respectively according to the license plate training image characteristics to obtain the license plate state recognition model and the license plate number recognition model.
By adopting the technical scheme, a large number of training samples are input into the license plate state recognition model to be trained and the license plate number recognition model to be trained, the characteristics of the training samples are obtained firstly, the obtained image characteristics are input into the license plate state recognition model to be trained and the license plate number recognition model to be trained to obtain the license plate state recognition model and the license plate number recognition model, the trained models are obtained by training the models firstly, the images to be recognized can be directly input into the trained models in the application stage, the recognition results can be obtained, and the practicability of the models is enhanced.
In a possible implementation manner, training a license plate state recognition model to be trained according to each license plate training image feature to obtain a license plate state recognition model includes:
inputting the feature of each license plate training image into the license plate state recognition model to be trained to obtain each license plate training state;
obtaining a function loss value by utilizing a cross entropy function according to each license plate training state and each corresponding standard label;
if the function loss value is within a preset loss value range, determining to obtain the license plate state recognition model;
and if the function loss value is not in the preset loss value range, performing iterative training on the license plate state recognition model to be trained according to the function loss value and the license plate training sample until the function loss value is in the preset loss value range.
By adopting the technical scheme, the license plate training state recognition model to be trained is input with the feature of each license plate training image to obtain each license plate training state, the loss value is obtained by utilizing the cross entropy function according to the training state and each license plate standard label, when the loss value is larger than the preset loss value range, the model is expressed to be required to be subjected to iterative training, when the obtained loss value is in the preset loss value range, the license plate state recognition model is determined to be trained, the loss value is calculated through the cross entropy function, relevant workers can know the degree of model training more visually, and the accuracy of the model training is improved through the iterative training.
In a second aspect, the present application provides a license plate recognition device, which adopts the following technical scheme:
a license plate recognition device comprises
The acquisition module is used for acquiring a license plate image to be recognized of a vehicle to be recognized;
the extracting module is used for extracting the license plate image characteristics of the license plate image to be recognized by utilizing a convolutional neural network;
the state recognition module is used for carrying out license plate state recognition on the license plate image characteristics through a license plate state recognition model to obtain the license plate state of the vehicle to be recognized, wherein the license plate state at least comprises illegal shielding, non-illegal shielding and no shielding;
the number recognition module is used for carrying out license plate number recognition on the license plate image characteristics through a license plate number recognition model to obtain the license plate number of the vehicle to be recognized;
and the output result module is used for outputting a license plate recognition result, and the recognition result comprises the license plate state of the vehicle to be recognized and the license plate number of the vehicle to be recognized.
By adopting the technical scheme, after the license plate image to be recognized of the vehicle to be recognized is obtained, the license plate image characteristics of the license plate image to be recognized are extracted through the convolutional neural network, the characteristics of the license plate image to be recognized are recognized through the license plate state recognition model, and the license plate state is obtained, wherein the license plate state at least comprises illegal shielding, non-illegal shielding and no shielding, the license plate image characteristics to be recognized are recognized through the license plate number recognition model, the license plate number of the vehicle to be recognized is obtained, finally the license plate state and the license plate number of the vehicle to be recognized are output together, the recognition result is more visual, meanwhile, related workers do not need to further judge whether the behavior of illegally shielding the license plate exists through the recognized license plate number, and the work burden of the related workers is reduced.
In a third aspect, the present application provides an electronic device, which adopts the following technical solutions:
an electronic device, comprising:
at least one processor;
a memory;
at least one application, wherein the at least one application is stored in the memory and configured to be executed by the at least one processor, the at least one application configured to: the license plate recognition method is executed.
In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solutions:
a computer-readable storage medium, comprising: a computer program is stored which can be loaded by a processor and which performs the above-described license plate recognition method.
In summary, the present application includes at least one of the following beneficial technical effects:
1. after a to-be-recognized license plate image of a to-be-recognized vehicle is obtained, license plate image features of the to-be-recognized license plate image are extracted through a convolutional neural network, the features of the to-be-recognized license plate image are recognized through a license plate state recognition model to obtain a license plate state, wherein the license plate state at least comprises illegal shielding, non-illegal shielding and no shielding, the license plate features of the to-be-recognized license plate image are recognized through the license plate number recognition model to obtain a license plate number of the to-be-recognized vehicle, the license plate state of the to-be-recognized vehicle and the license plate number are finally output together, a recognition result is more visual, meanwhile, related workers do not need to further judge whether illegal license plate shielding behaviors exist through the recognized license plate number, and work burden of related workers is relieved.
2. The method comprises the steps of obtaining an initial to-be-recognized image only containing a license plate image by cutting a vehicle image, and obtaining a final ground recognition license plate image by carrying out image processing on the obtained initial to-be-recognized license plate image, wherein the image processing mode at least comprises any one or more of the following modes: graying, binaryzation, expansion processing and corrosion, and the license plate image to be recognized after image processing is beneficial to feature extraction.
3. The clear license plate digits of the vehicle to be recognized are obtained by inputting the license plate image features into the license plate clear digit model for digit recognition, or the clear license plate digits of the vehicle to be recognized are determined by counting the license plate numbers of the vehicle to be recognized, so that related workers can know license plate information more visually.
Drawings
Fig. 1 is a schematic flowchart of a license plate recognition method in an embodiment of the present application;
FIG. 2 is a schematic structural diagram of a license plate recognition device according to an embodiment of the present disclosure;
fig. 3 is a schematic structural diagram of an electronic device in an embodiment of the present application.
Detailed Description
The present application is described in further detail below with reference to figures 1-3.
A person skilled in the art, after reading the present specification, may make modifications to the present embodiments as necessary without inventive contribution, but only within the scope of the claims of the present application are protected by patent laws.
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
In order to provide a more intuitive recognition result for relevant workers and reduce the burden of the relevant workers, after the license plate image to be recognized of the vehicle to be recognized is acquired by the embodiment of the application, extracting the license plate picture characteristics of the license plate picture to be recognized through a convolutional neural network, recognizing the characteristics of the license plate picture to be recognized through a license plate state recognition model to obtain the license plate state, wherein the license plate state at least comprises illegal shielding, non-illegal shielding and no shielding, the license plate number recognition model is used for recognizing the characteristics of the license plate image to be recognized to obtain the license plate number of the vehicle to be recognized, and finally the license plate state and the license plate number of the vehicle to be recognized are output together, the recognition result is more intuitive, meanwhile, related workers do not need to further judge whether the license plate is illegally shielded through the recognized license plate number, and the workload of the related workers is reduced.
Specifically, the embodiment of the application provides a license plate identification method, which is executed by an electronic device, where the electronic device may be a server or a terminal device, where the server may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing cloud computing services. The terminal device may be a smart phone, a tablet computer, a notebook computer, a desktop computer, and the like, but is not limited thereto, and the terminal device and the server may be directly or indirectly connected through wired or wireless communication, and the embodiment of the present application is not limited thereto.
Referring to fig. 1, fig. 1 is a schematic flowchart of a license plate recognition method in an embodiment of the present application, where the method includes step S110, step S120, step S130, step S140, and step S150, where:
step S110: and acquiring a license plate image to be recognized of the vehicle to be recognized.
Specifically, a plurality of image acquisition devices are arranged in the area to be recognized, so that the image acquisition devices can conveniently acquire the license plate images of the vehicles in the area to be recognized, and the acquired images can be acquired by the electronic device. The area to be identified may be a parking lot, a toll station, a traffic light intersection, or the like, and the vehicle to be identified is each vehicle passing through the area to be identified. Wherein, the image acquisition equipment can be any one of the following: integration camera, spherical camera, infrared camera, waterproof type camera.
Step S120: and extracting the license plate image characteristics of the license plate image to be recognized by using the convolutional neural network.
Specifically, a Convolutional Neural Network (CNN) is a feed-forward Neural network with a depth structure and including convolution calculation, a license plate image to be recognized is input into the Convolutional Neural network and subjected to convolution and pooling for a plurality of times to obtain license plate image features of the license plate image to be recognized, wherein convolution layer parameters include convolution kernel size, step length and padding, the size of the convolution layer output feature image is determined by the convolution kernel size, the step length and the padding, a user can set the convolution kernel size, the step length and the padding in a user-defined mode, after feature extraction is performed on the convolution layer, the output features are transmitted to a pooling layer for feature selection and information filtering, and finally the license plate image features of the license plate image to be recognized are obtained.
Step S130: and (3) carrying out license plate state recognition on the license plate image characteristics through a license plate state recognition model to obtain the license plate state of the vehicle to be recognized, wherein the license plate state at least comprises illegal shielding, non-illegal shielding and no shielding.
Specifically, the license plate state recognition model is obtained by training according to the characteristics of each license plate training image, wherein the characteristics of each license plate training image are obtained by extracting each license plate training image by using a convolutional neural network. The license plate state is used for representing the state of the surface of the license plate, for example, the collected license plate surface has the states of being shielded by leaves, reflected light, wires, safety helmets and the like, and belongs to non-illegal shielding; and if the surface of the license plate is smeared, damaged, covered and the like, the license plate belongs to illegal shielding.
Specifically, the license plate state recognition model comprises a convolution layer and a full connection layer, wherein the license plate state recognition model can change a characteristic channel value of a license plate image to be recognized through a convolution neural network, the modified channel value is multiplied by the license plate image characteristic to obtain the number of all neurons, all the neurons pass through the full connection layer and then output final neurons representing the license plate state, the number of the final neurons corresponds to the type number of the license plate state, the number of the final neurons is not specifically limited, and the final neurons can be modified according to self requirements.
For example, in the scheme, the channel value of the license plate map feature is 512, the channel value is changed into 128 after the convolution neural network passes through, the changed channel value 128 is multiplied by the license plate map feature 26 to obtain 3328 neurons, and since the types of the license plate map state to be identified include illegal occlusion, non-illegal occlusion and no occlusion, 3328 neurons pass through the full connection layer to obtain 3 neurons which finally represent the license plate state. And inputting the characteristic value of the license plate image to be recognized into the full-connection layer to obtain a corresponding neuron, comparing the neuron with a final neuron representing the license plate state, and determining the state type corresponding to the final neuron with the maximum similarity value as the state type of the license plate to be recognized.
Step S140: and identifying the license plate number of the vehicle to be identified by the license plate number identification model according to the license plate number characteristics to obtain the license plate number of the vehicle to be identified.
Specifically, the license plate number recognition model is obtained by training according to the characteristics of each license plate training image, wherein the characteristics of each license plate training image are obtained by extracting each license plate training image by using a convolutional neural network.
Specifically, the license plate number recognition model comprises a convolution cyclic neural network, and the convolution cyclic neural network is used for solving the problem of image-based sequence recognition, particularly the problem of scene character recognition. License plate number information is determined by comparing license plate map features with all character types, and for some locations without characters, it is denoted by "-".
Step S150: and outputting a license plate recognition result, wherein the recognition result comprises the license plate state of the vehicle to be recognized and the license plate number of the vehicle to be recognized.
Specifically, the final output recognition result includes the license plate state of the vehicle to be recognized and the license plate number of the vehicle to be recognized, and the output result is convenient for relevant workers to check, for example, the output result is: cyan Q98789 has no occlusion.
According to the embodiment of the application, after the license plate image to be recognized of the vehicle to be recognized is obtained, the license plate image characteristics of the license plate image to be recognized are extracted through the convolutional neural network, the characteristics of the license plate image to be recognized are recognized through the license plate state recognition model, the license plate state is obtained, the license plate state at least comprises illegal shielding, non-illegal shielding and no shielding, the license plate image characteristics to be recognized are recognized through the license plate number recognition model, the license plate number of the vehicle to be recognized is obtained, finally, the license plate state and the license plate number of the vehicle to be recognized are output together, the recognition result is more visual, meanwhile, related workers do not need to further judge whether behaviors of illegally shielding the license plate exist through the recognized license plate number, and the work burden of the related workers is reduced.
In order to reduce the possibility of incomplete cropping of a license plate to be recognized, an embodiment of the present application provides a license plate recognition method, where the step S110 of obtaining a license plate image to be recognized of a vehicle to be recognized may specifically include: step S1101 (not shown in the drawings), step S1102 (not shown in the drawings), step S1103 (not shown in the drawings), wherein:
step S1101: and acquiring an image of the vehicle to be identified.
Specifically, an image of the vehicle to be identified is acquired through an image acquisition device, and the image acquisition device captures images of a static or moving vehicle.
Step S1102: and cutting a license plate image area of the vehicle image to be recognized to obtain an initial license plate image to be recognized of the vehicle to be recognized.
Specifically, because the image acquisition equipment captures images of static or moving vehicles, license plates may only occupy a small part of the images of the vehicles, and are not beneficial to identifying license plate information.
Step S1103: sequentially carrying out image processing on the initial license plate image to be recognized to obtain the license plate image to be recognized, wherein the image processing mode at least comprises any one or more of the following modes: graying, binaryzation, expansion treatment and corrosion.
Specifically, after graying, binaryzation, expansion processing and corrosion are carried out on the initial license plate image to be recognized, extraction of image features is facilitated, and the license plate image to be recognized is obtained after the initial license plate image to be recognized is subjected to image processing. Wherein, the banding distortion can be avoided after the image is grayed; the image binarization can set the gray value of a pixel point on the initial license plate image to be recognized as 0 or 255, namely the whole initial license plate image to be recognized shows an obvious black-and-white effect, and the binarization greatly reduces the data volume in the initial license plate image to be recognized, so that the outline of the license plate number can be highlighted; filling the edge of the license plate number or the inner pit by image expansion processing, and expanding the edge of the original license plate image to be recognized; the image corrosion is that the highlight part in the initial license plate image to be recognized is corroded, the field is reduced, the effect image has a highlight area smaller than that of the initial license plate image to be recognized, the adjacent area is replaced by a minimum value during operation, and the highlight area is reduced.
According to the embodiment of the application, an initial to-be-recognized image only containing a license plate image is obtained by cutting a vehicle image, the obtained initial to-be-recognized license plate image is subjected to image processing to obtain a final to-be-recognized license plate image, and the image processing mode at least comprises any one or more of the following modes: graying, binaryzation, expansion processing and corrosion, and the license plate image to be recognized after image processing is beneficial to feature extraction.
In order to reduce the probability of incomplete cropping in the license plate image to be recognized, the embodiment of the present application provides a license plate recognition method, where the step S1102 is to crop the license plate image region of the vehicle image to be recognized, so as to obtain an initial license plate image to be recognized of the vehicle to be recognized, and the method specifically includes: step S1102a (not shown in the drawings), step S1102b (not shown in the drawings), step 1102c (not shown in the drawings), wherein:
step S1102 a: and determining a license plate image region from the vehicle image to be recognized.
Specifically, a cutting tool is used for obtaining a key area image of the vehicle to be recognized in the vehicle image, wherein the key area image can be a head image of the vehicle to be recognized or a tail image of the vehicle to be recognized, and the vertex coordinates of the license plate image are obtained from the license plate image of the key area by using a target detection algorithm, so that a license plate image area in the vehicle image to be recognized is obtained.
Step S1102 b: and expanding the edges of the license plate image region according to a preset proportion to obtain an expanded license plate image.
Specifically, according to the obtained license plate image area in the vehicle image to be recognized, the edge of the area is subjected to edge expansion according to a preset proportion, so that the edge of the license plate image is expanded, and the preset proportion can be modified according to requirements.
Step S1102 c: and determining the expanded license plate image as an initial license plate image to be recognized.
Specifically, after the license plate image area is subjected to edge expanding processing, cutting is carried out according to the outline of the edge expanding, and an initial license plate image to be recognized is obtained.
According to the method and the device, the edge of the license plate image area is expanded by expanding the license plate to be recognized according to the preset proportion, the expanded license plate image is cut to obtain the initial license plate image to be recognized, and the probability of incomplete image cutting in the image cutting process is reduced.
In order to effectively distinguish unclear license plates from shielded license plates in the license plate identification process, an embodiment of the present application provides a license plate identification method, and in step S140, the license plate number identification is performed on the license plate image features through a license plate number identification model to obtain the license plate number of the vehicle to be identified, which may specifically include: step S1401 (not shown in the drawings), step S1402 (not shown in the drawings), step S1403 (not shown in the drawings), in which:
step S1401: and performing initial recognition of the license plate number according to the license plate image characteristics of the vehicle to be recognized to obtain a character recognition result of the license plate to be recognized, wherein the character recognition result comprises each recognition character, unrecognized characters and respective corresponding positions.
Specifically, the license plate image of the vehicle to be recognized can be divided into a plurality of areas according to the license plate image characteristics of the vehicle to be recognized, each divided area is compared with all character types to determine characters, and all the character types comprise '0123456789 abcdefghjkmnpkntqrstywxyz, jinji, jin, jeao, heighu, su, thunberg, min, xi, yin, Guangxi, Shanxi, Ganling, new Gui wall police, a first special symbol and a second special symbol', and the character recognition result of the license plate to be recognized is obtained through comparison.
Step S1402: an unclear type character in the unrecognized character is replaced with a first character, and an occlusion type character in the unrecognized character is replaced with a second character.
Specifically, according to the obtained character result of the license plate to be recognized, the unrecognized characters are classified, if the unrecognized characters have edge outlines, the unrecognized characters are determined to be of an unclear type and are replaced by first special symbols, if the unrecognized characters do not have character outlines, the unrecognized characters are determined to be characters of a shielding type and are replaced by second special symbols, wherein the first special symbols and the second special symbols are not specifically limited, and the unrecognized characters and the shielding type can be distinguished. Preferably, the first special symbol is "+" and the second special symbol is "+".
Step S1403: and updating the character recognition result according to the substitution result to obtain the license plate number.
Specifically, the substitution result at least includes any one or more of the following: the method comprises the steps that recognized characters, first special symbols and second special symbols are obtained, character recognition results comprise recognized characters, unrecognized characters and corresponding positions of the recognized characters and the unrecognized characters, and the unrecognized characters in the character recognition results are replaced by the first special symbols or the second special symbols through replacing results, so that the license plate number is obtained.
According to the license plate number recognition method and device, before license plate number recognition is carried out, initial recognition of license plate numbers is carried out on license plate image features of vehicles to be recognized, recognition results comprise recognition characters, unrecognized characters and corresponding positions, the types of the unrecognized characters are distinguished, unclear character types are replaced by first special symbols, shielded character types are replaced by second special symbols, initial recognition results of license plates to be recognized are updated according to the replacement results, the license plate numbers are obtained, the license plate numbers at least comprise recognized characters, first special symbols and second special symbols, and accordingly the recognition effects of the license plate numbers are clearer.
In order to facilitate related workers to more intuitively know the license plate information of the vehicle to be recognized, the application provides a license plate recognition method, which includes step S1 (not shown in the attached drawings) and step S2 (not shown in the attached drawings), wherein:
step S1: and carrying out digit recognition on the license plate image characteristics to obtain clear license plate digits of the vehicle to be recognized.
Specifically, the license plate image to be recognized is divided into a plurality of areas according to the characteristics of the license plate image to be recognized, the number of the divided areas is the same as the characteristics of the license plate image to be recognized, the divided areas are respectively matched with the types of characters to obtain recognition characters, and the clear number of the license plate of the vehicle to be recognized is determined by counting the number of the recognition characters.
Step S2: and outputting the clear license plate number of the vehicle to be recognized.
Specifically, clear license plate digit is output, and the output result can be displayed through a display device, so that related workers can further know license plate information of the vehicle to be recognized.
And further, carrying out digit recognition on the license plate image characteristics to obtain the clear license plate digits of the vehicle to be recognized, and in a possible implementation mode, carrying out digit recognition on the license plate image characteristics through a license plate clear digit recognition model to obtain the clear license plate digits of the vehicle to be recognized.
Specifically, the characteristics of a license plate image to be recognized of a vehicle to be recognized are input into a license plate clear digit recognition model, wherein the license plate clear digit recognition model comprises a full connection layer, the input characteristics of the license plate image to be recognized of the vehicle to be recognized are convoluted through the full connection layer, a characteristic channel value of the license plate to be recognized is changed, the modified channel value is multiplied by the characteristic value to obtain the number of all neurons, all the neurons pass through the full connection layer, the final neurons are output, the number of the final neurons is not specifically limited, the final neurons can be modified according to self requirements, the number of the final neurons corresponds to the clear number of license plate digits, and the clear number of license plates is less than or equal to the total length of the license plates.
In another possible implementation manner, the clear license plate digit of the vehicle to be recognized is determined according to the license plate number of the vehicle to be recognized.
Specifically, the number of special symbols in the output license plate number is counted, and the number of the special symbols is subtracted from the fixed length of the license plate number to obtain a clear digit.
In the embodiment of the application, the number recognition is carried out by inputting the characteristics of the license plate image into the clear number model of the license plate, so that the clear number of the license plate of the vehicle to be recognized is obtained, or the clear number of the license plate of the vehicle to be recognized is determined by counting the number of the license plate of the vehicle to be recognized, so that the related workers can know the license plate information more visually.
In the embodiment of the present application, the training process of the license plate state and license plate number recognition model includes steps Sa (not shown in the drawings), Sb (not shown in the drawings), Sc (not shown in the drawings), and Sd (not shown in the drawings), wherein:
sa: and obtaining a license plate training sample, wherein the license plate training sample comprises a plurality of license plate training images and standard labels corresponding to the license plate training images.
Specifically, the license plate training samples comprise a large number of license plate training images, the license plate training images are manually labeled and correspond to respective standard labels, and the standard labels are labeled with the license plate states of each vehicle according to the license plate images of the vehicles, such as illegal shielding, illegal shielding and no shielding.
And Sb: and acquiring a license plate state recognition model to be trained and a license plate number recognition model to be trained.
Specifically, the license plate state recognition model to be trained comprises a convolutional neural network and a state classifier; the license plate number recognition model to be trained comprises a convolution neural network and a circulation neural network.
And step Sc, extracting the characteristics of each license plate training image by using a convolutional neural network.
Specifically, the convolutional neural network inputs the license plate training image features into the convolutional neural network, and the license plate image features of the training images are obtained through multiple convolution and pooling.
Step Sd: and training a license plate state recognition model to be trained and a license plate number recognition model to be trained respectively according to the license plate training image characteristics to obtain a license plate state recognition model and a license plate number recognition model.
Specifically, the extracted license plate image features of the training image are input into a to-be-trained license plate state recognition model and a to-be-trained license plate number recognition model for iterative training to obtain a trained license plate state recognition model and a trained license plate number recognition model, wherein the trained license plate state recognition model and the trained license plate number recognition model can output corresponding license plate states and license plate numbers according to the randomly input license plate image features.
In the embodiment of the application, a large number of training samples are input into the license plate state recognition model to be trained and the license plate number recognition model to be trained, the characteristics of the training samples are firstly obtained, the obtained image characteristics are input into the license plate state recognition model to be trained and the license plate number recognition model to be trained for training, the license plate state recognition model and the license plate number recognition model are obtained, the model is trained firstly, the trained model is obtained, the image to be recognized can be directly input into the trained model in the application stage, the recognition result can be obtained, and the practicability of the model is enhanced.
In the embodiment of the application, the license plate state recognition model to be trained is trained according to the feature of each license plate training image in step Sd to obtain the license plate state recognition model, which specifically includes steps Sd1 (not shown in the drawings), Sd2 (not shown in the drawings), Sd3 (not shown in the drawings), and Sd4 (not shown in the drawings), wherein:
sd 1: and inputting the characteristics of each license plate training image into a license plate state recognition model to be trained to obtain each license plate training state.
Specifically, license plate training image features are obtained after convolution and pooling for a plurality of times through a convolution neural network, the features are input into a license plate state recognition model to be trained for training, wherein the training state of the license plate is an output recognition result, and the recognition result comprises illegal shielding, illegal shielding and no shielding.
Step Sd 2: and obtaining a function loss value by utilizing a cross entropy function according to each license plate training state and each corresponding standard label.
Specifically, the real sample labels are 0,1 and 2, represent three occlusion types, correspond to the last 3 neurons in the network, obtain a loss value through a cross entropy loss function, and reversely propagate and update network parameters. Positive and negative classes: 3 neurons correspond to 3 classes, positive class 1 negative class 0. Three neurons and only 1 is positive. For example, (0.1,0.5, 0.3), the real label is 1, the positive and negative labels in the cross entropy loss are (0, 1, 0), the cross entropy loss is (0-0) log (|0-0.1|) + 0log (|0.1|) + (1-0) log (|1-0.5|) + 0log (|0.5|) + (2-0) log (|2-0.3|) + 0log (|0.3|), the full connection layer is used to distinguish 3 neurons of the shielding type, the probability that the sample corresponds to 3 classes can be obtained after normalization, wherein the maximum value corresponds to the index of the neuron, which is the prediction label of the sample.
Step Sd 3: and if the function loss value is within the preset loss value range, determining to obtain a license plate state recognition model.
Specifically, the preset loss value can be modified according to the user requirement, the output prediction result is compared with the real sample to obtain the loss value, when the loss value is within the preset loss index range, the license plate state recognition model is determined to be obtained, and the smaller the loss value is, the more excellent the performance of the obtained license plate state recognition model is.
Step Sd 4: and if the function loss value is not in the preset loss value range, performing iterative training on the license plate state recognition model to be trained according to the function loss value and the license plate training sample until the function loss value is in the preset loss value range.
Specifically, if the loss value of the output prediction result is outside the preset loss value range compared with the real sample, it is indicated that the training of the license plate state recognition model is not completed. And if the difference between the prediction result output by the training for multiple times and the standard value is larger, continuing to carry out repeated training for multiple times, namely, carrying out iterative training until the loss value is within the preset loss value range.
According to the embodiment of the application, each license plate training state is obtained by inputting each license plate training image feature into a to-be-trained license plate state recognition model, a loss value is obtained by using a cross entropy function according to the training state and each license plate standard label, when the loss value exceeds a preset loss value range, the model is represented to be subjected to iterative training, when the obtained loss value is within the preset loss value range, the license plate state recognition model is determined to be trained completely, the loss value is calculated through the cross entropy function, relevant workers can know the degree of model training more intuitively, and the accuracy of the model training is improved beneficially through the iterative training.
The embodiments described above introduce a license plate recognition method from the perspective of a method flow, and the following embodiments describe a license plate recognition device from the perspective of a virtual module or a virtual unit, which are described in detail in the following embodiments.
Referring to fig. 2, fig. 2 is a schematic structural diagram of a license plate recognition device in an embodiment of the present application, including: an obtaining module 210, an extracting module 220, a state identifying module 230, a number identifying module 240, and an outputting result module 250, wherein:
the obtaining module 210 is configured to obtain a license plate map to be recognized of a vehicle to be recognized;
the extracting module 220 is configured to extract a license plate image feature of a license plate image to be recognized by using a convolutional neural network;
the state recognition module 230 is configured to perform license plate state recognition on the license plate image features through a license plate state recognition model to obtain a license plate state of the vehicle to be recognized, where the license plate state at least includes illegal occlusion, non-illegal occlusion, and no occlusion;
the number recognition module 240 is used for recognizing the license plate number of the license plate image through the license plate number recognition model to obtain the license plate number of the vehicle to be recognized;
and the output result module 250 is configured to output a license plate recognition result, where the recognition result includes a license plate state of the vehicle to be recognized and a license plate number of the vehicle to be recognized.
In one possible implementation manner, the obtaining module 210 includes:
the image acquisition unit is used for acquiring an image of the vehicle to be identified;
the cutting unit is used for cutting the license plate image area of the vehicle image to be recognized to obtain an initial license plate image to be recognized of the vehicle to be recognized;
the processing unit is used for sequentially carrying out image processing on the initial license plate image to be recognized to obtain the license plate image to be recognized, wherein the image processing mode at least comprises any one or more of the following modes: graying, binaryzation, expansion treatment and corrosion.
In one possible implementation, the cutting unit comprises:
the determining subunit is used for determining a license plate image region from the vehicle image to be recognized;
the edge expanding subunit is used for expanding the edge of the license plate image region according to a preset proportion to obtain an edge expanded license plate image;
and the determining subunit is used for determining the expanded license plate image as an initial license plate image to be recognized.
In one possible implementation, the number identification module includes:
the license plate number identification device comprises a license plate number identification determining unit, a license plate number identification determining unit and a license plate number identification determining unit, wherein the license plate number identification determining unit is used for initially identifying a license plate number according to the license plate image characteristics of a vehicle to be identified to obtain a character identification result of the license plate to be identified, and the character identification result comprises each identification character, an unrecognized character and a position corresponding to each identification character;
a symbol replacing unit for replacing an unclear type character in the unrecognized character with a first character and replacing an occlusion type character in the unrecognized character with a second character;
and the result updating unit is used for updating the character recognition result according to the replacing result so as to obtain the license plate number.
In one possible implementation manner, the method further includes:
a digit identification module: the digital recognition is carried out on the license plate image characteristics to obtain the clear license plate number of the vehicle to be recognized;
and the output digit module is used for outputting clear license plate digits of the vehicle to be recognized.
In one possible implementation, the bit number recognition module includes:
the first recognition unit is used for carrying out digit recognition on the characteristics of the license plate image through a license plate clear digit recognition model to obtain clear license plate digits of the vehicle to be recognized;
and the second identification unit is used for determining the clear license plate number of the vehicle to be identified according to the license plate number of the vehicle to be identified.
In one possible implementation manner, the method further includes:
the license plate training system comprises an acquisition sample module, a storage module and a display module, wherein the acquisition sample module is used for acquiring a license plate training sample, and the license plate training sample comprises a plurality of license plate training images and standard labels corresponding to the license plate training images;
the acquisition model module is used for acquiring a license plate state recognition model to be trained and a license plate number recognition model to be trained;
the training image characteristic acquisition module is used for extracting the characteristics of each license plate training image by using a convolutional neural network;
and the model determining module is used for respectively training the license plate state recognition model to be trained and the license plate number recognition model to be trained according to the license plate training image characteristics to obtain the license plate state recognition model and the license plate number recognition model.
In one possible implementation, the model determination module includes:
the license plate state recognition model comprises a training state acquisition unit, a training state recognition unit and a training state recognition unit, wherein the training state acquisition unit is used for inputting the characteristics of each license plate training image into a license plate state recognition model to be trained to obtain each license plate training state;
the loss value obtaining unit is used for obtaining a function loss value by utilizing a cross entropy function according to each license plate training state and each corresponding standard label;
the model determining unit is used for determining to obtain a license plate state recognition model if the function loss value is within a preset loss value range;
and the iterative training unit is used for performing iterative training on the license plate state recognition model to be trained according to the function loss value and the license plate training sample if the function loss value is not in the preset loss value range until the function loss value is in the preset loss value range.
The following embodiments provide an electronic device, and are in part consistent with the above method, and will be described in detail in the following embodiments.
An electronic device is provided in an embodiment of the present application, as shown in fig. 3, fig. 3 is a schematic structural diagram of an electronic device in an embodiment of the present application, and an electronic device 30 shown in fig. 3 includes: a processor 301 and a memory 303. Wherein processor 301 is coupled to memory 303, such as via bus 302. Optionally, the electronic device 30 may also include a transceiver 304. It should be noted that the transceiver 304 is not limited to one in practical applications, and the structure of the electronic device 30 is not limited to the embodiment of the present application.
The Processor 301 may be a CPU (Central Processing Unit), a general-purpose Processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array) or other Programmable logic device, a transistor logic device, a hardware component, or any combination thereof. Which may implement or perform the various illustrative logical blocks, modules, and circuits described in connection with the disclosure. The processor 301 may also be a combination of computing functions, e.g., comprising one or more microprocessors, a combination of a DSP and a microprocessor, or the like.
The Memory 303 may be a ROM (Read Only Memory) or other type of static storage device that can store static information and instructions, a RAM (Random Access Memory) or other type of dynamic storage device that can store information and instructions, an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical Disc storage, optical Disc storage (including Compact Disc, laser Disc, optical Disc, digital versatile Disc, blu-ray Disc, etc.), a magnetic Disc storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited to these.
The memory 303 is used for storing application program codes for executing the scheme of the application, and the processor 301 controls the execution. The processor 301 is configured to execute application program code stored in the memory 303 to implement the aspects illustrated in the foregoing method embodiments.
Among them, electronic devices include but are not limited to: mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and the like, and fixed terminals such as digital TVs, desktop computers, and the like. But also a server, etc. The electronic device shown in fig. 3 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.
The present application provides a computer-readable storage medium, on which a computer program is stored, which, when running on a computer, enables the computer to execute the corresponding content in the foregoing method embodiments. Compared with the prior art, in the embodiment of the application, after the license plate image to be recognized of the vehicle to be recognized is obtained, the license plate image characteristics of the license plate image to be recognized are extracted through the convolutional neural network, the characteristics of the license plate image to be recognized are recognized through the license plate state recognition model, and the license plate state is obtained, wherein the license plate state at least comprises illegal shielding, non-illegal shielding and no shielding, the license plate image characteristics to be recognized are recognized through the license plate number recognition model, the license plate number of the vehicle to be recognized is obtained, finally, the license plate state of the vehicle to be recognized and the license plate number are output together, the recognition result is more visual, meanwhile, related workers do not need to further judge whether the license plate illegal shielding behaviors exist through the recognized license plate number, and the work burden of the related workers is reduced.
It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and may be performed in other orders unless explicitly stated herein. Moreover, at least a portion of the steps in the flow chart of the figure may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed alternately or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
The foregoing is only a partial embodiment of the present application, and it should be noted that, for those skilled in the art, several modifications and decorations can be made without departing from the principle of the present application, and these modifications and decorations should also be regarded as the protection scope of the present application.
Claims (11)
1. A license plate recognition method is characterized by comprising the following steps:
acquiring a license plate image to be recognized of a vehicle to be recognized;
extracting the license plate image characteristics of the license plate image to be recognized by using a convolutional neural network;
carrying out license plate state recognition on the license plate image characteristics through a license plate state recognition model to obtain the license plate state of the vehicle to be recognized, wherein the license plate state at least comprises illegal shielding, non-illegal shielding and no shielding;
identifying the license plate number of the vehicle to be identified by the license plate number identification model according to the license plate number characteristics;
and outputting a license plate recognition result, wherein the recognition result comprises the license plate state of the vehicle to be recognized and the license plate number of the vehicle to be recognized.
2. The license plate recognition method of claim 1, wherein the obtaining of the license plate image to be recognized of the vehicle to be recognized comprises:
acquiring a vehicle image to be identified;
cutting a license plate image area of the vehicle image to be recognized to obtain an initial license plate image to be recognized of the vehicle to be recognized;
sequentially carrying out image processing on the initial license plate image to be recognized to obtain the license plate image to be recognized, wherein the image processing mode at least comprises any one or more of the following modes: graying, binaryzation, expansion treatment and corrosion.
3. The license plate recognition method of claim 2, wherein the step of cropping the license plate image area of the vehicle image to be recognized to obtain an initial license plate image to be recognized of the vehicle to be recognized comprises the steps of:
determining a license plate image region from the vehicle image to be recognized;
expanding the license plate image area according to a preset proportion to obtain an expanded license plate image;
and determining the expanded license plate image as an initial license plate image to be recognized.
4. The license plate recognition method of claim 1, wherein the license plate number recognition of the license plate image features through a license plate number recognition model to obtain the license plate number of the vehicle to be recognized comprises:
performing initial recognition of the license plate number according to the license plate image characteristics of the vehicle to be recognized to obtain a character recognition result of the license plate to be recognized, wherein the character recognition result comprises each recognized character, unrecognized characters and respective corresponding positions;
replacing an unclear type character in the unrecognized character with a first character, and replacing an occlusion type character in the unrecognized character with a second character;
and updating the character recognition result according to the substitution result to obtain the license plate number.
5. The license plate recognition method of claim 1, further comprising:
carrying out digit recognition on the license plate image characteristics to obtain clear license plate digits of the vehicle to be recognized;
and outputting the clear license plate number of the vehicle to be recognized.
6. The license plate recognition method of claim 5, wherein the performing bit recognition on the license plate image features to obtain the clear license plate bits of the vehicle to be recognized comprises:
performing digit recognition on the license plate image characteristics through a license plate clear digit recognition model to obtain clear license plate digits of the vehicle to be recognized;
or determining the clear license plate digit of the vehicle to be recognized according to the license plate number of the vehicle to be recognized.
7. The license plate recognition method of claim 1, further comprising:
acquiring a license plate training sample, wherein the license plate training sample comprises a plurality of license plate training images and standard labels corresponding to the license plate training images;
acquiring a license plate state recognition model to be trained and a license plate number recognition model to be trained;
extracting the license plate training image characteristics of each license plate training image by using the convolutional neural network;
and training the license plate state recognition model to be trained and the license plate number recognition model to be trained respectively according to the license plate training image characteristics to obtain the license plate state recognition model and the license plate number recognition model.
8. The license plate recognition method of claim 7, wherein training a license plate state recognition model to be trained according to the license plate training image features to obtain a license plate state recognition model, comprises:
inputting the feature of each license plate training image into the license plate state recognition model to be trained to obtain the training state of each license plate;
obtaining a function loss value by utilizing a cross entropy function according to each license plate training state and each corresponding standard label;
if the function loss value is within a preset loss value range, determining to obtain the license plate state recognition model;
and if the function loss value is not in the preset loss value range, performing iterative training on the license plate state recognition model to be trained according to the function loss value and the license plate training sample until the function loss value is in the preset loss value range.
9. A license plate recognition device, comprising:
the acquisition module is used for acquiring a license plate image to be recognized of a vehicle to be recognized;
the extracting module is used for extracting the license plate image characteristics of the license plate image to be recognized by utilizing a convolutional neural network;
the state recognition module is used for carrying out license plate state recognition on the license plate image characteristics through a license plate state recognition model to obtain the license plate state of the vehicle to be recognized, wherein the license plate state at least comprises illegal shielding, non-illegal shielding and no shielding;
the number recognition module is used for carrying out license plate number recognition on the license plate image characteristics through a license plate number recognition model to obtain the license plate number of the vehicle to be recognized;
and the output result module is used for outputting a license plate recognition result, and the recognition result comprises the license plate state of the vehicle to be recognized and the license plate number of the vehicle to be recognized.
10. An electronic device, comprising:
at least one processor;
a memory;
at least one application, wherein the at least one application is stored in the memory and configured to be executed by the at least one processor, the at least one application configured to: a method of performing license plate recognition according to any one of claims 1-8.
11. A computer-readable storage medium, comprising: a computer program which can be loaded by a processor and which performs the method according to any of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210275239.4A CN114724128B (en) | 2022-03-21 | 2022-03-21 | License plate recognition method, device, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210275239.4A CN114724128B (en) | 2022-03-21 | 2022-03-21 | License plate recognition method, device, equipment and medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114724128A true CN114724128A (en) | 2022-07-08 |
| CN114724128B CN114724128B (en) | 2023-10-10 |
Family
ID=82237991
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210275239.4A Active CN114724128B (en) | 2022-03-21 | 2022-03-21 | License plate recognition method, device, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114724128B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116311568A (en) * | 2023-05-16 | 2023-06-23 | 广州铭创通讯科技有限公司 | ETC-based parking lot face recognition quick fee deduction method and device |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101246120B1 (en) * | 2012-07-26 | 2013-03-22 | 주식회사 비스타씨엔씨 | A system for recognizing license plate using both images taken from front and back faces of vehicle |
| US20140301606A1 (en) * | 2013-04-05 | 2014-10-09 | Xerox Corporation | Data augmentation method and system for improved automatic license plate recognition |
| CN104951784A (en) * | 2015-06-03 | 2015-09-30 | 杨英仓 | Method of detecting absence and coverage of license plate in real time |
| CN106980854A (en) * | 2017-03-29 | 2017-07-25 | 珠海习悦信息技术有限公司 | Number-plate number recognition methods, device, storage medium and processor |
| CN110909692A (en) * | 2019-11-27 | 2020-03-24 | 北京格灵深瞳信息技术有限公司 | Abnormal license plate recognition method and device, computer storage medium and electronic equipment |
| CN111209905A (en) * | 2020-04-23 | 2020-05-29 | 北京慧智数据科技有限公司 | Defect shielding license plate recognition method based on combination of deep learning and OCR technology |
| CN111639598A (en) * | 2020-05-29 | 2020-09-08 | 济南博观智能科技有限公司 | License plate recognition method, license plate recognition device, license plate recognition equipment and storage medium |
| CN112580754A (en) * | 2021-02-25 | 2021-03-30 | 湖南华辰悦科技有限公司 | Vehicle cleanliness judgment method and device suitable for construction site and storage medium |
| CN112861845A (en) * | 2021-04-23 | 2021-05-28 | 北京欣博电子科技有限公司 | License plate recognition method and device, computer equipment and storage medium |
| CN113869317A (en) * | 2021-09-10 | 2021-12-31 | 北京精英路通科技有限公司 | License plate recognition method and device, electronic equipment and storage medium |
-
2022
- 2022-03-21 CN CN202210275239.4A patent/CN114724128B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101246120B1 (en) * | 2012-07-26 | 2013-03-22 | 주식회사 비스타씨엔씨 | A system for recognizing license plate using both images taken from front and back faces of vehicle |
| US20140301606A1 (en) * | 2013-04-05 | 2014-10-09 | Xerox Corporation | Data augmentation method and system for improved automatic license plate recognition |
| CN104951784A (en) * | 2015-06-03 | 2015-09-30 | 杨英仓 | Method of detecting absence and coverage of license plate in real time |
| CN106980854A (en) * | 2017-03-29 | 2017-07-25 | 珠海习悦信息技术有限公司 | Number-plate number recognition methods, device, storage medium and processor |
| CN110909692A (en) * | 2019-11-27 | 2020-03-24 | 北京格灵深瞳信息技术有限公司 | Abnormal license plate recognition method and device, computer storage medium and electronic equipment |
| CN111209905A (en) * | 2020-04-23 | 2020-05-29 | 北京慧智数据科技有限公司 | Defect shielding license plate recognition method based on combination of deep learning and OCR technology |
| CN111639598A (en) * | 2020-05-29 | 2020-09-08 | 济南博观智能科技有限公司 | License plate recognition method, license plate recognition device, license plate recognition equipment and storage medium |
| CN112580754A (en) * | 2021-02-25 | 2021-03-30 | 湖南华辰悦科技有限公司 | Vehicle cleanliness judgment method and device suitable for construction site and storage medium |
| CN112861845A (en) * | 2021-04-23 | 2021-05-28 | 北京欣博电子科技有限公司 | License plate recognition method and device, computer equipment and storage medium |
| CN113869317A (en) * | 2021-09-10 | 2021-12-31 | 北京精英路通科技有限公司 | License plate recognition method and device, electronic equipment and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| 杨英仓: "用于卡口***的车牌遮挡检测方法", 《湖南科技大学学报（自然科学版）》, vol. 33, no. 02, pages 60 - 65 * |
| 聂文真: "出租汽车车牌遮挡行为判定与图像取证技术研究", 《中国优秀硕士学位论文全文数据库（工程科技Ⅱ辑）》, no. 03, pages 034 - 987 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116311568A (en) * | 2023-05-16 | 2023-06-23 | 广州铭创通讯科技有限公司 | ETC-based parking lot face recognition quick fee deduction method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114724128B (en) | 2023-10-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107944450B (en) | License plate recognition method and device | |
| JP5775225B2 (en) | Text detection using multi-layer connected components with histograms | |
| EP3907660A1 (en) | Method, apparatus, electronic device, and storage medium for recognizing license plate | |
| Türkyılmaz et al. | License plate recognition system using artificial neural networks | |
| CN109740415A (en) | Vehicle attribute recognition methods and Related product | |
| Jo | Cumulative dual foreground differences for illegally parked vehicles detection | |
| CN112215179B (en) | In-vehicle face recognition method, device, apparatus and storage medium | |
| CN111414888A (en) | Low-resolution face recognition method, system, device and storage medium | |
| CN110852311A (en) | Three-dimensional human hand key point positioning method and device | |
| CN111931683B (en) | Image recognition method, device and computer readable storage medium | |
| CN114170516B (en) | Vehicle weight recognition method and device based on roadside perception and electronic equipment | |
| CN106682092A (en) | Target retrieval method and terminal | |
| CN114038004A (en) | Certificate information extraction method, device, equipment and storage medium | |
| CN107578011A (en) | The decision method and device of key frame of video | |
| CN110852327A (en) | Image processing method, image processing device, electronic equipment and storage medium | |
| CN111860219A (en) | High-speed road occupation judging method and device and electronic equipment | |
| CN117218622A (en) | Road condition detection method, electronic equipment and storage medium | |
| CN114724128B (en) | License plate recognition method, device, equipment and medium | |
| CN108960246B (en) | Binarization processing device and method for image recognition | |
| CN112906671B (en) | Method and device for identifying false face-examination picture, electronic equipment and storage medium | |
| CN114168768A (en) | Image retrieval method and related equipment | |
| CN111709377B (en) | Feature extraction method, target re-identification method and device and electronic equipment | |
| CN113378837A (en) | License plate shielding identification method and device, electronic equipment and storage medium | |
| CN112949584A (en) | Image screening method and device, electronic equipment and storage medium | |
| CN112215188A (en) | Traffic police gesture recognition method, device, equipment and storage 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |