CN113095311A

CN113095311A - License plate number recognition method and device and storage medium

Info

Publication number: CN113095311A
Application number: CN202110242765.6A
Authority: CN
Inventors: 徐青; 周云鹏; 龙如意
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2021-03-05
Filing date: 2021-03-05
Publication date: 2021-07-09

Abstract

The embodiment of the application discloses a license plate number identification method, a license plate number identification device and a storage medium, and belongs to the field of image identification. In the embodiment of the application, a license plate amplification number in a vehicle image is identified, wherein the license plate amplification number is an identifier which is located on a target vehicle and is different from a license plate; acquiring a recognition result of the enlarged license plate number; and obtaining the license plate number of the target vehicle according to the recognition result of the enlarged license plate number, wherein the license plate number can be obtained based on the correspondence between the number of the enlarged license plate number and the license plate number, and the license plate number recognition rate is improved.

Description

License plate number recognition method and device and storage medium

Technical Field

The present disclosure relates to the field of image recognition technologies, and in particular, to a license plate number recognition method, a license plate number recognition device, and a storage medium.

Background

License plate number identification is one of main services in the traffic field, the existing license plate number identification methods are all realized based on the license plate of a vehicle, and the flow comprises the following steps: and detecting the position of the license plate in the vehicle image, then identifying the license plate, and outputting an identification result to obtain the license plate number. However, when the license plate of the vehicle is dirty or shielded, the license plate number cannot be recognized effectively, and the result may be output in a non-recognition manner.

Disclosure of Invention

The embodiment of the application provides a license plate number identification method, a license plate number identification device and a storage medium, and the license plate number identification rate is improved by identifying the license plate number of an enlarged license plate number.

In a first aspect, the present application provides a license plate number recognition method, including: identifying a license plate enlarged number in a vehicle image, wherein the license plate enlarged number is an identifier which is positioned on a target vehicle and is different from a license plate; acquiring a recognition result of the enlarged license plate number; and obtaining the license plate number of the target vehicle according to the identification result of the amplified license plate number.

The identification used for identifying the identity of the vehicle comprises a license plate amplification number and a license plate. The national regulation states that the rear parts of bodies or compartments of heavy and medium-sized trucks and trailers thereof, tractors and trailers thereof should be sprayed with enlarged marks, and characters should be correct and kept clear. The size of the enlarged mark sprayed on the rear part of the body or the rear compartment of the heavy and medium-sized truck and the trailer thereof is 2.5 times of the character of the registration number of the small-sized truck, and the enlarged mark is clear and complete and has obvious contrast with the color of the body. The method is characterized in that vehicles such as heavy trucks, medium trucks and trailers thereof (including small trucks, medium buses, buses and trucks with more than 10 seats) are required to be sprayed with a license plate number character enlarged version which is consistent with the corresponding license plate number at the rear part or rear carriage of the vehicle body, and the license plate enlarged version is called a license plate printing body in some scenes. The license plate enlarged number may include a registration number of the vehicle, a registration area, or other relevant information.

In the embodiment of the application, when the license plate number of the license plate cannot be effectively identified, the enlarged license plate number in the vehicle image is identified to obtain the number for identifying the enlarged license plate number, and the number based on the enlarged license plate number is correspondingly consistent with the license plate number, so that the license plate number can be obtained, and the license plate number identification rate is improved.

In one possible design, the obtaining of the recognition result of the enlarged license plate number specifically includes: and when the characters output by the license plate enlarged number in the vehicle image are recognized to comprise license plate characters and non-license plate characters, generating a recognition result of the license plate enlarged number by using the license plate characters only. And when the license plate enlarged number in the vehicle image is identified, the output characters possibly comprise license plate characters and non-license plate characters, and the identification result of the license plate enlarged number is generated only by using the license plate characters in the identification output. And when the license plate enlarged number in the vehicle image is identified, outputting characters only including license plate characters, and generating an identification result of the license plate enlarged number by using the license plate characters in the identification output.

In the embodiment of the application, when the characters in the vehicle image are recognized, the recognized and output characters can be processed according to the preset rule, so that the recognition result of the license plate enlarged number is obtained. The preset rule is used for recombining the characters in the output vehicle image so that the characters of the output license plate enlarged number are matched with the license plate number.

In the embodiment of the present application, the preset rules include, but are not limited to, the following: and (2) screening according to the character position of the enlarged license plate number to screen out proper license plate characters as the license plate number, and/or screening according to the character content of the enlarged license plate number to screen out proper license plate characters as the license plate number. The number length of the license plate is generally 7-8 digits and the like.

In one possible design, the method further includes: identifying the license plate in the vehicle image; acquiring a recognition result of the license plate; the obtaining of the license plate number of the target vehicle according to the recognition result of the enlarged license plate number specifically includes: and obtaining the license plate number of the target vehicle according to the recognition result of the license plate and the recognition result of the license plate enlarged number.

In the embodiment of the application, the amplified license plate number of the target vehicle is comprehensively recognized, the license plate number of the original license plate is recognized, and the final license plate number recognition result is comprehensively output, so that when the license plate cannot be effectively recognized due to dirt, shielding and the like, the integral license plate number recognition rate is further improved by fusing the amplified license plate number recognition.

In one possible design, the method further includes: and presenting a local image of the license plate enlargement number area in the vehicle image on a display interface. In one possible implementation manner, one or more of the license plate number, the license plate enlarged number image and the license plate image are presented on a display interface, wherein the license plate enlarged number image is a local image of a license plate enlarged number region in the vehicle image and is also an image of a license plate enlarged number region of the target vehicle, and the license plate image is a local image of a license plate region in the vehicle image and is also a local image of a license plate region of the target vehicle.

Wherein the vehicle image includes a target vehicle, which may be an overall image of the target vehicle. The license plate magnified number image and the license plate image may be images of a certain local area of a vehicle image, respectively, that is, the license plate magnified number image and the license plate image may be images of a certain local area of a target vehicle, such as an image of a body or a tail of the target vehicle. The license plate magnified image is different from the license plate image.

In the embodiment of the application, when the final license plate number of the target vehicle is output, one or more of a vehicle image, a license plate amplified number image and a license plate image of the target vehicle can be output, so that a user can judge the identification result.

In one possible design, the identifying the license plate in the vehicle image specifically includes: identifying the license plate according to a first character database, wherein the first character database comprises license plate characters associated with license plate numbers; the method for recognizing the enlarged license plate number in the vehicle image and acquiring the recognition result of the enlarged license plate number specifically comprises the following steps: and recognizing characters in the vehicle image according to a second character database, wherein the recognized and output characters comprise license plate characters and non-license plate characters, and the recognition result of the enlarged license plate number is generated only by using the license plate characters which are recognized and output, wherein the second character database comprises the license plate characters and the non-license plate characters which are not related to the license plate number.

The characters in the first character database comprise characters of the license plate number, such as characters or letters. The characters in the second character database comprise license plate characters and non-license plate characters, and the characters in the first character database are more than the characters in the second character database. The characters in the vehicle image may include words and/or letters.

In the embodiment of the application, the license plate in the vehicle image is identified according to the first character database, the characters in the vehicle image are identified according to the second character database, and the license plate characters comprising the enlarged license plate number are output, so that the enlarged license plate number is identified.

In one possible implementation, the license plate enlarged image includes an image of a rear or side area of the target vehicle. And spraying the amplified license plate on the rear part of the body or the carriage of the target vehicle, and acquiring an image of the body or the rear part of the carriage of the target vehicle, wherein the image is an amplified license plate image.

In one possible implementation manner, the obtaining of the license plate number of the target vehicle according to the recognition result of the license plate and the recognition result of the license plate enlarged number specifically includes: and selecting one of the recognition result of the license plate and the recognition result of the enlarged license plate number as the license plate number of the target vehicle.

In one possible implementation manner, the selecting one of the recognition result of the license plate and the recognition result of the enlarged license plate number as the license plate number of the target vehicle specifically includes: comparing the confidence coefficient of the recognition result of the license plate with the confidence coefficient of the recognition result of the enlarged number of the license plate; and taking the recognition result with high confidence as the license plate number of the target vehicle.

In the embodiment of the application, when the number identification of the license plate enlarged number and the license plate number is integrated, the final license plate number of the target vehicle is output according to the confidence coefficient, and the identification of the license plate number of the target vehicle is realized.

In one possible design, the selecting one of the recognition result of the license plate and the recognition result of the enlarged license plate number as the license plate number of the target vehicle specifically includes: the method comprises the following steps: obtaining the confidence of the recognition result of the license plate; and when the confidence coefficient is lower than a preset threshold value, taking the recognition result of the license plate enlarged number as the license plate number of the target vehicle.

In one possible design, the obtaining of the license plate number of the target vehicle according to the recognition result of the license plate and the recognition result of the license plate enlarged number specifically includes: acquiring a first confidence coefficient of each license plate character in the recognition result of the license plate; acquiring a second confidence coefficient of each license plate character in the recognition result of the license plate enlarged number; comparing the first confidence level and the second confidence level; and selecting the license plate characters with high confidence coefficient to form the license plate number of the target vehicle.

In this embodiment of the application, obtaining the license plate number of the target vehicle according to the recognition result of the license plate and the recognition result of the enlarged license plate number may include: selecting one of the recognition result of the license plate and the recognition result of the enlarged license plate number as the license plate number of the target vehicle; or fusing the recognition result of the license plate and the recognition result of the enlarged license plate number, and selecting one part of the recognition result of the license plate and one part of the recognition result of the enlarged license plate number to form the license plate number of the target vehicle.

In one possible design, after the acquiring the vehicle image including the enlarged license plate number, the method further includes: judging whether the target vehicle has a license plate enlarged number; if yes, judging whether the license plate enlarged number needs to be identified or not; and if so, identifying the license plate amplified number.

In the embodiment of the application, the existence of the license plate enlarged number is judged, whether the license plate enlarged number needs to be identified is judged, extra calculation overhead can be reduced to the maximum extent, and meanwhile, the identification capability of the license plate enlarged number can be provided under the scene needing to be identified most.

In a second aspect, the present application provides a license plate number recognition method, including: providing a display interface, wherein the display interface comprises a first identification item; detecting the triggering of the first recognition item, and recognizing a license plate amplification number in a vehicle image, wherein the license plate amplification number is an identifier which is located on a target vehicle and is different from a license plate; acquiring a recognition result of the enlarged license plate number; and obtaining the license plate number of the target vehicle according to the identification result of the amplified license plate number. The first identification item is used for starting identification of the license plate enlarged number.

In one possible design, the obtaining of the recognition result of the enlarged license plate number specifically includes: and when the characters output by the license plate enlarged number in the vehicle image are recognized to comprise license plate characters and non-license plate characters, generating a recognition result of the license plate enlarged number by using the license plate characters only.

In one possible design, the method further includes: identifying the license plate in the vehicle image, and acquiring the identification result of the license plate; the obtaining of the license plate number of the target vehicle according to the recognition result of the enlarged license plate number specifically includes: and obtaining the license plate number of the target vehicle according to the recognition result of the license plate and the recognition result of the license plate enlarged number.

In one possible design, the method further includes: and presenting a local image of the license plate enlargement number area in the vehicle image on a display interface. In one possible implementation manner, one or more of the license plate number, the enlarged license plate number image and the license plate image are presented on a display interface, wherein the enlarged license plate number image is an image of a license plate enlarged number region of the target vehicle, and the license plate image is an image of the license plate region of the target vehicle.

In one possible design, the display interface further includes a second identification item, and the second identification item is used for a user to select a shooting direction of the target vehicle, where the shooting direction includes a front direction, a side direction, and a back direction. The second identification item is used for determining that the vehicle image of the target vehicle is an image of the tail direction or the side direction of the vehicle; the method further comprises the following steps: detecting that the target vehicle conforms to a preset vehicle type and the second identification item is triggered, and determining that the target vehicle has a license plate enlarged number; judging whether the license plate enlarged number needs to be identified or not; and if so, identifying the license plate amplified number.

The method comprises the steps that on the basis of the fact that the rear portions of bodies or compartments of heavy-duty trucks and medium-duty trucks and trailers thereof, tractors and trailers thereof are sprayed with enlarged license plates, a target vehicle is a preset vehicle type such as the heavy-duty trucks and the medium-duty trucks and the trailers thereof, the tractors and the trailers thereof, and the second identification item is triggered, namely a vehicle image of the target vehicle is an image in the tail direction or the side direction of the vehicle, the license plate enlarged number on the target vehicle can be determined, and then whether the license plate enlarged number needs to be identified or not is judged.

In one possible design, the method further includes: identifying the license plate in the vehicle image according to a first character database, wherein the first character database comprises license plate characters associated with license plate numbers; and recognizing characters in the vehicle image according to a second character database, wherein the recognized and output characters comprise license plate characters and non-license plate characters, and the recognition result of the license plate enlarged number is generated only by using the license plate characters which are recognized and output. Wherein the second character database comprises the license plate characters and non-license plate characters not associated with license plate numbers; when the license plate characters output by recognition are used for generating a recognition result of the enlarged license plate number, the characters in the vehicle image can be processed according to a preset rule, and the characters of the enlarged license plate number are output, wherein the preset rule is used for enabling the output characters to be matched with the license plate number.

In one possible design, the license plate enlarged image includes an image of a rear or side area of the target vehicle.

In a third aspect, the present application provides a license plate number recognition apparatus for performing the method as described above.

In a fourth aspect, the present application provides a computer storage medium comprising computer instructions which, when run on an electronic device, cause the electronic device to perform the method of any of the above.

The technical effects obtained by the above second, third and fourth aspects are similar to the technical effects obtained by the corresponding technical means in the first aspect, and are not described herein again.

The beneficial effect that technical scheme that this application provided brought includes at least:

Drawings

FIG. 1a is a schematic view of an image of a vehicle according to an embodiment of the present disclosure;

FIG. 1b is a schematic view of another vehicle image provided by an embodiment of the present application;

fig. 2 is a schematic structural diagram of a license plate number recognition device according to an embodiment of the present disclosure;

fig. 3 is a schematic structural diagram of a license plate number recognition system according to an embodiment of the present disclosure;

fig. 4 is a schematic structural diagram of a computing device according to an embodiment of the present application;

fig. 5 is a schematic flowchart of a license plate number recognition method according to an embodiment of the present disclosure;

FIG. 6a is a schematic view of an image of a vehicle according to an embodiment of the present disclosure;

FIG. 6b is a schematic view of another vehicle image provided by the embodiments of the present application;

FIG. 7 is a schematic diagram of a display interface provided in an embodiment of the present application;

FIG. 8a is a schematic view of another enlarged license plate image according to an embodiment of the present disclosure;

FIG. 8b is a schematic view of another enlarged license plate image according to an embodiment of the present disclosure;

fig. 9 is a schematic flowchart of another license plate enlarged number recognition method according to an embodiment of the present disclosure;

FIG. 10a is a schematic view of another display interface provided in the embodiments of the present application;

FIG. 10b is a schematic view of another display interface provided in the embodiments of the present application;

fig. 11 is a schematic flowchart of another license plate number recognition method according to an embodiment of the present disclosure;

FIG. 12 is a schematic view of another display interface provided in an embodiment of the present application;

FIG. 13 is a schematic view of another vehicle image provided by an embodiment of the present application;

fig. 14 is a schematic flowchart of another license plate number recognition method according to an embodiment of the present disclosure;

fig. 15 is a schematic view of another display interface provided in the embodiment of the present application.

Detailed Description

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.

Before explaining the embodiments of the present application in detail, an application scenario related to the embodiments of the present application will be described.

The scheme of the application can be applied to various fields of road charging, parking management, traffic guidance, traffic enforcement, traffic monitoring, traffic inspection, vehicle scheduling, vehicle detection, community vehicle management and the like. In a traffic monitoring scene, the angle of a monitoring camera is shot from top to bottom, and for large vehicles such as a cement tank truck, a muck truck, a large truck and the like, because the vehicle body is large, the license plate is often not easy to identify through the camera due to the fact that the installation position of the license plate is close to the lower position. As shown in fig. 1a, the numbers on the license plate obtained by shooting are not easy to be identified. Meanwhile, in a part of large construction vehicles, the license plate is stained, covered by soil or ropes and cannot be identified by the camera, and as shown in fig. 1b, the license plate is partially shielded by other objects.

Based on this, in order to improve the recognition rate of the license plate number, the license plate number recognition method provided by the embodiment of the application obtains the number of the license plate enlarged number which is recognized and output by recognizing the license plate enlarged number of the target vehicle, and obtains the license plate number by recognizing the license plate enlarged number based on the correspondence between the number of the license plate enlarged number and the license plate number, thereby improving the recognition rate of the license plate number. Furthermore, after license plate recognition is carried out, license plate amplified number recognition is fused, the license plate number of the target vehicle is obtained by integrating the recognition result of the license plate and the recognition result of the license plate amplified number, and the license plate number recognition accuracy is further improved.

In addition, it should be noted that the identifier for identifying the identity of the vehicle includes a license plate enlarged number and a license plate. The national regulation states that the rear parts of bodies or compartments of heavy and medium-sized trucks and trailers thereof, tractors and trailers thereof should be sprayed with enlarged marks, and characters should be correct and kept clear. The size of the enlarged mark sprayed on the rear part of the body or the rear compartment of the heavy and medium-sized truck and the trailer thereof is 2.5 times of the character of the registration number of the small-sized truck, and the enlarged mark is clear and complete and has obvious contrast with the color of the body. The method is characterized in that for heavy-duty and medium-duty trucks and trailers (including small-sized trucks, small trucks, medium buses, buses and trucks with more than 10 seats) and other vehicles, license plate number characters are required to be sprayed on the body (such as the side part) or the rear part of a carriage of the heavy-duty and medium-duty trucks and trailers, wherein the license plate number characters are correspondingly consistent with license plate numbers, and the license plate number characters are also called as license plate printing bodies in some scenes. The license plate enlarged number and the license plate may include a registration number of the vehicle, a registration area, or other relevant information. As shown in fig. 1a, an enlarged license plate number "zhe FMXXXX" is sprayed on the rear portion of the carriage of the vehicle, and the "zhe FMXXXX" sprayed on the rear portion of the rear carriage of the vehicle is a license plate enlarged number, wherein the license plate enlarged number of the "zhe FMXXXX" is consistent with the characters of the license plate "zhe FMXXXX". As shown in fig. 1b, the enlarged license plate number "sobxxxx" is sprayed on the rear portion of the vehicle compartment, and the "sobxxxx" sprayed on the rear portion of the vehicle compartment is the enlarged license plate number. Wherein the enlarged license plate number of the 'SuBXXXXX' is consistent with the characters of the license plate 'SuBXXXXX'. Wherein the third to seventh bits "X" of the license plate can be any number from 0 to 9. For the number plate of China mainland, except the current optional individual number plate, the currently used nine-two type motor vehicle number plate has the first number plate of Chinese characters: representing the provincial administrative district where the vehicle entrance is located, which is the abbreviation of each (province, direct prefecture city, autonomous district), the second digit of the license plate is English letters (except I and O): representing the regional administrative district in which the vehicle user is located, which is typically a regional city, region, municipality, union letter code. The third and seventh digits of the license plate are numbers.

The nine-two type automobile number plate is exemplified below, wherein the third to seventh bits "X" of the plate may be any integer number from 0 to 9.

The embodiment of the application provides a license plate number recognition method, which is executed by a license plate number recognition device. The function of the license plate number recognition device can be realized by a software system, can also be realized by hardware equipment, and can also be realized by the combination of the software system and the hardware equipment.

When the license plate number recognition device is a software device, please refer to fig. 2, and fig. 2 is a schematic structural diagram of a license plate number recognition device according to an embodiment of the present disclosure. The license plate number recognition apparatus 200 may be logically divided into a plurality of modules, each of which may have a different function, the functions of each module being implemented by a processor in a computing device reading and executing instructions in a memory, and the computing device structure may be as the computing device 400 shown in fig. 4 below. For example, the license plate number recognition apparatus may include a first recognition module 201, a first obtaining module 202, and a obtaining module 203. In one specific implementation, the license plate number recognition device 200 may perform the contents described in steps S501-S503, steps S901-S905, and steps S111-S119 described below, or perform the contents described in steps S141-S144 described below. It should be noted that, in the embodiment of the present application, only the structure and the functional modules of the license plate number recognition device 200 are exemplarily divided, but the specific division is not limited at all.

The first recognition module 201 is configured to recognize a license plate enlarged number in a vehicle image, where the license plate enlarged number is an identifier located on a target vehicle and different from a license plate. And the number of the enlarged license plate number is correspondingly consistent with the number of the license plate. The target vehicle is provided with a license plate amplification number, and the vehicle image comprises the license plate amplification number of the target vehicle, namely the vehicle image comprises a license plate amplification number area of the target vehicle.

In one possible implementation manner, the requirement on normativity of spraying based on the license plate enlarged number is not high, the randomness is relatively high, the license plate enlarged number is not as integral as the license plate in many cases, when the license plate enlarged number in the vehicle image is recognized, the first recognition module 201 can detect each character in the vehicle image as a target, then based on the position of each character detected and output, the intelligent algorithm recognizes the character at each character position to output the recognition result of the character corresponding to each position, and the output recognition result may include the character in the vehicle image. The first recognition module 201 detects each character in the vehicle image as a target, that is, includes detecting a character in a license plate enlarged number region of the target vehicle, such as a character (characters, letters or numbers) sprayed on the rear portion of a vehicle body or a carriage of the target vehicle, as a target, and outputting a position of the license plate enlarged number character, and the intelligent algorithm recognizes the character at each character position, including recognizing the character at the license plate enlarged number character position, thereby recognizing the license plate enlarged number in the vehicle image.

The body of the target vehicle may include license plate characters associated with a license plate number and non-license plate characters not associated with the license plate number. The license plate characters include characters for composing a license plate number when the license plate number is set, and in one possible implementation manner, the license plate characters only refer to the characters composing the license plate number. Non-license plate characters may be understood as characters other than license plate characters. Taking the nine-two type motor vehicle license plate as an example, the license plate characters comprise characters such as Chinese characters, English letters (except I and O), numbers and the like, and specifically, the license plate characters comprise characters for short, English letter characters (except I and O) and numbers (0-9) of each province, the municipality in the direct jurisdiction and the municipality in the autonomous region. The non-license plate characters are characters which are not related to the constituent license plate numbers, namely the non-license plate characters comprise characters which are not related to the constituent license plate numbers, can be characters of Chinese characters except short characters of each (province, direct prefecture city and autonomous region) sprayed on a vehicle body, such as 'cold' characters and 'machine' characters and the like, and can also be operators and the like. The non-license plate characters can be set according to characters, except for license plate numbers, which are usually sprayed on the vehicle, and the non-license plate characters are not particularly limited in the present application. Characters refer to glyph-like units or symbols, including letters, numbers, operator symbols, punctuation marks and other symbols, as well as some functional symbols. The intelligent algorithm includes, but is not limited to, a deep learning algorithm.

When the first recognition module 201 recognizes a license plate enlarged number in a vehicle image, the output characters include license plate characters and non-license plate characters, or only license plate characters.

In one possible implementation manner, the first recognition module 201 may recognize characters in the vehicle image according to a second character database, where the second character database includes the license plate characters associated with the license plate number and non-license plate characters not associated with the license plate number.

The first obtaining module 202 is configured to obtain a recognition result of the enlarged license plate number. The first obtaining module 202 is in communication connection with the first recognition module 201, so as to obtain a recognition result of the enlarged license plate number according to the recognition of the enlarged license plate number of the vehicle image by the first recognition module 201. After the first recognition module 201 recognizes the license plate enlarged number of the vehicle image, the first obtaining module 202 obtains a recognition result of the license plate enlarged number by the first recognition module 201.

In this embodiment of the application, the first obtaining module 202 obtains a recognition result of the enlarged license plate number from a result of recognition and output by the first recognition module 201, directly uses the license plate character to generate the recognition result of the enlarged license plate number if the character recognized and output by the first recognition module 201 only includes the license plate character, and only uses the license plate character recognized and output by the first recognition module 201 to generate the recognition result of the enlarged license plate number if the character recognized and output by the first recognition module 201 includes the license plate character and a non-license plate character.

In one possible implementation manner, when the characters in the vehicle image only include the license plate characters, for example, only include the characters of the enlarged license plate number, the first recognition module 201 recognizes the output characters as the characters of the enlarged license plate number, and the first obtaining module 202 obtains the characters recognized and output by the first recognition module 201, so as to obtain the characters of the enlarged license plate number.

In one possible implementation manner, because the body of the target vehicle may include license plate characters and non-license plate characters, and the enlarged license plate number is random in the vehicle, the first obtaining module 202 needs to perform corresponding processing according to the recognition result obtained by the first recognition module 201 by recognizing the enlarged license plate number of the vehicle image, so as to obtain the recognition result of the enlarged license plate number. Specifically, after the first recognition module 201 recognizes and outputs all characters in the vehicle image, the first obtaining module 202 needs to perform corresponding processing on all the characters that are recognized and output, so as to obtain the characters with the enlarged license plate number, and obtain the recognition result of the enlarged license plate number.

In one possible implementation manner, the corresponding processing, by the first obtaining module 202, of the characters recognized and output by the first recognition module 201 includes that the first obtaining module 202 needs to match the obtained recognition result of the enlarged license plate number with the license plate number of the target vehicle.

The obtaining module 203 is configured to obtain the license plate number of the target vehicle according to the recognition result of the enlarged license plate number. The obtaining module 203 is in communication connection with the first obtaining module 202, and obtains the license plate number according to the license plate amplified number recognition result of the target vehicle in the first obtaining module 202.

Optionally, the license plate number recognition device 200 may further include a vehicle image acquisition module 204. The vehicle image acquisition module 204 is used for acquiring a vehicle image of the target vehicle. The vehicle image obtaining module 204 may be in communication connection with the first recognition module 201 to provide the first recognition module 201 with a vehicle image of the target vehicle, so that the first recognition module 201 recognizes the vehicle image and recognizes the enlarged license plate number in the vehicle image.

In one possible implementation, the vehicle image acquisition module 204 acquires an image captured by a camera, and acquires a vehicle image according to the image captured by the camera. For example, the vehicle image acquiring module 204 may be communicatively connected to a camera for capturing a target vehicle to acquire an image captured by the camera and including the target vehicle, wherein the image is a vehicle image when the vehicle in the image only includes the target vehicle, and the vehicle image of the target vehicle position area is intercepted from the image when a plurality of vehicles are included in the image. Still alternatively, the vehicle image acquiring module 204 may process the video of the target vehicle captured by the camera to obtain each image frame, find the image frame including the target vehicle, where the image frame includes only the target vehicle, the image frame is a vehicle image, and when the image frame includes a plurality of vehicles, the image frame intercepts a vehicle image of the target vehicle position area from the image frame.

Optionally, the license plate number recognition device 200 may further include a first determination module 205. The first determination module 205 is used for determining whether a license plate enlarged number is included in the vehicle image. Specifically, when the intelligent algorithm detects that the target vehicle is a large truck or a van and the shooting direction is the back direction, that is, the vehicle image is an image of the tail of the target vehicle shot by the camera, the first determination module 205 considers that the vehicle image includes the enlarged license plate number. When the intelligent algorithm detects that the vehicle image is the vehicle image of the car, the vehicle image does not include the license plate amplification number, or when the vehicle image is detected as the truck and the shooting direction is the front direction, namely the vehicle image is the image of the head of the target vehicle shot by the camera, the vehicle image does not include the license plate amplification number. The first determination module 205 is connected to the vehicle image acquisition module 204 in a communication manner to determine whether the vehicle image acquired by the vehicle image acquisition module 204 includes a license plate enlarged number.

Optionally, the license plate number recognition device 200 may further include a second determination module 206. The second judging module is used for judging whether the license plate enlarged number needs to be identified or not. The first determination module 205 is in communication connection with the second determination module 206, and after the first determination module 205 determines that the vehicle image includes the enlarged license plate number, the second determination module 206 may determine whether the enlarged license plate number needs to be recognized.

In one possible implementation manner, the second determination module 206 may perform the determination based on analyzing the number recognition result of the license plate, and according to different selection tendencies, completing the analysis of the determination. If a threshold value is set, when the confidence coefficient of the license plate recognition result reaches the threshold value, the license plate number obtained by license plate recognition is considered to be accurate, and license plate amplified number recognition is not needed.

Optionally, the license plate number recognition device 200 may further include a second recognition module 207. The second recognition module 207 is used for recognizing the license plate in the vehicle image. The license plate is a mark which is positioned on a target vehicle and is different from the enlarged license plate number. And the number of the enlarged license plate number is correspondingly consistent with the number of the license plate. The license plate is also called as license plate, also refers to vehicle number plate, is a plate material hung on the front and back of the vehicle, the material used is aluminum, iron sheet, plastic or paper, on which the registration number, registration area or other relevant information of the vehicle is marked. The target vehicle is provided with a license plate, and the vehicle image comprises the license plate of the target vehicle, namely the vehicle image comprises the license plate area of the target vehicle. The second recognition module 207 may be in communication connection with the vehicle image acquisition module 204, and the vehicle image acquisition module 204 provides the second recognition module 207 with a vehicle image of the target vehicle, so that the second recognition module 207 recognizes the vehicle image and recognizes a license plate in the vehicle image.

In one possible implementation manner, the second recognition module 207 may recognize characters in the vehicle image according to a first character database, where the first character database includes license plate characters associated with license plate numbers. Specifically, the second recognition module determines the position of the license plate in the vehicle image, recognizes characters on the position of the license plate according to the first character database based on the position of the license plate, obtains characters of the license plate, and outputs a recognition result of the license plate.

The second obtaining module 208 is configured to obtain a recognition result of the license plate. The characters of the license plate number on the basis of the license plate are all arranged on one plate, the license plate can be regarded as a whole, and the recognition result of the license plate can be directly output on the basis of the recognition of the license plate. The second obtaining module 208 may be in communication connection with the second identifying module 207, and obtain the identification result of the license plate from the identification result of the second identifying module 207.

In one possible implementation manner, the second obtaining module 208 may further be in communication connection with the second determining module 206, so that after the second obtaining module 208 obtains the recognition result of the license plate, the second determining module 206 obtains the confidence of the recognition result of the license plate, and further determines whether the license plate enlarged number needs to be recognized.

In one possible implementation manner, the second obtaining module 208 may further perform communication connection with the obtaining module 203, so that the obtaining module 203 may obtain a recognition result of the license plate and a recognition result of the enlarged license plate number, and further the obtaining module 203 obtains the license plate number of the target vehicle according to the recognition result of the license plate and the recognition result of the enlarged license plate number, obtains a recognition result of the license plate and recognition of the enlarged license plate number by the module 203, and improves the accuracy of license plate number recognition.

Optionally, the license plate number recognition device 200 may further include a presentation module (not shown) for presenting a display interface, and the presentation module is in communication connection with the obtaining module 203 to obtain the license plate number of the target vehicle. The local image of the license plate enlarged number region in the vehicle image can be presented on the display interface, and in one possible implementation manner, one or more of the vehicle image of the target vehicle, the license plate number, the license plate enlarged number image and the license plate image can be presented on the display interface, wherein the license plate enlarged number image is the local image of the license plate enlarged number region in the vehicle image, namely the local image of the license plate enlarged number region of the target vehicle. The license plate image is a local image of a license plate region in the vehicle image, namely a local image of the license plate region of the target vehicle. After the license plate amplified number is identified and output, the position of the license plate amplified number in the vehicle image can be determined according to the identification result of the license plate amplified number, so that a license plate amplified number area can be obtained, and the license plate amplified number area is intercepted from the vehicle image to obtain a license plate amplified number image. After the license plate is recognized and output, the position of the license plate in the vehicle image can be determined according to the recognition result of the license plate, so that a license plate area can be obtained, and the license plate area is intercepted from the vehicle image to obtain the license plate image.

The first identification module 201 and the second identification module 207 may be different independent modules. If the first recognition module 201 can recognize the characters in the vehicle image according to the second character database, the second recognition module 207 can recognize the characters in the vehicle image according to the first character database.

In addition, in some possible cases, some of the above-mentioned license plate number recognition devices 200 may also be combined into one module, for example, the above-mentioned first determination module 205 and the second determination module 206 may be combined into a recognition determination module, that is, the recognition determination module integrates the functions of the first determination module 205 and the second determination module 206. The first identification module 201 and the second identification module 207 may be combined into an identification module, that is, the identification module combines the functions of the first identification module 201 and the second identification module 207.

Please refer to fig. 3, which is a schematic structural diagram of a license plate number recognition system according to an embodiment of the present disclosure. The license plate number recognition system includes a camera 301, a network transmission device 302, a server 303, and a client 304.

The camera 301 is used for shooting the target vehicle to obtain a vehicle image of the target vehicle, and may be a camera 301 which is rotated in both horizontal and vertical directions and realizes multi-angle shooting. Illustratively, the camera 301 will take an image or video of the subject vehicle as the vehicle passes by the camera disposed on the aisle. The camera 301 may further process the captured vehicle image, including but not limited to image processing and the license plate number recognition process of the present application.

The network transmission device 302 is used for data transmission and may be a wireless network or a wired network.

The server 303 may be in communication connection with the camera 301 through the network transmission device 302, and may be a high-performance computer that controls the pan-tilt camera to rotate and zoom through control instructions. The server 303 may also be configured to process the vehicle images captured by the camera, including but not limited to image processing and license plate number recognition processing of the present application.

The client 304 may be a terminal such as a desktop computer and a tablet computer. The client 304 may be communicatively connected to the camera 301 through the network transmission device 302, and may receive and store information transmitted by the server 303. The client 304 may be configured to display the license plate number recognition result, for example, the client 304 may be configured to present a display interface, and present one or more of the vehicle image, the license plate number, the license plate enlarged number image, and the license plate image of the target vehicle on the display interface, and a user may view related information through the display interface of the client 304. The client 304 may also be used to process the vehicle images captured by the camera, including but not limited to image processing, and license plate number recognition processing of the present application.

In the embodiment of the present application, the license plate number recognition apparatus 200 described above can be flexibly deployed. The license plate number recognition device 200 may be deployed in the camera 303, or may be deployed in the server 303, and in one possible implementation, the license plate number recognition device 200 may be deployed in the client 304.

When the license plate number recognition apparatus is a hardware device, the license plate number recognition apparatus may be a computing device in a license plate number recognition system, for example, the license plate number recognition apparatus may be a camera, a server, or a client described above. Referring to fig. 4, fig. 4 is a schematic structural diagram of a computing device according to an embodiment of the present disclosure. The computing device 400 includes a processor 401, a communication bus 402, a memory 403, and at least one communication interface 404.

The processor 401 may be a Central Processing Unit (CPU), an application-specific integrated circuit (ASIC), a Graphics Processing Unit (GPU), or any combination thereof. The processor 401 may include one or more chips, and the processor 401 may include an AI accelerator, such as: a neural Network Processor (NPU).

Communication bus 402 may include a path that transfers information between components of computing device 400 (e.g., processor 401, memory 403, communication interface 404).

The Memory 403 may be a Read-Only Memory (ROM) or other type of static storage device that can store static information and instructions, a Random Access Memory (RAM) or other type of dynamic storage device that can store information and instructions, an Electrically Erasable Programmable Read-Only Memory (EEPROM), a Compact Disc Read-Only Memory (CD-ROM) or other optical Disc storage, optical Disc storage (including Compact Disc, laser Disc, optical Disc, digital versatile Disc, blu-ray Disc, etc.), magnetic disk storage media or other magnetic storage devices, 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 403 may be separate and coupled to the processor 401 through a communication bus 402. The memory 403 may also be integrated with the processor 401. The memory 403 may store computer instructions, and when the computer instructions stored in the memory 403 are executed by the processor 401, the license plate number recognition method of the present application may be implemented. In addition, the memory 403 may also store data required by the processor in performing the methods described below, as well as intermediate data and/or result data generated thereby.

The communication interface 404 may be any device, such as a transceiver, for communicating with other devices or communication Networks, such as ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.

In particular implementations, processor 401 may include one or more CPUs, as one embodiment.

In particular implementations, a computer device may include multiple processors, as one embodiment. Each of these processors may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. A processor herein may refer to one or more devices, circuits, and/or processing cores for processing data (e.g., computer program instructions).

Next, a license plate number recognition method provided in the embodiment of the present application is introduced.

Flow charts are used herein to illustrate operations performed by an apparatus according to embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in the exact order in which they are performed. Rather, the various steps may be processed in reverse order or simultaneously, as desired. Meanwhile, other operations may be added to the processes, or a certain step or several steps of operations may be removed from the processes.

Referring to fig. 5, fig. 5 is a schematic flow chart illustrating a license plate number recognition method according to an embodiment of the present disclosure. The license plate number recognition method can be executed by the license plate number recognition device, and comprises the following steps:

step S501: and identifying a license plate enlarged number in the vehicle image, wherein the license plate enlarged number is an identifier which is positioned on a target vehicle and is different from the license plate.

In the embodiment of the application, a vehicle image is input, the vehicle image comprises a license plate amplification number, the vehicle image is processed, and the license plate amplification number in the vehicle image is identified. Recognizing a license plate enlargement number in a vehicle image includes recognizing characters in the vehicle image, the layout of the sprayed characters on a target vehicle has great randomness, and the layout of the sprayed characters has great diversity among vehicles, for example, please refer to fig. 6a, fig. 6a is a vehicle image, such as fig. 6a, the characters in the vehicle image include license plate characters (zhe fxxxx) associated with license plate numbers and non-license plate characters (heavy machinery) not associated with the license plate numbers, the license plate enlargement number in fig. 6a is not a whole, other text characters also exist between the license plate enlargement number "zhe fxxxx", and text characters of "heavy machinery" also exist between "zhe FX" and "XXXX". Please refer to fig. 6b, fig. 6b is a vehicle image, the characters in the vehicle image include license plate characters (wan KQXXXX) associated with the license plate number and non-license plate characters (refrigerated) not associated with the license plate number, and the distance of the characters of the enlarged license plate number "wan KQXXXX" in the tail of the vehicle body of the target vehicle in fig. 6b is not as uniform as the characters on the license plate, nor as the characters on the license plate are framed by the license plate. Therefore, the license plate enlarged number recognition method of the application is required to recognize the license plate enlarged number, recognize each character in the vehicle image, for example, recognize the license plate character in the vehicle image to obtain the character of the license plate enlarged number, or recognize the license plate character and the non-license plate character in the vehicle image, then process the recognized and output character to obtain the character of the license plate enlarged number, recognize the license plate enlarged number in the vehicle image, and obtain the recognition result of the license plate enlarged number.

In the embodiment of the application, the characters in the vehicle image can be identified according to the second character database, so that the license plate amplification number in the vehicle image can be identified. Specifically, characters in the vehicle image are detected, each character is used as a target for detection, namely each character on the enlarged license plate number on the vehicle is also used as a target for detection, the intelligent algorithm determines the position of each character on the vehicle body by detecting each character on the vehicle body, and based on each character position, the intelligent algorithm identifies the character on each position and identifies the character according to the second character database. The second character database comprises license plate characters and non-license plate characters, wherein the license plate characters are associated with license plate numbers, and the license plate characters can be specifically set according to the specifications of the license plate numbers of various regions, acquired and stored. The non-license plate characters are characters which are not related to the license plate number, and can be determined according to characters sprayed on a vehicle in an actual situation, for example, trademarks, company names or store names can be sprayed on the vehicle, the non-license plate characters comprise characters of the trademarks, the company names or the store names, and the non-license plate characters are obtained and stored.

In this embodiment of the application, the second character database may be a local database, and if the license plate number recognition device is deployed in the camera, the second character database is a database on the camera, and the second character database may also be a cloud database. This is not a particular limitation of the present application.

In a possible implementation manner, the license plate number recognition device may provide a display interface for a user, where the display interface includes a first recognition item, and when the user triggers the first recognition item, the license plate number recognition device triggers the recognition of the enlarged license plate number. Referring to fig. 7, a license plate number recognition mode can be provided for a user in the display interface, and the license plate number recognition mode has two options, namely, license plate recognition and enlarged license plate number recognition. And the license plate number recognition device can recognize the license plate through the second recognition module. When a user selects a license plate amplified number recognition option, namely a first recognition item is triggered, and the license plate number recognition device can recognize the license plate amplified number through the first recognition module.

In a possible implementation manner, the client provides the display interface, the user operates the client to input the license plate enlarged number recognition mode, and the server, the camera or the client performs license plate enlarged number recognition to acquire a vehicle image including the license plate enlarged number so as to recognize the license plate enlarged number in the vehicle image.

Step S502: and acquiring the recognition result of the enlarged license plate number.

In this embodiment, step S502 may obtain a recognition result of the enlarged license plate number according to the recognition of the enlarged license plate number in the vehicle image in step S501. In step S502, the recognition result output in step S501 may be processed to obtain the recognition result of the enlarged license plate number, and if the character in the vehicle image may be processed according to a preset rule, the character of the enlarged license plate number may be output, where the preset rule is used to match the output character with the license plate number.

In the embodiment of the application, step S502 obtains a recognition result of the enlarged license plate number from the result of the recognition output in step S501, if the characters of the recognition output in step S501 only include license plate characters, the license plate characters are directly used to generate the recognition result of the enlarged license plate number, and if the characters of the recognition output in step S501 include license plate characters and non-license plate characters, the recognition result of the enlarged license plate number is generated only by using the license plate characters of the recognition output.

In the implementation of the application, the characters of the target vehicle may include license plate characters associated with a license plate number and non-license plate characters not associated with the license plate number, and for this reason, the characters output by the step S501 may be processed according to a preset rule, and if the characters output by the step S501 include the non-license plate characters, license plate characters meeting the license plate number are selected from the characters output by the step S501, and the non-license plate characters are removed. If the number of the license plate characters output by recognition is too large, if the license plate enlarged number region of the target vehicle also comprises a telephone number (digital characters), the characters in the vehicle image output by recognition need to be processed, the telephone number is removed, the output characters are matched with the license plate number, the output characters and the arrangement sequence or the number of the characters accord with the license plate number rule according to the preset rule, the preset rule can be set according to the actual condition if the setting rules of the license plate numbers in different regions are different, and the method is not specifically limited by the application.

In one possible implementation manner, the preset rule includes screening according to the positions of characters, and screening out the characters which are located on the same horizontal line and in the same line based on the attributes of the license plate with the enlarged number and the fact that the characters are sprayed on the same horizontal line in the common spraying process. The preset rule may further include filtering according to the content of the character. The license plate number is composed according to a certain rule, some characters cannot be composed into the license plate number, or the characters composing the license plate number have certain quantity limitation, so that a preset rule can be determined according to the content of the characters, and the characters meeting the license plate number rule are selected according to the preset rule.

In the embodiment of the application, the position of the enlarged license plate number can be determined based on the recognition result of the enlarged license plate number, the enlarged license plate number region can be determined based on the position of the enlarged license plate number, and after the recognition result of the enlarged license plate number is obtained in step S502, the character position of the enlarged license plate number can be determined based on the recognition result of the enlarged license plate number, so that the enlarged license plate number region can be determined according to the character position of the enlarged license plate number. The enlarged license plate number region may be a region where characters are displayed on the rear portion of the vehicle body or the rear portion of the carriage, such as a region where characters are sprayed, printed or engraved on the rear portion of the vehicle body or the rear portion of the carriage, and the enlarged license plate number region may include license plate characters (such as the enlarged license plate number characters) associated with a license plate number and non-license plate characters not associated with the license plate number, that is, the characters in the enlarged license plate number region include license plate characters and non-license plate characters. Referring to fig. 8a, fig. 8a is an enlarged license plate number image obtained by capturing a screen of a character area sprayed on the tail of the vehicle body in fig. 6a, fig. 8b is a reference to fig. 8b, and fig. 8b is an enlarged license plate number image obtained by capturing a screen of a character area sprayed on the tail of the vehicle body in fig. 6 b.

In one possible implementation manner, please refer to fig. 9, and fig. 9 is a schematic flowchart of another license plate enlarged number recognition method according to an embodiment of the present disclosure. Taking the identification of the sprayed characters in the vehicle image as an example, the license plate number identification method can be executed by the license plate number identification device, and the method comprises the following steps:

step S901: and detecting the sprayed characters.

In the embodiment of the application, spraying characters in a vehicle image are detected, each spraying character is used as a target for detection, namely, each spraying character on an enlarged number area on a vehicle is used as a target for detection, and the intelligent algorithm determines the position of each spraying character by detecting each spraying character on a vehicle body.

Illustratively, if the painted characters in the vehicle image of fig. 6a include "zhe fxxxx" and "heavy machinery", then the need to detect painted characters includes the following: zhe, F, X, X, X, X, X, weight, type, machine, and further determine the position of each sprayed character, and if the position of the "Zhe" character is "1", the position of the "F" character is "2", the position of the "X" character is "3", the position of the "heavy" character is "4", the position of the "type" character is "5", the position of the "machine" character is "6", the position of the "mechanical" character is "7", the position of the first numeric character "X" is "8", the position of the second numeric character "X" is "9", the position of the third numeric character "X" is "10", and the position of the fourth numeric character "X" character is "11", in this order. If the painted characters in the vehicle image of fig. 6b include "wan KQXXXX" and "cold storage," then the need to detect painted characters includes the following: in this case, if the position of the "cold" character is "1", the position of the "hidden" character is "2", the position of the "wan" character is "3", the position of the "K" character is "4", the position of the "Q" character is "5", the position of the first numeric character "X" is "6", the position of the second numeric character "X" is "7", the position of the third numeric character "X" is "8", and the position of the fourth numeric character "X" is "9", in this order.

Step S902: each spray character is identified.

In the embodiment of the present application, based on the position of each spraying character detected and output in step S901, the intelligent algorithm identifies the spraying character at each position, and outputs the identification result of the spraying character at each position. Illustratively, for the vehicle image of fig. 6a, location "1" outputs the character "zhe", location "2" outputs the character "F", location "3" outputs the character "X", location "4" outputs the character "heavy", location "5" outputs the character "type", location "6" outputs the character "machine", location "7" outputs the character "mechanical", location "8" outputs the character "X", location "9" outputs the character "X", location "10" outputs the character "X", and location "11" outputs the character "X". For the vehicle image of fig. 6b, the character "cold" is output at the position "1", the character "hidden" is output at the position "2", the character "wan" is output at the position "3", the character "K" is output at the position "4", the character "Q" is output at the position "5", the character "X" is output at the position "6", the character "X" is output at the position "7", the character "X" is output at the position "8", and the character "X" is output at the position "9".

In the embodiment of the application, the characters in the vehicle image are identified according to the second character database, and the characters associated with the license plate number (such as the characters "Zhe FXXXXX" in FIG. 6a and "Wan KQXXXX" in FIG. 6 b) and the characters not associated with the license plate number (such as the characters "heavy machinery" in FIG. 6a and "cold storage" in FIG. 6 b) are obtained.

Step S903: and screening according to the position of the character.

In the embodiment of the present application, since the body of the target vehicle may contain many characters, including chinese characters, arabic numerals, letters, and the like, and the recognized characters may also include non-license plate characters other than the license plate number, such as the characters "heavy, type, mechanical, and mechanical" in fig. 6a, and the characters "cold and hidden" in fig. 6b, it is necessary to recombine the output of the characters to screen out the characters that meet the license plate number.

In one possible implementation manner, the characters on the same horizontal line are screened out based on the attribute of the license plate with the enlarged license plate number and the spraying in the general spraying process, as shown in fig. 6b, the character "cold storage" is not in the same line as the character "wan KQXXXX", so that it can be determined that one line of the character "cold storage" and the character "wan KQXXXX" is the enlarged license plate number.

Step S904: and screening according to the content of the characters.

In the embodiment of the application, the composition based on the license plate number has a certain rule, and characters which accord with the license plate number are screened out according to the matching degree of the content contained in the characters and the license plate number, for example, the Chinese in the license plate number is only abbreviated as province, and other Chinese characters do not exist. Based on the content screening of the characters, the characters "heavy, type, mechanical" in fig. 6b and the characters "cold, hidden" in fig. 6b can be eliminated.

In one possible implementation mode, the number length of the license plate is generally 7-8 digits and the like, and when the number of the license plate characters output by recognition exceeds the number of the conventional license plate numbers, corresponding rejection can be carried out. The preset rules can preferably take the characters related to the province abbreviation as the characters of the enlarged license plate number, and remove the characters far away from the province abbreviation. Illustratively, characters in the recognized vehicle image include 'wan KQXXXX 13123456789', the number length of the license plate is generally 7-8 bits, the recognized character 'wan KQXXXX 13123456789' is 18 bits, for this purpose, characters 'KQXXXX' and 'wan' close to the character 'wan' are output as characters of the license plate enlarged size, and characters '13123456789' far from the character 'wan' are removed.

Exemplarily, as shown in fig. 6b, for two lines of characters "cold storage" and "wan KQXXXX", screening is performed according to the content of the characters, and the character "wan KQXXXX" is obtained as an enlarged license plate number.

Step S905: and obtaining a final identification result.

In this embodiment of the application, one of the characters identified in step S902 may be processed in step S903 and step S904, or the characters identified in step S902 may be processed in step S903 and step S904 in combination, and characters of an enlarged license plate number are output to obtain a final identification result. The enlarged number characters of the license plate output in fig. 6a include "zhe FMXXXX", and the enlarged number characters of the license plate output in fig. 6b include "wan KQXXXX".

Step S503: and obtaining the license plate number of the target vehicle according to the identification result of the amplified license plate number.

In the embodiment of the present application, the license plate number of the target vehicle may be obtained according to the recognition result of the enlarged license plate number, the recognition result of the enlarged license plate number may be obtained based on step S502, and the license plate number of the target vehicle may be obtained according to the recognition result of the enlarged license plate number based on the fact that the number of the enlarged license plate number is identical to the number of the license plate. In one possible implementation manner, the identification result can be provided for the user on the display interface. Referring to fig. 10a together, fig. 10a is a representation of the recognition result of fig. 6a, including the enlarged license plate number image of fig. 8a and the license plate number "zhe FXXXXX" on the display interface of fig. 10 a. Referring to fig. 10b together, fig. 10b is a representation of the recognition result of fig. 6b, and the enlarged license plate number image and the license plate number "wan KQXXXX" of fig. 8b are included in the display interface of fig. 10 b.

In the embodiment of the application, further, the identification of the license plate and the identification of the enlarged number of the license plate can be integrated, so that the identification accuracy of the license plate number is improved.

Referring to fig. 11, fig. 11 is a schematic flowchart illustrating another license plate number recognition method according to an embodiment of the present disclosure. The license plate number recognition method can be executed by the license plate number recognition device, and comprises the following steps:

step S111: an image of the vehicle is input.

In this embodiment, the vehicle image may include a license plate region of the target vehicle and a license plate enlarged number region of the target vehicle, that is, the vehicle image includes the license plate and the license plate enlarged number of the target vehicle, as shown in fig. 1a, 1b, 6a, and 6 b. The vehicle image may be an image obtained by the camera shooting the target vehicle alone, or may be an image frame in a video shot by the camera shooting the target vehicle alone. When the camera shoots a plurality of vehicles, an image including the plurality of vehicles is obtained, the position of the target vehicle in the image is located, and the target vehicle is intercepted from the image based on the position area of the target vehicle so as to obtain a vehicle image of the target vehicle.

Step S112: and (5) identifying the license plate.

In the embodiment of the application, when the license plate is identified, the license plate in the vehicle image can be identified, and the identification result of the license plate is obtained. The license plate is also called as license plate, also called as vehicle number plate, is a plate material respectively hung in front of and behind the vehicle, and is usually made of aluminum, iron sheet, plastic or paper, on which the registration number, registration area or other relevant information of the vehicle is engraved.

In one possible implementation mode, a license plate in a vehicle image is detected through a deep learning license plate detection algorithm, the position of the license plate is determined, then characters on the license plate are recognized according to a first character database based on the position of the license plate, and finally the recognized characters are directly output to obtain a recognition result of the license plate. The characters in the first character database comprise license plate characters associated with license plate numbers, such as province short names, city letter codes of various levels and numbers.

It can be understood that the identification of the license plate is different from the identification of the enlarged license plate, and first, the object of the identification is different, the character composition of the enlarged license plate is different from the conventional license plate composition, the conventional license plate is formed by imprinting the registration number, registration area or other relevant information of the vehicle on the plate, and the enlarged license plate is formed by presenting the registration number, registration area or other relevant information of the vehicle on the vehicle body. Characters on the license plate are a whole, while characters of the enlarged number of the license plate are not a whole in general, as shown in fig. 1b and fig. 6 a. Secondly, the detection modes are different, and when the license plate is identified, the characters of the license plate can be identified based on the position of the license plate by detecting the position of the license plate. When the license plate enlarged number is identified, each character in the vehicle image needs to be used as a target for detection, and the position of each character is determined. Secondly, the character databases used in the identification are different, a first character database used by the license plate is identified, the first character database comprises license plate characters associated with the license plate number, a second character database used by the enlarged license plate number is identified, and the second character database comprises license plate characters associated with the license plate number and non-license plate characters not associated with the license plate number. Secondly, the recognition modes are different, and when the license plate is recognized, characters on the plate of the license plate are recognized based on the position of the license plate, and the recognized characters are directly output. And when the license plate enlarged number is identified, after the position of each character in the vehicle image is determined, the character at each position is identified by using the second character database based on the position of each character in the vehicle image. And finally, outputting different recognition results, and obtaining the recognition result of the license plate directly according to the characters output by the recognition of the license plate during the recognition of the license plate. And when the license plate enlarged number is identified, the characters output by identification need to be processed, for example, the identified characters are screened according to a preset rule so that the output characters are matched with the license plate number, and when the characters output by identification comprise license plate characters and non-license plate characters, the license plate characters are only used for generating an identification result of the license plate enlarged number.

Step S113: and judging whether the vehicle image comprises a license plate enlarged number.

In the embodiment of the present application, when the license plate number of the target vehicle is identified by using the enlarged license plate number, it is required to determine whether the target vehicle has the enlarged license plate number, determine whether the target vehicle in the vehicle image includes the enlarged license plate number, execute step S114 when there is no enlarged license plate number, and execute step S115 when there is an enlarged license plate number.

In one possible implementation manner, whether the vehicle image includes the enlarged license plate number is determined according to the vehicle type and the shooting direction, and for example, when the intelligent algorithm detects that the target vehicle in the vehicle image is a vehicle type with the enlarged license plate number, such as a large truck, a van, a cement tank truck or a muck truck, the license plate number recognition device determines whether the enlarged license plate number is included according to the angle direction of the shooting target vehicle. The shooting direction includes a front direction, an oblique side direction, a side direction and a back direction, wherein the front direction refers to a shooting direction for shooting the front of the object, for example, the camera shoots the front of the object facing the head of the target vehicle. The oblique side direction refers to a shooting direction deviating from the front direction or shooting is performed between the left or right surrounding object moving to the side direction, that is, shooting the front part of the object and shooting the side part, for example, shooting towards the head of the target vehicle obliquely. The side direction refers to a shooting direction in which shooting is performed facing the side of the subject, such as shooting is performed facing the side of the subject vehicle. The rear direction refers to a shooting direction in which shooting is performed facing the rear of the subject, such as shooting facing the rear of the subject vehicle.

In the embodiment of the application, when the type of the target vehicle in the vehicle image is detected to be a vehicle type with an enlarged license plate number, and the shooting direction is an oblique side direction, a side direction and a back direction, the vehicle image can be determined to have the enlarged license plate number. On the basis of detecting that the type of the target vehicle in the vehicle image is a vehicle type with an enlarged license plate number, under certain conditions, the target vehicle only presents the enlarged license plate number on the back side, namely the rear part of a carriage, the target vehicle shot based on the oblique side direction and the side direction possibly does not include the enlarged license plate number, and when the shooting direction is the back side direction only, the vehicle image includes the enlarged license plate number. In some cases, the enlarged license plate number in the target vehicle shot in the oblique direction cannot be effectively presented, and it cannot be determined whether the shot target vehicle includes the enlarged license plate number, so that when the shooting direction is the side direction and the back direction, the vehicle image includes the enlarged license plate number. For example, when it is detected that the target vehicle in the vehicle image is a vehicle type with an enlarged license plate number, if the shooting direction is a back direction, and if the shot vehicle image is a tail of the target vehicle, it is determined that the vehicle image includes the enlarged license plate number. If the shooting direction is the front direction, if the shot vehicle image is the head of the target vehicle, the license plate amplification number is generally positioned at the tail or the side of the vehicle based on the license plate amplification number, and then the vehicle image is determined not to include the license plate amplification number.

In the embodiment of the application, the license plate number recognition device may obtain a shooting direction of the target vehicle shot by the camera from the camera, and further may determine the shooting direction according to an angle direction of the target vehicle shot by the camera, for example, the camera is arranged on the lane, and the lens faces a vehicle driving direction when the camera is arranged on the lane, and when the vehicle passes through the camera, the shooting direction of the camera is a back direction, and a tail image of the target vehicle is shot.

In one possible implementation manner, please refer to fig. 12, the display interface may further include a second recognition item, where the second recognition item is used for the user to select a shooting direction of the target vehicle, where the shooting direction includes a front direction, a side direction, and a back direction, and according to the selection of the user, it is determined whether the target vehicle includes a license plate enlarged number, and whether the vehicle image of the target vehicle includes the license plate enlarged number. The intelligent algorithm determines the type of the vehicle, when a user selects the shooting direction to be the side direction or the back direction on the display interface, a second recognition item is triggered, and the license plate number recognition device determines that the vehicle image comprises the license plate enlarged number according to the second recognition item triggered by the user and the fact that the current target vehicle is a vehicle with the license plate enlarged number. When the type and the shooting direction of the vehicle not including the enlarged license plate number are in the front direction, the target vehicle in the vehicle image can be determined not to include the enlarged license plate number.

Step S114: and outputting a license plate recognition result.

In the embodiment of the present application, when it is determined in step S113 that the vehicle image does not have the enlarged license plate number, the recognition result of the license plate is output, and the recognition result of the license plate is presented on the display interface.

Step S115: and judging whether the license plate enlarged number needs to be identified or not.

In the embodiment of the present application, if it is determined in step S115 that the license plate enlarged number does not need to be recognized, step S116 is performed, and if it is determined in step S115 that the license plate enlarged number needs to be recognized, step S117 is performed.

In the embodiment of the application, the main principle of judging whether the license plate enlarged number of the target vehicle needs to be identified is to analyze the license plate enlarged number based on the current license plate number identification result, and finish the judgment analysis according to different selection trends. In one of the ways of determination: if the confidence of the license plate recognition result exceeds a certain threshold value, such as 95%, the current license plate number recognition can be considered to be correct, and the license plate amplified number recognition is not needed. Otherwise, the license plate amplification number recognition is required. The confidence coefficient is the basis for judging the correctness of license plate number recognition by a deep learning license plate detection algorithm or an intelligent algorithm. In another judgment mode, if the user selects the license plate number amplification recognition, and the confidence coefficient of the detected license plate number recognition result does not reach a certain threshold value, such as 99%, the current license plate number recognition result still has an error, and the license plate number amplification recognition needs to be performed. In another judgment mode, if the user does not select the license plate enlarged number recognition, and the confidence coefficient of the license plate recognition result is detected to be lower than a certain threshold value, such as 85%, a prompt can be presented on a display interface to inquire whether the user needs to perform the license plate enlarged number recognition.

And S116, outputting a license plate recognition result.

In the embodiment of the application, when it is determined in step S115 that the enlarged license plate number does not need to be recognized, the recognition result of the license plate is output, and the recognition result of the license plate is presented on the display interface.

Step S117: and (5) identifying the amplified number of the license plate.

In the embodiment of the present application, the identification of the enlarged license plate number may refer to the identification process of the enlarged license plate number, which is not described in detail herein.

In one possible implementation manner, after the license plate of the vehicle image is recognized, the region of the license plate can be determined according to the recognition result of the license plate, so that the region of the license plate in the vehicle image can be shielded during the enlarged number recognition of the license plate, the region of the license plate is prevented from being recognized, and further, when each character in the vehicle image is used as a target for detection and recognition, the license plate can not be detected and recognized.

Step S118: and fusing the recognition result of the license plate and the recognition result of the enlarged license plate number.

In this embodiment of the application, obtaining the license plate number of the target vehicle according to the recognition result of the license plate and the recognition result of the enlarged license plate number, that is, fusing the recognition result of the license plate and the recognition result of the enlarged license plate number, may include: selecting one of the recognition result of the license plate and the recognition result of the enlarged license plate number as the license plate number of the target vehicle; or combining the recognition result of the license plate and the recognition result of the enlarged license plate number, and respectively selecting one part of the recognition result of the license plate and one part of the recognition result of the enlarged license plate number to form the license plate number of the target vehicle.

In the embodiment of the application, the recognition result of the license plate and the recognition result of the enlarged license plate number are fused, the output of the final recognition result is completed, and the license plate number recognition rate and the accuracy rate are improved. The recognition result of the license plate and the recognition result of the enlarged license plate can be fused in various ways. In practical use, the fusion can be realized by adopting a mode based on graph convolution and the like, and the specific use method is not limited in the scheme.

In one possible implementation manner, a recognition result of a license plate and a recognition result of a license plate enlarged number are fused, and one of the recognition result of the license plate and the recognition result of the license plate enlarged number is selected as the license plate number of the target vehicle.

In one possible implementation manner, the selecting one of the recognition result of the license plate and the recognition result of the enlarged license plate number as the license plate number of the target vehicle specifically includes: and acquiring the confidence coefficient of the recognition result of the license plate, and taking the recognition result of the license plate with the enlarged number as the final license plate number codebook application of the target vehicle when the confidence coefficient is detected to be lower than a preset threshold value, wherein the application is not particularly limited. The confidence of the recognition result of the license plate can be the overall confidence of the license plate.

Referring to table 1, table 1 shows the confidence of the license plate recognition result and the confidence of the license plate enlarged number recognition result.

License plate	Zhejiang province	A	1	2	3	4	5	Integral body
									Confidence level	0.7	0.9	0.9	0.9	0.9	0.5	0.9	0.81
License plate enlarged number	Hu (Chinese character of 'Hu')	A	1	2	3	5	6
									Confidence level	0.8	0.8	0.8	0.8	0.8	0.6	0.8	0.77

TABLE 1

As a first example, fusion is performed based on the overall confidence, when recognized characters are output by a license plate and a license plate enlarged number, the overall confidence of each recognition result is carried, an intelligent algorithm determines which recognition result has higher confidence, and selects a recognition result with higher confidence. Such as the license plate output characters in table 1: zhe A12345, the overall confidence of the characters output by the license plate is 81%, and the characters output by the license plate with enlarged numbers are as follows: the overall confidence of the characters output by Hu A12356, license plate enlargement is 77%. At the moment, according to the overall confidence coefficient, Zhe A12345 is finally selected as the recognition result of the fused license plate number.

As a second example, the confidence level of the characters at each position is fused, and when the characters of the license plate and the enlarged license plate are output, the corresponding confidence level is output for the characters at each position, as shown in table 1, the confidence level of the character "zhe" in the license plate recognition result is 0.7. The confidence of the character "Shanghai" in the license plate amplified number recognition result is 0.8. And selecting the character with high confidence coefficient of each position as a fused license plate number recognition result according to the confidence coefficient of the character at each position, and finally selecting Hu A12355 as the fused license plate number recognition result.

As a third example, the license plate prior tendency is fused, it is considered that the recognition effect of the license plate is relatively better, in the fusion, the recognition result of the license plate is preferred in a general scene, and only when the confidence of the license plate is smaller than a certain threshold (for example, 0.6), the recognition result of the enlarged number of the license plate is selected, as shown in table 1, if the license plate is trusted preferentially, the recognition result of the license plate number after fusion is finally selected as zhe a 12345.

Step S119: and outputting the fused license plate number recognition result.

In the embodiment of the application, the fused license plate number recognition result is output according to the selected fusion mode. As in the first example, the merged license plate number recognition result is output as a thunb a 12345. In the second example, the recognition result of the license plate number after fusion is output as the Shanghai A12355. In the third example, the merged license plate number recognition result is output as Zhe A12345.

The following specifically describes two application scenarios, i.e., a completely blocked license plate and a partially blocked license plate.

In a first application scenario, please refer to fig. 13 together, where the license plate of the target vehicle in fig. 13 is completely blocked and may not be recognized, the license plate recognition device may execute steps S111 to S119, complete the enlarged license plate number recognition based on the completion of the license plate detection and recognition, and finally output the recognized license plate number comprehensively according to the recognition result of the license plate and the recognition result of the enlarged license plate number. Specifically, the image shown in fig. 13 is taken as an input image, the license plate in the image is identified by the deep learning license plate detection algorithm, and the coordinate position of the image is given, because the license plate cannot be identified in fig. 13, the license plate cannot be identified at a high probability, and the license plate identification result is not output. And then detecting that the first recognition item is triggered, setting the first recognition item as default, recognizing the license plate enlarged number in the image 13 to obtain a recognition result of the license plate enlarged number, outputting the recognition result of the license plate enlarged number as a fused license plate number recognition result based on the fact that the license plate recognition result is not output, presenting the recognition result of the license plate enlarged number on a display interface, giving a license plate number snapshot image, namely a license plate enlarged number image, according to the position of the license plate enlarged number, and presenting the license plate enlarged number image and the license plate number obtained according to the recognized license plate enlarged number on the display interface.

In a second application scenario, which is illustrated in fig. 1a as an example, the license plate in fig. 1a cannot be completely recognized due to the shooting angle, so that the individual characters cannot be recognized, or even if the individual characters are recognized, the recognition effect is poor. But the enlarged number on the vehicle body is relatively clearly visible, and a final good identification effect can be formed through the identification of the enlarged number of the license plate. The license plate recognition device can execute the steps from S111 to S119, complete the enlarged license plate number recognition based on the detection and recognition of the license plate, and finally comprehensively output the recognized license plate number according to the recognition result of the license plate and the recognition result of the enlarged license plate number. Specifically, with fig. 1a as an input picture, the deep learning license plate detection algorithm identifies the license plate in fig. 1a and provides the coordinate position thereof, because in fig. 1a, the license plate is partially blocked, and finally, the license plate can only be partially correctly output with a high probability, such as correctly outputting abbreviated provinces or each level city letter codes or partial number plate numbers, as in fig. 1a, the first province code and the second city code thereof are invisible, and although the identification confidence is generally low. And then detecting that the first recognition item is triggered, setting the first recognition item as default, recognizing the enlarged license plate number in the graph 1a to obtain a recognition result of the enlarged license plate number, wherein the license plate part is visible in the graph 1a, the enlarged license plate number is completely visible, the recognition result of the enlarged license plate number and the recognition result of the license plate number are combined, the recognition result of the license plate number and the recognition result of the enlarged license plate number are integrated, the fused license plate number recognition result is output, a license plate number snapshot image, namely an enlarged license plate number image, can be selected according to the position of the enlarged license plate number, and the license plate image and the enlarged license plate number image can also be given according to the position of the license plate number and the enlarged license plate number.

In the embodiment of the application, the overall license plate recognition rate can be improved by recognizing the enlarged license plate number, and when the original license plate cannot be recognized, the license plate number recognition result can be provided through the enlarged license plate number, so that the recognition rate is improved. When the original license plate part cannot be identified, the enlarged license plate number can supplement the license plate number identification result, and the license plate number identification accuracy is improved. When the original license plate number can be identified, the license plate number can be further identified through the enlarged number of the license plate, and the accuracy of identifying the whole license plate number is improved.

Furthermore, the existence of the license plate enlarged number is judged, whether the license plate enlarged number needs to be identified is judged, extra calculation overhead can be reduced to the maximum extent, and the license plate enlarged number identification capability can be provided under the scene needing identification most. The license plate amplified number recognition method provided by the method integrates the license plate and the license plate amplification method, and is used as a part of an organic whole, so that the whole license plate recognition rate is improved by the recognition based on the license plate amplified number.

The foregoing embodiment mainly introduces an implementation process of license plate number recognition by a license plate number recognition device, and then introduces an implementation flow of the license plate number method from a viewpoint of interaction between a user and the license plate number recognition device, for example, please refer to fig. 14, where the method may be executed by the license plate number recognition device, and the method includes the following steps:

step S141: providing a display interface, wherein the display interface comprises a first identification item.

In the embodiment of the application, the license plate number recognition device may send the relevant information of the display interface to the client corresponding to the user, and the client displays the display interface so as to provide the display interface for the user, thereby facilitating the user to set the corresponding recognition item, and presenting the license plate number recognition result for the user, including but not limited to the license plate number, the relevant image, and the like.

The display interface may be implemented by referring to the display interfaces described in the embodiments shown in fig. 7, fig. 10a, fig. 10b, and fig. 12, which is not limited in this application.

Step S142: and detecting the triggering of the first recognition item, and recognizing a license plate amplification number in the vehicle image, wherein the license plate amplification number is an identifier which is located on the target vehicle and is different from the license plate.

In the embodiment of the application, the license plate number recognition device can acquire information input by a user on a display interface, for example, the user triggers the first recognition item to acquire information corresponding to the first recognition item, and starts recognition of the enlarged license plate number. In one possible implementation manner, when the identification of the license plate enlarged number is started, only the license plate enlarged number identification can be performed, and the license plate identification and the license plate enlarged number identification can also be performed.

Step S143: and acquiring the recognition result of the enlarged license plate number.

In the embodiment of the present application, the license plate number recognition device may execute the above steps of the license plate number recognition method, which is not described herein again.

Step S144: and obtaining the license plate number of the target vehicle according to the identification result of the amplified license plate number.

After obtaining the recognition result of the enlarged license plate number, the license plate number recognition device may feed back the recognition result of the enlarged license plate number to the client, and optionally, the license plate number recognition device may feed back the enlarged license plate number image, the license plate image, the recognition results of each image, the confidence of each image recognition result, the final license plate number, and the like to the client.

In one possible implementation manner, when the license plate number recognition device recognizes the license plate and the enlarged license plate number at the same time, please refer to fig. 15, and the display interface presents the enlarged license plate number image and the license plate image and the finally output license plate number. The obtaining of the license plate number of the target vehicle according to the recognition result of the license plate and the recognition result of the license plate enlarged number specifically includes: and selecting one of the recognition result of the license plate and the recognition result of the enlarged license plate number as the license plate number of the target vehicle.

The descriptions of the flows corresponding to the above-mentioned figures have respective emphasis, and for parts not described in detail in a certain flow, reference may be made to the related descriptions of other flows.

In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. A computer program product for implementing license plate number recognition comprises one or more computer program instructions for performing license plate number recognition, which when loaded and executed on a computer, result in all or in part in the processes or functions described in fig. 5, 9, and 11, or in all or in part in the processes or functions described in fig. 14, embodiments of the present application.

The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored on a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website, computer, server, or data center to another website, computer, server, or data center via wire (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., Digital Versatile Disk (DVD)), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.

It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing relevant hardware, where the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, etc.

The above-mentioned embodiments are provided not to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A license plate number recognition method is characterized by comprising the following steps:

identifying a license plate enlarged number in a vehicle image, wherein the license plate enlarged number is an identifier which is positioned on a target vehicle and is different from a license plate;

acquiring a recognition result of the enlarged license plate number;

and obtaining the license plate number of the target vehicle according to the identification result of the amplified license plate number.

2. The method of claim 1, wherein the obtaining of the recognition result of the license plate enlarged number specifically comprises:

and when the characters output by the license plate enlarged number in the vehicle image are recognized to comprise license plate characters and non-license plate characters, generating a recognition result of the license plate enlarged number by using the license plate characters only.

3. The method of claim 1, wherein the method further comprises:

identifying the license plate in the vehicle image;

acquiring a recognition result of the license plate;

the obtaining of the license plate number of the target vehicle according to the recognition result of the enlarged license plate number specifically includes:

and obtaining the license plate number of the target vehicle according to the recognition result of the license plate and the recognition result of the license plate enlarged number.

4. The method of claim 3,

the identifying the license plate in the vehicle image specifically includes:

identifying the license plate according to a first character database, wherein the first character database consists of license plate characters;

the method for recognizing the enlarged license plate number in the vehicle image and acquiring the recognition result of the enlarged license plate number specifically comprises the following steps:

recognizing characters in the vehicle image according to a second character database, wherein the recognized and output characters comprise license plate characters and non-license plate characters, and the second character database consists of the license plate characters and the non-license plate characters;

and generating a recognition result of the license plate enlarged number only by using the license plate characters output by recognition.

5. The method of any of claims 1 to 4, further comprising:

and presenting a local image of the license plate enlargement number area in the vehicle image on a display interface.

6. The method of claim 3 or 4, wherein obtaining the license plate number of the target vehicle according to the recognition result of the license plate and the recognition result of the license plate enlarged number comprises:

and selecting one of the recognition result of the license plate and the recognition result of the enlarged license plate number as the license plate number of the target vehicle.

7. The method of claim 6, wherein the selecting one of the recognition result of the license plate and the recognition result of the enlarged license plate number as the license plate number of the target vehicle comprises:

comparing the confidence coefficient of the recognition result of the license plate with the confidence coefficient of the recognition result of the enlarged number of the license plate;

and taking the recognition result with high confidence as the license plate number of the target vehicle.

8. The method of claim 6, wherein the selecting one of the recognition result of the license plate and the recognition result of the enlarged license plate number as the license plate number of the target vehicle comprises:

obtaining the confidence of the recognition result of the license plate;

and when the confidence coefficient is lower than a preset threshold value, taking the recognition result of the license plate enlarged number as the license plate number of the target vehicle.

9. The method of claim 3 or 4, wherein obtaining the license plate number of the target vehicle according to the recognition result of the license plate and the recognition result of the license plate enlarged number comprises:

acquiring a first confidence coefficient of each license plate character in the recognition result of the license plate;

acquiring a second confidence coefficient of each license plate character in the recognition result of the license plate enlarged number;

comparing the first confidence level and the second confidence level;

and selecting the license plate characters with high confidence coefficient to form the license plate number of the target vehicle.

10. A license plate number recognition method is characterized by comprising the following steps:

providing a display interface, wherein the display interface comprises a first identification item;

detecting the triggering of the first recognition item, and recognizing a license plate amplification number in a vehicle image, wherein the license plate amplification number is an identifier which is located on a target vehicle and is different from a license plate;

acquiring a recognition result of the enlarged license plate number;

11. The method of claim 10, wherein the obtaining of the recognition result of the license plate enlarged number specifically comprises:

12. The method of claim 10, wherein the method further comprises:

identifying the license plate in the vehicle image;

acquiring a recognition result of the license plate;

13. The method of claim 12, wherein obtaining the license plate number of the target vehicle according to the recognition result of the license plate and the recognition result of the license plate enlarged number comprises:

14. The method of any of claims 10 to 13, further comprising:

15. The method of any of claims 10 to 14, wherein the display interface further comprises a second identification item for a user to select a shooting direction for the subject vehicle, wherein the shooting direction comprises a front direction, a side direction, and a back direction.

16. The method of any one of claims 12 or 13,

the identifying the license plate in the vehicle image specifically includes:

recognizing characters in the vehicle image according to a second character database, wherein the recognized and output characters comprise license plate characters and non-license plate characters, and the second character database consists of the license plate characters and the non-license plate characters; and generating a recognition result of the license plate enlarged number only by using the license plate characters output by recognition.

17. A license plate number recognition device, wherein the device is configured to perform the method of any one of claims 1-9 or 10-16.

18. A computer storage medium comprising computer instructions that, when executed on an electronic device, cause the electronic device to perform the method of any of claims 1-9 or 10-16.