License plate image classification method and device, electronic equipment and storage medium

Abstract

This invention discloses a method, apparatus, electronic device, and storage medium for classifying license plate images. The method includes: in response to an image classification instruction for a target license plate image, determining image feature data of the target license plate image; inputting the image feature data into a pre-trained license plate classification model; and outputting an execution result for classifying the target license plate image, the execution result including at least two classification items and the probability of each classification item, wherein the at least two classification items are distributed in a tree-like structure. The technical solution of this invention achieves more accurate classification of license plate images, and the classification process is interpretable.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a method, apparatus, electronic device, and storage medium for classifying license plate images. Background Technology

[0002] For license plate image classification, related technologies generally employ manual methods, resulting in a lack of interpretability throughout the process. If the number of license plate images is large, classification errors are common, leading to low accuracy. Summary of the Invention

[0003] This invention provides a method, apparatus, electronic device, and storage medium for classifying license plate images, so as to achieve more accurate classification of license plate images and the classification process is interpretable.

[0004] According to one aspect of the present invention, a license plate image classification method is provided, the method comprising:

[0005] In response to an image classification instruction for a target license plate image, determine the image feature data of the target license plate image;

[0006] The image feature data is input into a pre-trained license plate classification model, and the execution result of classifying the target license plate image is output. The execution result includes at least two classification items and the probability of each classification item, and the at least two classification items are distributed in a tree-like form.

[0007] According to another aspect of the present invention, a license plate image classification device is provided. The device includes:

[0008] The image feature data determination module is used to determine the image feature data of the target license plate image in response to an image classification instruction for the target license plate image;

[0009] A license plate image classification model is used to input the image feature data into a pre-trained license plate classification model and output the execution result of classifying the target license plate image. The execution result includes at least two classification items and the probability of each classification item, and the at least two classification items are distributed in a tree-like form.

[0010] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0011] At least one processor; and

[0012] A memory communicatively connected to the at least one processor; wherein,

[0013] The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform any of the license plate image classification methods described above in this invention.

[0014] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement any of the above-described license plate image classification methods of the present invention.

[0015] The technical solution of this invention involves receiving an image classification instruction for a target license plate image, determining the image feature data of the target license plate image, inputting the image feature data into a pre-trained license plate classification model, and outputting the execution result of classifying the target license plate image. The execution result includes at least two classification items and the probability of each classification item, wherein the at least two classification items are distributed in a tree-like structure. This technical solution achieves more accurate classification of license plate images, and the classification process is interpretable.

[0016] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating a license plate image classification method provided in Embodiment 1 of the present invention;

[0019] Figure 2 An example diagram illustrating the classification items of a license plate image classification method provided in Embodiment 1 of the present invention;

[0020] Figure 3 This is a schematic diagram of the structure of a license plate image classification device provided in Embodiment 2 of the present invention;

[0021] Figure 4 This is a schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0023] It should be noted that the terms "comprising" and "having" and any variations thereof in the specification, claims and accompanying drawings of this invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such processes, methods, products or devices.

[0024] It is understood that before using the technical solutions disclosed in the various embodiments of this disclosure, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this disclosure in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0025] It is understood that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0026] Example 1

[0027] Figure 1 This is a flowchart illustrating a license plate image classification method according to Embodiment 1 of the present invention. This embodiment is applicable to the classification of license plate images. The method can be executed by a license plate image classification device, which can be implemented in hardware and / or software. The license plate image classification device can be configured in an electronic device such as a computer or server.

[0028] like Figure 1 As shown, the method in this embodiment includes:

[0029] S110. In response to an image classification instruction for a target license plate image, determine the image feature data of the target license plate image.

[0030] The target license plate image can be understood as a license plate image that needs to be classified. License plate images can include standard license plate images and non-standard license plate images. It is understood that a standard license plate image can be understood as a standardized license plate image. A non-standard license plate image can be understood as a non-standard license plate image. In practical applications, standard license plate images can include tram license plate images and gasoline vehicle license plate images, etc. That is, standardized license plate images can include tram license plate images and gasoline vehicle license plate images, etc. Non-standard license plate images can include at least one of blank license plate images, damaged license plate images, and obscured license plate images. Image feature data can be feature data obtained after feature extraction from the target license plate image. In this embodiment of the invention, the image classification instruction can be understood as an instruction for classifying the target license plate image. The image classification instruction can carry an identifier for the target license plate image. The identifier for the target license plate image can be used to distinguish different license plate images.

[0031] Specifically, upon receiving an image classification instruction for a target license plate image, features can be extracted from the target license plate image based on the image classification instruction. This yields image feature data of the target license plate image. In this embodiment of the invention, the feature extraction of the target license plate image based on the image classification instruction includes: responding to the image classification instruction by invoking a preset algorithm for feature extraction of the target license plate image, and performing feature extraction on the target license plate image. The preset algorithm can be a feature extraction algorithm pre-set according to actual needs, and is not specifically limited here.

[0032] In this embodiment of the invention, the method for obtaining the target license plate image may specifically be to obtain vehicle video data, and for each video frame of the vehicle video data, determine whether the video frame contains a license plate; if so, then the video frame is used as the target license plate image.

[0033] It should be noted that there are multiple ways to acquire vehicle video data in this embodiment of the invention. For example, vehicle video data can be collected based on multiple cameras; or, vehicle video data can be acquired from storage space used to store vehicle video data; or, vehicle video data can be read from a dashcam. Determining whether the video frame contains a license plate can be understood as determining whether the video frame contains the license plate of the front or rear of the vehicle.

[0034] In this embodiment of the invention, the image classification instruction can be obtained by receiving an image classification instruction input by a user for a target license plate image. Specifically, it can be receiving an image classification instruction input by a user based on an input device for a target license plate image; wherein, the input device can be a physical input device (e.g., a keyboard and / or mouse) or a virtual input device (e.g., a touch area).

[0035] S120. Input the image feature data into a pre-trained license plate classification model and output the execution result of classifying the target license plate image. The execution result includes at least two classification items and the probability of each classification item. The at least two classification items are distributed in a tree-like form.

[0036] The pre-trained license plate classification model can be understood as a model used to classify target license plate images. In other words, the pre-trained license plate classification model can be used to classify target license plate images, that is, it can be used to identify whether a target license plate image is a standard license plate image. The pre-trained license plate classification model can also be used to identify whether a target license plate image is for a tram or a gasoline vehicle when it is a standard license plate image; or, if the target license plate image is not a standard license plate image, it can identify whether it is a damaged license plate or an obscured license plate.

[0037] In this embodiment of the invention, the image feature data is input into a pre-trained license plate classification model, and the output is the classification result for the target license plate image. This can be understood as inputting the image feature data into a pre-trained license plate classification model. The target license plate image is then analyzed based on the pre-trained license plate classification model. This process of analyzing the target license plate image yields the classification result for that image.

[0038] In this embodiment of the invention, the execution result may include at least two classification items and the probability of each classification item, and the at least two classification items are distributed in a tree-like form. The classification items may include license plate type and / or license plate information (e.g., license plate number). See also Figure 2 License plate types can include standard license plates, non-standard license plates, license plates for gasoline vehicles, license plates for electric vehicles, license plates that are obscured, and no license plates. Among them, no license plate can be understood as a blank license plate.

[0039] Specifically, after inputting the target license plate image into the pre-trained license plate classification model, the model can first output the probability that the target license plate image is a standard license plate image and the probability that the target license plate image is a non-standard license plate image. Then, for the case where the target license plate image is a standard license plate image, the standard license plate can be used as the parent classification item, and gasoline vehicle license plates and electric vehicle license plates can be used as child classification items. The model outputs the probability and license plate information of the target license plate image being a gasoline vehicle license plate image among standard license plate images, and the probability and license plate information of the target license plate image being an electric vehicle license plate image among standard license plate images. For the case where the target license plate image is a non-standard license plate image, the non-standard license plate can be used as the parent classification item, and occluded license plates and blank license plates can be used as child classification items. The model outputs the probability that the target license plate image is an occluded license plate image among non-standard license plate images, and the probability that the target license plate image is a blank license plate image among non-standard license plate images.

[0040] To improve the interpretability of license plate recognition, in embodiments of the present invention, the target license plate image is typically used as the root node, that is, the target license plate image is used as the classification item at the root of the tree.

[0041] In this embodiment of the invention, the pre-trained license plate classification model includes a trained recurrent residual neural network and a trained decision tree. The step of inputting the image feature data into the pre-trained license plate classification model and outputting the classification result for the target license plate image may include: inputting the image feature data into the trained recurrent residual neural network, thereby obtaining a classification feature vector of the target license plate image. After obtaining the classification feature vector, the classification feature vector can be input into the trained decision tree, thereby obtaining the classification result for the target license plate image.

[0042] Specifically, the image feature data is input into the trained recurrent residual neural network. A classification feature vector for the target license plate image can be obtained based on the fully connected layers of the recurrent residual neural network. After obtaining the classification feature vector, it can be input into the trained decision tree. This allows for the classification of the target license plate image.

[0043] The trained decision tree includes at least two nodes. The step of inputting the classification feature vector into the trained decision tree to obtain the classification result for the target license plate image can include: inputting the classification feature vector into the trained decision tree. This allows calculation of the dot product between the classification feature vector and the node attribute vector of each node. The dot product results are then obtained. Based on these dot product results, the classification result for the target license plate image can be determined.

[0044] In practical applications, a trained decision tree typically includes multiple nodes, and these nodes are interconnected. Node attributes can include license plate type. A node attribute vector can be understood as a feature vector representing the license plate type.

[0045] In this embodiment of the invention, the current node can be determined based on each node in the trained decision tree. The dot product of the classification feature vector and the node attribute vector of the current node can then be calculated. The result of the dot product of the classification feature vector and the node attribute vector of the current node can then be obtained. Based on this dot product result, the probability that the target license plate image is a license plate type image corresponding to the current node can be obtained. The current node and its probability are output and displayed. After displaying the current node and its probability, the process can return to performing the operation of determining the current node based on each node in the trained decision tree. The traversal of the trained decision tree ends after all nodes in the trained decision tree have been traversed.

[0046] The determination of the current node based on the nodes in the trained decision tree can include: determining the current node according to the hierarchical relationship of the nodes in the trained decision tree. It should be noted that the number of current nodes can be one, two, or more, and all current nodes can be located at the same level in the trained decision tree. In practical applications, the hierarchical relationship can represent the parent-child relationship between nodes; in this embodiment of the invention, the parent node can be used as the current node.

[0047] Based on the above embodiments, the method further includes: acquiring training sample data and expected output data corresponding to the training sample data. The training sample data consists of license plate images, and the expected output data consists of the license plate type and / or license plate number corresponding to each license plate image. After acquiring the sample data and the expected data, the training sample data can be input into a pre-constructed initial recurrent residual neural network. This allows the actual output result of the recurrent residual neural network to be obtained. Furthermore, based on the expected output data and the actual output result, the network parameters of the initial recurrent residual neural network can be adjusted to obtain the trained recurrent residual neural network.

[0048] The initial recurrent residual neural network can be understood as a pre-constructed recurrent residual neural network. In this embodiment of the invention, the method of obtaining training sample data and the expected output data corresponding to the training sample data can specifically include: acquiring multiple license plate images and setting labels for each license plate image. A dataset is then constructed based on the acquired license plate images and their labels. After obtaining the dataset, data with a preset sample data partitioning ratio can be obtained from the dataset as a training set. Then, training sample data and the expected output data corresponding to the training sample data can be obtained from the training set. The preset sample data partitioning ratio can be set according to actual needs. In practical applications, the preset sample data partitioning ratio is usually the ratio of data in the training set to data in the test set.

[0049] To improve the accuracy of the regressive residual neural network, this embodiment of the invention further includes: acquiring test sample data and the expected test results corresponding to the test sample data. The test sample data can then be input into the trained regressive residual neural network, thereby obtaining test output results. Based on the test output results and the expected test results, the accuracy of the trained regressive residual neural network can be determined. If the accuracy does not reach a preset accuracy threshold, the training sample data is optimized. The preset accuracy threshold can be set according to actual needs, such as 0.85, 0.8, or 0.9. Optimizing the training sample data may include: data cleaning of the training sample data.

[0050] Based on the above embodiments, the method further includes: constructing an initial decision tree based on a preset decision tree generation algorithm; training the initial decision tree using the training sample data and the expected output data corresponding to the training sample data to obtain a trained decision tree. The preset decision tree generation algorithm can be any one of the ID3 decision tree algorithm, C4.5 decision tree algorithm, and CART decision tree algorithm. The initial decision tree can be a decision tree constructed based on the preset decision tree generation algorithm.

[0051] Based on the above embodiments, after obtaining the trained recurrent residual neural network, the fully connected layers of the trained recurrent residual neural network include latent vectors for identifying license plate types. Then, hierarchical classification can be performed on the initial decision tree based on these latent vectors, and the hierarchical structure of the initial decision tree can be constructed using WordNet technology. After construction, the connection between the trained recurrent residual neural network and the initial decision tree can be obtained. The construction of the hierarchical structure of the initial decision tree using WordNet technology can include: calculating the cross-entropy between different license plate types, and constructing the hierarchical structure of the initial decision tree based on the cross-entropy.

[0052] Based on the above embodiments, after obtaining the execution result of classifying the target license plate image, the execution result of classifying the target license plate image can be displayed in a tree-like form.

[0053] It should be noted that, in this embodiment of the invention, after obtaining the probabilities, the classification result of the target license plate image can be determined based on these probabilities. It should also be noted that, when the target license plate image is a standardized license plate image, the pre-trained license plate classification model can identify the license plate information (such as the license plate number) in the target license plate image. After identifying the license plate information, it can be used as a classification item and displayed.

[0054] The technical solution of this invention involves receiving an image classification instruction for a target license plate image, determining the image feature data of the target license plate image, inputting the image feature data into a pre-trained license plate classification model, and outputting the execution result of classifying the target license plate image. The execution result includes at least two classification items and the probability of each classification item, wherein the at least two classification items are distributed in a tree-like structure. This technical solution achieves more accurate classification of license plate images, and the classification process is interpretable.

[0055] Example 2

[0056] Figure 3 This is a schematic diagram of the structure of a license plate image classification device provided in Embodiment 2 of the present invention. Figure 3 As shown, the device includes: an image feature data determination module 210 and a license plate image classification model 220.

[0057] The image feature data determination module 210 is used to determine the image feature data of the target license plate image in response to an image classification instruction for the target license plate image.

[0058] The license plate image classification model 220 is used to input the image feature data into a pre-trained license plate classification model and output the execution result of classifying the target license plate image. The execution result includes at least two classification items and the probability of each classification item, and the at least two classification items are distributed in a tree-like form.

[0059] The technical solution of this invention, through an image feature data determination module, determines the image feature data of the target license plate image in response to an image classification instruction for the target license plate image. The image feature data is input into a pre-trained license plate classification model, which outputs the classification result for the target license plate image. The result includes at least two classification items and the probability of each item, wherein the at least two classification items are distributed in a tree structure. This technical solution achieves more accurate license plate image classification, and the classification process is interpretable.

[0060] Optionally, the pre-trained license plate classification model includes a trained recurrent residual neural network and a trained decision tree; the license plate image classification model 220 is used to input the image feature data into the trained recurrent residual neural network to obtain the classification feature vector of the target license plate image; and input the classification feature vector into the trained decision tree to obtain the execution result of classifying the target license plate image.

[0061] Optionally, the trained decision tree includes at least two nodes; the license plate image classification model 220 is used to input the classification feature vector into the trained decision tree, calculate the vector dot product of the classification feature vector with the node attribute vector of each node, and obtain the vector dot product results; based on the vector dot product results, determine the execution result of classifying the target license plate image.

[0062] Optionally, the device further includes: a recurrent residual neural network training module, used to acquire training sample data and expected output data corresponding to the training sample data; wherein the training sample data is license plate images, and the expected output data is the license plate type and / or license plate number corresponding to each license plate image; the training sample data is input into a pre-constructed initial recurrent residual neural network to obtain the actual output result of the recurrent residual neural network; and the network parameters of the initial recurrent residual neural network are adjusted according to the expected output data and the actual output result to obtain the trained recurrent residual neural network.

[0063] Optionally, the device further includes: a residual neural network testing module, used to acquire test sample data and expected test results corresponding to the test sample data; input the test sample data into the trained residual neural network to obtain test output results; determine the accuracy of the trained residual neural network based on the test output results and the expected test results; and optimize the training sample data if the accuracy does not reach a preset accuracy threshold.

[0064] Optionally, it further includes: a decision tree training module, used to construct an initial decision tree based on a preset decision tree generation algorithm; and to train the initial decision tree using the training sample data and the expected output data corresponding to the training sample data to obtain a trained decision tree.

[0065] Optionally, the device further includes a target license plate image acquisition module, used to acquire vehicle video data, and for each video frame of the vehicle video data, determine whether the video frame contains a license plate; if so, then use the video frame as the target license plate image.

[0066] The license plate image classification device provided in this embodiment of the invention can execute the license plate image classification method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0067] It is worth noting that the various units and modules included in the above-mentioned license plate image classification device are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the protection scope of the embodiments of the present invention.

[0068] Example 3

[0069] Figure 4 A schematic diagram of an electronic device 10 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0070] like Figure 4As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0071] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0072] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as license plate image classification methods.

[0073] In some embodiments, the license plate image classification method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the license plate image classification method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the license plate image classification method by any other suitable means (e.g., by means of firmware).

[0074] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0075] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0076] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0077] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0078] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0079] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0080] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0081] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for classifying license plate images, characterized in that, include: In response to an image classification instruction for a target license plate image, determine the image feature data of the target license plate image; The image feature data is input into a pre-trained license plate classification model, which outputs the classification result for the target license plate image. The execution result includes at least two classification items and the probability of each classification item. The at least two classification items are distributed in a tree structure. The classification items include license plate type and / or license plate information. The license plate type includes standard license plate, non-standard license plate, gasoline vehicle license plate, electric vehicle license plate, obscured license plate, and blank license plate. The image feature data is feature data obtained after feature extraction of the target license plate image. The image classification instruction is an instruction for classifying the target license plate image. The image classification instruction includes the identifier of the target license plate image, which is used to distinguish different license plate images. The pre-trained license plate classification model includes a trained recurrent residual neural network and a trained decision tree. The trained recurrent residual neural network is used to obtain the classification feature vector of the target license plate image based on the image feature data. The trained decision tree is used to obtain the classification result for the target license plate image based on the classification feature vector. The output of the classification result for the target license plate image includes: Output the probability that the target license plate image is a standard license plate image, and the probability that the target license plate image is a non-standard license plate image; If the target license plate image is a standardized license plate image, the standardized license plate is taken as the parent classification item, and the license plates of oil vehicles and electric vehicles are taken as the child classification items. The probability and license plate information of the target license plate image being an oil vehicle license plate image in the standardized license plate image are output, as well as the probability and license plate information of the target license plate image being an electric vehicle license plate image in the standardized license plate image. When the target license plate image is a non-standard license plate image, non-standard license plates are used as the parent classification item, and occluded license plates and blank license plates are used as the child classification items. The output is the probability that the target license plate image is an occluded license plate image among non-standard license plate images, and the output is the probability that the target license plate image is a blank license plate image among non-standard license plate images.

2. The method according to claim 1, characterized in that, The trained decision tree includes at least two nodes; the step of inputting the classification feature vector into the trained decision tree to obtain the classification result for the target license plate image includes: The classification feature vector is input into the trained decision tree, and the dot product of the classification feature vector with the node attribute vector of each node is calculated to obtain the dot product results. Based on the inner product results of each vector, the execution result of classifying the target license plate image is determined.

3. The method according to claim 1, characterized in that, The method further includes: Acquire training sample data and expected output data corresponding to the training sample data; wherein, the training sample data is license plate images, and the expected output data is the license plate type and / or license plate number corresponding to each of the license plate images; The training sample data is input into a pre-constructed initial recurrent residual neural network to obtain the actual output result of the recurrent residual neural network; Based on the expected output data and the actual output results, the network parameters of the initial recurrent residual neural network are adjusted to obtain the trained recurrent residual neural network.

4. The method according to claim 3, characterized in that, The method further includes: Obtain test sample data and the expected test results corresponding to the test sample data; The test sample data is input into the trained residual neural network to obtain the test output results; Based on the test output results and the expected test results, the accuracy of the trained residual neural network is determined. If the accuracy does not reach the preset accuracy threshold, the training sample data is optimized.

5. The method according to claim 3, characterized in that, The method further includes: An initial decision tree is constructed based on a pre-defined decision tree generation algorithm. The initial decision tree is trained using the training sample data and the expected output data corresponding to the training sample data to obtain a trained decision tree.

6. The method according to claim 1, characterized in that, The method further includes: Acquire vehicle video data, and for each video frame of the vehicle video data, determine whether the video frame contains a license plate. If so, use the video frame as the target license plate image.

7. An image classification device, characterized in that, include: The image feature data determination module is used to determine the image feature data of the target license plate image in response to an image classification instruction for the target license plate image; A license plate image classification model is used to input the image feature data into a pre-trained license plate classification model and output the execution result of classifying the target license plate image. The execution result includes at least two classification items and the probability of each classification item. The at least two classification items are distributed in a tree structure. The classification items include license plate type and / or license plate information. The license plate type includes standard license plates, non-standard license plates, gasoline vehicle license plates, electric vehicle license plates, obscured license plates, and blank license plates. The image feature data is obtained after feature extraction from the target license plate image. The image classification instruction is a command for classifying target license plate images. The image classification instruction includes an identifier for the target license plate image, which is used to distinguish different license plate images. The pre-trained license plate classification model includes a trained recurrent residual neural network and a trained decision tree. The trained recurrent residual neural network is used to obtain a classification feature vector of the target license plate image based on the image feature data. The trained decision tree is used to obtain the execution result of classifying the target license plate image based on the classification feature vector. The output of the classification result for the target license plate image includes: Output the probability that the target license plate image is a standard license plate image, and the probability that the target license plate image is a non-standard license plate image; If the target license plate image is a standardized license plate image, the standardized license plate is taken as the parent classification item, and the license plates of oil vehicles and electric vehicles are taken as the child classification items. The probability and license plate information of the target license plate image being an oil vehicle license plate image in the standardized license plate image are output, as well as the probability and license plate information of the target license plate image being an electric vehicle license plate image in the standardized license plate image. When the target license plate image is a non-standard license plate image, non-standard license plates are used as the parent classification item, and occluded license plates and blank license plates are used as the child classification items. The output is the probability that the target license plate image is an occluded license plate image among non-standard license plate images, and the output is the probability that the target license plate image is a blank license plate image among non-standard license plate images.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the license plate image classification method according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the license plate image classification method according to any one of claims 1-6.

Images