An automobile sign detection method and device, a terminal device, and a storage medium

Abstract

The application discloses a kind of automobile sign detection method, device, terminal equipment and storage medium, method includes: based on the image of the automobile to be detected, at least one sign label area in the image of the automobile to be detected is carried out coordinate normalization processing, obtains at least one first sign label area;According to the coordinate data of at least one first sign label area, the first sign label area is classified, and at least one second sign label area is obtained;The image corresponding to each second sign label area is input into sign identification model to carry out sign identification, and the sign identification information corresponding to each second sign label area is obtained;Based on the vehicle model number of the automobile to be detected and the preset area number corresponding to each second sign label area, the vehicle model number, the preset area number, the sign identification information are compared with sign contrast table, and the automobile sign detection result is obtained.The present application can detect that sign is pasted wrong area and sign is pasted wrong vehicle and so on sign mispasting condition.

Description

Technical Field

[0001] This invention relates to the field of vehicle inspection technology, and in particular to a method, apparatus, terminal equipment, and computer-readable storage medium for detecting vehicle license plates. Background Technology

[0002] During the automobile production process, it is usually necessary to affix labels to the vehicle body. Since different vehicle models have different labels, and the placement of the labels on the vehicle body is also different, it is necessary to inspect the labels on each vehicle during the automobile production process to ensure that the labels are affixed correctly.

[0003] Existing vehicle sign inspection methods can usually only detect whether there is a sign at the designated signage location on the vehicle body, but they are not good at detecting misplaced signs. Summary of the Invention

[0004] This invention provides a method, apparatus, terminal device, and storage medium for detecting vehicle signs, in order to solve the technical problem that existing vehicle sign detection methods are unable to detect misplaced signs and missing sign elements. By comparing the vehicle model number, area number, and sign identification information corresponding to each sign marking area with a sign comparison table, it is possible to detect misplaced signs, such as signs being placed in the wrong area or on the wrong vehicle.

[0005] To address the aforementioned technical problems, a first aspect of the present invention provides a method for detecting vehicle license plates, comprising the following steps:

[0006] Based on the pre-acquired vehicle image to be detected, coordinate normalization processing is performed on at least one pre-defined sign marking area in the vehicle image to be detected to obtain at least one first sign marking area.

[0007] Based on the coordinate data of the at least one first sign marking area, the at least one first sign marking area is classified to obtain at least one second sign marking area; wherein, the second sign marking area includes at least one first sign marking area;

[0008] The image corresponding to each of the second sign marking areas is input into a preset sign recognition model for sign recognition, thereby obtaining sign recognition information corresponding to each of the second sign marking areas;

[0009] Based on the pre-acquired vehicle model number of the vehicle to be detected and the preset area number corresponding to each of the second sign marking areas, the vehicle model number, the preset area number, and the sign recognition information are compared with a preset sign lookup table to obtain the vehicle sign detection result;

[0010] The sign comparison table includes several sign comparison information entries, and each sign comparison information entry includes at least a vehicle model comparison number, a region comparison number, and sign comparison information.

[0011] As a preferred embodiment, the method specifically obtains the sign recognition model through the following steps:

[0012] A pre-acquired sign marking image is used to train a pre-set deep convolutional neural network to obtain a pre-trained deep convolutional neural network; wherein, the sign marking image includes at least one car sign, and each car sign is pre-marked with sign marking information;

[0013] In the sign marking image, the sign elements in the car sign are removed to obtain a defective sign, and the sign marking information of the defective sign is modified to mark the defective sign and obtain a defect marking image;

[0014] The defect-marked images are used to train the pre-trained deep convolutional neural network to obtain the sign recognition model.

[0015] As a preferred embodiment, the step of comparing the vehicle model number, the preset area number, and the sign recognition information with a preset sign lookup table to obtain the vehicle sign detection result specifically includes the following steps:

[0016] The vehicle model number, the preset area number, and the sign identification information are compared with each sign comparison information in the sign comparison table to determine whether there is any sign comparison information that is the same as the vehicle model number, the preset area number, and the sign identification information.

[0017] If so, the vehicle sign inspection result is deemed qualified;

[0018] If not, the vehicle sign inspection result is deemed unqualified.

[0019] As a preferred embodiment, the step of classifying the at least one first sign marking area based on the coordinate data of the at least one first sign marking area to obtain at least one second sign marking area specifically includes the following steps:

[0020] Based on the coordinate data of the at least one first sign marking area, the at least one first sign marking area is classified using the nearest neighbor algorithm to obtain at least one area category;

[0021] Based on the coordinate data of the first sign marking area in each of the area categories, at least one second sign marking area is obtained.

[0022] As a preferred embodiment, the step of performing coordinate normalization processing on at least one pre-defined sign marking region in the vehicle image to be detected to obtain at least one first sign marking region specifically includes the following steps:

[0023] Based on the width and height of the vehicle image to be detected, the coordinates of the sign marking area are normalized using the following expression to obtain the first sign marking area:

[0024]

[0025] Where x1 represents the x-coordinate of the center point of the first sign marking area, x represents the x-coordinate of the center point of the sign marking area, W represents the width of the vehicle image to be detected, y1 represents the y-coordinate of the center point of the first sign marking area, y represents the y-coordinate of the center point of the sign marking area, H represents the height of the vehicle image to be detected, width1 represents the width of the first sign marking area, width represents the width of the sign marking area, height1 represents the height of the first sign marking area, and height represents the height of the sign marking area.

[0026] As a preferred embodiment, obtaining at least one second sign marking area based on the coordinate data of the first sign marking area in each of the area categories specifically includes the following steps:

[0027] Based on the coordinate data of the first sign marking area in each area category, the second sign marking area corresponding to the current area category is represented by the following expression: [min(x1), min(y1), max(x1+width1)-min(x1), max(y1+height1)-min(y1)].

[0028] As a preferred embodiment, the method specifically obtains the vehicle model number of the car to be detected through the following steps:

[0029] The vehicle model of the vehicle to be inspected is obtained by scanning the vehicle identification code pre-deployed on the vehicle to be inspected using a barcode scanner.

[0030] Based on the vehicle model to be tested, determine the vehicle model number.

[0031] A second aspect of the present invention provides a vehicle license plate detection device, comprising:

[0032] The coordinate normalization module is used to perform coordinate normalization processing on at least one pre-set sign marking area in the pre-acquired vehicle image to be detected, so as to obtain at least one first sign marking area.

[0033] The region classification module is used to classify the at least one first sign marking region according to the coordinate data of the at least one first sign marking region to obtain at least one second sign marking region; wherein, the second sign marking region includes at least one first sign marking region;

[0034] The sign recognition module is used to input the image corresponding to each of the second sign marking areas into a preset sign recognition model for sign recognition, and obtain the sign recognition information corresponding to each of the second sign marking areas;

[0035] The vehicle sign detection module is used to compare the vehicle model number, the preset area number, and the sign recognition information with a preset sign lookup table based on the pre-acquired vehicle model number and the preset area number corresponding to each of the second sign marking areas, in order to obtain the vehicle sign detection result.

[0036] The sign comparison table includes several sign comparison information entries, and each sign comparison information entry includes at least a vehicle model comparison number, a region comparison number, and sign comparison information.

[0037] A third aspect of the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the vehicle license plate detection method as described in any of the first aspects.

[0038] A fourth aspect of the present invention provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the vehicle license plate detection method as described in any of the first aspects.

[0039] Compared with the prior art, the beneficial effect of the embodiments of the present invention is that by comparing the vehicle model number, area number and sign identification information corresponding to each sign marking area with the sign comparison table, it is possible to detect whether the correspondence between the vehicle model number, area number and sign identification information is correct, thereby detecting sign misplacement situations such as signs being placed in the wrong area or signs being placed on the wrong vehicle. Attached Figure Description

[0040] Figure 1 This is a flowchart illustrating the vehicle sign detection method in an embodiment of the present invention;

[0041] Figure 2 This is a classification diagram of the first sign marking area in an embodiment of the present invention;

[0042] Figure 3This is a schematic diagram of the structure of the vehicle license plate detection device in an embodiment of the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 are within the scope of protection of the present invention.

[0044] See Figure 1 The first aspect of this invention provides a method for detecting vehicle license plates, comprising the following steps S1 to S4:

[0045] Step S1: Based on the pre-acquired vehicle image to be detected, perform coordinate normalization processing on at least one pre-defined sign marking area in the vehicle image to be detected to obtain at least one first sign marking area.

[0046] Step S2: Based on the coordinate data of the at least one first sign marking area, classify the at least one first sign marking area to obtain at least one second sign marking area; wherein, the second sign marking area includes at least one first sign marking area;

[0047] Step S3: Input the image corresponding to each of the second sign marking areas into a preset sign recognition model for sign recognition, and obtain the sign recognition information corresponding to each of the second sign marking areas;

[0048] Step S4: Based on the pre-acquired vehicle model number of the vehicle to be detected and the preset area number corresponding to each of the second sign marking areas, compare the vehicle model number, the preset area number, and the sign recognition information with a preset sign lookup table to obtain the vehicle sign detection result;

[0049] The sign comparison table includes several sign comparison information entries, and each sign comparison information entry includes at least a vehicle model comparison number, a region comparison number, and sign comparison information.

[0050] It is worth noting that, considering that car signs may be affixed to various locations on a car, the car image to be detected in this embodiment includes the front image, the body image, and the rear image. After obtaining the car image to be detected, the staff marks the car signs in the image, for example, by representing the marking frame of the car sign in the following coordinate form: [x, y, width, height].

[0051] Furthermore, coordinate normalization is performed on at least one sign marking region in the vehicle image to be detected, obtaining at least one first sign marking region. Based on the coordinate data of at least one first sign marking region, the at least one first sign marking region is classified to obtain at least one second sign marking region; wherein, the second sign marking region includes at least one first sign marking region. It is worth noting that for multiple vehicle signs that are close together, such as... Figure 2 As shown, the black rectangle is the first sign marking area, and the white rectangle surrounding several black rectangles is the second sign marking area. In order to reduce the number of images input to the sign recognition model, this embodiment classifies the first sign marking area according to the coordinate data of the first sign marking area, divides the first sign marking areas that are close to each other into the same type, and obtains the second sign marking area corresponding to each type.

[0052] Furthermore, the image corresponding to each second sign marking area is input into a preset sign recognition model for sign recognition to obtain sign recognition information corresponding to each second sign marking area.

[0053] As one optional embodiment, in order to distinguish different signs, this embodiment assigns different signs a number, such as b1, b2, b3, etc. The sign recognition model is used to recognize the car signs in the second sign marking area, and the sign number corresponding to each car sign can be identified as the sign recognition information.

[0054] Furthermore, based on the pre-acquired vehicle model number of the car to be tested and the preset area number corresponding to each second sign marking area, the vehicle model number, preset area number, and sign recognition information are compared with a preset sign lookup table to obtain the vehicle sign detection result. The sign lookup table includes several sign lookup information entries, each of which includes at least a vehicle model lookup number, an area lookup number, and sign lookup information. It is worth noting that in order to detect misplaced signs, it is necessary to bind the sign recognition information with the pasting area and vehicle model to establish a correspondence among the three. Understandably, the sign lookup table is pre-designed according to the sign placement plan for each vehicle model. Therefore, by comparing the pre-acquired vehicle model number of the car to be tested, the preset area number corresponding to each second sign marking area, and the sign recognition information with the sign lookup table, it is possible to detect whether there are misplaced signs in each second sign marking area.

[0055] For example, the above signage comparison table is shown in Table 1 below:

[0056] Table 1 Signage Comparison Table

[0057] Model number Area code Signage Number c1 1 b1 c1 1 b2 c1 2 b3 c2 1 b1 c2 1 b4 c2 2 b5

[0058] Assuming the vehicle model number of the car to be tested is c1, the preset area number corresponding to the current second sign marking area is 1, and the sign comparison information corresponding to the current second sign marking area is b2 and b3, that is, there are two different signs in the current second sign marking area, namely b2 and b3. Comparing with the sign comparison table in Table 1, it is detected that the vehicle model number and area number corresponding to sign b2 are the same as those in the sign comparison table, while the area number corresponding to sign b3 is different from the standard comparison table. Therefore, sign b3 is placed in the wrong area.

[0059] The present invention provides a vehicle sign detection method that can detect misplaced signs, such as signs being placed in the wrong area or on the wrong vehicle, by comparing the vehicle model number, area number, and sign identification information corresponding to each sign marking area with a sign comparison table.

[0060] As a preferred embodiment, the method specifically obtains the sign recognition model through the following steps:

[0061] A pre-acquired sign marking image is used to train a pre-set deep convolutional neural network to obtain a pre-trained deep convolutional neural network; wherein, the sign marking image includes at least one car sign, and each car sign is pre-marked with sign marking information;

[0062] In the sign marking image, the sign elements in the car sign are removed to obtain a defective sign, and the sign marking information of the defective sign is modified to mark the defective sign and obtain a defect marking image;

[0063] The defect-marked images are used to train the pre-trained deep convolutional neural network to obtain the sign recognition model.

[0064] It is worth noting that before training the deep convolutional neural network, it is necessary to acquire car images through a camera, such as images of the front, body, and rear of the car. The staff will then label the car signs in the images, for example, by representing the label box of the car sign in the following coordinate form: [x, y, width, height]. Information is marked on each car sign in the car image, for example, by using numbers. The image corresponding to the label area in the car image after information marking is cropped to obtain several label images.

[0065] In this embodiment, a deep convolutional neural network is trained using the sign marking image to obtain a pre-trained deep convolutional neural network. Further, to enable the deep convolutional neural network to identify defective signs, the sign elements in the car sign are first removed from the sign marking image. For example, image processing techniques such as Photoshop can be used to remove a letter from the car sign to obtain a defective sign. The sign marking information of the defective sign is then modified to mark it as defective, resulting in a defect-marked image. For instance, defective signs lacking sign elements are marked as background, thus preventing defective signs from being judged as valid signs.

[0066] It is worth noting that signage elements include, but are not limited to, graphic elements, character elements, and text elements.

[0067] Then, the defect-marked images are used to train the pre-trained deep convolutional neural network, enabling the deep convolutional neural network to learn the features of each sign element in the car sign, thereby obtaining a sign recognition model.

[0068] As one alternative implementation, the deep convolutional neural network uses EfficientNet.

[0069] As a preferred embodiment, the step of comparing the vehicle model number, the preset area number, and the sign recognition information with a preset sign lookup table to obtain the vehicle sign detection result specifically includes the following steps:

[0070] The vehicle model number, the preset area number, and the sign identification information are compared with each sign comparison information in the sign comparison table to determine whether there is any sign comparison information that is the same as the vehicle model number, the preset area number, and the sign identification information.

[0071] If so, the vehicle sign inspection result is deemed qualified;

[0072] If not, the vehicle sign inspection result is deemed unqualified.

[0073] It is worth noting that since each sign comparison information in the sign comparison table contains a set of vehicle model comparison information, region comparison information, and sign comparison information, in order to check whether the correspondence between the vehicle model number, the preset region number, and the sign recognition information is correct, the vehicle model number, the preset region number, and the sign recognition information are compared with each sign comparison information in the sign comparison table to determine whether there is any sign comparison information that is the same as the vehicle model number, the preset region number, and the sign recognition information.

[0074] If so, it indicates that there are no defects in the sign elements and no misplacement, therefore the vehicle sign inspection result is deemed qualified;

[0075] If not, the vehicle sign inspection result is deemed unqualified. Furthermore, if the vehicle model number and / or preset area number differ from the sign comparison information, it indicates that the sign is misplaced; if the sign identification information is a defect mark, it indicates that the sign has defective elements.

[0076] As a preferred embodiment, the step of classifying the at least one first sign marking area based on the coordinate data of the at least one first sign marking area to obtain at least one second sign marking area specifically includes the following steps:

[0077] Based on the coordinate data of the at least one first sign marking area, the at least one first sign marking area is classified using the nearest neighbor algorithm to obtain at least one area category;

[0078] Based on the coordinate data of the first sign marking area in each of the area categories, at least one second sign marking area is obtained.

[0079] As a preferred embodiment, the step of performing coordinate normalization processing on at least one pre-defined sign marking region in the vehicle image to be detected to obtain at least one first sign marking region specifically includes the following steps:

[0080] Based on the width and height of the vehicle image to be detected, the coordinates of the sign marking area are normalized using the following expression to obtain the first sign marking area:

[0081]

[0082] Where x1 represents the x-coordinate of the center point of the first sign marking area, x represents the x-coordinate of the center point of the sign marking area, W represents the width of the vehicle image to be detected, y1 represents the y-coordinate of the center point of the first sign marking area, y represents the y-coordinate of the center point of the sign marking area, H represents the height of the vehicle image to be detected, width1 represents the width of the first sign marking area, width represents the width of the sign marking area, height1 represents the height of the first sign marking area, and height represents the height of the sign marking area.

[0083] As a preferred embodiment, obtaining at least one second sign marking area based on the coordinate data of the first sign marking area in each of the area categories specifically includes the following steps:

[0084] Based on the coordinate data of the first sign marking area in each area category, the second sign marking area corresponding to the current area category is represented by the following expression: [min(x1), min(y1), max(x1+width1)-min(x1), max(y1+height1)-min(y1)].

[0085] It is worth noting that the range of all first sign marking areas in the current area category can be calculated using the expression [min(x1), min(y1), max(x1+width1)-min(x1), max(y1+height1)-min(y1)], which can then be used to represent the second sign marking area.

[0086] As a preferred embodiment, the method specifically obtains the vehicle model number of the car to be detected through the following steps:

[0087] The vehicle model of the vehicle to be inspected is obtained by scanning the vehicle identification code pre-deployed on the vehicle to be inspected using a barcode scanner.

[0088] Based on the vehicle model to be tested, determine the vehicle model number.

[0089] It is worth noting that in this embodiment, a vehicle identification number (VIN), such as a QR code, is generated in advance based on the vehicle model of each vehicle to be inspected. This VIN is printed onto a specification sheet, which is then placed on the corresponding vehicle to be inspected. During the inspection of the vehicle's license plate, the VIN deployed on the vehicle to be inspected is scanned using a barcode scanner to obtain the vehicle model. Then, based on the pre-set model number corresponding to each model, the model number of the vehicle to be inspected is determined.

[0090] See Figure 3 A second aspect of the present invention provides a vehicle license plate detection device, comprising:

[0091] The coordinate normalization module 301 is used to perform coordinate normalization processing on at least one pre-set sign marking area in the pre-acquired vehicle image to be detected, so as to obtain at least one first sign marking area.

[0092] The region classification module 302 is used to classify the at least one first sign marking region according to the coordinate data of the at least one first sign marking region to obtain at least one second sign marking region; wherein, the second sign marking region includes at least one first sign marking region;

[0093] The sign recognition module 303 is used to input the image corresponding to each of the second sign marking areas into a preset sign recognition model for sign recognition, and obtain sign recognition information corresponding to each of the second sign marking areas;

[0094] The vehicle sign detection module 304 is used to compare the vehicle model number, the preset area number, and the sign recognition information with a preset sign lookup table based on the pre-acquired vehicle model number and the preset area number corresponding to each of the second sign marking areas, in order to obtain the vehicle sign detection result.

[0095] The sign comparison table includes several sign comparison information entries, and each sign comparison information entry includes at least a vehicle model comparison number, a region comparison number, and sign comparison information.

[0096] As a preferred embodiment, the device further includes a sign recognition model acquisition module, used for:

[0097] A pre-acquired sign marking image is used to train a pre-set deep convolutional neural network to obtain a pre-trained deep convolutional neural network; wherein, the sign marking image includes at least one car sign, and each car sign is pre-marked with sign marking information;

[0098] In the sign marking image, the sign elements in the car sign are removed to obtain a defective sign, and the sign marking information of the defective sign is modified to mark the defective sign and obtain a defect marking image;

[0099] The defect-marked images are used to train the pre-trained deep convolutional neural network to obtain the sign recognition model.

[0100] As a preferred embodiment, the vehicle sign detection module 304 is used to compare the vehicle model number, the preset area number, and the sign recognition information with a preset sign lookup table to obtain the vehicle sign detection result, specifically including:

[0101] The vehicle model number, the preset area number, and the sign identification information are compared with each sign comparison information in the sign comparison table to determine whether there is any sign comparison information that is the same as the vehicle model number, the preset area number, and the sign identification information.

[0102] If so, the vehicle sign inspection result is deemed qualified;

[0103] If not, the vehicle sign inspection result is deemed unqualified.

[0104] As a preferred embodiment, the region classification module 302 is used to classify the at least one first sign marking region according to the coordinate data of the at least one first sign marking region to obtain at least one second sign marking region, specifically including:

[0105] Based on the coordinate data of the at least one first sign marking area, the at least one first sign marking area is classified using the nearest neighbor algorithm to obtain at least one area category;

[0106] Based on the coordinate data of the first sign marking area in each of the area categories, at least one second sign marking area is obtained.

[0107] As a preferred embodiment, the coordinate normalization module 301 is used to perform coordinate normalization processing on at least one preset sign marking region in the vehicle image to be detected, to obtain at least one first sign marking region, specifically including:

[0108] Based on the width and height of the vehicle image to be detected, the coordinates of the sign marking area are normalized using the following expression to obtain the first sign marking area:

[0109]

[0110] Where x1 represents the x-coordinate of the center point of the first sign marking area, x represents the x-coordinate of the center point of the sign marking area, W represents the width of the vehicle image to be detected, y1 represents the y-coordinate of the center point of the first sign marking area, y represents the y-coordinate of the center point of the sign marking area, H represents the height of the vehicle image to be detected, width1 represents the width of the first sign marking area, width represents the width of the sign marking area, height1 represents the height of the first sign marking area, and height represents the height of the sign marking area.

[0111] As a preferred embodiment, the region classification module 302 is used to obtain at least one second sign marking region based on the coordinate data of the first sign marking region in each region category, specifically including:

[0112] Based on the coordinate data of the first sign marking area in each area category, the second sign marking area corresponding to the current area category is represented by the following expression: [min(x1), min(y1), max(x1+width1)-min(x1), max(y1+height1)-min(y1)].

[0113] As a preferred embodiment, the device further includes a vehicle model number acquisition module, used for:

[0114] The vehicle model of the vehicle to be inspected is obtained by scanning the vehicle identification code pre-deployed on the vehicle to be inspected using a barcode scanner.

[0115] Based on the vehicle model to be tested, determine the vehicle model number.

[0116] It should be noted that the vehicle sign detection device provided in this embodiment of the invention can realize all the processes of the vehicle sign detection method described in any of the above embodiments. The functions and technical effects of each module in the device are the same as those of the vehicle sign detection method described in the above embodiments, and will not be repeated here.

[0117] A third aspect of the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the vehicle license plate detection method as described in any embodiment of the first aspect.

[0118] The terminal device can be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and memory. The terminal device may also include input / output devices, network access devices, buses, etc.

[0119] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0120] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0121] A fourth aspect of the present invention provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform the vehicle license plate detection method as described in any embodiment of the first aspect.

[0122] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary hardware platforms, and of course, it can also be implemented entirely by hardware. Based on this understanding, all or part of the technical solution of the present invention that contributes to the background art can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0123] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for detecting vehicle license plates, characterized in that, Includes the following steps: Based on the pre-acquired vehicle image to be detected, coordinate normalization processing is performed on at least one pre-defined sign marking area in the vehicle image to be detected to obtain at least one first sign marking area. Based on the coordinate data of the at least one first sign marking area, the at least one first sign marking area is classified to obtain at least one second sign marking area; wherein, the second sign marking area includes at least one first sign marking area; The image corresponding to each of the second sign marking areas is input into a preset sign recognition model for sign recognition, thereby obtaining sign recognition information corresponding to each of the second sign marking areas; Based on the pre-acquired vehicle model number of the vehicle to be detected and the preset area number corresponding to each of the second sign marking areas, the vehicle model number, the preset area number, and the sign recognition information are compared with a preset sign lookup table to obtain the vehicle sign detection result; The sign comparison table includes several sign comparison information entries, and each sign comparison information entry includes at least a vehicle model comparison number, a region comparison number, and sign comparison information. The step of classifying the at least one first sign marking area based on the coordinate data of the at least one first sign marking area to obtain at least one second sign marking area specifically includes the following steps: Based on the coordinate data of the at least one first sign marking area, the at least one first sign marking area is classified using the nearest neighbor algorithm to obtain at least one area category; Based on the coordinate data of the first sign marking area in each of the area categories, at least one second sign marking area is obtained.

2. The vehicle nameplate detection method as described in claim 1, characterized in that, The method specifically obtains the sign recognition model through the following steps: A pre-acquired sign marking image is used to train a pre-set deep convolutional neural network to obtain a pre-trained deep convolutional neural network; wherein, the sign marking image includes at least one car sign, and each car sign is pre-marked with sign marking information; In the sign marking image, the sign elements in the car sign are removed to obtain a defective sign, and the sign marking information of the defective sign is modified to mark the defective sign and obtain a defect marking image; The defect-marked images are used to train the pre-trained deep convolutional neural network to obtain the sign recognition model.

3. The vehicle nameplate detection method as described in claim 1, characterized in that, The step of comparing the vehicle model number, the preset area number, and the sign recognition information with a preset sign lookup table to obtain the vehicle sign detection result specifically includes the following steps: The vehicle model number, the preset area number, and the sign identification information are compared with each sign comparison information in the sign comparison table to determine whether there is any sign comparison information that is the same as the vehicle model number, the preset area number, and the sign identification information. If so, the vehicle sign inspection result is deemed qualified; If not, the vehicle sign inspection result is deemed unqualified.

4. The vehicle nameplate detection method as described in claim 1, characterized in that, The step of performing coordinate normalization processing on at least one preset sign marking region in the vehicle image to be detected to obtain at least one first sign marking region specifically includes the following steps: Based on the width and height of the vehicle image to be detected, the coordinates of the sign marking area are normalized using the following expression to obtain the first sign marking area: in, This represents the x-coordinate of the center point of the area marked by the first sign. This represents the x-coordinate of the center point of the area marked on the sign. This indicates the width of the image of the car to be detected. This represents the ordinate of the center point of the area marked by the first sign. This represents the ordinate of the center point of the area marked on the sign. Indicates the height of the image of the car to be detected. This indicates the width of the area marked on the first sign. Indicates the width of the area marked on the sign. Indicates the height of the area marked by the first sign. Indicates the height of the area marked on the sign.

5. The vehicle nameplate detection method as described in claim 4, characterized in that, The step of obtaining at least one second sign marking area based on the coordinate data of the first sign marking area in each area category specifically includes the following steps: Based on the coordinate data of the first sign marking area in each of the aforementioned area categories, the second sign marking area corresponding to the current area category is represented by the following expression: [min( x 1), min( y 1), max( x 1+ width 1)-min( x 1), max( y 1+ height 1)-min( y 1).

6. The vehicle nameplate detection method as described in claim 1, characterized in that, The method specifically obtains the vehicle model number of the car to be tested through the following steps: The vehicle model of the vehicle to be inspected is obtained by scanning the vehicle identification code pre-deployed on the vehicle to be inspected using a barcode scanner. Based on the vehicle model to be tested, determine the vehicle model number.

7. A vehicle license plate detection device, characterized in that, include: The coordinate normalization module is used to perform coordinate normalization processing on at least one pre-set sign marking area in the pre-acquired vehicle image to be detected, so as to obtain at least one first sign marking area. The region classification module is used to classify the at least one first sign marking region according to the coordinate data of the at least one first sign marking region to obtain at least one second sign marking region; wherein, the second sign marking region includes at least one first sign marking region; The sign recognition module is used to input the image corresponding to each of the second sign marking areas into a preset sign recognition model for sign recognition, and obtain the sign recognition information corresponding to each of the second sign marking areas; The vehicle sign detection module is used to compare the vehicle model number, the preset area number, and the sign recognition information with a preset sign lookup table based on the pre-acquired vehicle model number and the preset area number corresponding to each of the second sign marking areas, in order to obtain the vehicle sign detection result. The sign comparison table includes several sign comparison information entries, and each sign comparison information entry includes at least a vehicle model comparison number, a region comparison number, and sign comparison information. The region classification module is used to classify the at least one first sign marking region according to the coordinate data of the at least one first sign marking region to obtain at least one second sign marking region, specifically including: Based on the coordinate data of the at least one first sign marking area, the at least one first sign marking area is classified using the nearest neighbor algorithm to obtain at least one area category; Based on the coordinate data of the first sign marking area in each of the area categories, at least one second sign marking area is obtained.

8. A terminal device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the vehicle sign detection method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device on which the computer-readable storage medium is located to perform the vehicle license plate detection method as described in any one of claims 1 to 6.

Images