Traffic overload control intelligent law enforcement platform based on pre-training model architecture

Abstract

The invention, which relates to the field of road safety management, discloses a pre-training model architecture-based traffic overload control intelligent law enforcement platform comprising an uploading module used for uploading historical vehicle overload related information and storing the vehicle overload related information; the extraction module is used for traversing the vehicle overrun related information stored in the uploading module and extracting overrun vehicle characteristic parameters based on the vehicle overrun related information; the output module is used for receiving the overrun vehicle characteristic parameters extracted by the extraction module and outputting an overrun judgment threshold value based on the overrun vehicle characteristic parameters; on the basis of an image recognition technology, the truck overrun problem is supervised, in the supervision process, related information of historical overrun trucks is used as prior data for pre-training, so that an overrun judgment threshold value is obtained, and then a wheel image and a wheel arch image are used as main recognition targets; and carrying out distinguished overrun supervision on the trucks of different models based on the overrun judgment threshold in combination with the truck specifications and sizes.

Description

Technical Field

[0001] The present invention relates to the technical field of road safety management, and more specifically to a traffic overload control intelligent law enforcement platform based on a pre-trained model architecture. Background Art

[0002] Traffic overload control, that is, the governance of over-limit and overloaded transportation, is a key measure to ensure road traffic safety and maintain the integrity of highway infrastructure. It strictly controls the load weight, size, etc. of vehicles to prevent vehicles from driving with overloaded, over-height, or over-width loads. By means of checkpoint inspections, technological monitoring, etc., it severely cracks down on illegal over-limit behaviors to safeguard road smoothness and public travel safety.

[0003] The invention patent application with the application number 202122083860.4 discloses a highway overload control joint law enforcement system, which includes an overload control platform and several overload control points. Each overload control point includes a gantry, a dynamic weighing module, a capture module, a video recording module, and an outdoor cabinet. The gantries are set across above both ends of the highway of the joint law enforcement system. There are at least two gantries, which are erected above the front and rear ends of the highway. The capture module is set on the gantry. An over-limit detection area is formed between the two gantries. The dynamic weighing module is set on the road surface of the over-limit detection area. The video recording module is set at the top of one side column of the gantry. The outdoor cabinets are set on both sides of the road in the over-limit detection area. A workbench computer is set in the outdoor cabinet. The workbench computer is composed of a precision circuit board and a chip, and is afraid of water and dust. The outdoor cabinet can provide waterproof and dustproof conditions for the workbench computer. The workbench computer software is installed in the workbench computer. The workbench computer is connected to the dynamic weighing module, the capture module, and the video recording module. The workbench computer acquires and analyzes the signals transmitted by the dynamic weighing module, the capture module, and the video recording module, and transmits the data to the overload control platform. This application aims to solve the problems of "in the highway overload control scenario, illegal vehicles often cover their license plates and the number is increasing day by day, resulting in difficulty in using photos, videos, etc. for law enforcement, damaging highways, endangering public safety, and causing losses of state-owned assets. Because overload control belongs to the transportation bureau and license plate covering belongs to the traffic police, there are obstacles in department law enforcement, data is difficult to share, and it is also difficult to catch overweight vehicles by transmitting information afterwards."

[0004] Currently, for the weight over-limit management of transport trucks, monitoring is often implemented through technologies such as dynamic weighing. Its upfront deployment cost is relatively high and the equipment installation process affects road smoothness to a certain extent, so it is difficult to popularize to all sections of the road.

[0005] Therefore, we propose a traffic overload control intelligent law enforcement platform based on a pre-trained model architecture. Summary of the Invention

[0006] Aiming at the above-mentioned shortcomings of the prior art, the present invention provides a traffic overload control intelligent law enforcement platform based on a pre-trained model architecture, which can effectively solve the problems of the prior art.

[0007] To achieve the above object, the present invention is realized through the following technical solutions;

[0008] The present invention discloses a traffic overload intelligent law enforcement platform based on a pre-trained model architecture, including:

[0009] An upload module for uploading historical vehicle over-limit related information and storing the vehicle over-limit related information; an extraction module for traversing the vehicle over-limit related information stored in the upload module and extracting over-limit vehicle feature parameters based on the vehicle over-limit related information; an output module for receiving the over-limit vehicle feature parameters extracted by the operation of the extraction module and outputting an over-limit determination threshold based on the over-limit vehicle feature parameters; a collection module for real-time collecting vehicle size information and wheel and wheel arch image data on the over-limit management road and converting the wheel and wheel arch image data into vehicle feature parameters; a determination module for receiving the vehicle feature parameters in the collection module, obtaining the over-limit determination threshold corresponding to the vehicle with the same size information in the output module based on the size information of the vehicle from which the vehicle feature parameters are derived, and comparing the vehicle feature parameters with the over-limit determination threshold to determine whether the vehicle from which the vehicle feature parameters are derived is over-limit; an interaction module for receiving the determination result in the determination module and sending the determination result to a preset computer device when the determination result is yes; an early warning module for prompting the over-limit vehicle to pull over to the side of the road and wait for re-inspection:

[0010] Furthermore, the historical vehicle over-limit related information uploaded in the upload module includes: vehicle size, wheel and wheel arch image data;

[0011] Among them, all the historical vehicle over-limit related information uploaded in the upload module is derived from vehicles determined to be overloaded based on dynamic weighing in historical road driving scenarios. When the upload module stores the vehicle over-limit related information, it distinguishes and stores the vehicle over-limit related information based on the vehicle size in the vehicle over-limit related information. The vehicle and wheel arch image data in the vehicle over-limit related information comes from all wheel arch positions of the vehicle.

[0012] Furthermore, the extraction module runs continuously when extracting over-limit vehicle feature parameters, and each time it runs, it uses the wheel and wheel arch image data belonging to all vehicle over-limit related information in a differentiated storage interval in the upload module as the extraction target of over-limit vehicle feature parameters;

[0013] The over-limit vehicle feature parameter extraction logic in the extraction module is:

[0014] Obtain the wheel and wheel arch image data, perform detail enhancement processing on the wheel and wheel arch image data, and extract a contour image from the wheel and wheel arch image data after the detail enhancement processing;

[0015] Identify the upper, lower, left, and right edge points in the positive directions on the contour representing the wheel in the contour image, and identify the vertices of the contour representing the wheel arch in the contour image;

[0016] Connect the points corresponding to the left and lower positions among the edge points from the four positive directions to the points corresponding to the lower position to form an angle, and connect the edge points corresponding to the upper position among the edge points from the four positive directions to the vertices of the contour representing the wheel arch to form a line segment;

[0017] Measure the angle of the angle and the length of the line segment;

[0018] Among them, the measured angle of the angle and the length of the line segment are the over-limit vehicle characteristic parameters.

[0019] Furthermore, the detail enhancement processing logic of the wheel and wheel arch image data is expressed as:

[0020] Edge enhancement:

[0021] Ledge(x,y) = I(x,y) + k1·L(I(x,y));

[0022] Perform adaptive histogram equalization on L edge (x,y) to obtain the image Lcontrast(x,y) with enhanced contrast. In the adaptive histogram equalization stage, the image is divided into multiple small blocks, and histogram equalization is performed on each small block respectively;

[0023] Gamma correction:

[0024]

[0025] In the formula: L edge (x,y) is the image after edge enhancement; I(x,y) is the original wheel and wheel arch image data; k1 is a constant; L(I(x,y)) is the Laplacian operator; L enhanced (x,y) is the wheel and wheel arch image data finally obtained by detail enhancement processing; γ is the gamma value;

[0026] Among them, the Laplacian operator L(I(x,y)) is calculated through convolution operation, and the Laplacian kernel is The constant k1 > 0. If the wheel and wheel arch image data is collected at night, the gamma value is in the range of 0 < γ < 1. If the wheel and wheel arch image data is collected during the day, the gamma value is in the range of γ > 1.

[0027] Further, during the operation phase of the output module, the characteristic parameters of over-limit vehicles extracted from the corresponding vehicle over-limit related information stored in each partition in the upload module are obtained. The minimum included angle value and the longest line segment length are captured from the characteristic parameters of over-limit vehicles corresponding to each partition interval, and are bound to the vehicle size corresponding to the partition interval. Then, the output module outputs the bound minimum included angle value and the longest line segment length, which are the two over-limit determination thresholds.

[0028] Further, the acquisition module is integrated by a camera. After the acquisition module captures the vehicle license plate image, the vehicle size information is identified from the vehicle license plate image, and the vehicle size information is obtained based on the vehicle license plate number. When the image data of the wheels and wheel arches are acquired, the acquisition height relative to the road surface and the acquisition angle relative to the road surface are the same as those when the image data of the wheels and wheel arches in the historical vehicle over-limit related information are acquired;

[0029] Among them, the image data of the wheels and wheel arches acquired by the acquisition module are synchronously subjected to detail enhancement processing. After the detail enhancement processing is completed, the contour image is further extracted. Then, the upper, lower, left, and right edge points in the four positive directions of the contour representing the wheels and the vertexes of the contour representing the wheel arches are identified in the contour image. The characteristic parameters of the vehicle of the image data of the wheels and wheel arches acquired by the acquisition module are determined based on the edge points and the vertexes.

[0030] Further, the determination logic for whether the vehicle is over-limit in the determination module is as follows:

[0031]

[0032] In the formula: L and a are the characteristic parameters of the vehicle acquired by the acquisition module during operation; L max , a min are the over-limit determination thresholds corresponding to the vehicle size;

[0033] L, L max represent the line segment length, and a, a min represent the included angle. When any one or more of formulas (1) and (2) are established, it is determined that the vehicle is over-limit. When both formulas (1) and (2) are not established, it is determined that the vehicle is not over-limit;

[0034] The content of the determination result sent by the interaction module to the computer device is: The vehicle size information in the acquisition module comes from the vehicle license plate image.

[0035] Further, during the operation of the system, several temperature intervals are defined by the system end user. When the ambient temperature in the area where the over-limit management road is located switches between temperature intervals, the system resets and runs;

[0036] Among them, each time the system resets and runs, the historical vehicle over-limit related information uploaded in the upload module all comes from the corresponding temperature range scenario.

[0037] Furthermore, the warning module establishes a connection with all the electronic screens deployed on the over-limit management road through a wireless network. During the operation stage of the interaction module, it synchronously scrolls and broadcasts the text information prompting the over-limit vehicles to pull over to the side of the road and wait for re-inspection.

[0038] Among them, the content of the text information scrolled and broadcast by the electronic screen is: "License plate number of the over-limit vehicle" + "Please ask the over-limit vehicle to pull over to the side of the road and wait for re-inspection".

[0039] Furthermore, the upload module is connected to an extraction module and an output module through wireless network interaction. The output module is connected to a collection module through wireless network interaction. The collection module is connected to a determination module and an interaction module through wireless network interaction. The interaction module is connected to the warning module through wireless network interaction.

[0040] Adopting the technical solution provided by the present invention, compared with the known prior art, it has the following beneficial effects.

[0041] 1. Based on image recognition technology, the present invention supervises the problem of truck over-limit. During the supervision process, the relevant information of historical over-limit trucks is used as prior data for pre-training to obtain the over-limit determination threshold. Then, taking the wheel image and the wheel arch image as the main recognition targets, different types of trucks are distinguished and supervised based on the over-limit determination threshold and the truck specifications, ensuring that the over-limit trucks on the road can be stably controlled.

[0042] 2. During the operation of the present invention, by customizing the temperature range to adapt to the characteristics of tire pressure changes, the operation output result of the system in the present invention is more accurate. At the same time, with the application of this image data technology, the implementation cost and difficulty of this technical solution are greatly reduced, making it more suitable for popularization.

[0043] 3. In the present invention, by interacting with the display devices installed on the road, the driving users with over-limit problems are prompted in real time to pull over and accept re-inspection, further improving the immediacy of the over-limit control of this technical solution on the road. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0045] Figure 1Schematic diagram of the traffic overload intelligent law enforcement platform based on the pre-trained model architecture;

[0046] Figure 2 Schematic diagram of the source example of vehicle characteristic parameters in the present invention. Specific implementation mode

[0047] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0048] The present invention will be further described below with reference to the embodiments.

[0049] Embodiment:

[0050] The traffic overload intelligent law enforcement platform based on the pre-trained model architecture in this embodiment, as Figure 1 shown, includes:

[0051] An upload module, configured to upload historical vehicle over-limit related information and store the vehicle over-limit related information;

[0052] The historical vehicle over-limit related information uploaded in the upload module includes: vehicle size, wheel and wheel arch image data;

[0053] Among them, all the historical vehicle over-limit related information uploaded in the upload module is sourced from vehicles determined to be overloaded based on dynamic weighing in the historical road driving scenario. When the upload module stores the vehicle over-limit related information, it distinguishes and stores the vehicle over-limit related information based on the vehicle size in the vehicle over-limit related information. The vehicle and wheel arch image data in the vehicle over-limit related information are sourced from all wheel arch positions of the vehicle;

[0054] An extraction module, configured to traverse the vehicle over-limit related information stored in the upload module and extract over-limit vehicle characteristic parameters based on the vehicle over-limit related information;

[0055] When extracting the over-limit vehicle characteristic parameters, the extraction module runs continuously, and each time it runs, it uses the wheel and wheel arch image data belonging to all the vehicle over-limit related information in a distinguished storage interval in the upload module as the target for extracting the over-limit vehicle characteristic parameters;

[0056] The over-limit vehicle characteristic parameter extraction logic in the extraction module is:

[0057] Obtain the image data of the wheel and the wheel arch, perform detail enhancement processing on the image data of the wheel and the wheel arch, and extract the contour image from the image data of the wheel and the wheel arch after the detail enhancement processing is completed;

[0058] Identify the upper, lower, left, and right edge points in the four positive directions on the contour representing the wheel in the contour image, and identify the vertex of the contour representing the wheel arch in the contour image;

[0059] Connect the points corresponding to the left and lower positions among the edge points from the four positive directions to the points corresponding to the lower position to form an angle, and connect the edge points corresponding to the upper position among the edge points from the four positive directions to the vertex of the contour representing the wheel arch to form a line segment;

[0060] Measure the angle of the angle and the length of the line segment;

[0061] Among them, the measured angle of the angle and the length of the line segment are the over-limit vehicle characteristic parameters;

[0062] The detail enhancement processing logic of the image data of the wheel and the wheel arch is expressed as:

[0063] Edge enhancement:

[0064] Ledge(x,y) = I(x,y) + k1·L(I(x,y));

[0065] Perform adaptive histogram equalization on L edge (x,y) to obtain the image Lcontrast(x,y) with enhanced contrast. In the adaptive histogram equalization stage, the image is divided into multiple small blocks, and histogram equalization is performed on each small block respectively;

[0066] Gamma correction:

[0067]

[0068] In the formula: L edge (x,y) is the image after edge enhancement; I(x,y) is the original image data of the wheel and the wheel arch; k1 is a constant; L(I(x,y)) is the Laplacian operator; L enhanced (x,y) is the image data of the wheel and the wheel arch obtained by the final detail enhancement processing; γ is the gamma value;

[0069] Among them, the Laplacian operator L(I(x,y)) is calculated through convolution operation, and the Laplacian kernel is The constant k1 > 0. If the image data of the wheel and the wheel arch is collected at night, then the gamma value is in the range of 0 < γ < 1. If the image data of the wheel and the wheel arch is collected during the day, then the gamma value is in the range of γ > 1;

[0070] The wheel and wheel arch image data is processed for detail enhancement through the above logical formula, so as to ensure that the data for the further operation and application of the system in this embodiment is more accurate, and to ensure that the determination of the over-limit judgment threshold is more accurate;

[0071] An output module, configured to receive the over-limit vehicle feature parameters extracted by the extraction module during operation, and output an over-limit judgment threshold based on the over-limit vehicle feature parameters;

[0072] During the operation stage of the output module, the over-limit vehicle feature parameters extracted from the over-limit vehicle-related information stored in the upload module in each area are obtained. The minimum included angle value and the longest line segment length are captured from the over-limit vehicle feature parameters in each area interval, and are bound to the vehicle size corresponding to the area interval. Then, the output module outputs the bound minimum included angle value and the longest line segment length, that is, the two over-limit judgment thresholds;

[0073] A collection module, configured to collect the vehicle size information of the vehicles on the over-limit management road and the wheel and wheel arch image data in real time, and convert the wheel and wheel arch image data into vehicle feature parameters;

[0074] The collection module is integrated by a camera. After the collection module captures the vehicle license plate image, it identifies the vehicle license plate number in the vehicle license plate image, and obtains the vehicle size information based on the vehicle license plate number. When collecting the wheel and wheel arch image data, the collection height and the collection angle with respect to the road surface are the same as those when collecting the wheel and wheel arch image data in the historical vehicle over-limit related information;

[0075] Among them, the wheel and wheel arch image data collected by the collection module is synchronously processed for detail enhancement. After the detail enhancement processing is completed, the contour image is further extracted. Then, the upper, lower, left, and right four positive direction edge points representing the wheel and the vertex representing the contour of the wheel arch are identified in the contour image. The vehicle feature parameters of the vehicle and wheel arch image data collected by the collection module are determined based on the edge points and the vertex;

[0076] A determination module, configured to receive the vehicle feature parameters in the collection module, obtain the over-limit judgment threshold corresponding to the vehicle with the same size information in the output module based on the vehicle size information from which the vehicle feature parameters are derived, and compare the vehicle feature parameters with the over-limit judgment threshold to determine whether the vehicle from which the vehicle feature parameters are derived is over-limit;

[0077] The determination logic for whether the vehicle is over-limit in the determination module is:

[0078]

[0079] In the formula: L and a are the vehicle feature parameters collected by the collection module during operation; L max 、a min are the over-limit judgment thresholds corresponding to the vehicle size;

[0080] L, L max That is, it represents the line segment length, a, a min That is, it represents the included angle. When any one or more of formulas (1) and (2) hold, it is determined that the vehicle is over-limit. When neither formula (1) nor formula (2) holds, it is determined that the vehicle is not over-limit;

[0081] Through the above judgment formula, the system in this embodiment provides the judgment logic for whether the vehicle is over-limit ultimately, ensuring the stable output of the judgment result of whether the vehicle is over-limit.

[0082] The content of the judgment result sent by the interaction module to the computer device during operation is: the vehicle size information in the acquisition module comes from the vehicle license plate image;

[0083] The interaction module is used to receive the judgment result in the judgment module and send the judgment result to a preset computer device when the judgment result is yes;

[0084] The warning module is used to prompt the over-limit vehicle to pull over and wait for re-inspection;

[0085] The warning module establishes a connection with all the electronic screens deployed on the over-limit management road through the wireless network. During the operation stage of the interaction module, it synchronously scrolls and broadcasts the text information prompting the over-limit vehicle to pull over and wait for re-inspection;

[0086] Among them, the content of the text information scrolled and broadcast by the electronic screen is: "License plate number of over-limit vehicle" + "Please ask the over-limit vehicle to pull over and wait for re-inspection";

[0087] During the operation of the system, several temperature intervals are defined by the system-side user. When the ambient temperature in the area where the over-limit management road is located switches between temperature intervals, the system resets and runs;

[0088] Among them, each time the system resets and runs, the historical vehicle over-limit related information uploaded in the upload module all comes from the corresponding temperature interval scenario;

[0089] The upload module is connected to the extraction module and the output module through wireless network interaction. The output module is connected to the acquisition module through wireless network interaction. The acquisition module is connected to the judgment module and the interaction module through wireless network interaction. The interaction module is connected to the warning module through wireless network interaction.

[0090] In this embodiment, the upload module is operated to upload the historical vehicle overload related information, and the vehicle overload related information is stored. The extraction module synchronously traverses the vehicle overload related information stored in the upload module, and extracts the overload vehicle characteristic parameters based on the vehicle overload related information. The output module is post-operated to receive the overload vehicle characteristic parameters extracted by the extraction module, and the overload judgment threshold is output based on the overload vehicle characteristic parameters. Then, the acquisition module collects the vehicle size information and wheel and wheel arch image data of the overload control management road in real time, and converts the wheel and wheel arch image data into vehicle characteristic parameters. The judgment module further receives the vehicle characteristic parameters in the acquisition module, obtains the vehicle corresponding to the overload judgment threshold of the same size information in the output module based on the size information of the vehicle from the vehicle characteristic parameters, compares the vehicle characteristic parameters with the overload judgment threshold, and determines whether the vehicle from the vehicle characteristic parameters is overloaded. The interaction module receives the judgment result in the judgment module, and when the judgment result is yes, sends the judgment result to a preset computer device, and finally prompts the overloaded vehicle to stop at the roadside through the early warning module and wait for re-inspection;

[0091] Through graphic recognition technology, characteristic angles and distances are obtained in the wheel and wheel arch images. They are further combined with prior data to compare characteristic angles and distances to determine whether the truck has an over-limit problem, thereby providing convenient and fast maintenance for road truck transportation safety.

[0092] See also Figure 2 As shown, the outer circle in the figure represents the wheel, and the arc above the outer circle represents the wheel arch. Referring to the point marks, angle indicators and connecting lines in the figure, the characteristic line segments and characteristic angles are shown as examples.

[0093] In summary, in the above-mentioned embodiment, the system supervises the overload problem of trucks based on image recognition technology. During the supervision process, relevant information of historical overloaded trucks is used as prior data for pre-training to obtain the overload judgment threshold, and then the wheel image and the wheel arch image are used as the main recognition targets. Different types of trucks are differentiated and supervised based on the overload judgment threshold combined with the truck specifications and dimensions, so as to ensure that overloaded trucks on the road can be stably controlled. During operation, the system adapts to the characteristics of tire pressure changes through customized temperature ranges, so that the output results of the system operation in the present invention are more accurate. At the same time, the application of such image data technology greatly reduces the implementation cost and difficulty of the technical solution, making it more suitable for popularization. At the same time, the system interacts with the road-mounted display screen equipment to prompt driving users with overload problems to stop for re-inspection in real time, further improving the immediacy of the technical solution for road overload control.

[0094] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A traffic overload control intelligent law enforcement platform based on a pre-trained model architecture, characterized in that, Including: An upload module, which is used to upload historical vehicle over-limit related information and store the vehicle over-limit related information. An extraction module, which is used to traverse the vehicle over-limit related information stored in the upload module and extract over-limit vehicle characteristic parameters based on the vehicle over-limit related information. An output module, which is used to receive the over-limit vehicle characteristic parameters extracted by the operation of the extraction module and output an over-limit determination threshold based on the over-limit vehicle characteristic parameters. A collection module, which is used to collect vehicle size information and wheel and wheel arch image data of the over-limit management road in real time, and convert the wheel and wheel arch image data into vehicle characteristic parameters. A determination module, which is used to receive the vehicle characteristic parameters in the collection module, obtain the over-limit determination threshold corresponding to the vehicle with the same size information in the output module based on the size information of the vehicle from which the vehicle characteristic parameters are derived, and compare the vehicle characteristic parameters with the over-limit determination threshold to determine whether the vehicle from which the vehicle characteristic parameters are derived is over-limit. An interaction module, which is used to receive the determination result in the determination module and send the determination result to a preset computer device when the determination result is yes. An early warning module, which is used to prompt the over-limit vehicle to pull over to the side of the road and wait for re-inspection.

2. The intelligent law enforcement platform for traffic overload control based on the pre-trained model architecture according to claim 1, characterized in that The historical vehicle over-limit related information uploaded in the upload module includes: vehicle size, wheel and wheel arch image data. Among them, the historical vehicle over-limit related information uploaded in the upload module all comes from vehicles determined to be overloaded based on dynamic weighing in the historical road driving scenario. When the upload module stores the vehicle over-limit related information, it stores the vehicle over-limit related information separately based on the vehicle size in the vehicle over-limit related information. The vehicle and wheel arch image data in the vehicle over-limit related information come from all wheel arch positions of the vehicle.

3. The traffic overload intelligent law enforcement platform based on the pre-trained model architecture according to claim 1, characterized in that, When extracting the over-limit vehicle characteristic parameters, the extraction module runs continuously, and each time it runs, it uses the wheel and wheel arch image data belonging to all vehicle over-limit related information in a separately stored interval in the upload module as the extraction target of the over-limit vehicle characteristic parameters. The over-limit vehicle characteristic parameter extraction logic in the extraction module is as follows: Obtain the wheel and wheel arch image data, perform detail enhancement processing on the wheel and wheel arch image data, and extract a contour image from the wheel and wheel arch image data after the detail enhancement processing is completed. Identify the upper, lower, left, and right four positive direction edge points on the contour representing the wheel in the contour image, and identify the vertex of the contour representing the wheel arch in the contour image. Connect the points corresponding to the left and lower positions among the edge points from the four positive directions to the points corresponding to the lower position to form an angle, and connect the edge points corresponding to the upper position among the edge points from the four positive directions to the vertex of the contour representing the wheel arch to form a line segment. Measure the angle of the angle and the length of the line segment. Among them, the measured angle of the angle and the length of the line segment are the over-limit vehicle characteristic parameters.

4. The traffic overload intelligent law enforcement platform based on the pre-trained model architecture according to claim 3, characterized in that The detail enhancement processing logic of the wheel and wheel arch image data is expressed as: Edge enhancement: Ledge(x,y) = I(x,y) + k1·L(I(x,y)); Perform adaptive histogram equalization on L edge (x, y) to obtain the image Lcontrast(x, y) with enhanced contrast. In the adaptive histogram equalization stage, the image is divided into multiple small blocks, and histogram equalization is performed on each small block separately; Gamma correction: Where: L edge (x, y) is the image after edge enhancement; I(x, y) is the original image data of the wheel and wheel arch; k1 is a constant; L(I(x, y)) is the Laplacian operator; L enhanced (x, y) is the image data of the wheel and wheel arch obtained by the final detail enhancement process; γ is the gamma value; Among them, the Laplacian operator L(I(x, y)) is calculated through a convolution operation, and the Laplacian kernel is The constant k1 > 0; if the image data of the wheel and the wheel arch is collected at night, the gamma value is in the range of 0 < γ < 1, and if the image data of the wheel and the wheel arch is collected during the day, the gamma value is in the range of γ > 1.

5. The intelligent law enforcement platform for traffic overload control based on the pre-trained model architecture according to claim 1, characterized in that, During the operation stage of the output module, the over-limit vehicle characteristic parameters extracted from the over-limit relevant information stored in each area of the upload module are obtained. The minimum included angle value and the longest line segment length are captured from the over-limit vehicle characteristic parameters corresponding to each area interval, and are bound to the vehicle size corresponding to the area interval. Then, the output module outputs the bound minimum included angle value and the longest line segment length, which are the two over-limit judgment thresholds.

6. The traffic overloading intelligent law enforcement platform based on the pre-trained model architecture according to claim 1, characterized in that, The acquisition module is integrated by a camera. After the acquisition module captures the vehicle license plate image, the vehicle size information is recognized from the vehicle license plate image, and the vehicle size information is obtained based on the vehicle license plate number. When the wheel and wheel arch image data are acquired, the acquisition height relative to the road surface and the acquisition angle relative to the road surface are the same as those in the historical vehicle over-limit relevant information when the wheel and wheel arch image data are acquired. Among them, the wheel and wheel arch image data acquired by the acquisition module are synchronously subjected to detail enhancement processing. After the detail enhancement processing is completed, the contour image is further extracted. Then, the upper, lower, left, and right edge points in the positive directions on the contour representing the wheel and the vertex of the contour representing the wheel arch are identified in the contour image. Based on the edge points and the vertex, the vehicle characteristic parameters of the vehicle and the wheel arch image data acquired by the acquisition module are determined.

7. The traffic overload control intelligent law enforcement platform based on the pre-trained model architecture according to claim 1, characterized in that, The judgment logic for whether the vehicle is over-limit in the judgment module is as follows: In the formula: L and a are the vehicle characteristic parameters acquired by the acquisition module during operation. L max and a min is the over-limit determination threshold corresponding to the vehicle size; L, L max That is, it represents the line segment length, a, a min That is, it represents the included angle. When any one or more of the formulas (1) and (2) hold, it is determined that the vehicle is over-limit. When neither of the formulas (1) and (2) holds, it is determined that the vehicle is not over-limit; The content of the judgment result sent by the interaction module to the computer device during operation is: The vehicle size information in the acquisition module comes from the vehicle license plate image.

8. The traffic overload control intelligent law enforcement platform based on the pre-trained model architecture according to claim 1, characterized in that, During the operation of the system, several temperature intervals are defined by the system user. When the environmental temperature in the area where the over-limit management road is located switches between temperature intervals, the system resets and runs. Among them, each time the system resets and runs, the historical vehicle over-limit relevant information uploaded in the upload module comes from the corresponding temperature interval scenario.

9. The intelligent law enforcement platform for traffic overload control based on the pre-trained model architecture according to claim 1, characterized in that, The warning module establishes a connection with all the electronic screens deployed on the over-limit management road through a wireless network. During the operation stage of the interaction module, it synchronously scrolls and broadcasts the text information prompting the over-limit vehicle to stop by the roadside and wait for re-inspection. Among them, the content of the text information scrolled and broadcast by the electronic screen is: "Over-limit vehicle license plate number" + "Please ask the over-limit vehicle to stop by the roadside and wait for re-inspection".

10. The intelligent law enforcement platform for traffic overload control based on the pre-trained model architecture according to claim 1, characterized in that, The upload module is connected to the extraction module and the output module through wireless network interaction. The output module is connected to the acquisition module through wireless network interaction. The acquisition module is connected to the judgment module and the interaction module through wireless network interaction. The interaction module is connected to the warning module through wireless network interaction.

Images