Expressway emergency charging method and system

Abstract

The invention discloses an expressway emergency charging method and system, relates to the field of expressway toll station management and control, and solves the problems that an existing expressway automatic charging system mainly depends on identification and comparison of license plate images and lacks further verification of vehicle identity authenticity, so that fake-licensed vehicles are organic and ridable, and the charging efficiency is low. And the charging order is seriously disturbed. Traffic information of a target expressway section is obtained, and the road condition of the target expressway section is analyzed; based on the road condition of the target expressway section, the number of toll gates needing to be opened by the target toll station is analyzed; acquiring a vehicle image of a to-be-passed vehicle; extracting vehicle information from the vehicle image; based on the vehicle information, analyzing an actual license plate number of the to-be-passed vehicle; and acquiring passing transaction data of the to-be-passed vehicle based on the actual license plate number. According to the invention, the traffic efficiency and the service quality of the expressway are greatly improved, and the effectiveness and the safety of traffic management are enhanced at the same time.

Description

Technical Field

[0001] The present invention belongs to the field of toll station control on expressways, and specifically relates to an emergency toll collection method and system for expressways. Background Art

[0002] Currently, the expressway toll collection system mainly relies on two modes: electronic toll collection without stopping (ETC) and manual toll collection (MTC); However, when traffic congestion occurs on the expressway section, in order to relieve the traffic pressure at the toll station and improve the traffic efficiency, some areas have started to pilot or promote an automatic toll deduction system based on license plate recognition technology. This system does not require vehicles to stop at the toll station to pay tolls. Instead, it uses high-definition cameras to recognize the license plates of passing vehicles, and matches the recognition results with the user account information in the background database to achieve automatic toll deduction afterwards. However, this system highly depends on the recognition and comparison of license plate images, lacking further verification of the authenticity of vehicle identities, resulting in opportunities for license plate cloning vehicles and seriously disrupting the toll collection order. Summary of the Invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art; for this purpose, the present invention provides an emergency toll collection method and system for expressways, which is used to solve the technical problem that the existing expressway automatic toll deduction system mainly depends on the recognition and comparison of license plate images, lacks further verification of the authenticity of vehicle identities, resulting in opportunities for license plate cloning vehicles and seriously disrupting the toll collection order.

[0004] To achieve the above object, the first aspect of the present invention provides an emergency toll collection method for expressways, including: Obtain the traffic information of the target expressway section and analyze the road conditions of the target expressway section; Among them, the traffic information includes traffic volume and vehicle speed; the target expressway section represents the monitored expressway section; Based on the road conditions of the target expressway section, analyze the number of toll gates that need to be opened at the target toll station; among them, the target toll station is the end of the target expressway section; Obtain the vehicle image of the vehicle to be passed; extract vehicle information from the vehicle image; among them, the vehicle to be passed is the vehicle that currently needs to pass through the target toll station; Based on the vehicle information, analyze the actual license plate number of the vehicle to be passed; based on the actual license plate number, obtain the toll transaction data of the vehicle to be passed.

[0005] Preferably, the obtaining the traffic information of the target expressway section and analyzing the road conditions of the target expressway section includes: At every analysis period, obtain the number of vehicles on the target expressway section, and divide the number of vehicles by the analysis period to obtain the traffic volume of the target expressway section during the analysis period; And obtain the driving speed of the vehicle through the interval speed measurement camera, and calculate the average value of the speeds of all vehicles within the analysis period as the vehicle speed of the target highway section; When the traffic flow of several analysis periods is greater than threshold one and the vehicle speed is less than threshold two, congestion occurs on the target highway section; otherwise, congestion does not occur on the target highway section.

[0006] The present invention updates the traffic information of the target highway section regularly, that is, every analysis period, obtains the real-time traffic flow and real-time vehicle speed data of the target highway section within the analysis period, analyzes the condition of the target highway section, and by accurately judging the congestion condition of the highway section in a timely manner, the traffic management department can take corresponding measures, such as emergency measures for the highway section, to relieve congestion and improve traffic efficiency; at the same time, the timely congestion judgment also helps to reduce traffic accidents caused by congestion and improve road safety.

[0007] Preferably, analyzing the number of toll gates to be opened at the target toll station includes: Obtain the current traffic volume of the toll gates already opened on the target highway section; Calculate the number of toll gates K to be opened at the target toll station through the formula K = [Q - (N Cf - ΣCi)] / Cf; Where Q is the average traffic flow of the target highway section, i is the serial number of the opened toll gate, i = 0, 1,..., N, N is a positive integer, Ci is the traffic volume currently passing through toll gate i, Cf is the traffic volume of the toll gate per unit time, Σ sums over i, K takes a value less than or equal to K0, and K0 is the total number of toll gates.

[0008] Preferably, the average traffic flow of the target highway section is the average value of the traffic flows of several analysis periods.

[0009] The present invention dynamically calculates the number of toll gates K to be opened at the target toll station through a formula, and can respond to the change of the traffic flow Q of the target highway section in real time. When the traffic flow increases, the number of toll gates to be opened is correspondingly increased, and vice versa, so as to ensure that the passing capacity of the toll station matches the traffic flow. The optimal number of toll gates K obtained through calculation can avoid waste of resources caused by opening too many toll gates, and at the same time prevent congestion caused by insufficient number of opened toll gates. And a reasonable number of toll gates can ensure that vehicles pass through the toll station quickly, reduce waiting time, and improve the overall passing efficiency.

[0010] Preferably, extracting vehicle information from the vehicle image includes: Input the vehicle image into a vehicle recognition model, and the image recognition model outputs vehicle information; wherein, the vehicle image includes the roof image of the vehicle and four side images; the vehicle information includes the license plate number, the model and color of the vehicle. The vehicle recognition model is constructed by an artificial intelligence model, and the construction process is as follows: Obtain vehicle images of different types of vehicles and corresponding vehicle information from historical data; Integrate the vehicle images and corresponding vehicle information into several groups of training data and test data; use the training data to train the artificial intelligence model; use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally obtain a vehicle recognition model with vehicle images as input and vehicle information as output; among them, the artificial intelligence model is a BP neural network model or an RBF neural network model.

[0011] Preferably, analyzing the actual license plate number of the vehicle to pass includes: Judge whether the license plate number of the vehicle to pass is abnormal; if so, obtain the payment information of the vehicle to pass; if not, analyze the actual license plate number of the vehicle to pass; When the license plate number of the vehicle to pass is abnormal, obtain vehicles of the same model and the same color as the vehicle to pass from the upstream toll station, mark them as the first vehicles, and obtain the first vehicle set; And, obtain vehicles of the same model and color as the vehicle to pass from the target toll station, mark them as target vehicles, and obtain the target vehicle set; among them, the target vehicle set includes the vehicle to pass; Compare the license plate numbers of the first vehicles with the license plate numbers of the target vehicles; Extract the vehicles in the first vehicle set whose license plate numbers are inconsistent with the license plate number of the vehicle to pass, and mark them as abnormal vehicles; If there is only one abnormal vehicle, the license plate number of the abnormal vehicle is the actual license plate number of the vehicle to pass; If there are at least two abnormal vehicles, obtain the ETC records of each abnormal vehicle from the upstream toll station and the target toll station, and extract the abnormal vehicles with only upstream toll station ETC records. If there is only one abnormal vehicle, the license plate number of the abnormal vehicle is the actual license plate number of the vehicle to pass; if there are at least two abnormal vehicles, extract the vehicle images of the abnormal vehicles and the vehicle to pass; Perform image similarity recognition on the same-side images of the abnormal vehicles and the vehicle to pass respectively to obtain the similarity rate between the same-side images; among them, the image similarity recognition includes color recognition, defect recognition, etc., and the result of the image similarity recognition is obtained by the similarity recognition model; When the similarity rate between each pair of the same-side images exceeds the preset similarity threshold, the abnormal vehicle image is similar to the vehicle to pass, and the license plate number of the abnormal vehicle is the actual license plate number of the vehicle to pass.

[0012] Preferably, judging whether the license plate number of the vehicle to pass is abnormal includes: Based on the license plate number of the vehicle to pass, determine whether the vehicle information of the vehicle to pass is consistent with the standard vehicle information; if so, the license plate number of the vehicle to pass is normal; if not, the license plate number of the vehicle to pass is abnormal; wherein, the standard vehicle information is the vehicle information of the vehicle with the license plate number retrieved from the traffic vehicle management center.

[0013] Preferably, the upstream toll station is the starting point of the target highway section.

[0014] The present invention verifies whether the license plate number is abnormal through multi-dimensional information such as vehicle model and color, effectively reducing the risks of misjudgment and missed judgment. By obtaining vehicles of the same model and the same color as the vehicle to pass from the upstream toll station and the target toll station, and marking them as the first vehicle set and the target vehicle set, the method can quickly locate the vehicles that may be related, providing convenience for subsequent comparison and analysis.

[0015] By comparing the license plate number of the first vehicle with the license plate number of the target vehicle, and extracting the vehicles with inconsistent license plate numbers as abnormal vehicles, the method can intelligently screen out the vehicles that may have problems, reducing the workload of manual investigation; and when there are at least two abnormal vehicles, further extract the images of the abnormal vehicles and the vehicle to pass, and perform image similarity recognition. Through multi-dimensional analysis such as color recognition and defect recognition, the method can more accurately determine whether the abnormal vehicle and the vehicle to pass are the same vehicle. By accurately identifying the abnormal license plate and quickly locating the actual license plate number, it helps to prevent illegal acts such as vehicle toll evasion and license plate cloning, thereby enhancing road safety and traffic order.

[0016] Preferably, the obtaining of the passing transaction data of the vehicle to pass includes: Based on the actual license plate number, the traffic vehicle management center retrieves the passing transaction data of the actual license plate number; the passing transaction data includes passing mileage, vehicle type and transaction terminal; The passing transaction data includes passing mileage, vehicle type and transaction terminal; The passing mileage and the vehicle type are encrypted and transmitted to the settlement platform; the settlement platform sends the settlement result to the transaction terminal, and the transaction terminal performs the deduction operation.

[0017] The present invention accurately retrieves the passing transaction data through the actual license plate number, ensuring the accuracy and integrity of the data. It avoids data chaos or loss caused by license plate recognition errors or information mismatches, providing a reliable basis for subsequent settlement and deduction operations. Sensitive data such as passing mileage and vehicle model are encrypted before being transmitted to the settlement platform, effectively preventing the risk of data being stolen or tampered with during transmission, and protecting user privacy and data security. In addition, the settlement platform is responsible for receiving the encrypted passing mileage and vehicle model data, calculating the fees, and sending the settlement results to the transaction terminal. This centralized processing method improves the settlement efficiency and reduces manual intervention and errors.

[0018] Preferably, in the second aspect of the present invention, there is provided a highway emergency toll collection system, including an analysis module and a transaction module; Analysis module: used to obtain the traffic information of the target highway section and analyze the road conditions of the target highway section; Based on the road conditions of the target highway section, analyze the number of toll gates that need to be opened at the target toll station; And, obtain the vehicle image of the vehicle to be passed; extract vehicle information from the vehicle image; Transaction module: based on the vehicle information, analyze the actual license plate number of the vehicle to be passed; based on the actual license plate number, obtain the passing transaction data of the vehicle to be passed.

[0019] Compared with the prior art, the beneficial effects of the present invention are: By obtaining and analyzing the traffic information of the target highway section, the analysis module of the present invention can accurately judge the road conditions and dynamically analyze the number of toll gates that need to be opened at the target toll station accordingly. This helps to optimize resource allocation, reduce congestion, and improve passing efficiency. In the vehicle recognition link, the analysis module can quickly and accurately extract vehicle information from the vehicle image of the vehicle to be passed and keenly analyze abnormal conditions of the license plate number. When abnormal situations such as license plate cloning are found, it will comprehensively compare and analyze based on the first vehicle set, that is, the set of vehicles of the same model and color at the upstream toll station, and the target vehicle set, that is, the set of vehicles of the same model and color at the target toll station, and accurately locate the actual license plate number of the vehicle to be passed. Based on the actual license plate number, the transaction module can quickly obtain the passing transaction data, realize fast settlement and deduction. This not only greatly improves the user experience and allows users to enjoy efficient and convenient passing services, but also enhances the transparency and traceability of the system. Especially in abnormal scenarios such as license plate cloning, it effectively avoids transaction disputes caused by license plate recognition errors, maintains the normal traffic toll collection order, and provides strong support for the intelligent and refined management of highways. Description of the Drawings

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0021] Figure 1 It is a schematic flowchart of the method of the present invention; Figure 2 It is a schematic flowchart of the method for analyzing the traffic conditions of the target highway section of the present invention; Figure 3 It is a schematic flowchart of the method for confirming the license plate number of the vehicle to be passed of the present invention. Specific embodiments

[0022] The following will clearly and completely describe the technical solutions of the present invention in combination with the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0023] Please refer to Figure 1 , an embodiment of the first aspect of the present invention provides a highway emergency toll collection method, including: Obtain the traffic information of the target highway section and analyze the traffic conditions of the target highway section; wherein, the traffic information includes traffic flow and vehicle speed; the target highway section represents the monitored highway section; Please refer to Figure 2 , specifically, every analysis period, obtain the number of vehicles on the target highway section, and the ratio of the number of vehicles to the analysis period to obtain the traffic flow of the target highway section within the analysis period; And obtain the driving speed of the vehicle through the interval speed measurement camera, and calculate the average value of the speeds of all vehicles within the analysis period as the vehicle speed of the target highway section; When the traffic flow in several consecutive analysis periods is greater than threshold one and the vehicle speed is less than threshold two, the target highway section is congested; otherwise, the target highway section is not congested.

[0024] It should be noted that the analysis period, threshold one, and threshold two can be arbitrarily set to a value according to actual needs; For example: to analyze the traffic conditions of a certain highway section, assume that the analysis period is set to 5 minutes, which can be adjusted to 1 minute, 15 minutes, etc. according to needs; threshold one is set to 15 vehicles / minute, threshold two is set to 60 km / h, and the number of consecutive analysis periods can be set to 3 analysis periods, that is, 15 minutes; The interval speed measurement camera calculates the speed through the time difference between the entrance and the exit. Suppose the speeds of 100 vehicles within 5 minutes are [55, 58, 62,..., 50] km / h respectively. Suppose the calculated average vehicle speed is 56 km / h and the traffic flow is 20 vehicles per minute. For three consecutive analysis periods: if the traffic flow > threshold one and the average vehicle speed < threshold two, then there is congestion on this highway section.

[0025] Based on the road conditions of the target highway section, analyze the number of toll gates that need to be opened at the target toll station; where the target toll station is the toll station of the target highway section; Specifically, calculate the mean value of the traffic flow in several analysis periods to obtain the average traffic flow of the target highway section; And, obtain the current traffic volume of the toll gates that have been opened on the target highway section; Through the formula K = [Q - (N Cf - ΣCi)] / Cf, calculate the number of toll gates K that need to be opened at the target toll station; Where, Q is the average traffic flow of the target highway section, i is the serial number of the opened toll gate, i = 0, 1,..., N, N is a positive integer, Ci is the traffic volume currently passing through toll gate i, Cf is the traffic volume of the toll gate per unit time, and Σ sums up for i; It should be noted that the value of K is less than or equal to K0, and K0 is the total number of toll gates; this constraint condition ensures the feasibility of the calculation result and avoids the unreasonable situation of opening more toll gates than the actual number.

[0026] It should be further noted that the unit time period can be per minute or per 30 minutes or per hour, etc.

[0027] For example: There are 6 toll gates in total at the toll station of a certain highway section, and the traffic capacity of a single toll gate is 10 vehicles per minute; in the current three analysis periods, if it is 15 minutes, the total traffic flow in these 15 minutes is 900 vehicles, then the average traffic flow is 60 vehicles per minute; If 2 toll gates are currently opened, and the traffic volume currently passing through these two toll gates is 8 vehicles, then the remaining traffic volume of these 2 toll gates can be calculated through the formula N Cf - ΣCi, that is, 2 10 - 8 = 12 vehicles; the number of toll gates that still need to be opened is (60 - 12) / 10 ≈ 5 toll gates.

[0028] It should be noted that when the calculation result is not an integer, it is generally rounded up but not exceeding the total number of toll gates; if it exceeds the number of toll gates, then all toll gates are opened.

[0029] Obtain the vehicle image of the vehicle to pass; Extract vehicle information from the vehicle image; wherein, the vehicle to pass is the vehicle that currently needs to pass through the target toll station, and the vehicle image includes the roof image and four side images of the vehicle; Specifically, input the vehicle image into the vehicle recognition model, and the image recognition model outputs vehicle information; wherein, the vehicle information includes the license plate number, the model and color of the vehicle; Based on the vehicle information, analyze the actual license plate number of the vehicle to pass; Specifically, based on the license plate number of the vehicle to pass, determine whether the vehicle information of the vehicle to pass is consistent with the standard vehicle information; if so, the license plate number of the vehicle to pass is normal; if not, the license plate number of the vehicle to pass is abnormal; wherein, the standard vehicle information is the vehicle information of the vehicle whose license plate number is retrieved from the traffic vehicle management center.

[0030] Please refer to Figure 3 , when the license plate number of the vehicle to pass is abnormal, obtain vehicles of the same model and color as the vehicle to pass from the upstream toll station, mark them as the first vehicles, and obtain the first vehicle set; wherein, the upstream toll station is the starting point of the target highway section; And, obtain vehicles of the same model and color as the vehicle to pass from the target toll station, mark them as target vehicles, and obtain the target vehicle set; wherein, the target vehicle set includes the vehicle to pass; Compare the license plate numbers of the first vehicles with the license plate numbers of the target vehicles; Extract the vehicles in the first vehicle set whose license plate numbers are inconsistent with the license plate number of the vehicle to pass, and mark them as abnormal vehicles; If there is only one abnormal vehicle, the license plate number of the abnormal vehicle is the actual license plate number of the vehicle to pass; If there are at least two abnormal vehicles, obtain the ETC records of each abnormal vehicle from the upstream toll station and the target toll station, extract the abnormal vehicles with only upstream toll station ETC records, if there is only one abnormal vehicle, the license plate number of the abnormal vehicle is the actual license plate number of the vehicle to pass; if there are at least two abnormal vehicles, extract the vehicle images of the abnormal vehicles and the vehicle to pass; Perform image similarity recognition on the same-side images of the abnormal vehicles and the vehicle to pass respectively to obtain the similarity rate between the same-side images; wherein, the image similarity recognition includes color recognition and defect recognition, etc., and the result of the image similarity recognition is obtained by the similarity recognition model; When the similarity rate between each same-side image exceeds the preset similarity threshold, the abnormal vehicle image is similar to the vehicle to pass, and the license plate number of the abnormal vehicle is the actual license plate number of the vehicle to pass.

[0031] For example: On a certain highway section from Toll Station A (the upstream toll station) to Toll Station B (the target toll station), vehicle R plans to drive from A to B. However, when passing through Toll Station B, the system detects and reads its license plate number. But upon retrieving the vehicle information from the traffic vehicle management center, it shows that the standard vehicle information corresponding to this license plate is inconsistent with the currently passing vehicle. Therefore, the license plate number of vehicle R is abnormal; Retrieve vehicles of the same type as vehicle R from Toll Station A, such as those with the same color and model; obtain the vehicles of the same type as vehicle R at Toll Station A to get the first vehicle set. Also, retrieve vehicles of the same type as vehicle R from Toll Station B to get the target vehicle set; By comparing the license plate numbers of the first vehicle set and the target vehicle set, as follows: The first case: Assume the first vehicle set is S0 = [C, D, E], and the target vehicle set, such as S1 = [C, M, E, R]; then only vehicle D in S0 has a different license plate number from vehicle R. Therefore, vehicle D may have engaged in license plate forgery during the driving process, that is, the license plate number of vehicle C is the actual license plate number of vehicle R; The second case: Assume the first vehicle set is S0 = [C, D, E, F], and the target vehicle set, such as S1 = [C, M, E, R]; then there are vehicles D and F in S0 with different license plate numbers from vehicle R; then obtain the ETC records of vehicles D and F from Toll Station A and Toll Station B; If vehicle D only has an ETC record from Toll Station A, then the license plate number of vehicle D is the actual license plate number of vehicle R; If both vehicle D and vehicle F only have ETC records from Toll Station A, then extract the images of vehicles D, F, and R; Perform image similarity recognition on the roof images of vehicles D and F respectively with the roof image of vehicle R, and perform image similarity recognition on the side images of vehicles D and F respectively with the corresponding side images of vehicle R; Assume that the similarity rate between the roof images of vehicle D and vehicle R is analyzed to be 99.6%; the similarity rates between the side images of vehicle D and vehicle R are 99.2%, 98.4%, 97.4%, and 98.5% respectively; Assume that the similarity rate between the roof images of vehicle F and vehicle R is analyzed to be 70%; the similarity rates between the side images of vehicle D and vehicle R are 78.3%, 68.7%, 50%, and 68.8% respectively; Assume that the preset similarity threshold is set to 97%. By comparison, it can be seen that the similarity rates between the roof image and side images of vehicle D and vehicle R are all greater than the preset similarity threshold. Therefore, the license plate number of vehicle D is the actual license plate number of vehicle R.

[0032] Based on the actual license plate number, the traffic vehicle management center retrieves the passing transaction data of the actual license plate number; The passing transaction data includes passing mileage, vehicle type, and transaction terminal; The passing mileage and vehicle type are encrypted and transmitted to the settlement platform; because for security reasons, key information such as passing mileage and vehicle type needs to be encrypted before being sent to the settlement platform. To prevent data from being stolen or tampered with during transmission. The encryption algorithm may include symmetric encryption (such as AES) or asymmetric encryption (such as RSA), depending on the security requirements of the system.

[0033] The settlement platform sends the settlement result to the transaction terminal, and the transaction terminal performs the deduction operation; the transaction terminal is a payable mobile terminal; For example: Suppose a small car travels from point W to point Z, with a total distance of 100 kilometers, and the vehicle is a "small car". This information will be encrypted and sent to the settlement platform; To ensure data security, the passing mileage and vehicle type information of the vehicle will be encrypted before being uploaded to the settlement platform, such as using AES or RSA encryption algorithms, to prevent data from being tampered with or leaked during transmission.

[0034] The encrypted data is securely transmitted to the background settlement platform. The settlement platform calculates the total fee of 50 yuan based on the charging standard per kilometer for small cars (such as 0.5 yuan / km). After the settlement platform completes the billing, it sends the settlement result including the fee details back to the transaction terminal used by the vehicle owner, such as the mobile application bound to the ETC device, the in-vehicle OBU terminal, or the bank account system, through a secure channel.

[0035] The vehicle owner can view the detailed information of this passing on the transaction terminal, including the passing route, driving mileage, vehicle type classification, and the amount to be paid. After confirmation, the system will automatically or manually complete the deduction operation to complete the entire passing transaction process.

[0036] Both the vehicle recognition model and the similarity recognition model of the present invention are constructed by artificial intelligence models; Among them, the construction process of the vehicle recognition model is as follows: Obtain vehicle images of different types of vehicles and corresponding vehicle information from historical image data; Integrate the vehicle images and corresponding vehicle information into several groups of training data and test data; use the training data to train the artificial intelligence model; use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally, obtain a vehicle recognition model with vehicle images as the input and vehicle information as the output; among them, the artificial intelligence model is a BP neural network model or an RBF neural network model.

[0037] The construction process of the similarity recognition model is as follows: Obtain the similarity rate between the vehicle images of the same type of vehicle and the co-planar images from the historical image data; Integrate the similarity rate between the vehicle images and the co-planar images into several groups of training data and test data; use the training data to train the artificial intelligence model; use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally obtain a similarity recognition model with the input being the vehicle images of the same type of vehicle and the output being the similarity rate between the co-planar images; among them, the artificial intelligence model is a BP neural network model or an RBF neural network model.

[0038] Among them, when obtaining the historical image data, it is necessary to uniformly adjust its size and grayscale. Uniformizing the size can maximize the utilization rate of the GPU video memory and improve the batch processing speed. For example, when processing 224×224 images with a Batch = 32, the GPU utilization rate can reach more than 90%, while the mixed size will drop to 40%. And grayscale normalization can eliminate the influence of light. The images taken by the toll station camera at night / against the light may be too dark or overexposed.

[0039] The second aspect of the present invention provides a highway emergency toll collection system, including an analysis module and a transaction module; The analysis module obtains the traffic information of the target highway section and analyzes the road conditions of the target highway section; Based on the road conditions of the target highway section, analyze the number of toll gates that need to be opened at the target toll station; And, obtain the vehicle image of the vehicle to be passed; extract vehicle information from the vehicle image; The transaction module analyzes the actual license plate number of the vehicle to be passed based on the vehicle information; and obtains the passing transaction data of the vehicle to be passed based on the actual license plate number.

[0040] Some of the data in the above formula is calculated by removing the dimension and taking its value. The formula is obtained by software simulation of a large amount of collected data to obtain a formula closest to the actual situation; the preset parameters and preset thresholds in the formula are set by those skilled in the art according to the actual situation or obtained through simulation of a large amount of data.

[0041] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An emergency toll collection method for expressways, characterized in that, Including: Obtain the traffic information of the target highway section and analyze the road conditions of the target highway section; wherein, the traffic information includes traffic flow and vehicle speed; the target highway section is the monitored highway section; Based on the road conditions of the target highway section, analyze the number of toll gates that need to be opened at the target toll station; wherein, the target toll station is the end point of the target highway section; Obtain the vehicle image of the vehicle to be passed; extract vehicle information from the vehicle image; Based on the vehicle information, analyze the actual license plate number of the vehicle to be passed; based on the actual license plate number, obtain the passing transaction data of the vehicle to be passed.

2. The highway emergency toll collection method according to claim 1, characterized in that The analysis of the road conditions of the target highway section includes: At each analysis period, obtain the number of vehicles on the target highway section, and divide the number of vehicles by the analysis period to obtain the traffic flow on the target highway section during the analysis period; And obtain the driving speed of the vehicle through the interval speed measurement camera, and calculate the average value of the speeds of all vehicles during the analysis period as the vehicle speed of the target highway section; When the traffic flow in several analysis periods is greater than threshold one and the vehicle speed is less than threshold two, the target highway section is congested; otherwise, the target highway section is not congested.

3. A highway emergency toll collection method according to claim 2, characterized in that: The analysis of the number of toll gates that need to be opened at the target toll station includes: Obtain the current traffic volume of the toll gates that have been opened on the target highway section; The number of toll gates K to be opened at the target toll station is calculated by the formula K = [Q - (N Cf - ΣCi)] / Cf; Wherein, Q is the average traffic flow of the target highway section, i is the serial number of the opened toll gate, i = 0, 1, …, N, N is a positive integer, Ci is the traffic volume currently passing through toll gate i, Cf is the traffic volume of the toll gate per unit time, Σ sums up i, K takes a value less than or equal to K0, and K0 is the total number of toll gates.

4. A highway emergency toll collection method according to claim 3, characterized in that: The average traffic flow of the target highway section is the average value of the traffic flows in several analysis periods.

5. The method for emergency toll collection on expressways according to claim 1, characterized in that: The extraction of vehicle information from the vehicle image includes: Input the vehicle image into the vehicle recognition model, and the image recognition model outputs vehicle information; wherein, the vehicle image includes the roof image and side image of the vehicle; the vehicle information includes the license plate number, vehicle model and color; The vehicle recognition model is constructed by an artificial intelligence model, and the construction process is as follows: Obtain the vehicle images and corresponding vehicle information of different types of vehicles from historical data; Integrate the vehicle images and corresponding vehicle information into several groups of training data and test data; use the training data to train the artificial intelligence model; use the test data to test the trained artificial intelligence model, and adjust the artificial intelligence model according to the test results; finally obtain a vehicle recognition model with the vehicle image as the input and the vehicle information as the output; wherein, the artificial intelligence model is a BP neural network model or an RBF neural network model.

6. The highway emergency toll collection method according to claim 5, characterized in that, The analysis of the actual license plate number of the vehicle to be passed includes: Judge whether the license plate number of the vehicle to be passed is abnormal; if so, obtain the payment information of the vehicle to be passed; if not, analyze the actual license plate number of the vehicle to be passed; When the license plate number of the vehicle to be passed is abnormal, obtain the vehicles of the same model and color as the vehicle to be passed from the upstream toll station, mark them as the first vehicles, and obtain the first vehicle set; and obtaining a vehicle of the same model and color as the vehicle to be passed from the target toll station, marking it as a target vehicle, and obtaining a target vehicle set; wherein the target vehicle set includes the vehicle to be passed; comparing the license plate number of the first vehicle to the license plate number of the target vehicle; Extract vehicles whose license plate numbers are inconsistent with those of the vehicles to be passed from the first vehicle set and mark them as abnormal vehicles; If there is only one abnormal vehicle, the license plate number of the abnormal vehicle will be the actual license plate number of the vehicle to be passed; If there are at least two abnormal vehicles, the ETC records of each abnormal vehicle are obtained from the upstream toll station and the target toll station, and the abnormal vehicle with only the ETC record of the upstream toll station is extracted. If there is only one abnormal vehicle, the license plate number of the abnormal vehicle is the actual license plate number of the vehicle to be passed; if there are at least two abnormal vehicles, the vehicle images of the abnormal vehicle and the vehicle to be passed are extracted; Perform image similarity recognition on the same-surface images of the abnormal vehicle and the vehicle to be passed, and obtain the similarity rate between the same-surface images; image similarity recognition includes color recognition and defect recognition, and the results of image similarity recognition are obtained by the similarity recognition model; The abnormal vehicles whose similarity between each same-face image exceeds the preset similarity threshold are extracted, and the license plate number of the abnormal vehicle is the actual license plate number of the vehicle to be passed.

7. A highway emergency toll collection method according to claim 6, characterized in that, The determining whether the license plate number of the vehicle to be passed is abnormal includes: Based on the license plate number of the vehicle to be passed, determine whether the vehicle information of the vehicle to be passed is consistent with the standard vehicle information; if yes, the license plate number of the vehicle to be passed is normal; if not, the license plate number of the vehicle to be passed is abnormal; among them, the standard vehicle information is the vehicle information of the license plate number retrieved from the traffic vehicle management center.

8. The method for emergency toll collection on expressways according to claim 6, characterized in that: The upstream toll station is the starting point of the target expressway section.

9. The method for emergency toll collection on expressways according to claim 1, characterized in that: The obtaining of the traffic transaction data of the vehicle to be passed includes: Based on the actual license plate number, the traffic vehicle management center retrieves the traffic transaction data of the actual license plate number; the traffic transaction data includes the mileage, vehicle type and transaction terminal; The mileage and vehicle model are encrypted and transmitted to the settlement platform; The settlement platform sends the settlement result to the transaction terminal, and the transaction terminal performs a fee deduction operation.

10. A highway emergency toll collection system, operating based on a highway emergency toll collection method according to any one of claims 1 to 9, characterized in that: Includes analysis module and transaction module; Analysis module: used to obtain traffic information of the target highway section and analyze the road conditions of the target highway section; Analyze the number of toll gates that need to be opened at the target toll station based on the traffic conditions of the target highway section; and, obtaining a vehicle image of the vehicle to be passed; extracting vehicle information from the vehicle image; Transaction module: Based on vehicle information, analyze the actual license plate number of the vehicle to be passed; Based on the actual license plate number, obtain the traffic transaction data of the vehicle to be passed.

Images