Highway cloud charging method and system

Abstract

The invention provides a highway cloud toll collection method and system, and the method achieves the centralized processing and storage of data through transferring the core function of a toll collection system to a cloud end, thereby improving the toll collection efficiency and accuracy, and is specifically characterized in that an identification image block of a target toll collection vehicle is determined; processing the color gradual change loss amount at the image edge transition zone in the identification image block to obtain a gradual change harmonic image block, and obtaining a plurality of license plate identification points of the target toll vehicle through the gradual change harmonic image block; determining an identification feature vector corresponding to each license plate identification point; and performing cloud charging on the target charging vehicle according to the confidence license plate codes determined by all the identification feature vectors. According to the scheme, based on the confidence license plate code, cloud toll collection of the target toll collection vehicle can be realized when the license plate code is interfered by color gradual change of the edge transition zone, so that the accuracy of highway cloud toll collection is improved, and optimization and upgrading of a highway networking toll collection system are facilitated.

Description

Technical Field

[0001] The present application relates to the technical field of highway toll collection, and more specifically, to a highway cloud toll collection method and system. Background Art

[0002] Highway toll collection is a system that uses cloud computing, the Internet of Things and advanced information technology to innovate the traditional highway toll collection model. Traditional manual toll collection has problems such as vehicle queuing and high time cost. The highway toll collection system has significantly improved the traffic efficiency and user experience through contactless payment and automatic identification technology. The highway toll collection system relies on a variety of advanced technologies, such as electronic toll collection (ETC), which realizes automatic vehicle passage and automatic fee deduction by installing equipment on vehicles and establishing a communication network between roadside facilities. In addition, the highway toll collection system also uses RFID radio frequency identification technology, which realizes rapid vehicle passage and fee settlement by installing RFID tags on vehicles or license plates. In addition, image recognition technology is also used to charge toll vehicles at manual toll stations on highways. The combination of these technologies makes the toll collection process more intelligent and efficient, greatly improving the management efficiency and service quality of highway toll collection. With the continuous advancement of information technology and the expansion of application scenarios, the highway toll collection system is becoming an important part of modern traffic management, providing innovative solutions to problems such as urban traffic congestion and environmental protection.

[0003] In the existing technology, highway tolls are first collected by installing various types of sensing equipment on the highway to achieve real-time identification and data collection of vehicles. Secondly, advanced image recognition algorithms and data mining technologies are used to quickly and accurately identify the unique identification number of the vehicle and match it with the pre-established vehicle information database. Then, once the vehicle information is confirmed to be correct, the system automatically calculates the fee according to the set charging standards and rates, and sends a payment request to the vehicle owner or related payment account. Among them, in manual toll collection on highways, a pass card is usually required for toll collection. When the pass card is lost or damaged, the toll collector needs to manually verify the vehicle information, calculate the toll, and record it. This greatly increases the complexity and time length of the toll collection operation, resulting in collection errors and delays. To address the congestion at toll booths and the increase in management costs, automatic license plate recognition technology can be used to identify vehicles entering and exiting the highway, and cloud charging can be performed based on the identification results. In this way, even if the toll card is lost, the toll can be calculated and deducted by identifying the license plate code. However, due to the influence of factors such as vehicle type, weather conditions, and lighting changes, the edge transition zone of the license plate code during the identification of the target toll vehicle will have a color gradient problem, making the identification content of the license plate code confused and invalid, resulting in erroneous charging of the target toll vehicle, thereby reducing the accuracy of cloud charging; therefore, how to achieve cloud charging of the target toll vehicle when the license plate code is interfered by the color gradient of the edge transition zone, thereby improving the accuracy of cloud charging on highways has become a difficult problem faced by the industry. Summary of the invention

[0004] The present application provides a highway cloud toll collection method and system, which can realize cloud toll collection of target toll vehicles when the license plate code is disturbed by the color gradient of the edge transition zone, thereby improving the accuracy of highway cloud toll collection.

[0005] In a first aspect, the present application provides a highway cloud toll collection method, comprising the following steps:

[0006] Collect the passing images of the target toll vehicles through the highway toll terminal station;

[0007] The traffic image is uploaded to the highway toll collection cloud to extract the license plate image area, and the extracted license plate image area is used as the identification image block of the target toll vehicle, and the image edge transition zone in the identification image block is subjected to color gradient analysis to obtain the color gradient loss amount at the image edge transition zone in the identification image block;

[0008] Performing gradient blending on the identification image block according to the color gradient loss amount to obtain a gradient blending image block, performing identification feedback on each pixel point in the gradient blending image block to obtain an identification feedback amount of each pixel point, and extracting multiple license plate identification points of the target toll vehicle based on all the identification feedback amounts;

[0009] Determine the identification feature vector corresponding to each license plate identification point, and then obtain the confidence license plate code of the target toll vehicle according to the identification feature vector corresponding to each license plate identification point;

[0010] The toll information of the target toll vehicle is extracted through the trusted license plate code. Based on the toll information, the highway toll cloud deducts the toll from the target toll vehicle and returns the deduction result to the highway toll terminal site.

[0011] In some embodiments, uploading the traffic image to the highway toll collection cloud to extract the license plate image area specifically includes:

[0012] Obtaining a license plate identification sample of a target toll vehicle;

[0013] According to the license plate identification sample, the traffic image uploaded to the highway toll cloud is traversed and matched, and each traversal and matching is performed to calculate the identification matching value corresponding to an image area;

[0014] The image area corresponding to the maximum identification matching value is extracted as the license plate image area of ​​the target toll vehicle.

[0015] In some embodiments, performing color gradient analysis on the image edge transition zone in the identification image block to obtain the color gradient loss amount at the image edge transition zone in the identification image block specifically includes:

[0016] Acquire all edge pixel points at the image edge transition zone in the identified image block;

[0017] Performing a sliding difference calculation on the pixel value of each edge pixel point to obtain all color gradients at the image edge transition zone in the identified image block;

[0018] The color gradient loss amount at the image edge transition zone in the identified image block is determined according to all color gradients.

[0019] In some embodiments, performing gradient blending on the identification image block according to the color gradient loss amount to obtain the gradient blended image block specifically includes:

[0020] Acquire all edge pixel points at the image edge transition zone in the identified image block;

[0021] The pixel value of each edge pixel point is enhanced according to the color gradient loss amount to obtain the pixel enhancement value of each edge pixel point, and the original pixel value of each edge pixel point is replaced with the pixel enhancement value to obtain a gradient harmonized image block.

[0022] In some embodiments, determining the identification feature vector corresponding to each license plate identification point specifically includes:

[0023] Select a license plate identification point as the selected license plate identification point;

[0024] Set up identification windows for target toll collection vehicles;

[0025] Acquire all local pixel points corresponding to the selected license plate identification point according to the identification window, perform pixel elimination on all local pixel points, and obtain the identification feature vector corresponding to the selected license plate identification point;

[0026] Continue to determine the identification feature vectors corresponding to the remaining license plate identification points, and then obtain the identification feature vectors corresponding to each license plate identification point.

[0027] In some embodiments, obtaining all local pixel points corresponding to the selected license plate identification point according to the identification window, performing pixel elimination on all local pixel points, and obtaining the identification feature vector corresponding to the selected license plate identification point specifically includes:

[0028] Taking the selected license plate identification point as the center and the identification window as the range, extracting all local pixel points corresponding to the selected license plate identification point within the range;

[0029] The pixel value corresponding to the selected license plate identification point is measured similarly with the pixel value corresponding to each local pixel point to obtain the pixel similarity corresponding to each local pixel point;

[0030] The pixel similarity corresponding to each local pixel point is compared and eliminated with the preset pixel similarity, thereby obtaining the identification feature vector corresponding to the selected license plate identification point.

[0031] In some embodiments, a camera installed at a toll terminal on a highway is used to collect a passing image of a target toll vehicle.

[0032] In a second aspect, the present application provides a highway cloud toll collection system, including:

[0033] A collection module, used to collect the passing images of target toll vehicles through the toll terminal station of the expressway;

[0034] A processing module is used to upload the passing image to the highway toll cloud to extract the license plate image area, and use the extracted license plate image area as the identification image block of the target toll vehicle, perform color gradient analysis on the image edge transition zone in the identification image block, and obtain the color gradient loss amount at the image edge transition zone in the identification image block;

[0035] The processing module is further used to perform gradient blending on the identification image block according to the color gradient loss amount to obtain a gradient blending image block, perform identification feedback on each pixel point in the gradient blending image block to obtain the identification feedback amount of each pixel point, and extract multiple license plate identification points of the target toll vehicle based on all the identification feedback amounts;

[0036] The processing module is also used to determine the identification feature vector corresponding to each license plate identification point, and then obtain the confidence license plate code of the target toll vehicle according to the identification feature vector corresponding to each license plate identification point;

[0037] The execution module is used to extract the toll information of the target toll vehicle through the trusted license plate code. Based on the toll information, the highway toll cloud deducts the toll from the target toll vehicle and returns the deduction result to the highway toll terminal site.

[0038] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores codes, and the processor is configured to obtain the codes and execute the above-mentioned highway cloud toll collection method.

[0039] In a fourth aspect, the present application provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the above-mentioned highway cloud toll collection method is implemented.

[0040] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects:

[0041] In the highway cloud toll collection method and system provided by the present application, first, a passing image of a target toll vehicle is collected through a highway toll terminal site; secondly, the passing image is uploaded to the highway toll cloud to extract the license plate image area, and the extracted license plate image area is used as the identification image block of the target toll vehicle, and a color gradient analysis is performed on the image edge transition zone in the identification image block to obtain the color gradient loss amount at the image edge transition zone in the identification image block; further, the identification image block is gradient blended according to the color gradient loss amount to obtain a gradient blending image block, and identification feedback is performed on each pixel point in the gradient blending image block to obtain the identification feedback amount of each pixel point, and multiple license plate identification points of the target toll vehicle are extracted based on all the identification feedback amounts; then, the identification feature vector corresponding to each license plate identification point is determined, and then the confident license plate code of the target toll vehicle is obtained according to the identification feature vector corresponding to each license plate identification point; finally, the toll information of the target toll vehicle is extracted through the confident license plate code, and based on the toll information, the highway toll cloud deducts the toll from the target toll vehicle and returns the deduction result to the highway toll terminal site.

[0042] It can be seen that in the present application, cloud charging of target toll vehicles can be realized when the license plate code is interfered by the color gradient of the edge transition zone, thereby improving the accuracy of cloud charging of highways; first, the identification area of ​​the target toll vehicle is extracted according to the license plate identification sample to realize the aggregation of the identification features of the target toll vehicle, thereby avoiding the interference caused by other background areas; secondly, the amount of color information loss of the identification area at the edge transition zone is determined, and the identification area is gradually blended according to the color information loss to obtain a gradient blending image block, so as to effectively ensure the smoothness and naturalness of the color transition area, reduce the amount of color loss caused by the color gradient, and thus avoid the confusion and failure of the identification content of the license plate code; further, the gradient blending image block after gradient blending is subjected to identification feedback, and multiple license plate identification points of the target toll vehicle are identified according to the feedback degree of the identification feedback, so as to more accurately locate and identify the license plate information, thereby reducing Reduce the identification interference caused by various external factors; further, determine the identification feature vector of each license plate identification point according to the identification window, so as to effectively identify the identification features around the current license plate identification point, and then ensure that the extracted features can effectively reflect the local information in the image; then, obtain the confident license plate code of the target toll vehicle from the identification feature vector of each license plate identification point, so as to effectively extract the driving information of the target toll vehicle, thereby avoiding misjudgment of the toll on the target toll vehicle; finally, extract the toll information of the target toll vehicle through the confident license plate code, and based on the toll information, the target toll vehicle is deducted by the highway toll cloud and the deduction result is returned to the highway toll terminal site; in summary, the technical solution provided by the present application can realize cloud charging of the target toll vehicle when the license plate code is interfered by the color gradient of the edge transition zone, thereby improving the accuracy of highway cloud charging. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 is an exemplary flow chart of a highway cloud toll collection method according to some embodiments of the present application;

[0044] Figure 2 is an exemplary flow chart for determining the amount of color gradient loss according to some embodiments of the present application;

[0045] Figure 3 is an exemplary flow chart for determining identification feedback amount according to some embodiments of the present application;

[0046] Figure 4 is a schematic diagram of exemplary hardware and / or software of a highway cloud toll collection system according to some embodiments of the present application;

[0047] Figure 5It is a structural diagram of a computer device for implementing a highway cloud toll collection method according to some embodiments of the present application. DETAILED DESCRIPTION

[0048] The core of the present application is to first collect the passing image of the target toll vehicle through the highway toll terminal station; secondly, upload the passing image to the highway toll cloud to extract the license plate image area, and use the extracted license plate image area as the identification image block of the target toll vehicle, perform color gradient analysis on the image edge transition zone in the identification image block, and obtain the color gradient loss at the image edge transition zone in the identification image block; further, perform gradient blending on the identification image block according to the color gradient loss to obtain a gradient blending image block, perform identification feedback on each pixel point in the gradient blending image block, and obtain each pixel point identification feedback, and extract multiple license plate identification points of the target toll vehicle based on all the identification feedback; then, determine the identification feature vector corresponding to each license plate identification point, and then obtain the confident license plate code of the target toll vehicle according to the identification feature vector corresponding to each license plate identification point; finally, extract the toll information of the target toll vehicle through the confident license plate code, and based on the toll information, deduct the toll from the target toll vehicle by the highway toll cloud and return the deduction result to the highway toll terminal site, so that cloud toll collection of the target toll vehicle can be realized when the license plate code is interfered by the color gradient of the edge transition zone, thereby improving the accuracy of highway cloud toll collection.

[0049] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods. Figure 1 , which is an exemplary flow chart of a highway cloud charging method according to some embodiments of the present application. The highway cloud charging method 100 mainly includes the following steps:

[0050] In step 101, a passing image of a target toll vehicle is collected through a highway toll terminal station.

[0051] In specific implementation, the passing image of the target toll vehicle can be collected by a camera installed at a highway toll terminal site. The passing image represents the vehicle image captured by the surveillance camera when the target toll vehicle passes through the highway toll exit. In addition, in other embodiments, other image acquisition devices can also be used to collect the passing image of the target toll vehicle, which is not limited here.

[0052] It should be noted that the target toll vehicle in this embodiment refers to the vehicle currently leaving the highway toll gate and waiting to pay. The collection of the passing image is to identify the toll information of the vehicle, so as to accurately calculate the toll collection. In addition, the highway toll terminal site refers to the site set up on the highway for collecting vehicle tolls, usually called a toll station. The toll station is an important part of the highway management system and is responsible for collecting, recording and managing tolls for vehicles passing through the highway.

[0053] In step 102, the passing image is uploaded to the highway toll collection cloud to extract the license plate image area, and the extracted license plate image area is used as the identification image block of the target toll vehicle. The image edge transition zone in the identification image block is subjected to color gradient analysis to obtain the color gradient loss at the image edge transition zone in the identification image block.

[0054] In some embodiments, uploading the traffic image to the highway toll collection cloud to extract the license plate image area may specifically adopt the following steps, namely:

[0055] Obtaining a license plate identification sample of a target toll vehicle;

[0056] According to the license plate identification sample, the traffic image uploaded to the highway toll cloud is traversed and matched, and each traversal and matching is performed to calculate the identification matching value corresponding to an image area;

[0057] The image area corresponding to the maximum identification matching value is extracted as the license plate image area of ​​the target toll vehicle.

[0058] In specific implementation, the license plate identification sample of the target toll vehicle can be obtained through the highway cloud toll database. The license plate identification sample represents the identification reference template used for feature identification of the target toll vehicle in the highway toll system, that is, this application takes the license plate area of ​​the target toll vehicle as the identification object, and by obtaining the license plate identification sample, the identification block of the target toll vehicle can be effectively screened out.

[0059] It should be noted that the highway toll cloud in this application refers to a platform that uses cloud computing technology to achieve digitalization, intelligence and centralized management of the highway toll system, and provides more efficient, flexible and reliable toll services by migrating data processing, storage and management tasks in traditional toll systems to the cloud; in this embodiment, the highway cloud toll database represents a system for storing, managing and processing data related to highway tolls.

[0060] In a specific implementation, the traffic image uploaded to the highway toll cloud is traversed and matched according to the license plate identification sample, and an identification matching value corresponding to an image area is calculated each time the traversal is performed, that is, starting from the upper left corner area of ​​the traffic image uploaded to the highway toll cloud according to the scale size of the license plate identification sample, the traversal is performed in order from left to right and from top to bottom, and an image area is obtained each time the traversal is performed, and for each image area, a matching calculation is performed between the image area and the license plate identification sample to obtain the identification matching value corresponding to the image area.

[0061] In specific implementation, the image area and the license plate identification sample can be matched by the pixel mean square error matching method, that is, the mean square error of the pixel values ​​corresponding to the image area and the license plate identification sample is calculated, and the mean square error is used as the identification matching value. It will not be repeated here. In other embodiments, other matching calculation methods can also be used for matching calculation, for example, the image peak signal-to-noise ratio method and the image structure similarity method, which are not limited here.

[0062] It should be noted that, in this embodiment, the identification matching value represents a measure of the degree of similarity between image features, that is, the larger the identification matching value, the greater the degree of similarity between image features, and the smaller the identification matching value, the smaller the degree of similarity between image features. By determining the identification matching value, the identification area of ​​the target toll vehicle can be effectively identified, thereby improving the accuracy of toll collection.

[0063] It should also be noted that the license plate image area in the present application represents the image area where the license plate is located. By determining the license plate image area, the current target vehicle charging information can be effectively obtained.

[0064] In the specific implementation, the extracted license plate image area is used as the identification image block of the target toll vehicle. The identification image block in this application represents the identification feature block of the target toll vehicle, and the identification image block contains the key features of the target toll vehicle, namely the license plate code. In the manual toll collection of highway cloud toll collection, a pass card is usually required for toll collection. When the pass card is lost or damaged, the toll collector needs to manually verify the vehicle information, calculate the toll, and record it. This greatly increases the complexity and time length of the toll collection operation, causing congestion at the toll station and increased management costs. Therefore, by determining the identification image block, the highway cloud toll collection system can effectively determine the identity of the vehicle and perform corresponding toll collection operations.

[0065] It should be noted that the image edge transition zone in the identification image block represents the boundary portion from one area (such as a foreground object) to another area (such as the background) in the identification image block. In this area, the pixel value gradually transitions from an obvious value to another value, and this area usually contains a gradient color or grayscale value.

[0066] In some embodiments, reference Figure 2 As shown in FIG. 1 , this figure is an exemplary flow chart of determining the amount of color gradient loss according to some embodiments of the present application. In this embodiment, color gradient analysis is performed on the image edge transition zone in the identification image block to obtain the amount of color gradient loss at the image edge transition zone in the identification image block. The following steps can be used to achieve this:

[0067] First, in step 1021, all edge pixel points at the image edge transition zone in the identified image block are obtained;

[0068] Next, in step 1022, a sliding difference calculation is performed on the pixel value of each edge pixel point to obtain all color gradients at the image edge transition zone in the identified image block;

[0069] Finally, in step 1023, the amount of color gradient loss at the image edge transition zone in the marked image block is determined according to all color gradients.

[0070] In specific implementation, all edge pixel points at the image edge transition zone in the identification image block can be obtained through the image processing software Open CV. In addition, in other embodiments, other image processing software can also be used to obtain all edge pixel points at the image edge transition zone in the identification image block, such as ImageJ, Photoshop, Python, which is not limited here. It should be noted that the image edge transition zone in this embodiment represents the transition boundary of a specific area in the identification image block. For the identification image block, the edge transition zone generally refers to the edge portion between the license plate code and the background area in the identification image block.

[0071] In a specific implementation, a sliding difference calculation is performed on the pixel value of each edge pixel point to obtain all color gradients in the image edge transition zone in the identification image block, that is: for each edge pixel point, a difference calculation is performed between the pixel value of the edge pixel point and the pixel value of the next edge pixel point, and the difference calculation result is used as the color gradient corresponding to the edge pixel point, thereby obtaining all color gradients in the image edge transition zone in the identification image block.

[0072] It should be noted that the color gradient analysis in the present application refers to the process of analyzing the color changes in the transition zone between the background area and the target area in the image, wherein the color gradient analysis is performed on the image edge transition zone in the identification image block, that is: all edge pixel points at the image edge transition zone in the identification image block are obtained; the pixel value of each edge pixel point is slipped and calculated to obtain all color gradients at the image edge transition zone in the identification image block; the color gradient loss at the image edge transition zone in the identification image block is determined based on all the color gradients, that is, the color gradient analysis of the image edge transition zone in the identification image block is achieved.

[0073] It should be noted that the color gradient in this embodiment represents the amount of hue change of the edge pixel points at the image edge transition zone in the identification image block, that is, it is determined by the pixel value deviation of adjacent edge pixel points in the present application. In highway cloud toll collection, it is usually necessary to identify the specific identification area of ​​the target toll vehicle. However, during the identification process, the edge pixel points at the image edge transition zone will undergo uncertain changes in hue, thereby affecting the identification effect of the specific identification area. Therefore, the influence of the uncertainty change in hue of the edge pixel points can be avoided by determining the color gradient.

[0074] In a specific implementation, the amount of color gradient loss at the image edge transition zone in the identification image block is determined based on all color gradients, that is, the standard deviation of all color gradients is calculated, and the standard deviation calculation result is used as the amount of color gradient loss at the image edge transition zone in the identification image block. In addition, in other embodiments, other calculation methods can also be used to calculate the amount of color gradient loss at the image edge transition zone in the identification image block, which is not limited here.

[0075] It should be noted that the color gradient loss amount in the present application represents the amount of color information lost due to the color gradient at the image edge transition zone in the identification image block, that is, the larger the color gradient loss amount, the larger the color information lost due to the color gradient at the image edge transition zone in the identification image block, and the smaller the color gradient loss amount, the smaller the color information lost due to the color gradient at the image edge transition zone in the identification image block. The color gradient loss amount reflects the smoothness and fidelity of the color change at the image edge transition zone in the identification image block. By determining the color gradient loss amount, the identification features of the target toll vehicle can be correctly identified, thereby reducing misidentification caused by unnatural or discontinuous color transition.

[0076] In step 103, the identification image block is gradient-blended according to the color gradient loss amount to obtain a gradient-blended image block, identification feedback is performed on each pixel point in the gradient-blended image block to obtain the identification feedback amount of each pixel point, and multiple license plate identification points of the target toll vehicle are extracted based on all the identification feedback amounts.

[0077] In some embodiments, the identification image block is gradually blended according to the color gradient loss amount to obtain the gradient blended image block, which can be specifically obtained by the following steps, namely:

[0078] Acquire all edge pixel points at the image edge transition zone in the identified image block;

[0079] The pixel value of each edge pixel point is enhanced according to the color gradient loss amount to obtain the pixel enhancement value of each edge pixel point, and the original pixel value of each edge pixel point is replaced with the pixel enhancement value to obtain a gradient harmonized image block.

[0080] In a specific implementation, the pixel value of each edge pixel is enhanced according to the color gradient loss to obtain the pixel enhancement value of each edge pixel, that is, the color gradient loss is summed with the pixel value of each edge pixel to obtain the pixel enhancement value of each edge pixel. The pixel enhancement value represents the value obtained by enhancing the original pixel value by the color gradient loss. In other embodiments, other calculation methods can also be used to calculate the pixel enhancement value of each edge pixel, which is not limited here.

[0081] It should be noted that gradient blending in the present application refers to the process of enhancing the pixel values ​​of edge pixels by the amount of color loss. Gradient blending can make the image look more natural and continuous, so as to reduce the unnatural or abrupt visual effects of the image caused by color gradient, thereby improving the image quality and recognition accuracy.

[0082] It should also be noted that the gradient blending image block in the present application represents a new image obtained after gradient blending of the identification image block. By gradient blending the identification image block, the smoothness and naturalness of the color transition area can be effectively ensured, and the color loss caused by the color gradient can be reduced to improve the visual quality of the identification image block. Due to the influence of factors such as vehicle type, weather conditions, and lighting changes, the edge transition zone of the identification target (license plate code) of the target toll vehicle will have a color gradient problem, making the identification content of the license plate code confused and invalid, resulting in erroneous charging of the target toll vehicle. Therefore, by determining the gradient blending image block, the target toll vehicle can be effectively characterized to ensure the smooth progress of the toll operation.

[0083] In some embodiments, reference Figure 3As shown in FIG. 1 , this figure is an exemplary flow chart of determining the identification feedback amount according to some embodiments of the present application. In this embodiment, identification feedback is performed on each pixel point in the gradient blending image block, and the identification feedback amount of each pixel point can be obtained by the following steps:

[0084] First, in step 1031, a pixel point in the gradient blending image block is selected;

[0085] Next, in step 1032, all adjacent pixel points of the pixel point are obtained, and the opposite side pixel points are combined for all the adjacent pixel points to obtain a plurality of opposite side adjacent pixel point combinations;

[0086] Further, in step 1033, the pixel identification deviation of each adjacent pixel point combination on opposite sides is determined;

[0087] Furthermore, in step 1034, the central pixel identification deviation is determined according to the pixel value of the pixel point and the pixel values ​​of all adjacent pixels;

[0088] Then, in step 1035, the identification feedback amount of the pixel point is determined by the pixel identification deviations of all the adjacent pixel points on the opposite sides and the central pixel identification deviation;

[0089] Finally, in step 1036, the above steps are repeated to obtain the identification feedback amounts of the remaining pixels.

[0090] In a specific implementation, all neighboring pixel points of the pixel point are obtained, and the opposite side pixel points of all the neighboring pixel points are combined to obtain multiple opposite side adjacent pixel point combinations, that is: all neighboring pixel points of the pixel point are obtained through the eight-field method, all neighboring pixel points include the pixel points in eight directions of the pixel point, the upper and lower opposite side adjacent pixel points of the pixel point are combined, the left and right opposite side adjacent pixel points of the pixel point are combined, and the two diagonally opposite side adjacent pixel points of the pixel point are respectively combined to obtain multiple opposite side adjacent pixel point combinations. It should be noted that the opposite side adjacent pixel point combination in this embodiment represents a set of opposite side adjacent pixel points combined together.

[0091] In some embodiments, the pixel identification deviation of each adjacent pixel point combination on opposite sides may be determined by the following steps, namely:

[0092] For each pair of adjacent pixel points on opposite sides, the pixel values ​​of two adjacent pixel points in the pair of adjacent pixel points on opposite sides are difference calculated to obtain the pixel identification deviation of the pair of adjacent pixel points on opposite sides, and then the pixel identification deviation of each pair of adjacent pixel points on opposite sides is obtained.

[0093] It should be noted that in other embodiments, other calculation methods can also be used to calculate the pixel identification deviation of the adjacent pixel point combination on the opposite side, which is not limited here. In this embodiment, the pixel identification deviation represents the pixel identification deviation between the pixel values ​​of two adjacent pixel points in the adjacent pixel point combination on the opposite side.

[0094] In specific implementation, the central pixel identification deviation is determined based on the pixel value of the pixel point and the pixel values ​​of all adjacent pixels, that is: first, the absolute difference between the pixel value of the pixel point and the pixel values ​​of each adjacent pixel point is calculated, and the average value of all absolute difference calculation results is calculated to obtain the central pixel identification deviation.

[0095] It should be noted that in other embodiments, other calculation methods can also be used to calculate the center pixel identification deviation. In this embodiment, the center pixel identification deviation is used to measure the degree of deviation of the pixel identification concentration between the pixel point and all adjacent pixel points. That is, the larger the center pixel identification deviation, the greater the degree of deviation of the pixel identification concentration between the pixel point and all adjacent pixel points, and vice versa.

[0096] In some embodiments, the identification feedback amount of the pixel point can be determined by the pixel identification deviations of all adjacent pixel points on opposite sides and the central pixel identification deviation, and specifically the following steps can be adopted, namely:

[0097] Get the total number of combinations of all adjacent pixel points on opposite sides;

[0098] Compare the pixel identification deviation of each combination of adjacent pixel points on opposite sides with the central pixel identification deviation, and obtain the frequency of the pixel identification deviation being greater than the central pixel identification deviation;

[0099] The identification feedback amount of the pixel point is determined according to the ratio of the frequency to the total number of combinations.

[0100] It should be noted that when the pixel identification deviation is less than or equal to the central pixel identification deviation, no processing is performed.

[0101] It should also be noted that the identification feedback in the present application refers to the process of measuring the proximity of pixel values ​​between pixel points when identifying the target toll vehicle. In addition, the identification feedback amount in the present application refers to the process of measuring the degree of proximity between the pixel value of the current pixel point and the pixel value of the adjacent pixel point when identifying the target toll vehicle. That is, the larger the identification feedback amount, the greater the degree of proximity between the pixel value of the previous pixel point and the pixel value of the adjacent pixel point, and the smaller the identification feedback amount, the smaller the degree of proximity between the pixel value of the previous pixel point and the pixel value of the adjacent pixel point. By determining the identification feedback amount, the degree of proximity between the current pixel point and the surrounding adjacent pixel points can be effectively fed back.

[0102] In some embodiments, the following steps may be used to extract multiple license plate identification points of the target toll vehicle based on all identification feedback amounts, namely:

[0103] Compare each identification feedback amount with a preset identification feedback amount, extract all identification feedback amounts that are greater than the preset identification feedback amount, and use the pixel points corresponding to each extracted identification feedback amount as the license plate identification point of the target toll vehicle;

[0104] All identification feedback amounts that are less than or equal to the preset identification feedback amount will not be processed.

[0105] It should be noted that the preset identification feedback amount in this embodiment represents a pre-set standard identification feedback amount, which can be set according to actual needs. By setting the standard identification feedback amount, the valid license plate feature points of the target toll vehicle can be effectively identified.

[0106] It should also be noted that the license plate identification points in the present application represent key points for identifying license plate information. These points generally include characteristic positions of letters, numbers, symbols, etc. on the license plate. The license plate identification points have positioning accuracy and flexibility. By determining the license plate identification points, the vehicle identity can be effectively identified and the vehicle toll can be calculated. In the vehicle license plate code recognition, the letters, numbers, symbols and the background have obvious color differences, but there are also some letters, numbers, symbols and background colors mixed. Therefore, the key identification points (license plate identification points) can be extracted by setting a threshold value, thereby accurately identifying the license plate.

[0107] In step 104, the identification feature vector corresponding to each license plate identification point is determined, and then the confidence license plate code of the target toll vehicle is obtained according to the identification feature vector corresponding to each license plate identification point.

[0108] In some embodiments, the identification feature vector corresponding to each license plate identification point may be determined by the following steps, namely:

[0109] Select a license plate identification point as the selected license plate identification point;

[0110] Set up identification windows for target toll collection vehicles;

[0111] Acquire all local pixel points corresponding to the selected license plate identification point according to the identification window, perform pixel elimination on all local pixel points, and obtain the identification feature vector corresponding to the selected license plate identification point;

[0112] Continue to determine the identification feature vectors corresponding to the remaining license plate identification points, and then obtain the identification feature vectors corresponding to each license plate identification point.

[0113] In specific implementation, the identification window of the target toll vehicle can be set to a 3*3 window. In addition, in other embodiments, it can also be set according to actual needs, which is not limited here.

[0114] It should be noted that the identification window in the present application refers to a regional window used for local feature recognition. By determining the identification window, the feature distribution around the identification feature point can be effectively identified.

[0115] It should also be noted that each license plate identification point corresponds to multiple local pixel points, that is, for each license plate identification point, with the license plate identification point as the center and the identification window as the range, all pixel points within the range are regarded as local pixel points corresponding to the license plate identification point (except the license plate identification point), and the local pixel points represent the pixel points adjacent to the license plate identification point in the identification window.

[0116] In some embodiments, all local pixel points corresponding to the selected license plate identification point are obtained according to the identification window, and pixel elimination is performed on all local pixel points to obtain the identification feature vector corresponding to the selected license plate identification point. Specifically, the following steps can be adopted, namely:

[0117] Taking the selected license plate identification point as the center and the identification window as the range, extracting all local pixel points corresponding to the selected license plate identification point within the range;

[0118] The pixel value corresponding to the selected license plate identification point is measured similarly with the pixel value corresponding to each local pixel point to obtain the pixel similarity corresponding to each local pixel point;

[0119] The pixel similarity corresponding to each local pixel point is compared and eliminated with the preset pixel similarity, thereby obtaining the identification feature vector corresponding to the selected license plate identification point.

[0120] In specific implementation, the pixel value corresponding to the selected license plate identification point and the pixel value corresponding to each local pixel point can be measured for similarity using the Euclidean distance method to obtain the pixel similarity corresponding to each local pixel point. In addition, in other embodiments, other similarity measurement methods can be used for similarity measurement, such as Manhattan distance and Jaccard similarity coefficient, which are not limited here.

[0121] In specific implementation, the pixel similarity corresponding to each local pixel point is compared and eliminated with the preset pixel similarity, so as to obtain the identification feature vector corresponding to the selected license plate identification point, that is: all pixel similarities greater than the preset pixel similarity are extracted, and the pixel values ​​of the local pixel points corresponding to each extracted pixel similarity are eliminated to 0; all pixel similarities less than or equal to the preset pixel similarity are extracted, and the pixel values ​​of the local pixel points corresponding to each extracted pixel similarity are eliminated to 1, and the values ​​obtained after eliminating the pixel values ​​of each local pixel point are vectorized to obtain the identification feature vector corresponding to the selected license plate identification point.

[0122] It should be noted that the preset pixel similarity in this embodiment represents a pre-set standard pixel similarity, which can be set according to actual needs. By determining the preset pixel similarity, local pixel points with large differences can be effectively identified, and the pixel values ​​of local pixel points with large differences can be cancelled.

[0123] It should also be noted that the identification feature vector in the present application represents the identification feature within the corresponding range of the current license plate identification point, that is, one license plate identification point corresponds to the range of an identification window, and the identification feature vector contains the identification values ​​of each pixel point within the corresponding range of the current license plate identification point. The identification values ​​are 0 and 1, that is, the result of pixel elimination, and the identification value represents the value obtained after eliminating the pixel values ​​of each pixel point within the corresponding range of the current license plate identification point. In addition, the identification feature vector is arranged in a clockwise direction starting from the upper left corner in the identification window. By determining the identification feature vector, the identification features around the current license plate identification point can be effectively identified, thereby ensuring that the extracted features can effectively reflect the local information in the image and support the realization of related image analysis and application goals.

[0124] In some embodiments, the following steps may be used to obtain the trusted license plate code of the target toll vehicle according to the identification feature vector corresponding to each license plate identification point, namely:

[0125] Perform grayscale conversion on the identification feature vector corresponding to each license plate identification point to obtain the grayscale conversion value of each license plate identification point;

[0126] The grayscale conversion value of each license plate identification point is mapped correspondingly to the identification image block to obtain the trustworthy license plate code of the target toll vehicle.

[0127] In the specific implementation, the identification feature vector corresponding to each license plate identification point is converted into grayscale to obtain the grayscale conversion value of each license plate identification point, that is, for each license plate identification point, the corresponding identification feature vectors are combined into binary code, and the binary code is converted into decimal to obtain the grayscale conversion value of each license plate identification point. For example, the identification feature vector is (1, 0, 0, 1, 0, 1, 0, 0), then the binary code obtained by the combination of identification feature vectors is: 10010100, and the grayscale conversion value obtained after decimal conversion is: 148, wherein the grayscale conversion value in this embodiment represents the grayscale value obtained after local feature fusion.

[0128] In specific implementation, the grayscale conversion value of each license plate identification point is mapped to the identification image block to obtain the confident license plate code of the target toll vehicle, that is: for each license plate identification point, the corresponding grayscale conversion value is used to replace the grayscale value in the identification image block to obtain a confident license plate image with prominent license plate code, and the confident license plate code of the target toll vehicle is obtained from the confident license plate image.

[0129] In specific implementation, the trusted license plate code of the target toll vehicle is obtained from the trusted license plate image, that is, the trusted license plate code of the target toll vehicle is extracted through the code recognition technology in image recognition, such as OCR (optical character recognition) technology, which will not be repeated here.

[0130] It should be noted that in this embodiment, the trusted license plate code represents a valid license plate code used to describe the target toll vehicle, that is, the license plate number characteristics of the target toll vehicle. In the recognition of vehicle license plates, due to the influence of factors such as vehicle type, weather conditions, and lighting changes, the edge transition zone of the license plate code of the target toll vehicle may have a color gradient problem, making the recognition content of the license plate code confused and invalid, resulting in erroneous charging of the target toll vehicle. Therefore, by determining the trusted license plate code, the license plate number information of the target toll vehicle can be accurately identified, and the required toll is calculated based on the license plate number information of the target toll vehicle, thereby realizing automated toll management, which not only improves the toll efficiency, but also reduces human errors and waiting time in queues.

[0131] In step 105, the toll information of the target toll vehicle is extracted through the trusted license plate code. Based on the toll information, the highway toll cloud deducts the toll from the target toll vehicle and returns the deduction result to the highway toll terminal site.

[0132] In some embodiments, the toll information of the target toll vehicle is extracted through the trusted license plate code. Based on the toll information, the highway toll cloud deducts the toll from the target toll vehicle and returns the deduction result to the highway toll terminal site. Specifically, the following steps may be adopted, namely:

[0133] The highway toll cloud parses the trusted license plate code, and when the parsing is successful, the highway toll cloud automatically retrieves the toll information of the target toll vehicle, the toll information including the toll amount and user information of the target toll vehicle;

[0134] According to the toll amount and user information in the toll information, the highway toll cloud automatically deducts the corresponding fee from the user's payment account and returns the deduction result to the highway toll terminal site.

[0135] In specific implementation, the highway toll cloud parses the trusted license plate code, that is, the highway toll cloud automatically obtains the driving information of the vehicle corresponding to the trusted license plate code based on the information provided by the trusted license plate code, and calculates the toll information of the vehicle corresponding to the trusted license plate code through the driving information, thereby completing the parsing of the trusted license plate code.

[0136] It should be noted that the charging information includes the charging amount and user information of the target toll vehicle.

[0137] In specific implementation, the highway toll cloud parses the trusted license plate code. When the parsing is successful, the highway toll cloud automatically retrieves the toll information of the target toll vehicle. According to the toll amount and user information in the toll information, the highway toll cloud automatically deducts the corresponding fee from the user's payment account, and returns the deduction result to the highway toll terminal site to complete the cloud toll collection operation for the target toll vehicle.

[0138] In addition, in another aspect of the present application, in some embodiments, the present application provides a highway cloud toll collection system, referring to Figure 4 , which is a schematic diagram of exemplary hardware and / or software of a highway cloud toll collection system according to some embodiments of the present application. The highway cloud toll collection system 200 includes: a collection module 201, a processing module 202 and an execution module 203, which are described as follows:

[0139] The acquisition module 201 in this application is mainly used to collect the passing image of the target toll vehicle through the highway toll terminal station;

[0140] Processing module 202, in this application, processing module 202 is mainly used to upload the passing image to the highway toll cloud to extract the license plate image area, and use the extracted license plate image area as the identification image block of the target toll vehicle, perform color gradient analysis on the image edge transition zone in the identification image block, and obtain the color gradient loss at the image edge transition zone in the identification image block;

[0141] The processing module 202 is further used to perform gradient blending on the identification image block according to the color gradient loss amount to obtain a gradient blending image block, perform identification feedback on each pixel point in the gradient blending image block to obtain the identification feedback amount of each pixel point, and extract multiple license plate identification points of the target toll vehicle based on all the identification feedback amounts;

[0142] In addition, the processing module 202 is also used to determine the identification feature vector corresponding to each license plate identification point, and then obtain the confidence license plate code of the target toll vehicle according to the identification feature vector corresponding to each license plate identification point;

[0143] Execution module 203, in this application, execution module 203 is mainly used to extract the toll information of the target toll vehicle through the trusted license plate code. Based on the toll information, the highway toll cloud deducts the toll from the target toll vehicle and returns the deduction result to the highway toll terminal site.

[0144] In addition, the present application also provides a computer device, which includes a memory and a processor, the memory stores code, and the processor is configured to obtain the code and execute the above-mentioned highway cloud charging method.

[0145] In some embodiments, reference Figure 5 , which is a schematic diagram of the structure of a computer device for applying a highway cloud charging method according to some embodiments of the present application. The highway cloud charging method in the above embodiment can be Figure 5 The computer device 300 shown in the figure is implemented, and the computer device 300 includes at least one processor 301, a communication bus 302, a memory 303 and at least one communication interface 304.

[0146] Processor 301 can be a general-purpose central processing unit (CPU), or an application-specific integrated circuit (ASIC) or one or more for controlling the execution of the highway cloud charging method in the present application.

[0147] The communication bus 302 may be used to transmit information between the above-mentioned components.

[0148] The memory 303 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, an optical disc storage (including a compressed optical disc, a laser disc, an optical disc, a digital versatile disc, a Blu-ray disc, etc.), a magnetic disk or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 303 may exist independently and be connected to the processor 301 via the communication bus 302. The memory 303 may also be integrated with the processor 301.

[0149] Among them, the memory 303 is used to store the program code for executing the solution of the present application, and the execution is controlled by the processor 301. The processor 301 is used to execute the program code stored in the memory 303. The program code may include one or more software modules. The determination of the highway cloud charging method in the above embodiment can be implemented by the processor 301 and one or more software modules in the program code in the memory 303.

[0150] The communication interface 304 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0151] In a specific implementation, as an embodiment, a computer device may include multiple processors, each of which may be a single-CPU processor or a multi-CPU processor. The processor here may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0152] The above-mentioned computer device can be a general-purpose computer device or a special-purpose computer device. In a specific implementation, the computer device can be a desktop computer, a portable computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device or an embedded device. The embodiment of the present application does not limit the type of computer device.

[0153] In addition, the present application also provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the above-mentioned highway cloud toll collection method is implemented.

[0154] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0155] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A highway cloud toll collection method, characterized in that: The steps include: Collect the passing images of the target toll vehicles through the highway toll terminal station; The traffic image is uploaded to the highway toll collection cloud to extract the license plate image area, and the extracted license plate image area is used as the identification image block of the target toll vehicle, and the image edge transition zone in the identification image block is subjected to color gradient analysis to obtain the color gradient loss amount at the image edge transition zone in the identification image block; Performing gradient blending on the identification image block according to the color gradient loss amount to obtain a gradient blending image block, performing identification feedback on each pixel point in the gradient blending image block to obtain an identification feedback amount of each pixel point, and extracting multiple license plate identification points of the target toll vehicle based on all the identification feedback amounts; Determine the identification feature vector corresponding to each license plate identification point, and then obtain the confidence license plate code of the target toll vehicle according to the identification feature vector corresponding to each license plate identification point; The toll information of the target toll vehicle is extracted through the trusted license plate code. Based on the toll information, the highway toll cloud deducts the toll from the target toll vehicle and returns the deduction result to the highway toll terminal site.

2. The method according to claim 1, characterized in that The process of uploading the traffic image to the highway toll collection cloud to extract the license plate image area specifically includes: Obtaining a license plate identification sample of a target toll vehicle; According to the license plate identification sample, the traffic image uploaded to the highway toll cloud is traversed and matched, and each traversal and matching is performed to calculate the identification matching value corresponding to an image area; The image area corresponding to the maximum identification matching value is extracted as the license plate image area of ​​the target toll vehicle.

3. The method according to claim 1, characterized in that Performing color gradient analysis on the image edge transition zone in the identification image block to obtain the color gradient loss amount at the image edge transition zone in the identification image block specifically includes: Acquire all edge pixel points at the image edge transition zone in the identified image block; Performing a sliding difference calculation on the pixel value of each edge pixel point to obtain all color gradients at the image edge transition zone in the identified image block; The color gradient loss amount at the image edge transition zone in the identified image block is determined according to all color gradients.

4. The method according to claim 1, characterized in that Gradual blending is performed on the identification image block according to the color gradient loss amount to obtain the gradient blended image block, specifically comprising: Acquire all edge pixel points at the image edge transition zone in the identified image block; The pixel value of each edge pixel point is enhanced according to the color gradient loss amount to obtain the pixel enhancement value of each edge pixel point, and the original pixel value of each edge pixel point is replaced with the pixel enhancement value to obtain a gradient harmonized image block.

5. The method according to claim 1, characterized in that Determining the identification feature vector corresponding to each license plate identification point specifically includes: Select a license plate identification point as the selected license plate identification point; Set up identification windows for target toll collection vehicles; Acquire all local pixel points corresponding to the selected license plate identification point according to the identification window, perform pixel elimination on all local pixel points, and obtain the identification feature vector corresponding to the selected license plate identification point; Continue to determine the identification feature vectors corresponding to the remaining license plate identification points, and then obtain the identification feature vectors corresponding to each license plate identification point.

6. The method according to claim 5, characterized in that Acquiring all local pixel points corresponding to the selected license plate identification point according to the identification window, performing pixel elimination on all local pixel points, and obtaining the identification feature vector corresponding to the selected license plate identification point specifically includes: Taking the selected license plate identification point as the center and the identification window as the range, extracting all local pixel points corresponding to the selected license plate identification point within the range; The pixel value corresponding to the selected license plate identification point is measured similarly with the pixel value corresponding to each local pixel point to obtain the pixel similarity corresponding to each local pixel point; The pixel similarity corresponding to each local pixel point is compared and eliminated with the preset pixel similarity, thereby obtaining the identification feature vector corresponding to the selected license plate identification point.

7. The method according to claim 1, characterized in that The passing images of the target toll vehicles are collected by cameras installed at the toll terminal sites of the expressway.

8. A highway cloud toll collection system, characterized in that: include: A collection module, used to collect the passing images of target toll vehicles through the toll terminal station of the expressway; A processing module is used to upload the passing image to the highway toll cloud to extract the license plate image area, and use the extracted license plate image area as the identification image block of the target toll vehicle, perform color gradient analysis on the image edge transition zone in the identification image block, and obtain the color gradient loss amount at the image edge transition zone in the identification image block; The processing module is further used to perform gradient blending on the identification image block according to the color gradient loss amount to obtain a gradient blending image block, perform identification feedback on each pixel point in the gradient blending image block to obtain the identification feedback amount of each pixel point, and extract multiple license plate identification points of the target toll vehicle based on all the identification feedback amounts; The processing module is also used to determine the identification feature vector corresponding to each license plate identification point, and then obtain the confidence license plate code of the target toll vehicle according to the identification feature vector corresponding to each license plate identification point; The execution module is used to extract the toll information of the target toll vehicle through the trusted license plate code. Based on the toll information, the highway toll cloud deducts the toll from the target toll vehicle and returns the deduction result to the highway toll terminal site.

9. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores codes, and the processor is configured to obtain the codes and execute the highway cloud charging method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the highway cloud charging method as described in any one of claims 1 to 7 is implemented.

Images