Green channel traffic management method and device based on big data, equipment and storage medium

Abstract

The invention provides a green channel traffic management method and device based on big data, equipment and a storage medium, and belongs to the technical field of data processing, and the method comprises the steps: determining a green channel suspiciousness degree based on the transportation behavior of a target vehicle; determining a target inspection standard based on the green channel suspiciousness; and sending the target inspection standard to the target inspection equipment, obtaining a green channel inspection result based on the target inspection equipment, and determining whether the green channel inspection result meets a green channel condition based on the green channel inspection result. According to the big data-based green channel traffic management method and device, the equipment and the storage medium provided by the invention, the accuracy and reliability of green channel traffic management can be improved.

Description

Technical Field

[0001] The present disclosure belongs to the field of data processing technology, and more specifically, to a green pass management method and device, equipment, and storage medium based on big data. Background Art

[0002] In the modern transportation system, the green channel policy aims to ensure the fast and low-cost transportation of fresh agricultural products and other goods, which is of great significance to promoting the circulation of agricultural products, stabilizing prices and protecting people's livelihood. However, with the widespread implementation of the green channel policy, the phenomenon of counterfeit green channel vehicles occurs from time to time, which not only causes the loss of national fiscal revenue, but also affects the normal transportation order and the fairness of the policy. The traditional green channel vehicle inspection method has low accuracy and reliability.

[0003] Therefore, an accurate and reliable green pass management method is urgently needed. Summary of the invention

[0004] The purpose of the present disclosure is to provide a green pass management method and device, equipment, and storage medium based on big data to improve the accuracy and reliability of green pass management.

[0005] A first aspect of the disclosed embodiment provides a green pass management method based on big data, including: determining a green pass suspicion degree based on a transportation behavior of a target vehicle; Determine the target inspection standard based on the green pass suspicion level; The target inspection standard is sent to the target inspection device, a green pass inspection result is obtained based on the target inspection device, and whether the green pass condition is met is determined based on the green pass inspection result.

[0006] According to a second aspect of the disclosed embodiment, a green pass management device based on big data is provided, comprising: a suspicion level determination module, for determining the green pass suspicion level based on the transportation behavior of the target vehicle; An inspection standard determination module is used to determine the target inspection standard based on the green pass suspicion degree; The inspection module is used to send the target inspection standard to the target inspection device, obtain the green pass inspection result based on the target inspection device, and determine whether it meets the green pass condition based on the green pass inspection result.

[0007] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of the above-mentioned green pass management method based on big data are implemented.

[0008] According to a fourth aspect of the embodiments of the present disclosure, a computer-readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the steps of the above-mentioned green pass management method based on big data are implemented.

[0009] The green pass management method, device, equipment, and storage medium based on big data provided by the embodiments of the present disclosure have the following beneficial effects: The present disclosure determines the green pass suspiciousness based on the transportation behavior of the target vehicle, thereby more accurately identifying green pass vehicles that may have violations, and compared with the traditional manual inspection method, it improves the accuracy and efficiency of identifying fake green pass vehicles. The present disclosure determines the target inspection standard according to the green pass suspiciousness and sends it to the target inspection equipment, realizing the precision and intelligence of the inspection, which not only ensures the comprehensiveness of the inspection, but also avoids unnecessary delays, thereby improving the traffic efficiency of green pass vehicles and improving the accuracy and reliability of green pass traffic management. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1 A flowchart of a green pass management method based on big data provided by an embodiment of the present disclosure; Figure 2 A structural block diagram of a green pass management device based on big data provided by an embodiment of the present disclosure; Figure 3 A schematic block diagram of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0012] In the following description, specific details such as specific system structures and technologies are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present disclosure. However, it should be clear to those skilled in the art that the present disclosure may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obstructing the description of the present disclosure with unnecessary details.

[0013] In order to make the purpose, technical solutions and advantages of the present disclosure more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0014] Please refer to Figure 1 , Figure 1A flow chart of a green pass management method based on big data provided in an embodiment of the present disclosure, the method comprising: S101: Determine the green pass suspicion based on the transportation behavior of the target vehicle.

[0015] In this embodiment, the target vehicle refers to a specific transport vehicle that needs to undergo a green pass condition review during the green pass management process. Transport behavior refers to information covering multiple aspects of the vehicle during the transportation process, such as the route, traffic time period, historical green pass records, and whether there are any violations. Green pass suspicion refers to a value or level assessed based on the target vehicle's transport behavior, which is used to indicate the extent to which the vehicle may not comply with the green pass policy. By comparing and analyzing the various data of the vehicle's transport behavior with the normal green pass transport behavior pattern, it is concluded that the higher the value, the more suspicious it is, that is, the more likely it is that the green pass conditions will not be met.

[0016] The driving route in the transportation behavior can be obtained based on the Global Positioning System (GPS) or road monitoring cameras, the travel time period can be obtained based on the current time when the vehicle passes through the current toll station or green pass channel, and the historical green pass records can be obtained through the toll system database or the transportation management department platform.

[0017] The green pass suspicion degree can be obtained by calculating the suspicion degree of the driving route, the suspicion degree of the passing time period and the suspicion degree of the historical green pass record and performing weighted calculation. For example: In one embodiment of the present disclosure, the transportation behavior includes: driving route, passing time period and historical green pass record; Determine the green pass suspicion based on the target vehicle’s transportation behavior, including: determining a first suspicion level based on a driving route of the target vehicle; Determining a second suspiciousness based on a travel period of the target vehicle; Determine the third suspicion level based on the historical green pass record of the target vehicle; The green pass suspicion level is determined by performing a weighted calculation on the first suspicion level, the second suspicion level, and the third suspicion level.

[0018] In this embodiment, the first suspicious degree, the second suspicious degree and the third suspicious degree are quantitative evaluation indicators of the green pass suspicious degree of the target vehicle from different dimensions, considering that firstly, different green pass goods usually have relatively fixed and reasonable transportation routes, which are formed based on factors such as the origin, sales place and logistics cost of the goods. If the vehicle's driving route deviates too much from the standard route, it may mean that its transportation purpose or the nature of the goods is abnormal, such as avoiding certain checkpoints, or it is not a real green pass transportation at all, that is, a fake green pass vehicle; secondly, the planting of crops has relatively stable periodicity and regionality, and the planting scale and variety structure of the production area are relatively fixed within a certain period of time. Once the logistics transportation network and transportation habits are formed, they will remain stable for a certain period of time, or the night environment provides a certain cover for some violations. Some drivers may "switch" the transported goods at night, mix or replace the goods that do not meet the green pass standards with fresh agricultural products that originally meet the requirements to reduce the fees to be paid, etc. These violations are relatively more difficult to detect at night; secondly, the historical green pass records can reflect the vehicle's past integrity performance in green pass transportation. If a vehicle has a history of violations, such as being found to have faked green passes, falsely reported goods, etc., then the possibility of violating the rules again is relatively high. By considering the historical records, the credit status of the vehicle can be evaluated, and vehicles with bad credit can be inspected and supervised more strictly. Therefore, this embodiment considers the green pass suspicion by analyzing the three dimensions of the target vehicle's driving route, travel time, and historical green pass records.

[0019] The first suspicion is determined based on the target vehicle's route, reflecting the degree of deviation between the vehicle's actual driving trajectory and the normal green pass route. The greater the deviation, the higher the suspicion. It is used to determine whether there is an abnormality in the vehicle's driving path, so as to infer the authenticity of its green pass. The second suspicion is determined based on the target vehicle's travel time. By analyzing whether the vehicle's travel time conforms to the time pattern of normal green pass transportation, the suspicion of its green pass is measured, taking into account factors such as the rationality of different green pass cargo transportation in time. The third suspicion is determined based on the target vehicle's historical green pass record. It mainly examines the vehicle's past performance in the green pass transportation process, including whether there is a violation record, whether the green pass is frequently declared but the transported goods do not match the declaration, etc., and uses historical data to evaluate the suspicion of the current transportation behavior.

[0020] The first suspicion degree may be calculated by calculating the matching degree between the driving route of the target vehicle and the standard route. For example, in one embodiment of the present disclosure, the first suspicion degree is determined based on the driving route of the target vehicle, including: Determine a standard driving route based on the entry and exit sites of the target vehicle; Calculating a first matching degree between the driving route of the target vehicle and the standard driving route; A first suspicion level is determined based on the first matching level.

[0021] In this embodiment, the standard driving route refers to the normal driving route from the entry site to the exit site. According to the entry and exit site information of the target vehicle, combined with historical big data analysis or established industry standards, a reasonable driving route from the vehicle's departure site to the destination site under normal circumstances can be determined as a standard. For example, for green-pass vehicles transported from the vegetable and fruit production area of ​​place A to the wholesale market of place B, one or more conventional routes that most vehicles will choose are determined as standard driving routes based on the route data of a large number of similar transportations in the past. It should be noted that when there are multiple standard driving routes, the matching degree should be calculated separately, and the one with the highest matching degree should be used as the basis and premise for the subsequent judgment of suspicion. That is, in response to the fact that there are multiple standard driving routes, the matching degree of the target vehicle's driving route and each standard driving route is calculated separately, and the one with the highest matching degree is used as the first matching degree.

[0022] Both the standard driving route and the target vehicle driving route can be discretized into a series of path points. For example, the points passed by the vehicle can be recorded at a certain time interval or distance interval. Then, the matching degree is calculated by comparing the degree of overlap of the path points on the two routes. Specifically, the longest common subsequence algorithm in the dynamic programming algorithm can be used to find the longest common subsequence length of the two path point sequences, and the longest common subsequence length is divided by the number of path points of the standard driving route to obtain the matching degree. For example, the standard driving route is discretized into a path point sequence A=[A1,A2,A3,A4,A5,A6], and the target vehicle driving route is discretized into a path point sequence B=[A1,A3,A4,A6]. Using the LCS algorithm to calculate, the longest common subsequence length is 4, and the number of path points of the standard driving route is 6, then the first matching degree is 4÷6≈0.67. Or it can be calculated based on the section matching method, for example, the standard driving route is divided into several sections, for example, according to toll stations, intersections, etc. Then, it is determined whether the target vehicle's driving route passes through these sections, and the ratio of the number of sections passed through to the total number of sections of the standard driving route is counted as the matching degree. Specifically, the standard driving route is divided into 5 sections, namely R1, R2, R3, R4, and R5. The target vehicle's driving route passes through 3 sections of R1, R3, and R5, so the first matching degree is 3÷5=0.6.

[0023] In this embodiment, the first suspiciousness may be determined based on the first matching degree by linear function mapping, threshold-based mapping, or the like, for example, by threshold mapping in Table 1.

[0024] Table 1 First suspiciousness mapping table

[0025] It should be noted that the above table is only an example. For example, when the first matching degree is 0.95, its corresponding first suspicious degree may also be 10. In this case, the other corresponding first suspicious degrees should also be adjusted accordingly.

[0026] The second suspiciousness can also be obtained according to Table 2: Table 2 Second suspiciousness mapping table

[0027] The third suspiciousness can be obtained from Table 3: Table 3 The third suspiciousness mapping table

[0028] It should be noted that the judgment results of the above three tables should be considered consistent. For example, for a highly suspicious vehicle, the corresponding value of its suspicion should be 5, and the corresponding values ​​of the three tables should also be 5. If it is 50, the corresponding suspicion values ​​of the three tables should also be 50.

[0029] In this embodiment, the first suspicion degree, the second suspicion degree and the third suspicion degree obtained above can be weighted to obtain the final green pass suspicion degree. The corresponding weights can be set according to experience or practice.

[0030] S102: Determine the target inspection standard based on the green pass suspicion level.

[0031] In this embodiment, determining the target inspection standard based on the green pass suspicion level includes: in response to the green pass suspicion level being less than or equal to the target suspicion level threshold, using the conventional inspection standard as the target inspection standard; In response to the green pass suspicion being greater than the target suspicion threshold, the comprehensive inspection standard is used as the target inspection standard, and the inspection accuracy of the conventional inspection standard is lower than the inspection accuracy of the comprehensive inspection.

[0032] In this embodiment, the target suspicion threshold can be determined based on experience, and the routine inspection standard is to conduct a certain proportion of random inspections on the goods. For example, for a van, you can open the door of the compartment and inspect the surface layer of the goods to check whether the actual state of the goods is consistent with the declaration, and check whether the quantity of the goods is consistent with the transport documents. In addition to completing all items of the routine inspection, the comprehensive inspection standard requires a comprehensive inspection of the goods to ensure that the type, quality, quantity, etc. of the goods meet the requirements of the Green Pass policy. Check all parts of the vehicle to see if there are any hidden spaces or other signs of illegal modifications. At the same time, make detailed inquiries to the driver and accompanying personnel to verify the authenticity of information such as the transportation itinerary, source of goods, and destination. If necessary, the shipper or consignee can be contacted for information verification.

[0033] S103: Send the target inspection standard to the target inspection device, obtain the green pass inspection result based on the target inspection device, and determine whether it meets the green pass condition based on the green pass inspection result.

[0034] In this embodiment, the target inspection device may refer to various devices and systems that actually perform inspection operations at the green pass inspection point, for example, it may be the mobile phone of the relevant inspector, the central control platform, the inspection recorder, etc. After the inspection is completed, the relevant inspector can upload the green pass inspection results to the relevant equipment, and the green pass inspection results received by the relevant equipment will determine whether the target vehicle meets the green pass conditions. Whether the green pass conditions are met should be formulated and judged in accordance with the relevant local laws and regulations. The green pass inspection results may include: basic vehicle information, certificate information, cargo type, cargo quantity or cargo load, etc.

[0035] From the above, it can be concluded that the present disclosure determines the green pass suspicion based on the transportation behavior of the target vehicle, thereby more accurately identifying green pass vehicles that may have violations, and compared with the traditional manual inspection method, it improves the accuracy and efficiency of identifying fake green pass vehicles. The present disclosure determines the target inspection standard according to the green pass suspicion and sends it to the target inspection equipment, realizing the precision and intelligence of the inspection, which not only ensures the comprehensiveness of the inspection, but also avoids unnecessary delays, thereby improving the traffic efficiency of green pass vehicles and improving the accuracy and reliability of green pass traffic management.

[0036] In one embodiment of the present disclosure, the first suspicion level, the second suspicion level, and the third suspicion level may be weightedly calculated according to a first formula to determine the green pass suspicion level. The first formula may be:

[0037] in Indicates the suspiciousness of the green pass. Indicates the first degree of suspicion, Indicates the second level of suspicion, Indicates the third degree of suspicion, is the weight corresponding to the first suspicious degree, is the weight corresponding to the second suspicious degree, is the weight corresponding to the third suspicious degree, represents the first suspiciousness correction factor, Indicates the second suspicion correction factor.

[0038] The first suspiciousness and the second suspiciousness can be obtained according to the corresponding mapping table above. Considering that different violation types have different impacts on the implementation and fairness of the green pass policy, and the closer the violation time is to the current situation, the more likely the vehicle's recent behavior is related to the current situation, and the greater the impact on the current green pass suspiciousness. Therefore, the third suspiciousness can not only be obtained from Table 3, but also the impact of the violation type and violation time can be considered.

[0039] For example, , where the number of violations is , violation types are divided into minor, moderate and severe, They represent the basic suspiciousness corresponding to the violation type, which can be determined based on experiments. It is a time decay factor, which is used to reflect that the longer the violation time is, the smaller the impact on the current suspicion. It can be determined based on experiments. Indicates the number of years from the violation time to the current time, which can be determined based on the current time and the violation time. The violation type of light, moderate or severe can be determined based on the violation types in the historical green pass records, which can be obtained by presetting the violation types corresponding to each violation type.

[0040] The correction factor is a parameter used to fine-tune each suspicion level. For example, in response to the target vehicle driving route being a construction route and / or a traffic accident occurring, the first suspicion level correction factor is reduced based on the first correction step length. In response to the green pass vehicle's travel period being a holiday, a second suspicion correction factor is increased based on the second correction step length.

[0041] In this embodiment, considering that the road section where the target vehicle is traveling has recently been subjected to abnormal traffic conditions due to construction, accidents, etc., the vehicle's driving behavior may be different from normal. It can be appropriately lowered, and during holidays, the traffic policy or traffic habits of green pass vehicles will change. For example, the traffic volume of green pass vehicles increases significantly on certain holidays, and the traffic time periods may be more dispersed, so the consideration of the second suspiciousness is appropriately increased. The first correction step size and the second correction step size can be determined based on experiments or experience.

[0042] From the above, it can be concluded that this embodiment performs weighted calculation of the first suspicion, the second suspicion and the third suspicion through the first formula, and can comprehensively consider information from multiple dimensions to determine the green pass suspicion. This method is more comprehensive and accurate than a single-dimensional evaluation, and can better reflect the actual situation of the target vehicle. This embodiment improves the flexibility and adaptability of the present disclosure by introducing correction factors. According to different scenarios and needs, the values ​​of weights and correction factors can be adjusted to adapt to different evaluation criteria. In this embodiment, the calculation of the third suspicion not only considers the number of violations, but also the influence of violation types and time. This processing method is more reasonable and improves the accuracy and reliability of green pass management.

[0043] In one embodiment of the present disclosure, the green pass management method based on big data further includes: In response to the target vehicle's travel route including a plurality of toll stations, adjusting the green pass suspicion degree based on the historical green pass suspicion degrees of the plurality of toll stations; The historical green pass suspicion degree is the green pass suspicion degree determined by the toll stations in the target vehicle's travel route.

[0044] In this embodiment, considering that in the green pass management, there may be differences in the situations of green pass vehicles handled by different toll stations in the past, and that a counterfeit vehicle may not be suspicious in a certain section of the journey, but may be suspicious in another section of the journey, the suspicion of this section of the journey should also be appropriately adjusted. That is, the behavior of counterfeiting green pass vehicles may be staged or localized. A vehicle may behave normally in a certain section of the route, but when passing through a specific section of the road and related toll stations, it may show higher suspicion for various reasons (such as counterfeiting methods are easier to implement in this area, etc.). Therefore, the green pass suspicion determined only based on the vehicle's route, travel time and historical green pass records of this section may not be comprehensive and accurate enough, and it needs to be adjusted in combination with the historical green pass suspicion of each toll station in the route, so as to more accurately evaluate the vehicle's green pass suspicion level during the entire driving process.

[0045] For example, the green pass suspicion level is adjusted based on the historical green pass suspicion levels of multiple toll booths, including: In response to the historical green pass suspicion levels of multiple toll stations being less than or equal to the target suspicion level threshold, no adjustment is made; In response to at least one of the historical green pass suspicions of the plurality of toll booths being greater than a target suspicion threshold, the green pass suspicion is adjusted based on the historical green pass suspicions greater than the target suspicion threshold.

[0046] In this embodiment, the target suspicion threshold is consistent with the above. If there is at least one toll station whose historical green pass suspicion is greater than the target suspicion threshold, all historical green passes greater than the target suspicion threshold and the current green pass suspicion can be weighted calculated, that is, the green pass suspicion is adjusted to obtain the green pass suspicion. The weight distribution can be set based on experience.

[0047] From the above, it can be concluded that the present disclosure adjusts the green pass suspicion of the target vehicle by introducing the historical green pass suspicion of multiple toll stations, thereby achieving a comprehensive evaluation of the entire vehicle driving process. It not only takes into account the driving route, travel time and historical green pass records of the vehicle in this section, but also combines the historical data of each toll station in the driving route, so as to more comprehensively reflect the real situation of the vehicle and improve the accuracy of green pass management. This embodiment takes into account that the behavior of counterfeit green pass vehicles may be staged or localized. This embodiment combines the historical green pass suspicion of multiple toll stations to more accurately identify the suspicious behavior of vehicles in different sections. This helps to reduce the misjudgment rate caused by insufficient or misleading data on a single section and improve management efficiency.

[0048] In one embodiment of the present disclosure, determining the green pass suspicion based on the transportation behavior of the target vehicle includes: Determine the hyperparameters of the random forest algorithm model based on the standard green transport behavior dataset; The transportation behavior of the target vehicle is input into the random forest algorithm model to obtain the green pass suspicion degree.

[0049] In this embodiment, the green pass suspicion can also be calculated by a random forest algorithm model. Before using the random forest algorithm model, its parameters must first be set and adjusted. For example, in one embodiment of the present disclosure, the hyperparameters of the random forest algorithm model include: the maximum number of features, the minimum number of sample splits, and the minimum number of sample leaf nodes; The hyper parameters of the random forest algorithm model are determined based on the standard green transport behavior dataset, including: In response to the number of relevant feature clusters in the green pass transportation behavior data set being greater than the first relevant number, determining a compensation feature number based on the number of relevant feature clusters, and reducing a reference value of a maximum feature number based on the compensation feature number to obtain a maximum feature number; In response to the data distribution in the green transport behavior data set satisfying the first distribution condition, increasing the reference value of the minimum sample splitting number based on the first sample splitting number to obtain the minimum sample splitting number; In response to the data fluctuation value in the green pass transportation behavior data set being greater than the first fluctuation threshold, the reference value of the minimum sample leaf node number is increased based on the first leaf node number to obtain the minimum sample leaf node number.

[0050] In this embodiment, the hyperparameters of the random forest algorithm model also include: the number of decision trees; Determining the hyperparameters of the random forest algorithm model based on the standard green transport behavior data set also includes: in response to the number of data in the green transport behavior data set being greater than the preset number of data, increasing the reference value of the number of decision trees based on the first number of decision trees to obtain the number of decision trees.

[0051] In this embodiment, the reference value of the number of decision trees, the reference value of the maximum number of features, the reference value of the minimum number of sample splits, and the reference value of the minimum number of sample leaf nodes can be set according to the initial reference parameter settings of the random forest model, or the parameters for solving similar problems in the field.

[0052] Considering that a large amount of data requires enough decision trees to fully learn the characteristics and rules therein to avoid underfitting of the model, when the number of data in the green transport behavior data set is greater than the preset number of data, the reference value of the number of decision trees can be increased according to the first number of decision trees. The first number of decision trees can be determined based on experience or increased as the number of data increases.

[0053] Considering that if there are multiple features with high correlation in the data, when selecting the maximum number of features, it is necessary to avoid selecting too many related features at the same time. The features with high correlation can be screened by calculating the correlation coefficient between the features, and the most representative features can be selected to be included in the consideration range of the maximum number of features. For example, the features in the green pass vehicle data may be correlated, such as the vehicle cargo weight and cargo volume have a certain correlation, because usually, the weight of large cargo is also relatively large. When there is a high correlation between the features, the model may be overly dependent on these related features, resulting in a decrease in the generalization ability of the model and poor recognition of new data in practical applications. By limiting the maximum number of features, the model cannot consider all relevant features at the same time when constructing a decision tree, thereby reducing the negative impact of feature collinearity and improving the stability and generalization ability of the model. If the maximum number of features is not limited, the model may use all relevant features for complex fitting, thereby remembering the noise and details in the training data, resulting in overfitting. That is, in response to the number of related feature clusters in the green pass transportation behavior data set being greater than the first related number, the compensating feature number is determined based on the number of related feature clusters, and the reference value of the maximum feature number is reduced based on the compensating feature number to obtain the maximum feature number, and the first related number can be preset.

[0054] Therefore, the number of related feature clusters in the Green Pass transportation behavior data set can be calculated to determine whether the reference value of the maximum feature tree needs to be reduced and to what extent. The correlation matrix between the various features in the Green Pass vehicle data can be calculated, and the linear or nonlinear correlation between the features can be measured by statistical methods such as the Pearson correlation coefficient and the Spearman correlation coefficient. For example, the Pearson correlation coefficient can measure the degree of linear correlation between two features, and the value range is between [-1,1]. The closer the absolute value is to 1, the stronger the correlation. According to the correlation matrix, find the feature clusters with strong correlation. For example, if the correlation coefficients between the three features are found to be above 0.8, it means that the three features are highly correlated, and they can be regarded as a related feature cluster. For each related feature cluster, only one or several of the most representative features are selected to be included in the consideration range of the maximum number of features. For example, if there are 10 related feature clusters, 1 feature is selected for each cluster, plus some important features with weak correlation with other features, and a total of 30 features are selected as the maximum number of features. This ensures that the model can take into account important information from different aspects, avoids the reuse of related features, and improves the performance and efficiency of the model.

[0055] The more related clusters there are, the more interrelated features there are in the data. The number of compensating features is determined based on the number of related feature clusters in order to measure the degree of redundant features in the data.

[0056] The compensation characteristic number can be determined according to the second formula: ,in represents the number of compensation characteristics, is a tuning parameter that can be determined during the experiment based on data characteristics and model performance. represents the number of related feature clusters, and the total number of features is , Reflects the direct impact of the number of related feature clusters on the number of compensating features. It reflects the influence of the proportional relationship between the number of relevant feature clusters and the total number of features on the number of compensating features.

[0057] Taking into account the imbalanced distribution of categories, that is, if the ratio of normal green pass vehicle data to fake green pass vehicle data is seriously unbalanced, for example, normal green pass vehicle data accounts for 90% and fake green pass vehicle data accounts for 10%, the minimum sample split number and the minimum sample leaf node number can be adjusted based on data balancing techniques such as oversampling or undersampling. The minimum sample split number can be appropriately increased to prevent the model from over-splitting on a large number of majority class samples and ignoring the characteristics of minority class samples; the minimum sample leaf node number is also increased accordingly to ensure that minority class samples are sufficiently representative in the leaf nodes and avoid the model's poor recognition effect on minority classes. Therefore, when the data distribution in the green pass transportation behavior data set meets the first distribution condition, the reference value of the minimum sample split number can be increased based on the first sample split number. The first distribution condition can be that the proportion of normal green pass vehicle data is greater than or equal to the first proportion, and the first proportion and the first sample split number can be set based on experience.

[0058] Secondly, it is also considered that when the data distribution is extremely unreasonable, that is, the data distribution in the Green Pass transportation behavior dataset meets the second distribution condition, the second distribution condition can be that the proportion of normal Green Pass vehicle data is greater than or equal to the second proportion, and the second proportion is greater than the first proportion. At this time, it means that the number of fake vehicle data is too small. Even if the reference value of the minimum sample splitting tree is increased, the Green Pass suspicion cannot be accurately and reliably output. At this time, the number of fake vehicle data can be expanded based on the generative adversarial network.

[0059] For example, in one embodiment of the present disclosure, before determining the hyperparameters of the random forest algorithm model based on the standard green pass transportation behavior dataset, the method further includes: In response to the data distribution in the green pass transportation behavior data set satisfying the second distribution condition, generating vehicle transportation behavior data that does not meet the green pass condition based on the generative adversarial network; The first distribution condition and the second distribution condition have different judgment thresholds.

[0060] In this embodiment, it is also considered that the goal of the generator is to generate vehicle transportation behavior data that does not meet the green pass conditions as realistically as possible, so that it can deceive the discriminator. It needs to learn how to generate data similar to the real data distribution from the latent space. This process is relatively more exploratory and creative, and usually requires a slower learning rate to avoid the generated samples from deviating too much from the real data distribution, resulting in unstable generation results or problems such as mode collapse. The task of the discriminator is to accurately distinguish between real data and fake data generated by the generator. It needs to quickly learn the difference between real data and generated data in order to provide effective feedback to the generator and prompt the generator to continuously improve. Therefore, the discriminator can generally use a relatively high learning rate, so that it can update parameters faster, respond more keenly to the output of the generator, and better guide the training of the generator.

[0061] Therefore, when setting the learning rates of the generator and the discriminator, the upper limit of the learning rate of the generator should be set to the first proportion of the learning rate of the discriminator, that is, the generative adversarial network includes a generator and a discriminator, and the learning rate of the generator is less than or equal to the learning rate of the discriminator. The first proportion can be preset.

[0062] It is also considered that for complex green transport behavior data, the learning rate of the generator needs to be limited to a lower level, that is, the first ratio needs to be reduced, because for complex green transport behavior data, the generator needs to learn the data distribution more slowly and stably, while allowing the discriminator to quickly feedback and guide the generated data.

[0063] That is, in response to the data complexity in the green pass transportation behavior data set being greater than or equal to the first complexity, the first ratio is reduced based on the first complexity ratio.

[0064] In this embodiment, the data complexity can be obtained by calculating the information entropy, and the first complexity and the first complexity ratio can be determined according to actual conditions.

[0065] Finally, if the collected data fluctuates greatly, such as the vehicle load is unstable due to factors such as season and market demand, the minimum number of sample leaf nodes should be appropriately increased. This can ensure that each leaf node contains enough samples, improve the generalization ability of the model, and reduce sensitivity to local data fluctuations.

[0066] Therefore, when the data fluctuation value in the green transport behavior data set is greater than the first fluctuation threshold, the reference value of the minimum number of sample leaf nodes can be increased based on the first number of leaf nodes.

[0067] In this embodiment, the standard deviation or coefficient of variation of the data in the green transport behavior data set can be calculated as the fluctuation value of the data. The standard deviation reflects the degree of dispersion of the data relative to the mean. The larger the standard deviation or coefficient of variation, the greater the fluctuation of the data. The first fluctuation threshold and the first leaf node number can be determined based on experience.

[0068] In one embodiment of the present disclosure, determining the hyperparameters of the random forest algorithm model based on the standard green pass transportation behavior dataset also includes: In response to the number of relevant feature clusters in the green pass transportation behavior data set being less than or equal to the first relevant number, taking the reference value of the maximum feature number as the maximum feature number; In response to the data distribution in the green transport behavior data set not satisfying the first distribution condition, the reference value of the minimum sample split number is used as the minimum sample split number In response to the data fluctuation value in the green transport behavior data set being less than or equal to the first fluctuation threshold, taking the reference value of the minimum number of sample leaf nodes as the minimum number of sample leaf nodes; In response to the number of data in the green pass transportation behavior data set being less than or equal to the preset number of data, a reference value of the number of decision trees is used as the number of decision trees.

[0069] In this embodiment, if the above conditions are not triggered or met, the corresponding reference values ​​can be used as hyperparameters to train the model.

[0070] It can be concluded from the above that this embodiment optimizes the model by determining the hyperparameters of the random forest algorithm model based on the standard green pass transportation behavior data set. This embodiment adjusts the hyperparameters such as the maximum number of features, the minimum number of sample splits, the minimum number of sample leaf nodes and the number of decision trees, so that the model can more accurately capture the key features in the green pass transportation behavior, thereby improving the accuracy of the green pass suspicion. Considering the characteristics of correlation features, imbalanced category distribution and data fluctuations that may exist in the data, this embodiment proposes a corresponding adjustment strategy. For features with high correlation, the maximum number of features is limited to avoid the model from over-relying on these features, thereby improving the stability and generalization ability of the model. For the case of uneven category distribution, the minimum number of sample splits and the minimum number of sample leaf nodes are adjusted to prevent the model from over-splitting on a large number of majority class samples, while ensuring that the minority class samples are sufficiently representative in the leaf nodes. For cases with large data fluctuations, the minimum number of sample leaf nodes is increased to reduce the sensitivity of the model to local data fluctuations and improve the robustness of the model. When the data distribution is extremely unreasonable, that is, the number of fake vehicle data is too small, this embodiment proposes a solution for data expansion based on a generative adversarial network.

[0071] This embodiment uses a generator to generate vehicle transportation behavior data that does not meet the green pass conditions, which can effectively alleviate the data imbalance problem, provide the model with richer training samples, and further improve the recognition ability of the model. At the same time, the setting problem of the learning rate of the generator and the discriminator is considered. By limiting the learning rate of the generator to be less than or equal to the learning rate of the discriminator, and dynamically adjusting this ratio according to the complexity of the data, it can ensure that the generator learns the data distribution more stably, and at the same time allow the discriminator to quickly feedback and guide the generated data, thereby speeding up the training process and improving the quality of the generated data, and enhancing the accuracy and reliability of green pass management.

[0072] Corresponding to the green pass management method based on big data in the above embodiment, Figure 2 This is a structural block diagram of a green pass management device based on big data provided by an embodiment of the present disclosure. For the sake of convenience, only the parts related to the embodiment of the present disclosure are shown. Figure 2The green pass management device 20 based on big data includes: a suspicion determination module 21, an inspection standard determination module 22 and an inspection module 23.

[0073] The suspiciousness determination module 21 is used to determine the green pass suspiciousness based on the transportation behavior of the target vehicle; An inspection standard determination module 22, used to determine a target inspection standard based on the green pass suspicion level; The inspection module 23 is used to send the target inspection standard to the target inspection device, obtain the green pass inspection result based on the target inspection device, and determine whether it meets the green pass condition based on the green pass inspection result.

[0074] In one embodiment of the present disclosure, the transportation behavior includes: driving route, traffic time period and historical green pass records; A suspicion level determination module 21, specifically configured to determine a first suspicion level based on a driving route of a target vehicle; Determining a second suspiciousness based on a travel period of the target vehicle; Determine the third suspicion level based on the historical green pass record of the target vehicle; The green pass suspicion level is determined by performing a weighted calculation on the first suspicion level, the second suspicion level, and the third suspicion level.

[0075] In one embodiment of the present disclosure, the suspiciousness determination module 21 is further configured to determine a standard driving route based on the entry and exit sites of the target vehicle; Calculating a first matching degree between the driving route of the target vehicle and the standard driving route; A first suspicion level is determined based on the first matching level.

[0076] In one embodiment of the present disclosure, the green pass management device 20 based on big data further includes: a suspiciousness adjustment module; The suspicion adjustment module is used for adjusting the green pass suspicion based on the historical green pass suspicion of the multiple toll stations in response to the target vehicle's driving route containing multiple toll stations; The historical green pass suspicion degree is the green pass suspicion degree determined by the toll stations in the target vehicle's travel route.

[0077] In one embodiment of the present disclosure, the suspicion determination module 21 is further specifically used to determine the hyperparameters of the random forest algorithm model based on the standard green pass transportation behavior data set; The transportation behavior of the target vehicle is input into the random forest algorithm model to obtain the green pass suspicion degree.

[0078] In one embodiment of the present disclosure, the hyperparameters of the random forest algorithm model include: the maximum number of features, the minimum number of sample splits, and the minimum number of sample leaf nodes; The suspiciousness determination module 21 is further configured to determine a compensation feature number based on the number of relevant feature clusters in the green pass transportation behavior data set in response to the number of relevant feature clusters being greater than the first relevant number, and to reduce a reference value of the maximum feature number based on the compensation feature number to obtain the maximum feature number; In response to the data distribution in the green transport behavior data set satisfying the first distribution condition, increasing the reference value of the minimum sample splitting number based on the first sample splitting number to obtain the minimum sample splitting number; In response to the data fluctuation value in the green pass transportation behavior data set being greater than the first fluctuation threshold, the reference value of the minimum sample leaf node number is increased based on the first leaf node number to obtain the minimum sample leaf node number.

[0079] In one embodiment of the present disclosure, the green pass management device 20 based on big data further includes: a data expansion module; A data expansion module, configured to generate vehicle transportation behavior data that does not meet the green pass condition based on a generative adversarial network in response to the data distribution in the green pass transportation behavior data set satisfying a second distribution condition; The first distribution condition and the second distribution condition have different judgment thresholds.

[0080] See also Figure 3 , Figure 3 A schematic block diagram of an electronic device provided by an embodiment of the present disclosure. Figure 3 The electronic device 300 in the embodiment shown may include: one or more processors 301, one or more input devices 302, one or more output devices 303 and one or more memories 304. The processors 301, input devices 302, output devices 303 and memories 304 communicate with each other via a communication bus 305. The memory 304 is used to store computer programs, which include program instructions. The processor 301 is used to execute the program instructions stored in the memory 304. The processor 301 is configured to call the program instructions to execute the functions of each module / unit in the above-mentioned device embodiments, such as Figure 2 The functions of modules 21 to 23 are shown.

[0081] It should be understood that in the embodiment of the present disclosure, the processor 301 may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0082] The input device 302 may include a touch panel, a fingerprint collection sensor (for collecting the user's fingerprint information and fingerprint direction information), a microphone, etc., and the output device 303 may include a display (LCD, etc.), a speaker, etc.

[0083] The memory 304 may include a read-only memory and a random access memory, and provide instructions and data to the processor 301. A portion of the memory 304 may also include a non-volatile random access memory. For example, the memory 304 may also store information about the device type.

[0084] In a specific implementation, the processor 301, input device 302, and output device 303 described in the embodiments of the present disclosure can execute the implementation methods described in the first and second embodiments of the green pass management method based on big data provided in the embodiments of the present disclosure, and can also execute the implementation methods of the electronic device described in the embodiments of the present disclosure, which will not be repeated here.

[0085] In another embodiment of the present disclosure, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by the processor, all or part of the processes in the above-mentioned embodiment method are implemented, and the computer program can also be completed by instructing the relevant hardware through the computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by the processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0086] The computer-readable storage medium may be an internal storage unit of the electronic device of any of the aforementioned embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Furthermore, the computer-readable storage medium may also include both an internal storage unit of the electronic device and an external storage device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium may also be used to temporarily store data that has been output or is to be output.

[0087] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this disclosure.

[0088] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the electronic devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0089] In the several embodiments provided in the present application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces or units, or it can be an electrical, mechanical or other form of connection.

[0090] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present disclosure.

[0091] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0092] The above are only specific embodiments of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present disclosure, and these modifications or replacements should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

1. A green pass management method based on big data, characterized in that: include: Determine the green pass suspicion based on the target vehicle’s transportation behavior; Determining target inspection standards based on the green pass suspicion level; Sending the target inspection standard to the target inspection device, obtaining a green pass inspection result based on the target inspection device, and determining whether it meets the green pass condition based on the green pass inspection result; The transportation behavior includes: driving route, travel time and historical green pass records; The green pass suspicion level is determined based on the transportation behavior of the target vehicle, including: determining a first suspicion level based on a driving route of the target vehicle; Determining a second suspiciousness based on a travel period of the target vehicle; Determine the third suspicion level based on the historical green pass record of the target vehicle; A green pass suspicion degree is determined by performing a weighted calculation on the first suspicion degree, the second suspicion degree, and the third suspicion degree.

2. The green pass management method based on big data as claimed in claim 1, characterized in that: The determining of the first suspiciousness based on the driving route of the target vehicle includes: Determining a standard driving route based on the entry and exit sites of the target vehicle; Calculating a first matching degree between the driving route of the target vehicle and the standard driving route; A first suspicious degree is determined based on the first matching degree.

3. The green pass management method based on big data as claimed in claim 2, characterized in that: Also includes: In response to the target vehicle's travel route including a plurality of toll stations, adjusting the green pass suspicion level based on historical green pass suspicion levels of the plurality of toll stations; The historical green pass suspicion level is the green pass suspicion level determined by the toll stations in the driving route of the target vehicle.

4. The green pass management method based on big data as claimed in claim 1, characterized in that: The green pass suspicion level is determined based on the transportation behavior of the target vehicle, including: Determine the hyperparameters of the random forest algorithm model based on the standard green transport behavior dataset; The transportation behavior of the target vehicle is input into the random forest algorithm model to obtain the green pass suspicion degree.

5. The green pass management method based on big data as claimed in claim 4, characterized in that: The hyper parameters of the random forest algorithm model include: maximum number of features, minimum number of sample splits, and minimum number of sample leaf nodes; The hyper parameters of the random forest algorithm model are determined based on the standard green transport behavior data set, including: In response to the number of relevant feature clusters in the green pass transportation behavior data set being greater than the first relevant number, determining a compensation feature number based on the number of relevant feature clusters, and reducing a reference value of a maximum feature number based on the compensation feature number to obtain the maximum feature number; In response to the data distribution in the green transport behavior data set satisfying the first distribution condition, increasing the reference value of the minimum sample splitting number based on the first sample splitting number to obtain the minimum sample splitting number; In response to the data fluctuation value in the green pass transportation behavior data set being greater than a first fluctuation threshold, the reference value of the minimum sample leaf node number is increased based on the first leaf node number to obtain the minimum sample leaf node number.

6. The green pass management method based on big data as claimed in claim 5, characterized in that: Before determining the hyper parameters of the random forest algorithm model based on the standard green-pass transportation behavior data set, the method further includes: In response to the data distribution in the green pass transportation behavior data set satisfying a second distribution condition, generating vehicle transportation behavior data that does not meet the green pass condition based on a generative adversarial network; The first distribution condition and the second distribution condition have different judgment thresholds.

7. A green pass management device based on big data, characterized in that: include: A suspicion level determination module, used to determine the green pass suspicion level based on the transportation behavior of the target vehicle; An inspection standard determination module, used to determine a target inspection standard based on the green pass suspicion level; An inspection module, used for sending the target inspection standard to a target inspection device, obtaining a green pass inspection result based on the target inspection device, and determining whether the green pass condition is met based on the green pass inspection result; The transportation behavior includes: driving route, travel time and historical green pass records; A suspicion level determination module, specifically configured to determine a first suspicion level based on a driving route of a target vehicle; Determining a second suspiciousness based on a travel period of the target vehicle; Determine the third suspicion level based on the historical green pass record of the target vehicle; A green pass suspicion degree is determined by performing a weighted calculation on the first suspicion degree, the second suspicion degree, and the third suspicion degree.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images