CN117152973B

CN117152973B - Expressway real-time flow monitoring method and system based on ETC portal data

Info

Publication number: CN117152973B
Application number: CN202311406723.7A
Authority: CN
Inventors: 阮雪飞; 李大全; 朱秋实; 王志海; 杨唐; 吴政沅
Original assignee: Guizhou Hongxinda High New Technology Co ltd
Current assignee: Guizhou Hongxinda High New Technology Co ltd
Priority date: 2023-10-27
Filing date: 2023-10-27
Publication date: 2024-01-05
Anticipated expiration: 2043-10-27
Also published as: CN117152973A

Abstract

The application provides a real-time highway flow monitoring method and system based on ETC portal data, which are characterized in that a target monitoring section set is obtained, traffic congestion evaluation feature clusters corresponding to all target monitoring sections in the target monitoring section set are obtained, the scattered positions of traffic congestion evaluation features in the target feature clusters corresponding to congestion evaluation layers are analyzed based on monitoring section monitoring algorithms, and algorithm evaluation information corresponding to the target monitoring sections is determined according to the scattered position information of each monitoring section monitoring algorithm. According to the method, the scattered positions of the traffic congestion evaluation features are analyzed based on the congestion evaluation layers, the algorithm evaluation information corresponding to the monitoring section is obtained by combining the scattered position information of all the congestion evaluation layers, so that the congestion decision is carried out on the monitoring section by integrating the traffic congestion evaluation features of the monitoring section on each congestion evaluation layer, and the algorithm evaluation information output by the algorithm has higher reliability and accuracy.

Description

Expressway real-time flow monitoring method and system based on ETC portal data

Technical Field

The present disclosure relates to the field of data processing, and more particularly, to a real-time highway traffic monitoring method and system based on ETC portal data.

Background

ETC portal, i.e., electronic toll collection portal (Electronic Toll Collection), is a modern highway toll collection system. The wireless communication and vehicle-mounted equipment technology are adopted, and convenience that the vehicle does not need to park and pay when passing through a toll station is realized. Along with the integrated construction and promotion of expressway information, the expressway information is endowed with more comprehensive information collection functions, for example, vehicle information such as license plate numbers and vehicle types can be collected; traffic events and places, such as transit time stamps; traffic flow information such as lane identification, number of flows, etc. Therefore, the real-time traffic monitoring of the expressway is carried out according to the data collected by the ETC portal, the help can be provided for the early warning of the congestion of the expressway, and the accuracy and the efficiency of the monitoring can be understood, so that the method is a technical problem concerned by the traffic monitoring of the expressway.

Disclosure of Invention

In view of this, the embodiments of the present application at least provide a real-time highway traffic monitoring method and system based on ETC portal data.

According to an aspect of the disclosed embodiments, there is provided a real-time highway traffic monitoring method based on ETC portal data, the method including:

In response to the flow monitoring instruction, obtaining a flow monitoring dataset comprising traffic flow data for at least one target monitoring zone obtained based on the ETC portal;

acquiring traffic congestion evaluation feature clusters corresponding to each target monitoring section respectively, wherein the traffic congestion evaluation feature clusters comprise traffic congestion evaluation features corresponding to a plurality of congestion evaluation levels respectively;

determining a monitoring section monitoring algorithm array, wherein the monitoring section monitoring algorithm array comprises a plurality of monitoring section monitoring algorithms; the monitoring section monitoring algorithm in the monitoring section monitoring algorithm array is different in monitoring section evaluation modes according to the monitoring section monitoring algorithm;

acquiring scattered position information of each traffic congestion evaluation feature in the corresponding traffic congestion evaluation feature cluster of the target monitoring section in the corresponding target feature cluster of the corresponding congestion evaluation layer through each monitoring section monitoring algorithm;

determining algorithm evaluation information of the monitoring section monitoring algorithm on the target monitoring section based on the scattered position information obtained by the monitoring section monitoring algorithm;

and summing up algorithm evaluation information of the target monitoring section by each monitoring section monitoring algorithm in the monitoring section monitoring algorithm array to obtain traffic flow monitoring information of the target monitoring section.

According to an example of an embodiment of the present disclosure, the obtaining, by each monitoring section monitoring algorithm, the scattered location information of each traffic congestion evaluation feature in the traffic congestion evaluation feature cluster corresponding to the corresponding congestion evaluation level in the target feature cluster corresponding to the target monitoring section includes:

acquiring traffic congestion evaluation characteristics corresponding to congestion evaluation levels from traffic congestion evaluation characteristic clusters corresponding to the target monitoring sections to obtain target characteristic clusters corresponding to the congestion evaluation levels respectively;

acquiring a corresponding scattered position determination strategy of the target feature cluster in a monitoring section monitoring algorithm;

and carrying out scattered position determination on the traffic congestion evaluation characteristics in the target characteristic clusters of the corresponding congestion evaluation levels based on the scattered position determination strategy to obtain scattered position information of the traffic congestion evaluation characteristics in the target characteristic clusters of the corresponding congestion evaluation levels.

According to an example of the embodiment of the present disclosure, the determining a dispersion position of the target feature cluster includes determining a dispersion position of the target feature cluster according to a threshold, where the determining a dispersion position of the target feature cluster in the monitoring algorithm includes determining a dispersion position of the target feature cluster according to the threshold; performing dispersion position determination on the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels based on the dispersion position determination strategy to obtain dispersion position information of the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels, wherein the method comprises the following steps:

Acquiring a characteristic dispersion topological graph, wherein the characteristic dispersion topological graph comprises a plurality of component topological points;

determining an initial topological point of the feature dispersion topological graph as a current composition topological point corresponding to the target monitoring section, acquiring a current congestion evaluation level corresponding to the current composition topological point, and acquiring a current feature dispersion determination threshold value of a current target feature cluster corresponding to the current congestion evaluation level;

determining the dispersion position information of the traffic congestion evaluation feature in a current target feature cluster through a current feature dispersion determination threshold and the traffic congestion evaluation feature of the target monitoring section at the current congestion evaluation level;

and determining a next composition topological point corresponding to the target monitoring section based on the dispersion position information, determining the next composition topological point as an optimized current composition topological point, returning to a current congestion evaluation level corresponding to the current composition topological point, and acquiring a current characteristic dispersion determination threshold value of a current target characteristic cluster corresponding to the current congestion evaluation level until the composition topological point corresponding to the target monitoring section is optimized.

According to an example of an embodiment of the present disclosure, the determining algorithm evaluation information of the monitoring section monitoring algorithm on the target monitoring section based on the scattered location information obtained by the monitoring section monitoring algorithm includes:

Determining corresponding composition topological points of the target monitoring section based on the scattered position information;

counting corresponding composition topological points of the target monitoring section to obtain a dispersion scale of the target monitoring section in the characteristic dispersion topological graph;

determining a first congestion monitoring index corresponding to the target monitoring section based on the dispersion scale, wherein the congestion monitoring index is inversely related to the dispersion scale;

and determining algorithm evaluation information of the monitoring section monitoring algorithm to the target monitoring section based on the first congestion monitoring index.

According to an example of the embodiment of the present disclosure, the scattered location determining policy includes a manner of performing scattered location determination based on a scattered location interval, and the acquiring monitoring section monitors the scattered location determining policy corresponding to the target feature cluster in the algorithm; performing dispersion position determination on the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels based on the dispersion position determination strategy to obtain dispersion position information of the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels, wherein the method comprises the following steps:

acquiring a characteristic dispersion position determination interval cluster corresponding to the target characteristic cluster in a monitoring section monitoring algorithm, wherein the characteristic dispersion position determination interval cluster comprises a plurality of characteristic dispersion position determination intervals;

Acquiring the feature dispersion number of the traffic jam evaluation features in the target feature cluster in each feature dispersion position determination interval;

and determining a probability density function corresponding to the feature dispersion position determination interval based on the feature dispersion number, and taking the probability density function as the dispersion position information of the traffic congestion evaluation feature in the target feature cluster of the corresponding congestion evaluation level.

determining a characteristic congestion monitoring index corresponding to the traffic congestion evaluation characteristic based on the probability density function, wherein the characteristic congestion monitoring index is inversely related to the probability density function;

the characteristic congestion monitoring indexes corresponding to the traffic congestion evaluation characteristics in the traffic congestion evaluation characteristic clusters corresponding to the target monitoring sections are summed up, and a second congestion monitoring index corresponding to the target monitoring sections is obtained;

and determining algorithm evaluation information of the monitoring section monitoring algorithm to the target monitoring section based on the second congestion monitoring index.

According to an example of an embodiment of the present disclosure, the acquiring a scattered location determination policy corresponding to the target feature cluster in the monitoring algorithm of the monitoring section; performing dispersion position determination on the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels based on the dispersion position determination strategy to obtain dispersion position information of the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels, wherein the method comprises the following steps:

determining an edge dispersion determination threshold;

determining a first feature cluster corresponding to a feature tail side and a second feature cluster corresponding to a feature start side in the target feature cluster based on the edge dispersion determination threshold;

the first feature cluster is summed up, and a tail side feature sum result corresponding to the target feature cluster is obtained;

determining the confidence coefficient of the tail side dispersion position corresponding to the traffic congestion evaluation feature based on the tail side feature sum-up result corresponding to the target feature cluster;

the second feature cluster is summed up, and an initial side feature sum result corresponding to the target feature cluster is obtained;

determining the confidence coefficient of the scattered positions of the starting side corresponding to the traffic congestion evaluation features based on the starting side feature sum-up result corresponding to the target feature cluster;

And determining target scattered position confidence corresponding to the traffic congestion evaluation feature based on the tail side scattered position confidence and the initial side scattered position confidence, and taking the target scattered position confidence as scattered position information of the traffic congestion evaluation feature in a target feature cluster of a corresponding congestion evaluation layer.

According to an example of an embodiment of the present disclosure, the determining, based on the tail-side scattered location confidence and the start-side scattered location confidence, a target scattered location confidence corresponding to the traffic congestion evaluation feature, taking the target scattered location confidence as scattered location information of the traffic congestion evaluation feature in a target feature cluster of a corresponding congestion evaluation level, includes:

acquiring the average value of each traffic congestion evaluation feature in the traffic congestion evaluation feature cluster, and determining the dispersion deflection degree corresponding to the target monitoring section based on the average value corresponding to the traffic congestion evaluation feature cluster;

acquiring a dispersion deflection degree comparison value, and determining a numerical comparison result of the dispersion deflection degree and the dispersion deflection degree comparison value;

determining a skew misalignment position confidence from the tail-side misalignment position confidence and the originating-side misalignment position confidence based on a numerical comparison of the misalignment degree to the misalignment degree comparison;

Performing numerical ranking on the tail side scattered position confidence, the initial side scattered position confidence and the skew scattered position confidence, and determining target scattered position confidence corresponding to the traffic congestion evaluation feature in the tail side scattered position confidence, the initial side scattered position confidence and the skew scattered position confidence according to a numerical ranking result;

and taking the target scattered position confidence coefficient as scattered position information of the traffic congestion evaluation characteristics in target characteristic clusters of the corresponding congestion evaluation level.

According to an example of an embodiment of the present disclosure, the summing algorithm evaluation information of the target monitoring section for each monitoring section monitoring algorithm in the monitoring section monitoring algorithm array to obtain traffic flow monitoring information of the target monitoring section includes:

determining that algorithm evaluation information is the number of negative evaluation results of congestion through the algorithm evaluation information of the target monitoring section;

if the number of the negative evaluation results is larger than the number of the preset negative evaluation results, determining that the target monitoring section is a congestion monitoring section, and obtaining traffic monitoring information of the target monitoring section;

The method further comprises the steps of:

acquiring the passive mark recording time of the target monitoring section, wherein the passive mark recording time is the recording time of the passive mark recording of the target monitoring section by the mark recording monitoring section;

determining the monitoring time for obtaining the traffic flow monitoring information of the target monitoring section when the traffic flow monitoring information of the target monitoring section is estimated to be the congestion monitoring section;

determining a difference between the passive tag recording time and the monitoring time;

obtaining a control result based on the difference value, wherein the control result is a control result of the congestion monitoring section when the monitoring section is identified to be the congestion monitoring section based on the monitoring section monitoring algorithm;

the obtaining the traffic congestion evaluation feature clusters corresponding to each target monitoring section respectively comprises the following steps:

acquiring a monitoring section traffic data set corresponding to the target monitoring section, wherein the monitoring section traffic data set comprises a plurality of monitoring section traffic data;

summarizing congestion evaluation levels respectively corresponding to a plurality of traffic data types according to the traffic data types;

and the traffic data attributes of the traffic data sets of the monitoring sections corresponding to the same congestion evaluation level are summed up to obtain traffic congestion evaluation characteristics corresponding to the congestion evaluation levels respectively, and the traffic congestion evaluation characteristics corresponding to the congestion evaluation levels form traffic congestion evaluation characteristic clusters corresponding to the target monitoring sections.

According to another aspect of the disclosed embodiments, there is provided a flow monitoring system comprising:

one or more processors;

and one or more memories, wherein the memories have stored therein computer readable code, which when executed by the one or more processors, causes the one or more processors to perform the method described above.

The application at least comprises the following beneficial effects:

according to the expressway real-time flow monitoring method and system based on ETC portal data, the traffic congestion evaluation feature clusters corresponding to all the target monitoring sections in the target monitoring section set are obtained through the target monitoring section set, the scattered positions of the traffic congestion evaluation features in the target feature clusters corresponding to the congestion evaluation level are analyzed based on the monitoring section monitoring algorithm, and algorithm evaluation information corresponding to the target monitoring sections is determined through the scattered position information of each monitoring section monitoring algorithm. According to the method, the congestion evaluation layer is used for analyzing the scattered positions of the traffic congestion evaluation characteristics, the scattered position information of all the congestion evaluation layers is combined to obtain the algorithm evaluation information corresponding to the monitoring section, so that the traffic congestion evaluation characteristics of the comprehensive monitoring section on each congestion evaluation layer are used for carrying out congestion decision on the monitoring section, the algorithm evaluation information output by the algorithm has higher reliability and accuracy, the monitoring section monitoring algorithm is different in the monitoring section evaluation mode according to the monitoring section monitoring algorithm, the algorithm evaluation information corresponding to the monitoring section monitoring algorithms is summed up to obtain the traffic flow monitoring information of the target monitoring section, and the traffic flow monitoring information of the target monitoring section can be obtained according to various scattered position information, so that the reliability and accuracy of monitoring of the monitoring section can be greatly improved.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the aspects of the present application.

Drawings

The above and other objects, features and advantages of the presently disclosed embodiments will become more apparent from the more detailed description of the presently disclosed embodiments when taken in conjunction with the accompanying drawings. The accompanying drawings are included to provide a further understanding of embodiments of the disclosure, and are incorporated in and constitute a part of this specification, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure, without limitation to the disclosure. In the drawings, like reference numerals generally refer to like parts or steps.

Fig. 1 is a schematic diagram of an application scenario provided in an embodiment of the present application.

Fig. 2 is a schematic implementation flow chart of a real-time highway traffic monitoring method based on ETC portal data according to an embodiment of the present application.

Fig. 3 is a schematic diagram of a composition structure of a flow monitoring device according to an embodiment of the present application.

Fig. 4 is a schematic hardware entity diagram of a computer device according to an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the drawings in the embodiments of the present disclosure. It will be apparent that the described embodiments are merely embodiments of a portion, but not all, of the present disclosure. All other embodiments, which can be made by one of ordinary skill in the art without the need for inventive faculty, are intended to be within the scope of the present disclosure, based on the embodiments in this disclosure. For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application are further elaborated below in conjunction with the accompanying drawings and examples, which should not be construed as limiting the present application, and all other embodiments obtained by those skilled in the art without making inventive efforts are within the scope of protection of the present application.

In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is to be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments and can be combined with one another without conflict. The term "first/second/third" is merely to distinguish similar objects and does not represent a specific ordering of objects, it being understood that the "first/second/third" may be interchanged with a specific order or sequence, as permitted, to enable embodiments of the present application described herein to be practiced otherwise than as illustrated or described herein. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing the present application only and is not intended to be limiting of the present application.

The highway real-time flow monitoring method based on ETC portal data can be applied to an application environment shown in FIG. 1. Wherein the ETC gantry 102 communicates with the flow monitoring system 104 through a network. The data storage system may store data that the flow monitoring system 104 needs to process. The data storage system may be integrated on the flow monitoring system 104 or may be located on the cloud or other network server. The traffic monitoring system 104 may be implemented as a stand-alone server or as a cluster of servers.

The highway real-time flow monitoring method based on ETC portal data provided by the embodiment of the application is applied to a flow monitoring system, and specifically comprises the following steps:

s101, responding to a flow monitoring instruction, acquiring a flow monitoring data set, wherein the flow monitoring data set comprises traffic flow data of at least one target monitoring section acquired based on an ETC portal.

The flow monitoring instruction can be instruction information generated by automatically generating or receiving an external instruction according to the cycle by the system. The traffic flow data of the monitoring section in the traffic monitoring data set is data collected by an ETC portal (may be one or more) located in the monitoring section (the setting of the monitoring section is not limited, such as a fixed road section), and the traffic flow data of the monitoring section in the traffic monitoring data set includes the number of vehicles of the monitoring section collected by the ETC portal in the collecting period, a time stamp of each vehicle passing through the ETC portal, lane information of each vehicle (the number of vehicles contained in each lane), a vehicle type (such as a truck, a sedan, a bus) and the like.

S102, acquiring traffic jam assessment feature clusters corresponding to each target monitoring section respectively, wherein the traffic jam assessment feature clusters comprise traffic jam assessment features corresponding to a plurality of jam assessment layers respectively.

The traffic congestion evaluation features are corresponding features for carrying out traffic congestion evaluation (whether congestion is generated or not), and the traffic congestion evaluation features comprise features of a plurality of congestion evaluation levels, and each congestion evaluation level corresponds to at least one traffic congestion evaluation feature. Wherein the congestion evaluation level may include: traffic flow, travel speed, road capacity utilization, or inlet-outlet flow balance. It can be understood that relevant data is obtained from corresponding traffic flow data according to the need during the evaluation of different congestion evaluation levels. The traffic congestion evaluation feature cluster is a data cluster composed of a plurality of traffic congestion evaluation features, and one target monitoring section can correspond to a plurality of traffic congestion evaluation feature clusters. The traffic congestion assessment feature cluster may be stored in the traffic monitoring system. In practical application, the traffic congestion evaluation characteristics of each target monitoring section under a plurality of congestion evaluation levels can be respectively obtained in the flow monitoring system, and the traffic congestion evaluation characteristics corresponding to each target monitoring section form a traffic congestion evaluation characteristic cluster to obtain the traffic congestion evaluation characteristic cluster corresponding to each target monitoring section.

S103, determining a monitoring section monitoring algorithm array, wherein the monitoring section monitoring algorithm array comprises a plurality of monitoring section monitoring algorithms; the monitoring section monitoring algorithm in the monitoring section monitoring algorithm array is different in monitoring section evaluation mode according to the monitoring section monitoring algorithm.

The monitoring section monitoring algorithm array is an array (or a set) comprising a plurality of monitoring section monitoring algorithms, wherein the monitoring section monitoring algorithms are neural network algorithms or traditional machine learning algorithms, such as self-encoders, sparse encoders, generation countermeasure networks, random forest algorithms and the like, capable of completing monitoring section congestion decisions according to a characteristic monitoring section evaluation mode. The monitoring section evaluation mode is a mode for carrying out congestion decision basis, for example, a mode for classifying traffic congestion evaluation features or classifying features into barrels, and the monitoring section evaluation modes of each monitoring section monitoring algorithm are different.

S104, acquiring the scattered position information of each traffic congestion evaluation feature in the corresponding traffic congestion evaluation feature cluster of the target monitoring section in the corresponding target feature cluster of the corresponding congestion evaluation layer through each monitoring section monitoring algorithm.

The dispersion position information is a classification result of the dispersion position of each traffic congestion evaluation feature in the corresponding target feature cluster, the dispersion position can be understood as the distribution information of the feature, the dispersion position information can be expressed based on a classification label (such as congestion and smoothness), or the dispersion position information can be expressed by the probability corresponding to each classification (such as 0.3 and 0.8), or the dispersion position information can be a probability density function for obtaining the probability. The target feature cluster is a feature cluster formed by local or global traffic congestion evaluation features under a congestion evaluation level. And the traffic congestion evaluation features in the target feature cluster correspond to preset feature dispersion position conditions, and the traffic congestion evaluation features are subjected to dispersion position determination through feature dispersion position forms so as to obtain dispersion position information (namely a distribution result in the feature cluster). For example, the feature dispersion location case may be represented as a histogram.

As an implementation manner, the monitoring section monitoring algorithm may select a target feature cluster of a congestion evaluation level, classify the situation of the dispersed position of the target feature cluster of the congestion evaluation level according to a corresponding monitoring section evaluation mode, obtain dispersed position information corresponding to the congestion evaluation level, and then integrate the dispersed position information under each congestion evaluation level to obtain dispersed position information of the traffic congestion evaluation feature in the target feature cluster corresponding to the corresponding congestion evaluation level. The monitoring section monitoring algorithm may also perform global analysis on traffic congestion evaluation features of each congestion evaluation level, for example, determine global dispersion position conditions of traffic congestion evaluation features of all congestion evaluation levels, and obtain global dispersion position information of all congestion evaluation levels based on classification thereof.

S105, determining algorithm evaluation information of the monitoring section monitoring algorithm on the target monitoring section based on the scattered position information obtained by the monitoring section monitoring algorithm.

The algorithm evaluation information is a monitoring result of the monitoring section monitoring algorithm on the congestion condition corresponding to the target monitoring section (i.e. the corresponding congestion decision category), for example, the classification label based on the congestion condition represents the algorithm evaluation information of the monitoring section monitoring algorithm on the target monitoring section. The algorithm evaluation information may also contain an algorithm ID for each monitoring segment monitoring algorithm so that the system can determine the output source of each algorithm evaluation information.

The discrete location information for each monitoring segment monitoring algorithm may represent the classification result of each monitoring segment monitoring algorithm for each traffic congestion assessment feature. And analyzing the scattered position information based on the information to obtain algorithm evaluation information of each monitoring section monitoring algorithm on the target monitoring section. And mapping the result of the scattered position information expressed based on the classification label or the probability mode, and taking the mapping result as algorithm evaluation information. For example, the probability is compared with the preset probability, if the probability is greater than the preset probability, the algorithm evaluation information may be mapped to Y (or mapped to a numerical value, for example, 1), otherwise, the algorithm evaluation information is set to N.

As one embodiment, S105 specifically includes: the flow monitoring system determines the scattered position information of the traffic jam evaluation features determined by the monitoring section monitoring algorithm in the target feature clusters corresponding to the corresponding jam evaluation levels, and sums the scattered position information corresponding to the jam evaluation levels to obtain the scattered position information of the target monitoring section under the corresponding monitoring section monitoring algorithm; and mapping the scattered position information of the target monitoring section under the corresponding monitoring section monitoring algorithm, and determining the mapping result as algorithm evaluation information of the monitoring section monitoring algorithm on the target monitoring section.

S106, the algorithm evaluation information of the target monitoring section is summed up for each monitoring section monitoring algorithm in the monitoring section monitoring algorithm array, and the traffic flow monitoring information of the target monitoring section is obtained.

The traffic flow monitoring information may indicate whether the target monitoring section is a congestion monitoring section, and the traffic flow monitoring information may be that the target monitoring section is a congestion monitoring section, or that the target monitoring section is a clear monitoring section, or that the target monitoring section may be a retarded monitoring section according to actual needs. The method comprises the steps of summarizing algorithm evaluation information, namely counting the number of the algorithm evaluation information according to a preset statistical operation method, and determining a counted result as traffic flow monitoring information. For example, when the number of algorithm evaluation information reaches a preset number (set according to actual needs), the traffic monitoring information of the target monitoring section is determined as the congestion monitoring section. The process of obtaining the traffic flow monitoring information through the algorithm evaluation information of each monitoring section monitoring algorithm is the process of combining and learning the monitoring algorithms of each monitoring section, and the performance deficiency of a single classifier is overcome by combining a plurality of classifiers (algorithms), so that the performance, generalization capability and stability of the whole algorithm are improved.

According to the highway real-time flow monitoring method based on ETC portal data, the scattered positions of the traffic congestion assessment features are analyzed based on the congestion assessment layers, the algorithm assessment information corresponding to the monitoring section is obtained by combining the scattered position information of all the congestion assessment layers, so that the congestion decision is carried out on the monitoring section by integrating the traffic congestion assessment features of the monitoring section on each congestion assessment layer, the algorithm assessment information output by the algorithm has higher reliability and accuracy, the monitoring section assessment modes according to which the monitoring section monitoring algorithm is based are different, the algorithm assessment information corresponding to the monitoring section monitoring algorithm is integrated, the traffic flow monitoring information of the target monitoring section is obtained, the traffic flow monitoring information of the target monitoring section can be obtained according to various scattered position information, and the reliability and accuracy of monitoring by the monitoring section can be greatly improved. In addition, the scattered positions of the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation level are analyzed based on the monitoring section monitoring algorithm, the scattered position information can be obtained by the monitoring section monitoring algorithm under the condition that no training mark exists, the process of obtaining the training mark is saved on the premise that the reliability of traffic flow monitoring information is guaranteed, and the speed of obtaining the monitoring result of the monitoring section can be greatly improved.

As an implementation manner, obtaining, by using each monitoring section monitoring algorithm, information of a dispersed position of each traffic congestion evaluation feature in a corresponding traffic congestion evaluation feature cluster of a corresponding monitoring section in a corresponding target feature cluster of a corresponding congestion evaluation layer includes: acquiring traffic congestion evaluation characteristics corresponding to the congestion evaluation levels from the traffic congestion evaluation characteristic clusters corresponding to the target monitoring sections to obtain target characteristic clusters corresponding to the congestion evaluation levels respectively; acquiring a corresponding scattered position determination strategy of a target feature cluster in a monitoring section monitoring algorithm; and carrying out scattered position determination on the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels based on the scattered position determination strategy to obtain scattered position information of the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels.

The decentralized position determination strategy is a strategy for dividing traffic congestion evaluation features in the target feature cluster into matched feature ranges. The distributed position determination policy may be a policy of performing distributed position determination depending on a threshold value or a policy of performing distributed position determination based on a distributed position section.

The strategy for determining the scattered positions according to the threshold value is as follows: and comparing the traffic jam evaluation feature with a feature dispersion determination threshold, if the traffic jam evaluation feature is smaller than the feature dispersion determination threshold, dividing the traffic jam evaluation feature into one feature range, and if the traffic jam evaluation feature is not smaller than the feature dispersion determination threshold, dividing the traffic jam evaluation feature into other feature ranges.

The strategy for determining the scattered positions based on the scattered position intervals is as follows: determining a plurality of feature dispersion position determining sections, wherein each feature dispersion position determining section is matched with a feature value section corresponding to the traffic congestion evaluation feature, comparing the feature value of the traffic congestion evaluation feature with the feature value section, and dividing the traffic congestion evaluation feature into feature dispersion position determining sections of the feature value. The characteristic value of the traffic congestion estimation feature is, for example, a numerical value of the traffic congestion estimation feature, and may be, for example, the number of passing vehicles, a time stamp difference value of passing vehicles through two ETC portal frames, or the like.

As an implementation manner, traffic congestion evaluation features corresponding to each congestion evaluation level are selected from traffic congestion evaluation feature clusters corresponding to the target monitoring section, the traffic congestion evaluation features under one congestion evaluation level form a target feature cluster, the traffic congestion evaluation features in the target feature cluster are subjected to dispersion position determination based on a dispersion position determination strategy, so that each traffic congestion evaluation feature is divided into corresponding feature ranges, one feature range can correspond to one dispersion position information, and further the dispersion position information of the traffic congestion evaluation features in the target feature cluster corresponding to the congestion evaluation level is obtained.

As one embodiment, the decentralized position determination strategy may be determined by means of monitoring segment evaluations of each monitoring segment monitoring algorithm. Further, acquiring, by using the monitoring algorithm of each monitoring section, scattered location information of each traffic congestion evaluation feature in the corresponding traffic congestion evaluation feature cluster of the target monitoring section in the corresponding target feature cluster of the corresponding congestion evaluation layer, including: acquiring traffic congestion evaluation characteristics corresponding to the congestion evaluation levels from the traffic congestion evaluation characteristic clusters corresponding to the target monitoring sections to obtain target characteristic clusters corresponding to the congestion evaluation levels respectively; determining a corresponding scattered position determining strategy of the target feature cluster in the monitoring section monitoring algorithm through a corresponding monitoring section evaluating mode of each monitoring section monitoring algorithm, and determining the scattered position of the traffic jam evaluating feature in the target feature cluster of the corresponding jam evaluating level based on the scattered position determining strategy to obtain scattered position information of the traffic jam evaluating feature in the target feature cluster of the corresponding jam evaluating level.

The method has the advantages that the traffic congestion evaluation characteristics are subjected to dispersion position determination in the target characteristic clusters corresponding to the congestion evaluation levels, the dispersion position information of each congestion evaluation level can be determined, on the premise that the distance of the traffic congestion evaluation characteristics is not available, the dispersion position information corresponding to the traffic congestion evaluation characteristics can be acquired, the dispersion position information of the characteristics can be efficiently acquired, and the follow-up efficiency is improved.

As an implementation manner, the determining strategy of the scattered positions corresponding to the target feature clusters includes a manner of determining the scattered positions according to a threshold value, and the determining strategy of the scattered positions corresponding to the target feature clusters in the monitoring section monitoring algorithm is obtained; performing dispersion position determination on the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels based on the dispersion position determination strategy to obtain dispersion position information of the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels, wherein the method comprises the following steps: acquiring a characteristic dispersion topological graph (which can be a random tree), wherein the characteristic dispersion topological graph comprises a plurality of component topological points (or called topological nodes); determining an initial topological point of the feature dispersion topological graph as a current composition topological point corresponding to the target monitoring section, acquiring a current congestion evaluation level corresponding to the current composition topological point, and acquiring a current feature dispersion determination threshold value of a current target feature cluster corresponding to the current congestion evaluation level; determining a threshold value and traffic congestion evaluation characteristics of a target monitoring section at the current congestion evaluation level through current characteristic dispersion, and determining dispersion position information of the traffic congestion evaluation characteristics in a current target characteristic cluster; and determining the next composition topological point corresponding to the target monitoring section based on the dispersion position information, determining the next composition topological point as an optimized current composition topological point, returning to a current congestion evaluation level corresponding to the current composition topological point, and acquiring a current characteristic dispersion determination threshold value of the current target characteristic cluster corresponding to the current congestion evaluation level until the composition topological point corresponding to the target monitoring section is optimized.

The above individual traffic congestion assessment features are topological points in a feature dispersion topological graph, which can be considered as a monitoring segment monitoring algorithm. The feature dispersion topological graph can be one or more, and the feature dispersion topological graph of the monitoring section monitoring algorithm can simultaneously determine the dispersion positions of the traffic congestion evaluation features, so that the traffic congestion evaluation features corresponding to the target monitoring section are divided into corresponding composition topological points. In addition, the traffic congestion evaluation feature clusters corresponding to one target monitoring section can be respectively input into each feature dispersion topological graph, each feature dispersion topological graph outputs the dispersion position information of the target monitoring section, the dispersion position information of each feature dispersion topological graph is integrated, global dispersion position information corresponding to the feature dispersion topological graph is obtained, and the global dispersion position information is determined to be the dispersion position information of the monitoring algorithm corresponding to the monitoring section. Based on the same thought, traffic congestion evaluation feature clusters corresponding to other target monitoring sections are respectively input into each feature dispersion topological graph to obtain corresponding dispersion position information. The current feature dispersion determination threshold may be preset or may be obtained by obtaining a feature value of the traffic congestion estimation feature of the target monitoring section at the current congestion estimation level, for example, determining a variance of the feature value of the traffic congestion estimation feature of the target monitoring section at the current congestion estimation level as the current feature dispersion determination threshold. The condition that the optimization of the composition topological points is completed is that the composition topological points corresponding to the traffic congestion evaluation features do not have the next topological points, and the dispersion position information corresponding to the terminal topological points where the traffic congestion evaluation features are located can be determined as the dispersion position information in the target feature clusters of the congestion evaluation level where the traffic congestion evaluation features are located.

As one embodiment, determining, by the current feature dispersion determining threshold and the traffic congestion evaluation feature of the target monitoring section at the current congestion evaluation level, dispersion location information of the traffic congestion evaluation feature in the current target feature cluster includes: comparing the traffic congestion evaluation characteristics of the target monitoring section at the current congestion evaluation level with a current characteristic dispersion determination threshold; if the traffic congestion evaluation characteristics of the target monitoring section at the current congestion evaluation level are smaller than the current characteristic dispersion determination threshold, dividing the traffic congestion evaluation characteristics of the target monitoring section at the current congestion evaluation level into a first topological point; and if the traffic congestion evaluation characteristics of the target monitoring section at the current congestion evaluation level are not smaller than the current characteristic dispersion determination threshold value, classifying the traffic congestion evaluation characteristics of the target monitoring section at the current congestion evaluation level into a second topological point. Taking the first topological point as an example, determining the next composition topological point of the first topological point as the optimized current composition topological point, and circulating the above process.

As one embodiment, determining algorithm evaluation information of a monitoring section monitoring algorithm for a target monitoring section based on scattered position information obtained by the monitoring section monitoring algorithm includes: determining corresponding composition topological points of the target monitoring section based on the scattered position information; counting corresponding composition topological points of the target monitoring section to obtain a dispersion scale (which can be understood as a path extending in the dispersion topological diagram) of the target monitoring section in the characteristic dispersion topological diagram; determining a first congestion monitoring index corresponding to the target monitoring section based on the dispersion scale, wherein the congestion monitoring index is inversely related to the dispersion scale; and determining algorithm evaluation information of a monitoring algorithm of the monitoring section to the target monitoring section based on the first congestion monitoring index. The corresponding composition topological points of the target monitoring section are all topological points between the starting topological point and the terminal topological point of the traffic congestion evaluation feature of the target monitoring section, and the number of the composition topological points is used as the dispersion scale of the target monitoring section in the feature dispersion topological graph. Wherein the first congestion monitoring index is a value that can evaluate whether the target monitoring zone is a congestion monitoring zone.

As one embodiment, determining a first congestion monitoring index corresponding to the target monitoring section based on the dispersion scale includes: if the characteristic dispersion topological graph is one, determining a dispersion scale corresponding to the target monitoring section determined by the characteristic dispersion topological graph, taking the dispersion scale as an index, taking a predetermined constant as a base, and establishing an index function based on the dispersion scale. And inputting the dispersion scale corresponding to the target monitoring section into the first congestion monitoring index, and outputting the first congestion monitoring index. If the characteristic dispersion topological graph is a plurality of, determining a random variable average value of a dispersion scale corresponding to the target monitoring section through the dispersion scale, taking the random variable average value as an index, taking a predetermined constant as a base, establishing an exponential function based on the index function, inputting the random variable average value of the dispersion scale corresponding to the target monitoring section into the random variable average value, and outputting the random variable average value to obtain a first congestion monitoring index.

The corresponding congestion monitoring index is obtained based on the dispersion scale of the corresponding composition topological points of the target monitoring section in the characteristic dispersion topological graph, the smaller the dispersion scale value is, the less the traffic congestion monitoring index needs to be segmented, at the moment, the traffic congestion evaluation characteristics corresponding to the target monitoring section are more far away from the conventional points, and the larger the congestion monitoring index is.

As one embodiment, determining algorithm evaluation information of a monitoring segment monitoring algorithm for a target monitoring segment based on a first congestion monitoring index includes: comparing the first congestion monitoring index with a first congestion monitoring index threshold (set arbitrarily according to actual needs), and if the first congestion monitoring index is larger than the first congestion monitoring index threshold, determining the algorithm evaluation information of the target monitoring section as the congestion monitoring section.

Determining a first congestion monitoring index corresponding to the target monitoring section through the dispersion scale comprises: generating a feature dispersion topological graph through p learning samples, and determining an average dispersion scale S (p) of the feature dispersion topological graph, namely:

the specific value of k is set according to actual needs.

For the traffic congestion evaluation feature f of the target monitoring section, determining a first congestion monitoring index corresponding to the target monitoring section, specifically, the following formula may be adopted for determination:

wherein,to learn the random variable average of the dispersion scale of sample f in a plurality of feature dispersion topologies.

Therefore, when the congestion monitoring index is determined, the distance between the traffic congestion evaluation features is not determined, but accurate congestion decision can be completed through the topology point classification of the feature dispersion topology map, huge operation expenditure is not needed in the process, and the cost of a hardware environment is saved.

In the process of generating the feature dispersion topological graph, a plurality of feature dispersion topological graphs are generated through a plurality of learning samples, namely training samples, the process can adopt self-supervision to generate the feature dispersion topological graph without setting training marks, x learning samples are set during the generation of the feature dispersion topological graph, one traffic congestion evaluation feature f and a corresponding division coefficient w are arbitrarily selected, the learning samples corresponding to the current composition topological point are divided into a plurality of topological points based on the division coefficient w, and if the number of layers of the topological graph reaches a threshold value, or the topological points only comprise one learning sample, or the learning sample features of the topological points are stopped simultaneously.

As one implementation manner, the scattered location determining strategy includes a manner of performing scattered location determination based on a scattered location interval, and acquiring a scattered location determining strategy corresponding to a target feature cluster in a monitoring section monitoring algorithm; performing dispersion position determination on the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels based on the dispersion position determination strategy to obtain dispersion position information of the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels, wherein the method comprises the following steps: acquiring a characteristic dispersion position determination interval cluster corresponding to a target characteristic cluster in a monitoring section monitoring algorithm, wherein the characteristic dispersion position determination interval cluster comprises a plurality of characteristic dispersion position determination intervals; acquiring the feature dispersion number of the traffic jam evaluation features in the target feature cluster in each feature dispersion position determination interval; and determining a probability density function corresponding to the feature dispersion position determination interval based on the feature dispersion number, and taking the probability density function as the dispersion position information of the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation layers. Wherein each characteristic dispersion position determination section corresponds to a characteristic value section with traffic congestion evaluation characteristics.

As an implementation manner, target feature clusters corresponding to the congestion evaluation levels respectively can be determined, a probability density function corresponding to a feature dispersion position determination interval is determined based on traffic congestion evaluation features in the target feature clusters, and information of each dispersion position is obtained. After the dispersion position information corresponding to each congestion evaluation layer is obtained, the dispersion position information is summed up, and the global dispersion position information of the traffic congestion evaluation characteristics of the target monitoring section is obtained through the sum result.

In the method, the traffic congestion evaluation features in the target feature cluster are subjected to dispersion position determination through a plurality of feature dispersion position determination intervals, corresponding probability density functions are determined based on the feature dispersion number of the traffic congestion evaluation features in each feature dispersion position determination interval, so that dispersion position information is obtained, and when no marks exist, the traffic congestion evaluation features of the target monitoring section can be accurately divided through the probability density functions, so that accurate dispersion position information is obtained.

As one embodiment, determining algorithm evaluation information of a monitoring section monitoring algorithm for a target monitoring section based on scattered position information obtained by the monitoring section monitoring algorithm includes: determining a characteristic congestion monitoring index corresponding to the traffic congestion evaluation characteristic based on the probability density function, wherein the characteristic congestion monitoring index is inversely related to the probability density function; the characteristic congestion monitoring indexes corresponding to the traffic congestion evaluation characteristics in the traffic congestion evaluation characteristic clusters corresponding to the target monitoring sections are summed up, and a second congestion monitoring index corresponding to the target monitoring sections is obtained; and determining algorithm evaluation information of a monitoring algorithm of the monitoring section to the target monitoring section through the second congestion monitoring index. The characteristic congestion monitoring index is a value of whether the evaluation target monitoring section is a congestion monitoring section, for example, the characteristic congestion monitoring index is obtained by performing a statistical calculation in advance on a characteristic congestion monitoring index corresponding to a traffic congestion evaluation characteristic determined by a probability density function, for example, by calculating the reciprocal, and the second congestion monitoring index is a value of whether the evaluation target monitoring section is a congestion monitoring section.

After determining the characteristic congestion monitoring indexes corresponding to each traffic congestion evaluation characteristic in the traffic congestion evaluation characteristic cluster corresponding to the target monitoring section, determining the characteristic congestion monitoring indexes corresponding to each congestion evaluation level, and summing the characteristic congestion monitoring indexes of each congestion evaluation level to obtain a second congestion monitoring index corresponding to the target monitoring section.

For the traffic congestion evaluation feature cluster f corresponding to the target monitoring section, the feature congestion monitoring index is a probability density function p, and the number of congestion evaluation layers is i, when determining the second congestion monitoring index e corresponding to the target monitoring section, the method may be as follows:

as one embodiment, determining algorithm evaluation information of a monitoring segment monitoring algorithm for a target monitoring segment based on a second congestion monitoring index includes: comparing the second congestion monitoring index with a second congestion monitoring index threshold (set according to actual needs, specifically not limited), and if the second congestion monitoring index is greater than the second congestion monitoring index threshold, determining the algorithm evaluation information of the target monitoring section as the congestion monitoring section.

According to the method and the device, the congestion monitoring index is obtained through the probability density function, then the scattered position information corresponding to the monitoring algorithm of the monitoring section is obtained through the congestion monitoring index, the traffic congestion evaluation characteristics of the target monitoring section can be reliably classified through the probability density function on the premise that no marks exist, the traffic congestion evaluation characteristics with smaller probability density function are obtained through determination, and the reliable scattered position information is obtained.

As an implementation mode, acquiring a corresponding scattered position determining strategy of a target feature cluster in a monitoring section monitoring algorithm; performing dispersion position determination on the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels based on the dispersion position determination strategy to obtain dispersion position information of the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels, wherein the method comprises the following steps: determining an edge dispersion determination threshold; determining a first feature cluster corresponding to a feature tail side and a second feature cluster corresponding to a feature start side in the target feature cluster based on the edge dispersion determination threshold; the first feature cluster is summed up to obtain a tail feature sum result corresponding to the target feature cluster; determining the confidence level of the tail side dispersion position corresponding to the traffic jam evaluation feature based on the tail side feature sum result corresponding to the target feature cluster; the second feature cluster is summed up to obtain a starting side feature sum result corresponding to the target feature cluster; determining confidence level of a scattered position of the traffic jam assessment feature corresponding to the starting side based on the starting side feature sum result corresponding to the target feature cluster; and determining target scattered position confidence corresponding to the traffic congestion evaluation feature based on the tail scattered position confidence and the initial scattered position confidence, and taking the target scattered position confidence as scattered position information of the traffic congestion evaluation feature in a target feature cluster of the corresponding congestion evaluation layer.

The edge dispersion determination threshold is a value for dividing the traffic congestion evaluation feature at the edge of the target feature cluster, and may be set in advance as needed, and specifically includes a start side dispersion determination threshold (i.e., a threshold on the start side for dividing the traffic congestion evaluation feature at the edge of the start side) and a tail side dispersion determination threshold (i.e., a threshold on the end side for dividing the traffic congestion evaluation feature at the edge of the end side).

As an embodiment, traffic congestion estimation features smaller than the starting side dispersion determination threshold value in the target feature cluster may be regarded as the second feature cluster, and traffic congestion estimation features larger than or equal to the tail side dispersion determination threshold value in the target feature cluster may be regarded as the first feature cluster.

As an embodiment, the average value of the feature values of the traffic congestion estimation features in the first feature cluster may be calculated, determined as the tail-side feature summation result, and the average value of the feature values of the traffic congestion estimation features in the second feature cluster may be calculated as the tail-side feature summation result.

The confidence coefficient of the tail side scattered position is the confidence coefficient from the traffic jam evaluation characteristic scattered position to the first characteristic cluster; the initial side scattered location confidence is the confidence of one traffic congestion assessment feature scattered location to a second feature cluster. As one embodiment, determining a target discrete location confidence corresponding to a traffic congestion assessment feature based on a tail-side discrete location confidence and an originating-side discrete location confidence includes: and calculating the mean value of the tail side scattered position confidence and the initial side scattered position confidence as the target scattered position confidence. The congestion monitoring index is obtained through the situation of the edge dispersion positions, then the dispersion position information is obtained, the traffic congestion evaluation characteristics of the target monitoring section are classified through the normally distributed edge dispersion positions on the premise of no marking, and the traffic congestion evaluation characteristics with small probability are obtained, so that the reliable dispersion position information is obtained.

As one embodiment, determining a target discrete location confidence corresponding to a traffic congestion evaluation feature based on a tail-side discrete location confidence and an originating-side discrete location confidence, taking the target discrete location confidence as discrete location information of the traffic congestion evaluation feature in a target feature cluster of a corresponding congestion evaluation layer, including: calculating the average value of each traffic congestion evaluation feature in the traffic congestion evaluation feature cluster, and acquiring the dispersion deflection degree (namely, the deflection coefficient gamma) corresponding to the target monitoring section through the average value corresponding to the traffic congestion evaluation feature cluster; obtaining a dispersion deflection degree comparison value (preset according to actual needs, specifically without limitation), and determining a numerical comparison result of the dispersion deflection degree and the dispersion deflection degree comparison value; determining a skew dispersion position confidence (or a bias distribution confidence) from the tail side dispersion position confidence and the start side dispersion position confidence based on a numerical comparison result of the dispersion deflection degree and the dispersion deflection degree contrast value; numerical sorting is carried out on the tail side scattered position confidence coefficient, the initial side scattered position confidence coefficient and the skew scattered position confidence coefficient (sorting results are obtained through comparison of sizes), and target scattered position confidence coefficients corresponding to traffic jam evaluation features are determined from the tail side scattered position confidence coefficient, the initial side scattered position confidence coefficient and the skew scattered position confidence coefficient through the numerical sorting results; and taking the target scattered position confidence coefficient as scattered position information of the traffic congestion evaluation characteristics in the target characteristic clusters of the corresponding congestion evaluation level. The dispersion deflection degree comparison value is a value for comparing the dispersion deflection degrees.

The degree of dispersion deflection corresponding to the traffic congestion evaluation feature of the target monitoring section can be calculated by using a general deflection calculation formula, and will not be described here.

As one embodiment, determining the tail-side scatter position confidence, the beginning-side scatter position confidence, and the skew scatter position confidence includes: aiming at one traffic jam evaluation feature o in the traffic jam evaluation feature cluster, the starting side feature sum result is thatThe caudal feature summary result is expressed as +.>. Then:

G1＝

G2＝

G3＝G1，(Hr＜0)；

G3＝G2，(Hr≥0)

wherein Hr is the degree of dispersed skew corresponding to the r-th congestion evaluation level. If the dispersion deflection degree is less than 0, the dispersion position representing the target feature cluster is deflected to the initial side, and the initial side feature sum result is taken as G3; if the dispersion deflection degree is more than or equal to 0, the dispersion position representing the target feature cluster is deflected to the tail side, and the tail side feature sum result is taken as G3.

The initial side scattered position confidence coefficient is obtained by taking the negative logarithm of G1, the tail side scattered position confidence coefficient is obtained by taking the negative logarithm of G2, and the deflection scattered position confidence coefficient is obtained by taking the negative logarithm of G3.

As an embodiment, determining the target discrete location confidence corresponding to the traffic congestion evaluation feature from among the tail-side discrete location confidence, the beginning-side discrete location confidence, and the skew discrete location confidence by the numerical ranking result may be to take the maximum value of the tail-side discrete location confidence, the beginning-side discrete location confidence, and the skew discrete location confidence as the target discrete location confidence corresponding to the traffic congestion evaluation feature. And carrying out numerical comparison on the initial side scattered position confidence, the tail side scattered position confidence and the deflection scattered position confidence, and taking the maximum value of the initial side scattered position confidence, the tail side scattered position confidence and the deflection scattered position confidence as the target scattered position confidence.

According to the embodiment of the application, the corresponding scattered position confidence coefficient is obtained through the numerical comparison result of the scattered deflection degree and the scattered deflection degree comparison value on each congestion evaluation layer, and the accurate scattered position confidence coefficient can be obtained through the centroid deviation determination of the target feature cluster, so that the accurate scattered position information is obtained.

As an implementation manner, the method for summarizing algorithm evaluation information of a target monitoring section for each monitoring section monitoring algorithm in the monitoring section monitoring algorithm array to obtain traffic flow monitoring information of the target monitoring section includes: determining the number of negative evaluation results (namely, the algorithm evaluation information is that the target monitoring section is a congestion monitoring section) of which the algorithm evaluation information is congestion through the algorithm evaluation information of the target monitoring section; if the number of the negative evaluation results is larger than the number of the preset negative evaluation results (set according to actual needs), determining that the target monitoring section is a congestion monitoring section, and obtaining traffic flow monitoring information of the target monitoring section.

As one embodiment, if the number of negative evaluation results is greater than the preset number of negative evaluation results, the flow monitoring system considers the monitoring segment monitoring algorithm that is greater than the preset number of negative evaluation results to identify the target monitoring segment as a congestion monitoring segment. As an embodiment, if the number of target monitoring sections is plural and the number of negative evaluation results is greater than the preset number of negative evaluation results, all the target monitoring sections are determined as congestion monitoring sections, or the corresponding target monitoring sections whose number of negative evaluation results is greater than the preset number of negative evaluation results are determined as congestion monitoring sections.

In the embodiment of the application, when the algorithm evaluation information of which the evaluation number is greater than the preset negative evaluation result number is congestion, the target monitoring section is determined to be the congestion monitoring section, and the outputs of the monitoring algorithms of the plurality of monitoring sections are integrated to obtain the traffic flow monitoring information, so that compared with the accuracy of obtaining the traffic flow monitoring information based on the single monitoring section monitoring algorithm, the accuracy is higher and more reliable.

As one implementation, the algorithm evaluation information may be based on a confidence representation that the target monitoring segment is a congestion monitoring segment. The method for summarizing the algorithm evaluation information of the target monitoring section by each monitoring section monitoring algorithm in the monitoring section monitoring algorithm array to obtain the traffic flow monitoring information of the target monitoring section comprises the following steps: the method comprises the steps that algorithm evaluation information of each monitoring section monitoring algorithm on a target monitoring section is utilized to determine the algorithm confidence that each monitoring section monitoring algorithm is a congestion monitoring section on the target monitoring section; respectively obtaining comparison results of the algorithm confidence coefficient of each monitoring section monitoring algorithm relative to corresponding preset confidence coefficient (set according to actual needs, for example, taking the average value of the algorithm confidence coefficient corresponding to each monitoring section monitoring algorithm as the preset confidence coefficient), and converting the algorithm confidence coefficient of each monitoring section monitoring algorithm into a grading result of the target monitoring section as the congestion monitoring section based on the comparison results; and (3) carrying out sum-up (statistical operation) on scoring results corresponding to the monitoring algorithms of all the monitoring sections to obtain traffic flow monitoring information of the target monitoring sections.

As one embodiment, the scoring result is, for example, a scoring number of the congestion monitoring section is determined, if the scoring number is greater than a preset scoring number (set according to practical situations, for example, a fixed value), the traffic monitoring information of the target monitoring section is determined to be the congestion monitoring section, and if the scoring number is not greater than the preset scoring number, the traffic monitoring information of the target monitoring section is determined to be the clear monitoring section.

As an implementation manner, after the algorithm evaluation information of the target monitoring section is summed up by each monitoring section monitoring algorithm in the monitoring section monitoring algorithm array to obtain the traffic flow monitoring information of the target monitoring section, the method further includes: acquiring a negative mark corresponding to a target monitoring section from a preset mark set; comparing the traffic flow monitoring information of the target monitoring section with the negative marks corresponding to the target monitoring section, and determining the monitoring floating information corresponding to the monitoring algorithm array of the monitoring section according to the comparison result; the monitoring floating information comprises an alternative monitoring section with traffic flow monitoring information of which traffic is congested and negatively marked as unblocked; determining a monitoring section classification result of the alternative monitoring section based on the traffic congestion evaluation feature cluster corresponding to the alternative monitoring section based on the monitoring section classification algorithm; monitoring floating assessment information of the monitoring section monitoring algorithm array relative to the preset mark set is determined based on monitoring section classification results of the alternative monitoring sections.

The preset mark set is a mark including whether each monitoring section is a congestion monitoring section or not in the high-speed whole process, namely whether each monitoring section is a congestion monitoring section or not.

Optionally, the traffic congestion evaluation characteristics of the monitoring sections can be classified through a neural network algorithm, the marks of the monitoring sections of the congestion monitoring sections obtained through classification are determined to be the congestion monitoring sections, and the congestion monitoring sections are stored in a preset mark set. And after the monitoring section monitoring algorithm array obtains traffic flow monitoring information, acquiring passive marks in a preset mark set, analyzing the monitoring precision of the monitoring section monitoring algorithm array based on the passive marks, wherein the monitoring section monitoring algorithm array is obtained through self-supervision and debugging, and the passive marks in the preset mark set are obtained through supervised learning. The monitoring float evaluation information may be indicative of the accuracy of the monitoring segment monitoring algorithm array compared to the predetermined set of markers. The method for determining the monitoring floating evaluation information is, for example, to determine the number of monitoring segments which are accurately detected by the monitored segment monitoring algorithm array and are not accurately detected by the preset mark set, and the ratio of the number of monitoring segments to the total number of target monitoring segments is used as the monitoring floating evaluation information.

The monitoring section classification algorithm is a neural network algorithm capable of determining the category of the monitoring section, as an implementation mode, the traffic congestion evaluation feature cluster corresponding to the alternative monitoring section is input into the monitoring section classification algorithm, and feature mining analysis is carried out on the traffic congestion evaluation feature cluster through the monitoring section classification algorithm, so that a monitoring section classification result is obtained.

As an embodiment, the real-time highway traffic monitoring method based on the ETC portal data further includes: the method comprises the steps of obtaining the passive mark recording time of a target monitoring section, wherein the passive mark recording time is the recording time of the passive mark recording of the target monitoring section by the mark recording monitoring section; when the traffic flow monitoring information of the target monitoring section is estimated to be the congestion monitoring section, determining the monitoring time for obtaining the traffic flow monitoring information of the target monitoring section; determining the difference value between the passive mark recording time and the monitoring time; and obtaining a control result based on the difference value, wherein the control result is the control result of the congestion monitoring section when the monitoring section is identified to be the congestion monitoring section based on the monitoring section monitoring algorithm.

The tag record monitoring section is a monitoring section that adds a negative tag to each monitoring section. The difference is obtained by making a difference between the record time and the monitoring time of the electrolytic mark. The control result is to control traffic flow of the congestion monitoring section, such as limiting the duration of entering the monitoring section. As one embodiment, deriving the control result based on the difference value includes: obtaining control parameters (according to experience setting), and multiplying the difference value by the control parameters to obtain a control result.

As one embodiment, determining the difference between the passive marker registration time and the monitoring time; obtaining a control result based on the difference value, including: if the monitoring time is earlier than the passive marker recording time, determining a difference value between the passive marker recording time and the monitoring time; and obtaining a control result based on the difference value. Therefore, when the monitoring section monitoring algorithm array recognizes the corresponding congestion monitoring section earlier than the preset mark set, the congestion monitoring section can be controlled, and congestion is prevented from being aggravated.

As an implementation manner, obtaining a traffic congestion evaluation feature cluster corresponding to each target monitoring section, includes: acquiring a monitoring section traffic data set corresponding to a target monitoring section, wherein the monitoring section traffic data set comprises a plurality of monitoring section traffic data; summarizing congestion evaluation levels respectively corresponding to a plurality of traffic data types according to the traffic data types; and (3) summarizing the traffic data attributes of the traffic data sets of the monitoring sections corresponding to the same congestion evaluation levels to obtain traffic congestion evaluation features corresponding to the congestion evaluation levels respectively, wherein the traffic congestion evaluation features corresponding to the congestion evaluation levels form traffic congestion evaluation feature clusters corresponding to the target monitoring sections.

The traffic data type is collected traffic data of different layers (refer to step S101). The traffic data attribute is a value of the corresponding traffic data type, such as 345 vehicles. And vectorizing the numerical values according to other dimensions, such as lanes where vehicles are located, so as to obtain feature vectors serving as traffic jam assessment feature clusters.

As one embodiment, determining a monitoring segment monitoring algorithm array includes: acquiring an algorithm parameter set (such as weight, bias and learning rate) for establishing a monitoring section monitoring algorithm, wherein the algorithm parameter set comprises a plurality of algorithm parameters, acquiring parameter value ranges corresponding to all algorithm parameters in the algorithm parameter set, acquiring alternative parameter combination based on the parameter value ranges corresponding to all algorithm parameters in the algorithm parameter set, traversing the alternative parameter combination, performing accuracy detection on the monitoring section monitoring algorithm through the determined parameter combination to obtain a first detection result, determining a target parameter combination in the alternative parameter combination through the first detection result, performing algorithm parameter adjustment on the monitoring section monitoring algorithm based on the target parameter combination, and obtaining a monitoring section monitoring algorithm array based on the monitoring section monitoring algorithm after parameter adjustment.

As an implementation manner, the obtaining of the parameter value range corresponding to each algorithm parameter in the algorithm parameter set may be obtaining a plurality of initial parameter values corresponding to each algorithm parameter in the algorithm parameter set, so as to obtain a plurality of initial parameter combinations, performing accuracy detection on the monitoring section monitoring algorithm based on the plurality of initial parameter combinations, obtaining a second detection result, and determining the parameter value range corresponding to each algorithm parameter in the algorithm parameter set based on the second detection result.

As an implementation manner, the accuracy detection is performed on the monitoring algorithm of the monitoring section based on the selected parameter combination to obtain a first detection result, which may be based on the monitoring algorithm of the monitoring section, the traffic flow monitoring information of the monitoring section is obtained through the determined parameter combination, the negative mark of the monitoring section is obtained in the preset mark set, and the detection result corresponding to the monitoring algorithm of the monitoring section is determined through the similarity between the traffic flow monitoring information of the monitoring section and the negative mark of the monitoring section to obtain the first detection result.

Based on the same inventive concept, the embodiment of the application also provides a flow monitoring device for realizing the above-mentioned real-time highway flow monitoring method based on ETC portal data. The implementation of the solution provided by the device is similar to the implementation described in the above method, so the specific limitation of one or more embodiments of the flow monitoring device provided below may be referred to above for limitation of the real-time highway flow monitoring method based on ETC portal data, and will not be repeated here.

In one embodiment, as shown in FIG. 3, there is provided a flow monitoring device 300 comprising:

a data acquisition module 310 for acquiring a flow monitoring dataset comprising traffic flow data for at least one target monitoring zone acquired based on the ETC portal in response to the flow monitoring instruction;

the feature acquisition module 320 is configured to acquire a traffic congestion evaluation feature cluster corresponding to each of the target monitoring segments, where the traffic congestion evaluation feature cluster includes traffic congestion evaluation features corresponding to a plurality of congestion evaluation levels;

the algorithm determining module 330 is configured to determine a monitoring section monitoring algorithm array, where the monitoring section monitoring algorithm array includes a plurality of monitoring section monitoring algorithms; the monitoring section monitoring algorithm in the monitoring section monitoring algorithm array is different in monitoring section evaluation modes according to the monitoring section monitoring algorithm;

the distribution determining module 340 is configured to obtain, by using each monitoring section monitoring algorithm, information of a dispersed position of each traffic congestion evaluation feature in a target feature cluster corresponding to a corresponding congestion evaluation layer in a traffic congestion evaluation feature cluster corresponding to a target monitoring section;

an algorithm evaluation module 350, configured to determine algorithm evaluation information of the monitoring section monitoring algorithm on the target monitoring section based on the scattered location information obtained by the monitoring section monitoring algorithm;

And the evaluation and summation module 360 is configured to sum up algorithm evaluation information of the target monitoring section for each monitoring section monitoring algorithm in the monitoring section monitoring algorithm array, so as to obtain traffic flow monitoring information of the target monitoring section.

The respective modules in the tag processing apparatus described above may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.

In one embodiment, a flow monitoring system is provided, which may be a server, the internal structure of which may be as shown in fig. 4. The flow monitoring system includes a processor, a memory, an Input/Output interface (I/O) and a communication interface. The processor, the memory and the input/output interface are connected through a system bus, and the communication interface is connected to the system bus through the input/output interface. Wherein the processor of the flow monitoring system is configured to provide computing and control capabilities. The memory of the flow monitoring system includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the flow monitoring system is used for storing data including traffic flow data and the like. The input/output interface of the flow monitoring system is used to exchange information between the processor and the external device. The communication interface of the flow monitoring system is used for communicating with an external terminal through network connection. The computer program, when executed by the processor, implements a highway real-time flow monitoring method based on ETC portal data.

Those skilled in the art will appreciate that the structure shown in fig. 4 is merely a block diagram of a portion of the structure associated with the present application and is not limiting of the flow monitoring system to which the present application is applied, and that a particular flow monitoring system may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.

In one embodiment, there is also provided a flow monitoring system including a memory and a processor, the memory having stored therein a computer program, the processor implementing the steps of the method embodiments described above when the computer program is executed.

In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when executed by a processor, carries out the steps of the method embodiments described above.

In an embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the steps of the method embodiments described above.

It should be noted that, the object information (including, but not limited to, device information, corresponding personal information, etc. of the object) and the data (including, but not limited to, data for analysis, stored data, presented data, etc.) referred to in the present application are information and data authorized by the object or sufficiently authorized by each party, and the collection, use, and processing of the related data are required to comply with the related laws and regulations and standards of the related country and region.

Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in the various embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magnetic random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like. The databases referred to in the various embodiments provided herein may include at least one of relational databases and non-relational databases. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processors referred to in the embodiments provided herein may be general purpose processors, central processing units, graphics processors, digital signal processors, programmable logic units, quantum computing-based data processing logic units, etc., without being limited thereto.

The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.

The above examples only represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the present application. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application shall be subject to the appended claims.

Claims

1. The method for monitoring the real-time traffic of the expressway based on ETC portal data is characterized by comprising the following steps of:

acquiring traffic congestion evaluation feature clusters corresponding to each target monitoring section respectively, wherein the traffic congestion evaluation feature clusters comprise traffic congestion evaluation features corresponding to a plurality of congestion evaluation levels respectively; the congestion evaluation level comprises at least one of traffic flow, running speed, road capacity utilization rate or import-export flow balance, and relevant data are acquired from corresponding traffic flow data according to requirements when different congestion evaluation levels are evaluated;

Determining a monitoring section monitoring algorithm array, wherein the monitoring section monitoring algorithm array comprises a plurality of monitoring section monitoring algorithms; the monitoring section monitoring algorithm in the monitoring section monitoring algorithm array is different in monitoring section evaluation modes according to the monitoring section monitoring algorithm; the monitoring section monitoring algorithm is a neural network algorithm or a traditional machine learning algorithm for completing congestion decision of the monitoring section according to a monitoring section evaluation mode of traffic congestion evaluation characteristics;

acquiring scattered position information of each traffic congestion evaluation feature in the corresponding traffic congestion evaluation feature cluster of the target monitoring section in the corresponding target feature cluster of the corresponding congestion evaluation layer through each monitoring section monitoring algorithm; the target feature cluster is a feature cluster formed by local or global traffic congestion evaluation features under a congestion evaluation level, the scattered position information is a classification result of the scattered positions of the traffic congestion evaluation features in the corresponding target feature cluster, the scattered positions are distribution information of the traffic congestion evaluation features, the scattered position information is expressed based on classification labels, or the probability corresponding to each classification is expressed, or the scattered position information is a probability density function for obtaining the probability;

Determining algorithm evaluation information of the monitoring section monitoring algorithm on the target monitoring section based on the scattered position information obtained by the monitoring section monitoring algorithm; the algorithm evaluation information is a monitoring result of a monitoring section monitoring algorithm on the congestion condition corresponding to the target monitoring section;

and summing up algorithm evaluation information of the target monitoring section by each monitoring section monitoring algorithm in the monitoring section monitoring algorithm array to obtain traffic flow monitoring information of the target monitoring section, wherein the traffic flow monitoring information is whether the target monitoring section is a congestion monitoring section or not.

2. The method of claim 1, wherein the obtaining, by each monitoring segment monitoring algorithm, the scattered location information of each traffic congestion evaluation feature in the corresponding traffic congestion evaluation feature cluster of the target monitoring segment in the corresponding target feature cluster of the corresponding congestion evaluation level includes:

3. The method of claim 2, wherein the determining the discrete location strategy corresponding to the target feature cluster includes determining the discrete location strategy corresponding to the target feature cluster in the acquisition monitoring segment monitoring algorithm according to a threshold; performing dispersion position determination on the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels based on the dispersion position determination strategy to obtain dispersion position information of the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels, wherein the method comprises the following steps:

acquiring a characteristic dispersion topological graph, wherein the characteristic dispersion topological graph comprises a plurality of component topological points, each traffic congestion evaluation characteristic is a topological point in the characteristic dispersion topological graph, and the characteristic dispersion topological graph is a monitoring section monitoring algorithm;

4. The method of claim 3, wherein the determining algorithm evaluation information of the monitoring segment monitoring algorithm for the target monitoring segment based on the scattered location information obtained by the monitoring segment monitoring algorithm comprises:

counting corresponding composition topological points of the target monitoring section to obtain a dispersion scale of the target monitoring section in the characteristic dispersion topological graph, wherein the dispersion scale is a path extending in the dispersion topological graph;

5. The method of claim 2, wherein the scattered location determination strategy comprises a manner of scattered location determination based on a scattered location interval, and the obtained monitoring section monitors the corresponding scattered location determination strategy of the target feature cluster in the algorithm; performing dispersion position determination on the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels based on the dispersion position determination strategy to obtain dispersion position information of the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels, wherein the method comprises the following steps:

6. The method of claim 5, wherein the determining algorithm evaluation information of the monitoring segment monitoring algorithm for the target monitoring segment based on the scattered location information obtained by the monitoring segment monitoring algorithm comprises:

7. The method of claim 2, wherein the acquiring the corresponding discrete location determination strategy for the target feature cluster in the monitoring segment monitoring algorithm; performing dispersion position determination on the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels based on the dispersion position determination strategy to obtain dispersion position information of the traffic congestion evaluation features in the target feature clusters of the corresponding congestion evaluation levels, wherein the method comprises the following steps:

determining an edge dispersion determination threshold, wherein the edge dispersion determination threshold is a numerical value for dividing traffic congestion evaluation features at the edge of a target feature cluster, and specifically comprises a starting side dispersion determination threshold and a tail side dispersion determination threshold, wherein the starting side dispersion determination threshold is a starting side threshold and is used for dividing traffic congestion evaluation features at the edge of the starting side, and the tail side dispersion determination threshold is a finishing side threshold and is used for dividing traffic congestion evaluation features at the edge of the tail side;

8. The method of claim 7, wherein the determining the target discrete location confidence corresponding to the traffic congestion assessment feature based on the tail side discrete location confidence and the originating side discrete location confidence, taking the target discrete location confidence as the discrete location information of the traffic congestion assessment feature in a target feature cluster of a corresponding congestion assessment level comprises:

Acquiring the average value of each traffic congestion evaluation feature in a traffic congestion evaluation feature cluster, and determining the dispersion deflection degree corresponding to the target monitoring section based on the average value corresponding to the traffic congestion evaluation feature cluster, wherein the dispersion deflection degree is a bias coefficient;

9. The method of claim 1, wherein the summing the algorithm evaluation information of the target monitoring segment for each monitoring segment monitoring algorithm in the monitoring segment monitoring algorithm array to obtain the traffic monitoring information of the target monitoring segment, comprises:

determining the number of negative evaluation results of which the algorithm evaluation information is congestion according to the algorithm evaluation information of the target monitoring section, wherein the negative evaluation results are that the algorithm evaluation information is that the target monitoring section is the congestion monitoring section;

the method further comprises the steps of:

10. A flow monitoring system, comprising:

one or more processors;

and one or more memories, wherein the memories have stored therein computer readable code, which, when executed by the one or more processors, causes the one or more processors to perform the method of any of claims 1-9.