Traffic situation macro-analysis method based on highway video cloud networking platform

Abstract

This application relates to the field of intelligent transportation technology and discloses a traffic situation macro-analysis method based on a highway video cloud networking platform. The method includes data collection, vehicle detection and congestion event detection, statistical analysis, current traffic situation macro-analysis, historical pattern analysis, and traffic situation deduction. This application uses highway video data to perform macro-traffic situation analysis, making the data complexity far lower than traditional methods. It also reuses the video analysis capabilities of the highway video cloud networking platform, allowing for the rapid development and launch of traffic situation macro-analysis functions within the platform without the need to build a new traffic situation analysis system and front-end sensing equipment.

Description

Technical Field

[0001] The present application relates to the field of smart transportation technology, and specifically to a traffic situation macro-analysis method based on a highway video cloud networking platform. Background Art

[0002] Traffic situation analysis refers to the process of monitoring, evaluating, and predicting the status and changes of various traffic elements in a transportation system. Traditional traffic situation analysis technology is a multivariate data fusion analysis method that typically collects structured data such as ETC flow data, sign recognition flow data, traffic flow data, and instantaneous vehicle speed. Using big data technology, it monitors and analyzes real-time traffic flows, including indicators such as average speed, congestion delay index, total number of real-time congestion, total number of vehicles in transit, and congested road mileage ratio. Machine learning and deep learning methods (such as ARIMA, SVM, CNN, and LSTM algorithms) are then used to predict short-term traffic flow trends.

[0003] For example, the document "Zhang Tingting. Research on Highway Traffic Flow Parameter Prediction and Situation Assessment Based on ETC Gantry Data [D]. Chongqing: Chongqing Jiaotong University, 2024." records a method for predicting highway traffic flow parameters and evaluating traffic situation based on ETC gantry data, integrating toll station data and meteorological data, constructing multi-feature fusion features, and using deep learning and clustering algorithms. The main contents are as follows: (1) Based on highway ETC gantry data, toll station data and meteorological data, time series analysis and correlation analysis are carried out to reveal the periodic fluctuations and distribution characteristics of traffic flow in time and space dimensions. (2) Combining the spatial information extraction advantages of one-dimensional convolutional neural network (1DCNN), the time series prediction ability of bidirectional long short-term memory neural network (Bi-LSTM) and attention mechanism (ATT), a multi-feature extraction method that integrates spatial characteristics, temporal characteristics and external influencing factors is proposed, and a CNN-BiLSTM-ATT traffic flow parameter prediction model is constructed to predict traffic flow. (3) A traffic situation assessment model based on IPSO-WFCM is proposed to evaluate traffic situation.

[0004] However, road network operation monitoring data typically collects structured data such as ETC gantry data, sign recognition flow charts, and radar data. Prioritizing cross-system data integration and data collection from front-end sensing devices is crucial. Cross-system data collection is often difficult. For example, ETC gantry data, which is part of the toll collection system within the expressway toll network, is maintained on an independent network, isolated from other networks and systems, and requires regulatory approval for access. Data collection from front-end sensing devices presents challenges with integrating and maintaining platform compatibility with diverse hardware devices. The inconsistent access protocols of different hardware devices necessitate platform compatibility with these protocols, complicating platform integration, commissioning, and maintenance. Furthermore, varying data cleaning and processing methods for different data types increase data processing complexity, requiring the system to be capable of processing diverse traffic data. Appropriate multivariate data fusion and analysis methods are needed to understand the dynamic operation of the road network. Multivariate predictive modeling techniques (such as regression models and graph neural networks) can then be used to identify trends within this complex data.

[0005] Therefore, existing traffic situation analysis methods have difficulties in data collection, data processing and analysis. Summary of the Invention

[0006] The purpose of this application is to provide a traffic situation macro-analysis method based on a highway video cloud networking platform to solve the technical problems raised in the above background technology.

[0007] To achieve the above objectives, the present application discloses the following technical solution: a traffic situation macro-analysis method based on a highway video cloud networking platform, the method comprising the following steps:

[0008] Data collection: Video cloud gateways on all sections of the province capture video footage of the main highway at intervals and upload it to the provincial cloud platform;

[0009] Vehicle detection and congestion event detection: The YOLOv8s algorithm is used to detect vehicles in video screenshots and simultaneously detect mainline congestion events;

[0010] Statistical analysis: Count the number of trucks, buses, and vehicles in the video screenshots. Count the total mainline traffic volume, bus volume, truck volume, passenger and truck traffic ratio, total congestion mileage, and number of congestion events by road section. Calculate the mainline congestion rate for the road section and road network.

[0011] Macro-analysis of current traffic trends: Traffic flow visualization, including GIS-based mainline traffic heat maps, mainline passenger and freight vehicle distribution maps, and mainline congestion events and congestion rate distribution. Traffic operation status is identified using the k-means algorithm, categorizing traffic operation status into five levels: unimpeded, mostly unimpeded, slightly congested, moderately congested, and severely congested.

[0012] Historical pattern analysis: Based on historical mainline traffic data, the spatiotemporal correlation and fluctuation characteristics of traffic flow data are analyzed. The spatiotemporal correlation and fluctuation characteristics include monthly correlation, weekly correlation, daily correlation, holiday fluctuation characteristics, and seasonal fluctuation characteristics.

[0013] Traffic situation deduction: Based on the changing patterns of mainline traffic flow, the T-GCN algorithm is used to predict the mainline traffic flow of highways in the future time period, and the k-means algorithm is used to determine the traffic operation status within the prediction time period.

[0014] Preferably, the time interval is 5 minutes or 10 minutes.

[0015] Preferably, the congestion rate of the road section and the main line of the road network includes:

[0016] The congestion rate of the main line of a section is the ratio of the congested mileage of the main line of the section to the total mileage of the main line of the section;

[0017] The congestion rate of the main line of the road network is the ratio of the congested mileage of the main line of the road network to the total mileage of the main line of the road network.

[0018] Preferably, in the macro analysis of the current traffic situation, the z-score method is used to normalize the traffic operation data set of each road section.

[0019] Preferably, the normalization process specifically includes:

[0020] After normalizing each set of traffic operation data, a value in the range of [-1, 1] is obtained, where the mean and variance in the normalization process are the mean and variance of the traffic operation data, respectively.

[0021] Preferably, the traffic operation status identification by the k-means algorithm specifically includes:

[0022] Randomly select k indicator data from the data set D as the initial cluster center points, calculate the Euclidean distance of the non-class center data in the data set to the k cluster center points, and divide them into the cluster formed by the nearest center point;

[0023] Divide the non-central data of the data set into K clusters to form K clusters, and calculate the mean distance within the k clusters as the new cluster center point;

[0024] The error criterion function between the old cluster center point and the new cluster center point is continuously calculated iteratively. When it converges or reaches the maximum number of iterations, the iteration is stopped and the clustering result is generated.

[0025] Preferably, the calculation formula of the Euclidean distance is:

[0026]

[0027] in, is the traffic operation parameter dimension, is the corresponding dimension value of the traffic operation parameter, is the corresponding dimension value of the cluster center.

[0028] Preferably, in the historical regularity analysis:

[0029] The monthly correlation analysis is to draw a traffic time series diagram with a period of one month, with the time point as the x-axis, the number of days as the y-axis, and the traffic flow as the z-axis. The traffic time series diagrams of multiple months are observed to obtain the monthly traffic flow change pattern;

[0030] The weekly correlation analysis is to draw a traffic time series diagram with a cycle of one week, with the time point as the x-axis, the number of days as the y-axis, and the traffic flow as the z-axis. The traffic time series diagrams of multiple weeks are observed to obtain the weekly traffic flow change pattern.

[0031] The daily correlation analysis is to draw a traffic time series diagram of the same week X every week within a cycle with the time point as the horizontal axis and the traffic flow as the vertical axis, where X is the specific day of the week. The traffic time series diagrams of the same week X for multiple weeks are observed to obtain the daily traffic flow variation pattern;

[0032] The analysis of the holiday fluctuation characteristics is to draw a traffic time series diagram for the same holiday in a year with the time point as the horizontal axis and the traffic flow as the vertical axis, and observe the traffic time series diagrams for several years to obtain the holiday fluctuation characteristics;

[0033] The analysis of seasonal fluctuation characteristics is to draw a traffic time series diagram for the same season of the year with the time point as the x-axis, the number of days as the y-axis, and the traffic flow as the z-axis, and observe the traffic time series diagrams for several years to obtain seasonal fluctuation characteristics.

[0034] Preferably, in the traffic situation deduction, the T-GCN algorithm captures the road spatial topology structure through GCN to obtain spatial correlation, and captures the dynamic changes of node attributes through the GRU model to obtain time dependency. The input data set consists of an adjacency matrix and a feature matrix. The adjacency matrix is ​​used to describe the spatial relationship between roads, and the feature matrix is ​​used to describe the changes in traffic flow on each road over time.

[0035] Preferably, the traffic flow visualization also includes visualization of the macroscopic traffic operation status of provincial highways based on GIS maps.

[0036] Compared with the existing technology, the traffic situation macro-analysis method based on the highway video cloud networking platform of this application has the following beneficial effects:

[0037] By conducting macro-traffic situation analysis using highway video data, the data complexity is far lower than that of traditional methods. In addition, the video analysis capabilities of the highway video cloud networking platform can be reused, and the macro-traffic situation analysis function can be quickly developed and launched on the platform without the need to build a new traffic situation analysis system and front-end perception equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0039] Figure 1 A flowchart of a traffic situation macro-analysis method based on a highway video cloud networking platform provided in an embodiment of the present application;

[0040] Figure 2 A schematic diagram of a three-dimensional traffic sequence diagram provided in an embodiment of the present application. DETAILED DESCRIPTION

[0041] The following is a clear and complete description of the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0042] In this document, the term "comprising" is intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising..." does not preclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0043] Aiming to solve the problem of how to conduct macro traffic situation analysis on provincial highway main lines only through video data, this embodiment provides a technology for evaluating and predicting road network-level traffic operation status by analyzing video data. Specifically, it provides a method for Figure 1 The traffic situation macro-analysis method based on the highway video cloud networking platform shown in the figure includes the following steps:

[0044] S1-Data Collection: Video capture of the main highway lines is taken at intervals (e.g., every 5 or 10 minutes) through the video cloud gateways at all sections of the province, and uploaded to the provincial cloud platform (i.e., the provincial highway video cloud networking platform, a cloud computing-based platform designed to meet the application needs of ministerial-level video networking. It provides services such as equipment resource aggregation, video access, video control, and video sharing along the highway through cloud services).

[0045] S2 - Vehicle Detection and Congestion Event Detection: Use the YOLOv8s algorithm to detect vehicles in the video screenshots and simultaneously detect mainline congestion events. The YOLOv8s algorithm can be any existing technology.

[0046] S3-Statistical Analysis: Count the number of trucks, buses, and vehicles in the video screenshots. Count the total mainline traffic volume, bus traffic volume, truck traffic volume, passenger and truck traffic ratio, total congestion mileage, and number of congestion events by road section. Calculate the mainline congestion rate for the road section and the road network. The mainline congestion rate for a road section is the ratio of the congested mileage of the mainline of the road section to the total mileage of the mainline of the road section, and the mainline congestion rate for the road network is the ratio of the congested mileage of the mainline of the road network to the total mileage of the mainline of the road network.

[0047] S4 - Macro analysis of current traffic situation: Traffic flow visualization is performed, including a GIS-based mainline traffic flow heat map, a mainline passenger and freight vehicle distribution map, and mainline congestion events and congestion rate distribution. Traffic operation status is identified using the k-means algorithm, and is divided into five levels: unimpeded, basically unimpeded, slightly congested, moderately congested, and severely congested.

[0048] S5-Historical Pattern Analysis: Based on historical mainline traffic data, analyze the spatiotemporal correlation and fluctuation characteristics of traffic flow data, including monthly correlation, weekly correlation, daily correlation, holiday fluctuation characteristics, and seasonal fluctuation characteristics;

[0049] S6-Traffic Situation Simulation: Based on the changing patterns of mainline traffic flow, the T-GCN algorithm is used to predict the mainline traffic flow of highways in the future time period, and the k-means algorithm is used to determine the traffic operation status within the prediction time period.

[0050] In one embodiment, the z-score method is used to analyze the traffic operation data set of each road segment. Perform normalization processing, specifically including: normalizing each set of traffic operation data to obtain a value within the range of [-1,1] , the normalized values ​​are as follows:

[0051]

[0052] in, is the mean of traffic operation data, is the variance of traffic operation data.

[0053] Furthermore, the traffic operation status identification by the k-means algorithm specifically includes:

[0054] Randomly select k indicator data from the data set D as the initial cluster center points , calculate the non-class center data in the data set To k cluster centers Euclidean distance , and the non-class center data Divide into clusters formed by the nearest center point; the calculation formula of the Euclidean distance is:

[0055]

[0056] in, is the traffic operation parameter dimension, Traffic operation parameters The corresponding dimension value of The cluster center The corresponding dimension value of ;

[0057] Divide the non-central data of the data set into K clusters to form K clusters, and calculate the mean distance within the k clusters as the new cluster center point , then ,in, is the data volume of the cluster;

[0058] Continuously iteratively calculate the old cluster center points With the new cluster center When the error criterion function converges or reaches the maximum number of iterations, the iteration is stopped and the clustering results are generated;

[0059] When the traffic operation status is set to five levels: smooth, basically smooth, lightly congested, moderately congested and severely congested, the clustering results are obtained. ,in, Traffic operation status after traffic operation data is clustered;

[0060] By calculating the traffic operation status of each road section according to the above algorithm, the traffic operation status of each road section in the provincial road network can be obtained, and then the macro traffic operation status of provincial highways can be visualized based on the GIS map.

[0061] In one embodiment, in the historical regularity analysis:

[0062] The monthly correlation analysis is plotted with time points as the x-axis (step length is 1 hour), days as the y-axis (step length is 1 day), and traffic flow as the z-axis (unit is vehicle). Figure 2 The cycle shown is a traffic time series diagram for one month. By observing the traffic time series diagrams for multiple months, the monthly traffic flow variation pattern can be obtained;

[0063] The weekly correlation analysis is to draw a traffic time series diagram with a cycle of one week, with the time point as the x-axis (with a step length of 1 hour), the number of days as the y-axis (with a step length of 1 day), and the traffic flow as the z-axis (with a unit of vehicle). The traffic time series diagrams of multiple weeks are observed to obtain the weekly traffic flow variation pattern;

[0064] The daily correlation analysis is to draw a traffic time series diagram for the same day of the week X every week within a cycle, with the time point as the horizontal axis (with a step size of 1 hour) and the traffic flow as the vertical axis (in vehicles), where X is a specific day of the week (such as Monday, Tuesday, Wednesday, Thursday, Friday, Saturday or Sunday). The traffic time series diagrams for the same day of the week X for multiple weeks are observed to obtain the daily traffic flow variation pattern;

[0065] The analysis of the holiday fluctuation characteristics is to draw a traffic time series diagram for the same holiday in the year with the time point as the horizontal axis (with a step length of 1 hour) and the traffic flow as the vertical axis (unit: vehicle). The holiday fluctuation characteristics are obtained by observing the traffic time series diagrams for several years.

[0066] The analysis of seasonal fluctuation characteristics is to draw a traffic time series diagram for the same season of the year with the time point as the x-axis (with a step size of 1 hour), the number of days as the y-axis (with a step size of 1 day), and the traffic flow as the z-axis (unit: vehicles). The seasonal fluctuation characteristics are obtained by observing the traffic time series diagrams for several years.

[0067] In one embodiment, the traffic situation prediction algorithm, "T-GCN: A Temporal Graph Convolutional Network for Traffic Prediction," published in the 2019 IEEE Transactions on Intelligent Transportation Systems, is employed. The GCN captures the spatial topology of roads to obtain spatial correlation, and the GRU model captures the dynamic changes in node attributes to obtain temporal dependencies. Specifically, n historical time series data are used as input, and a graph convolutional network is used to capture the topology of the road network to obtain spatial features. Next, the obtained time series with spatial features are input into the GRU model, where dynamic changes are captured through inter-unit information transfer to obtain temporal features. Finally, the results are obtained through a fully connected layer. The input dataset consists of an adjacency matrix and a feature matrix. The adjacency matrix describes the spatial relationships between roads, with each row representing a road and the values ​​in the matrix indicating the connections between roads. The feature matrix describes the temporal changes in traffic flow on each road, with each row representing a road and each column representing the traffic flow on the road during different time periods.

[0068] In one embodiment, the traffic flow visualization further includes visualization of the macroscopic traffic operation status of provincial highways based on a GIS map.

[0069] From what has been said, we can see that:

[0070] Traditional traffic situation analysis technology usually uses various structured data such as ETC gantry data, license plate recognition flow data, traffic control data, radar data, etc. to analyze traffic operation conditions and predict traffic flow. It is necessary to collect data through front-end sensing equipment or connect to related systems to obtain data. There are objective situations such as multiple data types, high collection difficulty, and high data processing complexity. In addition, how to choose a suitable analysis and prediction algorithm for multivariate data is also a difficult problem.

[0071] The traffic situation macro-analysis method based on the highway video cloud networking platform of this embodiment only needs to process and analyze single data, which greatly reduces the difficulty of data collection and the complexity of data processing. It can also reuse the video analysis technology of the upgraded cloud platform for vehicle detection and congestion event detection, supporting the next step of traffic situation macro-analysis.

[0072] Secondly, most current traffic flow analysis methods require the use of several types of road network operation monitoring data, such as ETC gantry data, license plate recognition flow data, radar data, and traffic control data. In addition, a clear understanding of the above business data is required to accurately calculate the section-level traffic flow. This results in complex information interweaving and complex calculation models.

[0073] The traffic situation macro-analysis method based on the highway video cloud networking platform of this embodiment only needs to analyze the video data of the highway main line. It is simple and easy to understand in terms of business, low in data complexity and simple in function in development, and can quickly realize the business function online on the existing provincial cloud platform. In addition to conventional traffic flow analysis, the present invention also has the ability to visualize traffic situation, and can display the macro situation of the traffic situation of the provincial highway main line through visualization methods such as main line traffic heat map, main line passenger and freight vehicle distribution map, main line congestion events and congestion rate distribution map. Since the road network operation data has spatiotemporal correlation and volatility characteristics, the present invention uses a deep learning algorithm that can capture dynamic spatial dependency and time dependency to predict traffic flow, which improves the prediction accuracy compared to the machine learning method.

[0074] Therefore, in summary, the traffic situation macro-analysis method based on the highway video cloud networking platform of this embodiment can achieve relatively accurate traffic situation analysis and prediction based on a single data set. Compared with traditional traffic situation analysis methods, it reduces the difficulty of data collection, data processing and analysis complexity, and overall reduces the difficulty of system development. Furthermore, the traffic situation macro-analysis method based on the highway video cloud networking platform of this embodiment can fully utilize the existing video data sources, cloud network resources, and system platforms of the highway network. Simply adding a traffic situation analysis function to the provincial cloud platform can achieve macro-analysis of traffic situation on the main highway line. There is no need to add new front-end equipment or build new systems to meet business needs, which can greatly save construction costs.

[0075] In the embodiments provided herein, it should be understood that the embodiments described herein can be implemented using hardware, software, firmware, middleware, code, or any appropriate combination thereof. For hardware implementation, the processor may be implemented in one or more of the following: an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field-programmable gate array (FPGA), a processor, a controller, a microcontroller, a microprocessor, or other electronic units designed to implement the functionality described herein, or any combination thereof. For software implementation, some or all of the processes of the embodiments may be performed by a computer program instructing the relevant hardware. During implementation, the program may be stored in a computer-readable storage medium or transmitted as one or more instructions or codes on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein communication media includes any medium that facilitates the transmission of a computer program from one location to another. The storage medium may be any available medium that can be accessed by a computer. Computer-readable storage media may include, but are not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing the desired program code in the form of instructions or data structures and accessible by a computer.

[0076] Finally, it should be noted that the above is only a preferred embodiment of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or make equivalent replacements for some of the technical features therein. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A traffic situation macro-analysis method based on a highway video cloud networking platform, characterized in that: The method comprises the following steps: Data collection: Video cloud gateways on all sections of the province capture video footage of the main highway at intervals and upload it to the provincial cloud platform; Vehicle detection and congestion event detection: The YOLOv8s algorithm is used to detect vehicles in video screenshots and simultaneously detect mainline congestion events; Statistical analysis: Count the number of trucks, buses, and vehicles in the video screenshots. Count the total mainline traffic volume, bus volume, truck volume, passenger and truck traffic ratio, total congestion mileage, and number of congestion events by road section. Calculate the mainline congestion rate for the road section and road network. Macro-analysis of current traffic trends: Traffic flow visualization, including GIS-based mainline traffic heat maps, mainline passenger and freight vehicle distribution maps, and mainline congestion events and congestion rate distribution. Traffic operation status is identified using the k-means algorithm, categorizing traffic operation status into five levels: unimpeded, mostly unimpeded, slightly congested, moderately congested, and severely congested. Historical pattern analysis: Based on historical mainline traffic data, the spatiotemporal correlation and fluctuation characteristics of traffic flow data are analyzed. The spatiotemporal correlation and fluctuation characteristics include monthly correlation, weekly correlation, daily correlation, holiday fluctuation characteristics, and seasonal fluctuation characteristics. Traffic situation prediction: Based on the changing patterns of mainline traffic flow, the T-GCN algorithm is used to predict the mainline traffic flow of highways in the future time period, and the k-means algorithm is used to determine the traffic operation status within the predicted time period; The traffic operation status identification by the k-means algorithm specifically includes: From the traffic operation dataset of each road segment Randomly select k indicator data as the initial cluster center points, calculate the Euclidean distance from the non-class center data in the data set to the k cluster center points, and divide them into the cluster formed by the nearest center point; Divide the non-central data of the data set into K clusters to form K clusters, and calculate the mean distance within the k clusters as the new cluster center point; Continuously iterate and calculate the error criterion function between the old cluster center point and the new cluster center point. When it converges or reaches the maximum number of iterations, stop the iteration and generate the clustering result. The calculation formula of the Euclidean distance is: ； in, is the traffic operation parameter dimension, is the corresponding dimension value of the traffic operation parameter, is the corresponding dimension value of the cluster center.

2. The traffic situation macro-analysis method based on the highway video cloud networking platform according to claim 1 is characterized in that: The time interval is 5 minutes or 10 minutes.

3. The traffic situation macro-analysis method based on the highway video cloud networking platform according to claim 1 is characterized in that: The mainline congestion rates of the road sections and road networks include: The congestion rate of the main line of a section is the ratio of the congested mileage of the main line of the section to the total mileage of the main line of the section; The congestion rate of the main line of the road network is the ratio of the congested mileage of the main line of the road network to the total mileage of the main line of the road network.

4. The traffic situation macro-analysis method based on the highway video cloud networking platform according to claim 1 is characterized in that: In the macro analysis of the current traffic situation, the z-score method is used to normalize the traffic operation data set of each road section.

5. The traffic situation macro-analysis method based on the highway video cloud networking platform according to claim 4 is characterized in that: The normalization process specifically includes: After normalizing each set of traffic operation data, a value in the range of [-1, 1] is obtained, where the mean and variance in the normalization process are the mean and variance of the traffic operation data, respectively.

6. The traffic situation macro-analysis method based on the highway video cloud networking platform according to claim 1 is characterized in that: In the analysis of historical laws: The monthly correlation analysis is to draw a traffic time series diagram with a period of one month, with the time point as the x-axis, the number of days as the y-axis, and the traffic flow as the z-axis. The traffic time series diagrams of multiple months are observed to obtain the monthly traffic flow change pattern; The weekly correlation analysis is to draw a traffic time series diagram with a cycle of one week, with the time point as the x-axis, the number of days as the y-axis, and the traffic flow as the z-axis. The traffic time series diagrams of multiple weeks are observed to obtain the weekly traffic flow change pattern. The daily correlation analysis is to draw a traffic time series diagram of the same week X every week within a cycle with the time point as the horizontal axis and the traffic flow as the vertical axis, where X is the specific day of the week. The traffic time series diagrams of the same week X for multiple weeks are observed to obtain the daily traffic flow variation pattern; The analysis of the holiday fluctuation characteristics is to draw a traffic time series diagram for the same holiday in a year with the time point as the horizontal axis and the traffic flow as the vertical axis, and observe the traffic time series diagrams for several years to obtain the holiday fluctuation characteristics; The analysis of seasonal fluctuation characteristics is to draw a traffic time series diagram for the same season of the year with the time point as the x-axis, the number of days as the y-axis, and the traffic flow as the z-axis, and observe the traffic time series diagrams for several years to obtain seasonal fluctuation characteristics.

7. The traffic situation macro-analysis method based on the highway video cloud networking platform according to claim 1 is characterized in that: In the traffic situation deduction, the T-GCN algorithm captures the road spatial topology structure through GCN to obtain spatial correlation, and captures the dynamic changes of node attributes through the GRU model to obtain temporal dependency. The input data set consists of an adjacency matrix and a feature matrix. The adjacency matrix is ​​used to describe the spatial relationship between roads, and the feature matrix is ​​used to describe the changes in traffic flow on each road over time.

8. The traffic situation macro-analysis method based on the highway video cloud networking platform according to claim 1 is characterized in that: The traffic flow visualization also includes visualization of the macroscopic traffic operation status of provincial highways based on GIS maps.

Images