Expressway severe weather congestion node prediction method and emergency response system

Abstract

The invention relates to the technical field of intelligent traffic systems, in particular to a highway severe weather congestion node prediction method and an emergency response system, and the method comprises the following steps: data collection and preprocessing: obtaining real-time video streams from continuous cameras deployed along a highway; pre-optimization: performing pre-optimization on the video stream, including reducing the frame rate, compressing the image quality and reducing the data volume; key indexes are extracted, the key indexes comprise the vehicle density, the average vehicle speed and abnormal events, and the abnormal events comprise parking, retrograding and collision; and high-risk video screening: screening out a high-risk video region according to the key index. According to the invention, the processing capability of the traffic management system under the condition of massive video input can be obviously improved, and rapid and accurate traffic jam prediction can be realized, so that the road safety and the traffic efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent transportation systems, and in particular to a method for predicting highway congestion nodes in severe weather and an emergency response system. Background Art

[0002] With the rapid expansion of the highway network, traffic volume continues to increase, and traffic management systems face huge challenges. Especially for highway monitoring, the traffic volume is large and fluctuates frequently. The mileage of highways continues to extend, and the number of observation points is increasing, making the real-time monitoring and analysis tasks extremely heavy.

[0003] This situation has resulted in traditional manual monitoring methods being inefficient and labor-intensive, often resulting in staff being unable to fully cover all monitoring points. More seriously, manual inspections may miss detections and fail to detect congestion in a timely manner, thus causing a series of traffic safety issues, such as traffic accidents and travel delays, which directly affect road efficiency and traffic safety. Traditional video analysis methods are unable to cope with the growing demand for monitoring.

[0004] Therefore, it is particularly urgent to develop a method that can efficiently process massive video inputs and quickly predict congestion. Summary of the invention

[0005] The purpose of the present invention is to provide a method for predicting highway congestion nodes during severe weather and an emergency response system to solve the problem that the traditional manual method proposed in the above background technology is not only inefficient and labor-intensive, but also may miss detections, resulting in delayed congestion detection.

[0006] To achieve the above object, the present invention provides a method for predicting highway congestion nodes due to severe weather, comprising the following steps: S1. Data collection and preprocessing: real-time video streams are obtained from continuous cameras deployed along the highway; S2, pre-optimization: pre-optimize the video stream, including lowering the frame rate and compressing the image quality to reduce the data volume; S3, key indicator extraction: extract key indicators, including vehicle density, average vehicle speed and abnormal events, where abnormal events include parking, wrong-way driving and collision; S4. High-risk video screening: screening out high-risk video areas based on the key indicators; S5. Multi-video area correlation analysis: perform multi-video area correlation analysis to calculate the change of traffic flow; S6. Result output: Based on traffic flow changes and vehicle speed, output congestion scores and prediction results for high-risk areas.

[0007] As a further improvement of the technical solution, the data collection and preprocessing in step S1 specifically includes the following steps: S11. Camera deployment: Deploy continuous cameras along the highway to ensure coverage of all lanes and key intersections. The camera numbers are set as ; S12, data collection: collect vehicle flow data and video images through continuous cameras deployed along the highway; S13, data preprocessing: convert the acquired vehicle flow data and video images into a standard format and store them in a local or cloud database for subsequent processing.

[0008] As a further improvement of the technical solution, the pre-optimization in step S2 specifically includes the following steps: S21, frame rate optimization: set the frame rate of vehicle flow data and video images, and reduce the frame rate of the camera to 5fps; S22. Image quality optimization: Apply video compression algorithm to compress the video to reduce the resolution and image quality of the video.

[0009] As a further improvement of the present technical solution, the key indicator extraction in step S3 specifically includes the following steps: S31. Vehicle density index Extraction: Extract vehicle density indicators from each camera’s video image in real time; S32, average speed index Extraction: Extract the average vehicle speed index from the video image of each camera in real time; S33, abnormal event indicator extraction: extract abnormal event indicators from the video images of each camera in real time.

[0010] As a further improvement of the present technical solution, the formula for extracting the vehicle density index in step S31 is: ,

[0011] in, is the number of vehicles detected in the video frame, is the area of ​​the monitoring area.

[0012] As a further improvement of the present technical solution, the formula for extracting the average vehicle speed index in step S32 is: ,

[0013] in, is the speed of each vehicle.

[0014] As a further improvement of the present technical solution, the abnormal event indicator extraction in step S33 specifically includes the following steps: S331. Calculate parking rate :The parking rate is calculated using the following formula: ,

[0015] in, is the number of vehicles detected with a speed close to zero; S332, extracting the wrong-way vehicle event: extracting the event that a vehicle is detected to be moving in a direction opposite to the main lane; S333, extract collision events: extract events in which the detection frames of multiple vehicles overlap quickly and the overlapping time lasts.

[0016] As a further improvement of the technical solution, the high-risk video screening in step S4 specifically includes the following steps: S41, fast priority assessment: grade and score the video images, and define the priority formula based on the key indicators in step S3: ,

[0017] in, are weight parameters, is the weight factor for detecting abnormal events; S42. Filter high-risk videos: Filter the videos with the highest priority based on the scoring results. videos, if the vehicle density continues to exceed the threshold and the average speed is less than , marked as high risk of congestion.

[0018] As a further improvement of the technical solution, the multi-video region association analysis in step S5 specifically includes the following steps: S51, flow monitoring and change calculation: First, the vehicle flow is calculated, according to the set The vehicle flow information extracted from the two video sources or monitoring areas before and after a specific monitoring video and , then perform flow change detection and set a flow change threshold , and calculate the flow changes of the two monitoring areas before and after: ,

[0019] if ,It is believed that the traffic flow difference between the front and rear areas is too large, and there may be congestion; S52, dynamic data update and real-time analysis: using sliding window technology, the data of each video image will be updated regularly, and the new data will overwrite the old data to ensure real-time monitoring of the current traffic flow. After each sliding window update, the data will be recalculated. and , and analyze the flow changes. Considering that the changes in traffic flow on highways often have time series characteristics, a short time window of 5 minutes is used to calculate the flow changes, and the flow difference between the front and rear areas is evaluated in each time window.

[0020] As a further improvement of the present technical solution, the result output in step S6 specifically includes the following steps: S61, congestion score calculation: congestion score calculation is performed according to the flow change data and vehicle speed obtained in step S5; S62. Output prediction results: Sort by traffic flow change and vehicle speed, and output the congestion score, traffic flow change, vehicle speed, geographic location, and surveillance video number of all high-risk areas.

[0021] The present invention also provides an emergency response system for the highway severe weather congestion node prediction method, and the system is used for running the above method.

[0022] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention optimizes the frame rate and image quality of the video stream, thereby reducing the computational burden and further reducing the processing and storage requirements of the video data, thereby effectively improving the data processing efficiency.

[0023] 2. The present invention combines key indicators such as vehicle density, average vehicle speed and abnormal events, and performs correlation analysis through multiple video areas, so as to timely identify and predict potential congestion risks and provide effective support for traffic management.

[0024] 3. The present invention avoids the possible missed detection phenomenon in traditional manual monitoring by adopting an automated video analysis method, and improves the accuracy and response speed of detection.

[0025] 4. The present invention can not only monitor and analyze traffic conditions in real time, but also help traffic management departments respond quickly through accurate congestion prediction, alleviate traffic pressure, and improve road traffic efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a flow chart of the method for predicting highway severe weather congestion nodes of the present invention.

[0027] Figure 2 It is a schematic diagram of the overall steps of the present invention.

[0028] Figure 3 Schematic diagram of data collection and preprocessing steps of the present invention.

[0029] Figure 4 It is a schematic diagram of the pre-optimization steps of the present invention.

[0030] Figure 5 It is a schematic diagram of the key indicator extraction steps of the present invention.

[0031] Figure 6 It is a schematic diagram of the high-risk video screening steps of the present invention.

[0032] Figure 7 Schematic diagram of the multi-video region association analysis steps of the present invention.

[0033] Figure 8 It is a schematic diagram of the result output steps of the present invention. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] In a specific embodiment, Figure 1 and Figure 2 As shown, the present invention provides a method for predicting highway congestion nodes during severe weather, which specifically includes the following steps: The first step, such as Figure 3 As shown in Figure 2, data collection and preprocessing: real-time video streams are obtained from continuous cameras deployed along the highway.

[0036] 1. Camera deployment: Deploy continuous cameras along the highway to ensure coverage of all lanes and key intersections. The camera number is set as .

[0037] Site selection analysis: Conduct on-site surveys to identify key monitoring locations on the highway, including entry and exit lanes, lane divergence and merging points, and traffic congestion-prone sections. Use traffic flow models to assess the monitoring needs and importance of different locations and determine the number and layout of cameras.

[0038] Camera selection: Choose the right camera equipment according to your needs, such as a high-definition digital camera, a wide-angle lens with night vision and rain and fog performance. Determine the camera's resolution, frame rate, and field of view to ensure that clear real-time images are captured.

[0039] Installation and debugging: Install the camera at the selected location to ensure that its field of view covers all lanes and avoids interference from man-made or physical obstacles. Adjust the angle and height of the camera to ensure a clear image of the driver and vehicle without distortion or blur.

[0040] Number and network configuration: Set the camera number to , forming a unified numbering system for subsequent data management. Configure the network settings of the camera to ensure that the video stream can be transmitted to the data center in real time through the network.

[0041] 2. Data collection: Vehicle flow data and video images are collected through continuous cameras deployed along the highway.

[0042] Real-time video acquisition: Through the continuous cameras installed along the highway, the video acquisition software is used to capture vehicle flow data at regular intervals to generate video streams. When each camera collects the video stream in real time, the video timestamp and camera number are recorded to form a video record.

[0043] Data recording and storage: Real-time recording of each vehicle's passing time, license plate recognition information, vehicle type (such as car, truck, van), etc. The storage period of each frame image can be set (for example, one frame is stored every 10 seconds), and the corresponding data is summarized in the database. The statistical formula for vehicle flow data is: .

[0044] Data tagging: Each piece of data is tagged through a numbering system to ensure that the record can be traced back to a specific camera and time. The vehicle flow data and video images are combined to improve the integrity and accuracy of the data for subsequent analysis.

[0045] 3. Data preprocessing: Convert the acquired vehicle flow data and video images into a standard format and store them in a local or cloud database for subsequent processing.

[0046] Convert to standard format: Format the collected video images and vehicle flow data and convert them into a unified storage format, such as MP4, JPEG or CSV. Use data cleaning technology to remove missing values ​​and outliers to ensure data integrity and accuracy.

[0047] Data compression and storage: Apply video compression algorithms (such as H.264 or HEVC) to compress the video to reduce storage space requirements while ensuring that the image quality can be used for analysis. Vehicle flow data is stored in a relational database (such as MySQL) or a big data platform (such as HDFS) to support subsequent query and analysis.

[0048] Indexing and backup: Create indexes for data to quickly retrieve and query traffic data for a specific time period. Set up a regular backup mechanism to ensure data security and integrity and prevent accidental losses.

[0049] Data preparation: Divide the data into training and test sets according to the needs of subsequent steps so that they can be used for model training and verification. Prepare labels for image data and traffic data to ensure that the machine learning model can effectively identify and predict traffic flow conditions.

[0050] Step 2: Figure 4 As shown, pre-optimization: pre-optimize the video stream, including lowering the frame rate and compressing the image quality to reduce the amount of data.

[0051] 1. Frame rate optimization: Set the frame rate of vehicle traffic data and video images, and reduce the frame rate of the camera to 5fps.

[0052] Set frame rate: In order to ensure a certain video quality and reduce the data processing burden, the frame rate of the camera is set to 5fps. Considering the characteristics of vehicle flow, in some cases, 5fps is sufficient to capture relatively accurate traffic data.

[0053] Data collection mode: Using the 5fps setting means that only 5 image frames are collected per second. Consider a scenario with high average traffic volume, where the frequency of vehicles passing a certain location is If the video rate is 10000 vehicles / second, each video frame will be able to capture some vehicle information. The number of vehicles that can be represented in each frame can be expressed by the following formula: .

[0054] For example, if the width of the road being monitored by video is 3 meters, and the average length of each vehicle is 4 meters, in the case of high vehicle flow, assuming the flow rate is 3 vehicles / second, each 5fps frame may capture information of 7 vehicles.

[0055] Optimize storage and transmission: By reducing the frame rate to 5fps, the required storage and computing resources can be greatly reduced. For example, if the original frame rate is 30fps, reducing it to 5fps can reduce the amount of video data to the original amount while maintaining the same monitoring time: This means that the amount of data is significantly reduced, making the system more efficient.

[0056] 2. Image quality optimization: Apply video compression algorithm to compress the video to reduce the resolution and image quality of the video.

[0057] Necessity of video compression: Due to the huge amount of data generated by video surveillance, directly storing and processing high-resolution videos in real time will bring huge storage and computing burdens to computers. Therefore, it is very necessary to apply video compression algorithms to reduce storage requirements and improve transmission efficiency.

[0058] Apply compression algorithm: Select a suitable compression algorithm (such as H.264, H.265, etc.) to compress the video. The specific process is as follows: a. Select the compression type: select frame-based compression (such as motion compensation) or block-based compression.

[0059] b. Adjust resolution: You can set the video resolution to 640x480 (VGA) or lower.

[0060] c. Compression settings: Adjust the video encoding bit rate as needed to achieve a certain level of smoothness and visualization effect.

[0061] Reduce video resolution: The resolution of the video directly affects the storage size and video quality. Reducing the resolution of the video usually significantly reduces the amount of video data. For example, if the original video resolution is 1920x1080 (Full HD), and it is set to 640x480 after compression, the amount of storage data can be calculated as follows: .

[0062] This means that when compressing the video resolution from Full HD to VGA, the amount of data stored can be reduced to 1 / 6.74 of the original amount.

[0063] Step 3: Figure 5 As shown, key indicator extraction: extract key indicators, including vehicle density, average speed and abnormal events, where abnormal events include parking, wrong driving and collision.

[0064] 1. Vehicle density index Extraction: Extract vehicle density indicators from each camera’s video image in real time.

[0065] The formula for extracting vehicle density index is: .

[0066] in, is the number of vehicles detected in the video frame, is the area of ​​the monitoring area.

[0067] 2. Average speed index Extraction: Extract the average vehicle speed indicator from the video image of each camera in real time.

[0068] The formula for extracting the average speed index is: .

[0069] in, is the speed of each vehicle.

[0070] 3. Abnormal event indicator extraction: Abnormal event indicators are extracted from the video images of each camera in real time.

[0071] Calculate parking rate :The parking rate is calculated using the following formula: .

[0072] in, is the number of vehicles detected with a speed close to zero.

[0073] Extract wrong-way vehicle events: Extract events where a vehicle is detected moving in the opposite direction of the main lane.

[0074] Extract collision events: Extract events where multiple vehicle detection frames overlap quickly and the overlap time lasts for a long time.

[0075] Step 4: Figure 6 As shown, high-risk video screening: based on the key indicators, high-risk video areas are screened out.

[0076] 1. Quick priority assessment: grade and score the video images, and define the priority formula based on the key indicators in step S3: .

[0077] in, are weight parameters, is the weight factor for detecting abnormal events.

[0078] 2. Filter high-risk videos: Filter the highest priority videos based on the scoring results. videos, if the vehicle density continues to exceed the threshold and the average speed is less than , marked as high risk of congestion.

[0079] Set screening thresholds: Determine the screening conditions for high-risk videos, usually based on the thresholds of vehicle density and average speed. You can set the following conditions: Vehicle density threshold : For example, if the density exceeds 30 vehicles / km, it is considered high risk. Speed ​​threshold : For example, if the average speed is less than 20 km / h, it is considered high risk.

[0080] High-risk video screening: For the top videos (such as the first 10) for further screening. The specific steps are as follows: and , then mark this video as "high risk". Add videos that meet the above conditions to the high-risk video list.

[0081] Integrate screening results: Summarize videos marked as high-risk and generate a high-risk video set. On this basis, further analysis of dwell time and congestion periods can be performed to form data support for subsequent decision-making.

[0082] Continuous monitoring and dynamic updating: Design a dynamic monitoring mechanism to facilitate real-time updating of high-risk video lists to ensure that sudden changes in traffic conditions can be captured in a timely manner.

[0083] Step 5: Figure 7 As shown, multi-video area correlation analysis: perform multi-video area correlation analysis to calculate the change of traffic flow.

[0084] 1. Traffic flow monitoring and change calculation: First, calculate the vehicle flow. The vehicle flow information extracted from the two video sources or monitoring areas before and after a specific monitoring video and , then perform flow change detection and set a flow change threshold , and calculate the flow changes of the two monitoring areas before and after: .

[0085] if , it is believed that the traffic flow difference between the front and rear areas is too large and there may be congestion.

[0086] 2. Dynamic data update and real-time analysis: Using sliding window technology, the data of each video image will be updated regularly, and the new data will overwrite the old data to ensure real-time monitoring of the current traffic flow. After each sliding window update, the data will be recalculated. and , and analyze the flow changes. Considering that the changes in traffic flow on highways often have time series characteristics, a short time window of 5 minutes is used to calculate the flow changes, and the flow difference between the front and rear areas is evaluated in each time window.

[0087] Step 6: Figure 8 As shown, the result output is: based on the traffic flow changes and vehicle speed, the congestion score and prediction results of the high-risk area are output.

[0088] 1. Congestion score calculation: Congestion score calculation is performed based on the traffic change data and vehicle speed obtained in step S5.

[0089] Collect data: Get traffic change data from step 5 (Unit: vehicle / hour), and get the average speed of the vehicle from the third step (Unit: km / h).

[0090] Define the congestion score formula: The congestion score can be calculated based on the relationship between flow rate changes and average vehicle speed. Choose a linear or nonlinear model to derive the congestion score The definition formula is as follows: in, is the maximum allowed speed in the area (as indicated by a speed limit sign), and It is an empirical parameter used to adjust the importance of flow rate and vehicle speed to the congestion score.

[0091] Congestion score normalization: Calculated congestion score It can be further normalized to convert to a standard score range (such as 0 to 100). The following formula is used for normalization: .

[0092] Among them, and are the minimum and maximum possible values ​​in the score, respectively.

[0093] 2. Output prediction results: Sort by traffic flow changes and vehicle speeds, and output the congestion scores, traffic flow changes, vehicle speeds, geographic locations, and surveillance video numbers of all high-risk areas.

[0094] Data collation: Data of all high-risk areas are collated into a list or data table, including: area ID, congestion score , Traffic changes , average speed , geographic location (such as longitude and latitude or other identifiers), and surveillance video number (unique identifier of the video stream).

[0095] Sorting: Sort the data table by traffic change and congestion score. The sorting strategy is to first sort in descending order based on traffic change, and then sort in descending order based on congestion score for the same traffic change.

[0096] Result output: The sorted high-risk area information is output in a table format, allowing users to quickly access and understand the data.

[0097] The present invention also provides an emergency response system for the highway severe weather congestion node prediction method, and the system is used for running the above method.

[0098] The highway severe weather congestion node prediction method and emergency response system implemented through the above steps can significantly improve the processing capability of the traffic management system in the face of massive video inputs, and achieve fast and accurate traffic congestion prediction, thereby improving road safety and traffic efficiency. Compared with the existing technology, it has the following beneficial effects: 1. By optimizing the frame rate and image quality of the video stream, the computing burden can be reduced, thereby reducing the processing and storage requirements of video data, effectively improving data processing efficiency; 2. By combining key indicators such as vehicle density, average speed and abnormal events, and performing correlation analysis through multiple video areas, it can timely identify and predict potential congestion risks, providing effective support for traffic management; 3. By adopting automated video analysis methods, it avoids the possibility of missed detection in traditional manual monitoring, and improves the accuracy and response speed of detection; 4. It can not only monitor and analyze traffic conditions in real time, but also help traffic management departments respond quickly through accurate congestion predictions, alleviate traffic pressure, and improve road traffic efficiency.

[0099] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for predicting highway congestion nodes in severe weather, characterized in that: The following steps are involved: S1. Data collection and preprocessing: real-time video streams are obtained from continuous cameras deployed along the highway; S2, pre-optimization: pre-optimize the video stream, including lowering the frame rate and compressing the image quality to reduce the data volume; S3, key indicator extraction: extract key indicators, including vehicle density, average vehicle speed and abnormal events, where abnormal events include parking, wrong-way driving and collision; S4. High-risk video screening: screening out high-risk video areas based on the key indicators; S5. Multi-video area correlation analysis: perform multi-video area correlation analysis to calculate the change of traffic flow; S6. Result output: Based on traffic flow changes and vehicle speed, output congestion scores and prediction results for high-risk areas.

2. The method for predicting highway congestion nodes due to severe weather according to claim 1, characterized in that: The data collection and preprocessing in step S1 specifically include the following steps: S11. Camera deployment: Deploy continuous cameras along the highway to ensure coverage of all lanes and key intersections. The camera numbers are set as ; S12, data collection: collect vehicle flow data and video images through continuous cameras deployed along the highway; S13, data preprocessing: convert the acquired vehicle flow data and video images into a standard format and store them in a local or cloud database for subsequent processing.

3. The method for predicting highway congestion nodes due to severe weather according to claim 1, characterized in that: The pre-optimization in step S2 specifically includes the following steps: S21, frame rate optimization: set the frame rate of vehicle flow data and video images, and reduce the frame rate of the camera to 5fps; S22. Image quality optimization: Apply video compression algorithm to compress the video to reduce the resolution and image quality of the video.

4. The method for predicting highway congestion nodes due to severe weather according to claim 1, characterized in that: The key indicator extraction in step S3 specifically includes the following steps: S31. Vehicle density index Extraction: Extract vehicle density indicators from each camera’s video image in real time; S32, average speed index Extraction: Extract the average vehicle speed index from the video image of each camera in real time; S33, abnormal event indicator extraction: extract abnormal event indicators from the video images of each camera in real time.

5. The method for predicting highway congestion nodes due to severe weather according to claim 4 is characterized in that: The formula for extracting the vehicle density index in step S31 is: ， in, is the number of vehicles detected in the video frame, is the area of ​​the monitoring area.

6. The method for predicting highway congestion nodes due to severe weather according to claim 4, characterized in that: The formula for extracting the average vehicle speed index in step S32 is: ， in, is the speed of each vehicle.

7. The method for predicting highway congestion nodes due to severe weather according to claim 4, characterized in that: The abnormal event indicator extraction in step S33 specifically includes the following steps: S331. Calculate the parking rate S: The parking rate is calculated using the following formula: ， in, is the number of vehicles detected with a speed close to zero; S332, extracting the wrong-way vehicle event: extracting the event that a vehicle is detected to be moving in a direction opposite to the main lane; S333, extract collision events: extract events in which the detection frames of multiple vehicles overlap quickly and the overlapping time lasts.

8. The method for predicting highway congestion nodes due to severe weather according to claim 1, characterized in that: The high-risk video screening in step S4 specifically includes the following steps: S41, fast priority assessment: grade and score the video images, and define the priority formula based on the key indicators in step S3: ， in, are weight parameters, is the weight factor for detecting abnormal events; S42. Filter high-risk videos: Filter the videos with the highest priority based on the scoring results. videos, if the vehicle density continues to exceed the threshold and the average speed is less than , marked as high risk of congestion.

9. The method for predicting highway congestion nodes due to severe weather according to claim 1, characterized in that: The multi-video region association analysis in step S5 specifically includes the following steps: S51, flow monitoring and change calculation: First, the vehicle flow is calculated, according to the set The vehicle flow information extracted from the two video sources or monitoring areas before and after a specific monitoring video and , then perform flow change detection and set a flow change threshold , and calculate the flow changes of the two monitoring areas before and after: ， if ,It is believed that the traffic flow difference between the front and rear areas is too large, and there may be congestion; S52, dynamic data update and real-time analysis: using sliding window technology, the data of each video image will be updated regularly, and the new data will overwrite the old data to ensure real-time monitoring of the current traffic flow. After each sliding window update, the data will be recalculated. and , and analyze the flow changes. Considering that the changes in traffic flow on highways often have time series characteristics, a short time window of 5 minutes is used to calculate the flow changes, and the flow difference between the front and rear areas is evaluated in each time window.

10. The method for predicting highway congestion nodes due to severe weather according to claim 1, characterized in that: The result output in step S6 specifically includes the following steps: S61, congestion score calculation: congestion score calculation is performed according to the flow change data and vehicle speed obtained in step S5; S62. Output prediction results: Sort by traffic flow change and vehicle speed, and output the congestion score, traffic flow change, vehicle speed, geographic location, and surveillance video number of all high-risk areas.

11. The emergency response system of the highway severe weather congestion node prediction method according to any one of claims 1 to 10, characterized in that: The system is used to run the above method.

Images