Intelligent auxiliary decision method and system applied to expressway emergency control

Abstract

The embodiment of the application provides an intelligent auxiliary decision-making method and system applied to expressway emergency management and control, relates to the technical field of video monitoring, and first acquires continuous video stream data of global monitoring equipment; then constructs a video stream scene correlation network based on the space-time distribution characteristics of the continuous video stream data, and the video stream scene correlation network contains information such as road section scene units; then the abnormal scene units are located through the video stream scene correlation network, and an emergency transmission path is modeled; then the management and control resource basic information of the global expressway is collected, and a management and control resource adaptation scheme is dynamically adapted and generated; finally, intelligent auxiliary decision-making instructions are generated according to the management and control resource adaptation scheme, to guide the on-site and background cooperative disposal of the emergency. The application can comprehensively monitor the traffic condition, accurately model the transmission path, optimize the resource allocation, and realize intelligent management and control.

Description

Technical Field

[0001] This application relates to the field of video surveillance technology, and more specifically, to an intelligent auxiliary decision-making method and system for emergency management on highways. Background Technology

[0002] In the highway traffic management system, the effective control of emergencies is a key link in ensuring road safety and smooth operation. Highways are characterized by high traffic volume, high speed, and complex traffic conditions. Once an emergency occurs, such as a traffic accident, severe weather, or road construction, if it is not handled in a timely and accurate manner, it can easily cause traffic congestion and even secondary accidents, seriously threatening the lives and property of passing vehicles and people.

[0003] Currently, emergency management on highways primarily relies on manual monitoring and experience-based decision-making. Monitoring personnel acquire real-time footage from surveillance equipment distributed throughout the highway, assess the type and scope of the incident based on their experience, and then formulate corresponding control measures. However, this approach has several limitations. Firstly, manual monitoring struggles to comprehensively and in real-time grasp the traffic conditions across the entire highway network, easily leading to blind spots and delayed detection of emergencies. Secondly, experience-based decision-making lacks scientific rigor and precision, making it difficult to accurately predict the transmission path and impact range of emergencies, and hindering the rational allocation of control resources. This results in ineffective implementation of control measures, failing to effectively alleviate traffic congestion and reduce accident risks.

[0004] In addition, while some existing intelligent management and control systems can use some technical means to monitor and analyze emergencies, most of them only focus on the traffic situation of a single road segment or local area, without fully considering the traffic flow transmission and event impact correlation between different road segments of the entire highway, and cannot accurately model the transmission path of emergencies. Summary of the Invention

[0005] In view of this, the purpose of this application is to provide an intelligent auxiliary decision-making method and system for the management and control of emergencies on highways.

[0006] In conjunction with the first aspect of this application, an intelligent auxiliary decision-making method for highway emergency management is provided, which is applied to an intelligent auxiliary decision-making system for highway emergency management. The method includes: Acquire continuous video stream data collected by monitoring equipment deployed throughout the highway, wherein the continuous video stream data includes real-time traffic scene recordings of each road segment at different times; A video stream scene association network is constructed based on the spatiotemporal distribution characteristics of continuous video stream data. The video stream scene association network includes road segment scene units, scene unit association relationships, and association strength information. The scene unit association relationships include traffic flow transmission information and event impact association information between different road segment scene units. Abnormal scene units are located by video stream scene association network, and the transmission path of sudden events is modeled based on association relationship and association strength information. The transmission path of sudden events includes the diffusion path information and the order of influence information of abnormal events among scene units in different road segments. Collect basic information on control resources across the entire highway network, dynamically adapt control resources and transmission paths based on the spatiotemporal attributes of emergency transmission paths, and generate control resource adaptation schemes. Intelligent auxiliary decision-making instructions are generated based on the resource adaptation scheme. These instructions include resource scheduling instructions and road traffic control instructions, which are used to guide on-site control equipment and the back-end control system to coordinate and execute emergency response operations.

[0007] In conjunction with the second aspect of this application, an intelligent auxiliary decision-making system for highway emergency management is provided. The intelligent auxiliary decision-making system for highway emergency management includes a machine-readable storage medium and a processor. The machine-readable storage medium stores machine-executable instructions. When the processor executes the machine-executable instructions, the intelligent auxiliary decision-making system for highway emergency management implements the aforementioned intelligent auxiliary decision-making method for highway emergency management.

[0008] In conjunction with a third aspect of this application, a computer-readable storage medium is provided, wherein computer-executable instructions are stored therein, and when the computer-executable instructions are executed, the aforementioned intelligent auxiliary decision-making method for highway emergency management is implemented.

[0009] Combining any of the above aspects, firstly, continuous video stream data collected by highway-wide monitoring equipment is acquired. Based on the spatiotemporal distribution characteristics of this continuous video stream data, a video stream scene association network is constructed. This network includes road segment scene units, scene unit association relationships, and association strength information, comprehensively reflecting traffic flow transmission and event impact associations between different road segments across the entire highway. Abnormal scene units are located through the video stream scene association network, and the transmission path of emergencies is modeled, presenting the diffusion path and impact sequence of abnormal events among different road segment scene units. Basic information on highway-wide control resources is collected, and control resources are dynamically adapted to the transmission path based on the spatiotemporal attributes of the emergency transmission path, generating a control resource adaptation scheme. This optimizes the allocation of control resources, improves resource utilization efficiency, and ensures rapid and effective resource allocation for handling emergencies. Finally, intelligent auxiliary decision-making instructions are generated based on the control resource adaptation scheme, guiding on-site control equipment and the back-end control system to collaboratively execute emergency response operations. This significantly improves the timeliness, accuracy, and effectiveness of highway emergency control, effectively ensuring the safe and smooth operation of the highway. Attached Figure Description

[0010] Without requiring any creative effort, other related figures can be obtained by combining the above-mentioned figures.

[0011] Figure 1 This is a flowchart illustrating the intelligent auxiliary decision-making method for highway emergency management provided in this application embodiment. Detailed Implementation

[0012] Figure 1 This paper illustrates a flowchart of an intelligent auxiliary decision-making method for highway emergency management provided in an embodiment of this application, which includes the following details: Step S110: Obtain continuous video stream data collected by monitoring equipment deployed throughout the highway. The continuous video stream data includes real-time traffic scene recordings of different time periods for each road segment.

[0013] In this embodiment, the highway management system collects video data through monitoring equipment deployed along the highway. This equipment includes high-definition cameras, panoramic cameras, etc., distributed at various key locations along the highway, such as bridges, tunnels, interchanges, and accident-prone sections. Each monitoring device continuously collects video footage at a set sampling frequency, forming a continuous video stream. This video stream data is transmitted in real-time to the management center's server via a dedicated communication network and stored in a video database. The video stream data not only contains real-time traffic scene footage but also includes metadata such as the collection time, device number, and geographical location, enabling accurate association with specific road segments and times during subsequent processing. To ensure data security and privacy, encryption technology is used during video stream transmission to prevent data leakage. Simultaneously, the system preprocesses the video stream, removing invalid footage, such as periods without vehicle traffic or obstructed footage, to reduce the amount of data required for subsequent processing.

[0014] Step S120: Construct a video stream scene association network based on the spatiotemporal distribution characteristics of continuous video stream data. The video stream scene association network includes road segment scene units, scene unit association relationships, and association strength information. The scene unit association relationships include traffic flow transmission information and event impact association information between different road segment scene units.

[0015] To construct a video stream scene association network, the highway first needs to be segmented. Based on the actual conditions of the highway, it is divided into multiple segment scene units according to certain length standards. Each segment scene unit corresponds to a specific highway segment and includes the complete monitoring area of ​​that segment. Next, static and dynamic association elements are extracted from each segment scene unit. Static association elements include relatively stable features such as the number of lanes in the segment, road surface structure type, connection methods of adjacent segments, and deployment density of monitoring equipment; dynamic association elements include features that change over time, such as traffic flow direction, density trends, and changes in traffic conditions at different times.

[0016] Based on the extracted static correlation elements, the structural fit degree between scene units of different road segments is calculated. Structural fit degree reflects the degree of connection and matching of road segments in terms of road structure, such as whether the number of lanes matches and whether the road surface type is consistent. A static correlation matrix is ​​generated by comprehensively evaluating the above factors. Similarly, traffic flow transmission efficiency is calculated based on dynamic correlation elements. Traffic flow transmission efficiency reflects the smooth transmission of traffic flow between different road segments, such as whether the direction of travel is consistent and whether the traffic flow is stable, thereby generating a dynamic correlation matrix.

[0017] The static and dynamic correlation matrices are fused, and the fused correlation values ​​are calculated according to a certain weight ratio to obtain the fused correlation matrix. Using road segment scene units as network nodes, and the road segment scene unit pairs corresponding to non-zero elements in the fused correlation matrix as correlation objects, correlation relationships are established between scene units. The correlation strength is determined by the element values ​​in the fused correlation matrix. Then, the spatiotemporal attribute information of each road segment scene unit is embedded into the initially constructed network skeleton to strengthen the spatiotemporal identification of the nodes. Finally, the network skeleton, scene unit correlation relationships, and correlation strength information are integrated to form a complete video stream scene correlation network.

[0018] Step S121: According to the highway segment division standard, the monitoring coverage area corresponding to the continuous video stream data is divided into multiple independent segment scene units. Each segment scene unit corresponds to a fixed length of highway segment and the complete monitoring screen range of that highway segment.

[0019] When dividing road segment scene units, the mileage markers and geographical features of the highway are taken into account. Starting from the toll station at the beginning, division points are set at regular intervals along the road centerline. When encountering special road facilities such as interchanges, tunnels, or bridges, the positions of the division points are adjusted to ensure that these special facilities are completely included within a single road segment scene unit. Each road segment scene unit has clearly defined start and end mileage markers and corresponding geographical boundary coordinates. Simultaneously, the monitoring equipment included in each road segment scene unit is determined; the coverage areas of these monitoring devices collectively constitute the complete monitoring range of that road segment scene unit. Through this method, a continuous monitoring coverage area is divided into multiple independent road segment scene units with clearly defined boundaries.

[0020] Step S122: Extract the static association elements of each road segment scene unit. The static association elements include the number of road lanes, road surface structure type, connection method of adjacent road segments, and deployment density of fixed monitoring equipment corresponding to the road segment scene unit.

[0021] For each road segment scenario unit, relevant static information is obtained from the highway's engineering archives and management database. The number of lanes in a road segment refers to the total number of lanes in one or both directions, such as four lanes or six lanes in both directions. Pavement structure types include different types such as asphalt concrete pavement, cement concrete pavement, and composite pavement; this information can be obtained from road construction records. The connection method between adjacent road segments describes the connection form between the current road segment and its preceding and following adjacent road segments, such as direct connection, connection via interchange, or connection via toll station. The deployment density of fixed monitoring equipment is obtained by counting the number of monitoring devices within the road segment scenario unit and dividing by the road segment length, with the unit being units per kilometer. The above information is organized into static correlation elements, serving as the basis for subsequent calculations of structural adaptability.

[0022] Step S123: Extract the dynamic correlation elements of each road segment scene unit. The dynamic correlation elements include the traffic flow direction, traffic flow density change trend and traffic status change at different times within the road segment scene unit.

[0023] Extracting dynamic correlation elements requires analyzing and processing continuous video stream data. Traffic direction is determined based on road segment design and actual traffic conditions, divided into uphill and downhill directions. Traffic flow density variation trends are analyzed by examining video stream data over a period of time, counting the number of vehicles passing through the road segment per unit time, calculating traffic flow density, and using methods such as linear regression to analyze its trend and determine whether the density is increasing, decreasing, or remaining stable. Traffic status changes at different times are analyzed by dividing the day into multiple time periods, such as morning peak, off-peak, evening peak, and nighttime. Parameters such as average traffic speed and lane occupancy are collected for each time period, and based on these parameters, traffic status is classified into different levels such as smooth flow, slow flow, and congestion, and the proportion of each traffic status in each time period is calculated. By extracting these dynamic correlation elements, the dynamic characteristics of traffic flow within a road segment scene unit can be reflected.

[0024] Step S124: Calculate the structural fit degree between different road segment scene units based on static correlation elements. The structural fit degree includes the connection and matching data of the road segment scene units on the road structure. Generate a static correlation matrix. The matrix elements of the static correlation matrix contain the structural fit degree related data of the corresponding two road segment scene units.

[0025] Step S1241: Evaluate the matching relationships of multiple structural elements between different road segment scene units. The matching relationships of multiple structural elements include lane number matching relationships, road surface structure type matching relationships, adjacent road segment connection method matching relationships, and fixed monitoring equipment deployment density matching relationships.

[0026] For each pair of road segment scene units, the matching relationships of their various structural elements are evaluated one by one. Lane number matching is determined by comparing the differences in lane numbers between the two road segments; if the number of lanes is the same, the matching degree is high; if there are differences, a matching score is calculated based on the magnitude of the difference. Road surface structure type matching determines whether the road surface structures of the two road segments are the same; if they are the same, the matching degree is high, and if they are different, the matching degree is low. Adjacent road segment connection method matching is evaluated based on the type of connection; direct connections have a higher matching degree than connections through other facilities. Fixed monitoring equipment deployment density matching is evaluated by calculating the difference in monitoring equipment deployment density between the two road segments; the smaller the density difference, the higher the matching degree.

[0027] Step S1242: For each pair of road segment scene units, determine the lane matching level based on the lane number matching relationship, determine the road structure matching level based on the road structure type matching relationship, determine the connection method matching level based on the adjacent road segment connection method matching relationship, and determine the monitoring coverage matching level based on the fixed monitoring equipment deployment density matching relationship.

[0028] The evaluation results of the matching relationships of various structural elements are converted into levels. For example, lane number matching scores greater than or equal to 0.8 are level 3, scores between 0.5 and 0.8 are level 2, and scores less than 0.5 are level 1. Similarly, the matching relationships of pavement structure type, connection method, and monitoring coverage are also classified into similar levels. Through level classification, the degree of matching of various structural elements can be more intuitively represented.

[0029] Step S1243: Based on the lane matching level, road surface structure matching level, connection method matching level and monitoring coverage matching level, determine the structural adaptability between different road segment scene units according to the preset comprehensive evaluation rules.

[0030] The comprehensive evaluation rules consider the weights of each matching level. For example, the weight of lane matching level is 0.3, road structure matching level is 0.2, connection method matching level is 0.3, and monitoring coverage matching level is 0.2. Each level is converted into a corresponding weight value, such as 1.0 for level 3, 0.5 for level 2, and 0.2 for level 1. Then, the weight values ​​of each matching level are weighted and summed to obtain the structural fit. The structural fit value ranges from 0 to 1; a higher value indicates a higher degree of structural fit between the two road segment scene units.

[0031] Step S1244: Using the number of the road segment scene unit as the identifier of the matrix row and column, construct an initial empty matrix. The number of rows and columns of the initial empty matrix are equal to the total number of road segment scene units.

[0032] Create a square matrix based on the total number of segment scene units. The rows and columns of the matrix are identified by the segment scene unit numbers, and each cell represents the structural fit between a pair of segment scene units. Initially, all elements in the matrix are set to 0.

[0033] Step S1245: Fill the calculated structural fit degree between each pair of road segment scene units into the corresponding row and column positions in the initial empty matrix to generate an intermediate static correlation matrix.

[0034] For each pair of road segment scene units, the previously calculated structural fit value is filled into the corresponding cell in the matrix. For example, the structural fit between road segment scene unit i and road segment scene unit j is filled into the i-th row and j-th column of the matrix. For non-adjacent road segment scene units, their structural fit may be 0, or it may be calculated and filled according to the actual situation.

[0035] Step S1246: Normalize the matrix elements in the intermediate static correlation matrix so that the values ​​of all matrix elements are within the same preset range, eliminating the dimensional differences caused by different dimension coefficients.

[0036] Since the dimensions and value ranges of various structural elements may differ, normalization of the intermediate static correlation matrix is ​​necessary to ensure comparability of structural fit. Normalization employs the min-max normalization method, transforming each element value in the matrix to the range of 0 to 1. Specifically, the maximum and minimum values ​​in the matrix are first identified. Then, for each element x, the normalization is performed using the formula x' = (x - min) / (max - min), where x' is the normalized result, min is the minimum value in the matrix, and max is the maximum value.

[0037] Step S1247: Check the matrix elements in the normalized intermediate static correlation matrix. If there are matrix elements with abnormal values, recalculate the structural fit of the corresponding road segment scene unit pair and replace the abnormal values.

[0038] After normalization, the elements in the matrix need to be checked for outliers. Outliers may be caused by data acquisition errors, calculation errors, or other reasons. If the value of an element is found to deviate significantly from the normal range, such as being much greater than 1 or much less than 0, the structural element matching relationship evaluation and structural fit calculation process of the road segment scene unit needs to be re-examined to find the problem and correct it, replacing the outlier with the correct value.

[0039] Step S1248: Generate the final static association matrix based on the corrected intermediate static association matrix. The matrix elements of the static association matrix contain structural adaptation data related to the scene units of the two road segments.

[0040] After verification and correction, the final static correlation matrix is ​​obtained. This static correlation matrix fully reflects the structural adaptability between different road segment scene units, and each element in the static correlation matrix represents the degree of connection and matching between a pair of road segment scene units in the road structure.

[0041] Step S125: Calculate the traffic flow transmission efficiency between different road segment scene units based on dynamic correlation elements, and generate a dynamic correlation matrix. The traffic flow transmission efficiency includes relevant data on the smooth transmission of traffic flow between different road segment scene units, and the matrix elements of the dynamic correlation matrix include relevant data on the traffic flow transmission efficiency of corresponding two road segment scene units.

[0042] Step S1251: Evaluate multiple dynamic correlation elements between different road segment scene units, including traffic direction relationship, traffic flow stability, traffic status smoothness, and traffic flow transmission status.

[0043] For each pair of road segment scene units, the relationship between their traffic directions is analyzed. If the traffic directions of the two road segments are the same, the traffic flow transmission is relatively smooth; otherwise, conflicts may occur. The traffic flow stability is assessed by calculating the coefficient of variation of traffic flow parameters over a period of time; the smaller the coefficient of variation, the more stable the traffic flow. The traffic state stability is judged by statistically analyzing the number of changes in traffic state between adjacent time periods; the fewer the changes, the higher the stability. The traffic flow transmission state is determined by analyzing the correlation between traffic volume on upstream and downstream road segments; the higher the correlation, the better the transmission state.

[0044] Step S1252: For each pair of road segment scene units, determine the directional association level based on their traffic direction relationship, determine the stability level based on the traffic flow stability state, determine the stability level based on the traffic state stability, and evaluate the traffic flow transmission state level based on traffic flow conversion data.

[0045] The evaluation results of various dynamic correlation elements are converted into levels. For example, same-direction traffic is level 3, and opposite-direction traffic is level 1; the coefficient of variation of traffic flow stability is less than 0.3 for level 3, 0.3 to 0.5 for level 2, and greater than 0.5 for level 1; the number of changes in traffic stability is less than 3 times / hour for level 3, 3 to 5 times for level 2, and greater than 5 times for level 1; the correlation coefficient of traffic flow transmission status is greater than 0.7 for level 3, 0.3 to 0.7 for level 2, and less than 0.3 for level 1.

[0046] Step S1253: Based on the directional association level, stability level, smoothness level and traffic flow transmission status level, determine the traffic flow transmission efficiency between different road segment scenario units according to the preset comprehensive evaluation rules.

[0047] The comprehensive evaluation rule also employs a weighted summation method, with the weight of each level determined based on its impact on traffic flow transmission efficiency. For example, the weight of the directional correlation level is 0.4, the weight of the stability level is 0.2, the weight of the smoothness level is 0.1, and the weight of the traffic flow transmission state level is 0.3. After converting each level into its corresponding weight value, a weighted summation is performed to obtain the traffic flow transmission efficiency.

[0048] Step S1254: Using the number of the road segment scene unit as the identifier of the matrix row and column, construct an initial empty dynamic matrix. The number of rows and columns of the initial empty dynamic matrix are equal to the total number of road segment scene units.

[0049] Similar to constructing a static association matrix, create a square matrix with the same number of road segment scene units as the total number of units, with rows and columns identified by road segment scene unit numbers, and initial element values ​​of 0.

[0050] Step S1255: Fill the calculated traffic flow transmission efficiency between each pair of road segment scene units into the corresponding row and column positions in the initial empty dynamic matrix to generate an intermediate dynamic correlation matrix.

[0051] The calculated traffic flow transmission efficiency value is filled into the corresponding cell in the matrix to reflect the degree of smooth transmission of traffic flow between different road segment scene units.

[0052] Step S1256: Perform a rationality check on the matrix elements in the intermediate dynamic correlation matrix, remove abnormal values ​​that deviate significantly from the normal range by referring to historical traffic flow transmission data, and recalculate the traffic flow transmission efficiency corresponding to the abnormal values.

[0053] By combining historical traffic flow transmission data from the same period, check for any anomalies in the elements of the intermediate dynamic correlation matrix. If the value of a certain element deviates significantly from historical data, it may be due to current data anomalies or calculation errors, requiring a reassessment and recalculation of the traffic flow transmission efficiency of that road segment scenario unit.

[0054] Step S1257: Generate the final dynamic correlation matrix based on the intermediate dynamic correlation matrix after verification and correction. The matrix elements of the dynamic correlation matrix contain traffic flow transmission efficiency related data for the corresponding two road segment scene units.

[0055] After verification and correction, the final dynamic correlation matrix is ​​obtained, which reflects the actual transmission efficiency of traffic flow between different road segment scenario units.

[0056] Step S126: Element-level fusion of the static correlation matrix and the dynamic correlation matrix, weighted according to a preset ratio to calculate the fused correlation values, and generate a fused correlation matrix. The matrix elements of the fused correlation matrix contain structural adaptation and traffic flow transmission related data between road segment scene units.

[0057] To comprehensively consider the impact of structural adaptability and traffic flow transmission efficiency on the correlation between road segment scene units, the static and dynamic correlation matrices are fused. During fusion, elements of both matrices are assigned weights; for example, the static correlation matrix has a weight of 0.4, and the dynamic correlation matrix has a weight of 0.6. Then, the corresponding elements of the two matrices are weighted and summed to obtain the element values ​​in the fused correlation matrix. Larger element values ​​in the fused correlation matrix indicate a higher degree of comprehensive correlation between the two road segment scene units.

[0058] Step S127: Using each road segment scene unit as a network node, and the road segment scene unit pairs corresponding to the non-zero elements in the fusion correlation matrix as the association objects, establish scene unit association relationships between road segment scene units. The association strength of the scene unit association relationship is determined by the corresponding element values ​​in the fusion correlation matrix.

[0059] Each road segment scene unit is considered a node in the video stream scene association network. Node attributes include the road segment scene unit number, geographical range, static features, and dynamic features. Non-zero elements in the fusion association matrix indicate a relationship between corresponding two road segment scene units, which are then treated as association objects, and directed or undirected edges are established between the nodes. The association strength is determined by the value of the corresponding element in the fusion association matrix; the larger the value, the stronger the association.

[0060] Step S128: Construct an initial network skeleton based on the relationship between network nodes and scene units. The initial network skeleton includes all road segment scene unit nodes and their corresponding relationships.

[0061] Based on the determined relationships between network nodes and scene units, an initial skeleton of the video stream scene association network is constructed. This initial skeleton includes all road segment scene unit nodes and the relationships between nodes, forming a preliminary network structure.

[0062] Step S129: Embed the spatiotemporal attribute information of each road segment scene unit into the initial network skeleton. The spatiotemporal attribute information includes the geographical coordinate range of the road segment scene unit and the temporal coverage range of the video stream data, thereby strengthening the spatiotemporal identification of network nodes.

[0063] To enrich the information in network nodes, spatiotemporal attributes such as the geographic coordinate range of road segment scene units and the temporal coverage of video stream data are embedded into the initial network skeleton. The geographic coordinate range can be represented by latitude and longitude polygons, while the temporal coverage records the time period during which the video data of the road segment scene unit was collected. By embedding these spatiotemporal attributes, the spatiotemporal characteristics of the road segment scene units can be reflected more accurately.

[0064] Step S1210: Integrate the network skeleton with embedded spatiotemporal attributes with the association relationship and association strength information of scene units to generate a video stream scene association network. The video stream scene association network includes traffic flow transmission information and event impact association information between scene units of different road segments.

[0065] By integrating the network skeleton embedded with spatiotemporal attributes, the relationship between scene units, and the information on the strength of the relationship, a complete video stream scene association network is formed. This video stream scene association network not only includes the structural and dynamic relationships between road segment scene units, but also reflects the transmission information of traffic flow and the association information of event impact.

[0066] Step S130: Locate abnormal scene units through the video stream scene association network, and model the sudden event transmission path based on the association relationship and association strength information. The sudden event transmission path includes the diffusion path information and impact order information of the abnormal event among scene units in different road segments.

[0067] By utilizing a pre-constructed video stream scene association network, real-time monitoring and analysis of highway traffic scenes are performed. Key features of the traffic scene images for each road segment are extracted and compared with baseline features of normal traffic conditions to identify road segment scene units exhibiting anomalies. Then, based on the association relationships and strength information within the network, the possible propagation direction and path of abnormal events are analyzed. By tracking the transmission of feature difference information within associated road segment scene units, the transmission nodes and sequence of events are determined, constructing a model of the transmission path of sudden events.

[0068] Step S131: Traverse each road segment scene unit in the video stream scene association network, and extract the key features of the traffic scene in the continuous video stream data corresponding to each road segment scene unit. The key features of the traffic scene include the integrity of vehicle shape, the existence status of road obstacles, and the orderly status of traffic flow.

[0069] For each road segment scene unit in the video stream scene association network, the system periodically extracts keyframes from its corresponding continuous video stream data. Image analysis is then performed on these keyframes to extract key features of the traffic scene. Vehicle morphological integrity refers to detecting vehicles in the video using object detection algorithms and assessing whether their morphology is intact and whether there are any collisions or damage. Road obstacle presence status is determined using background subtraction and object recognition techniques to detect whether there are obstacles on the road surface other than normally traveling vehicles. Traffic flow order status is determined by analyzing parameters such as vehicle direction, speed, and spacing to assess whether the traffic flow is orderly and whether there are any abnormalities such as congestion or wrong-way driving.

[0070] Step S132: Retrieve the normal traffic state baseline features corresponding to each road segment scene unit. The normal traffic state baseline features are the statistical average of the key features of the traffic scene image of the road segment scene unit when no abnormal events occur.

[0071] The system pre-stores baseline feature data for each road segment scene unit under normal traffic conditions. These baseline features are obtained through analysis and statistics of historical video stream data, including the average level of vehicle morphological integrity, the probability of road obstacles, and the range of various parameters related to the orderly state of traffic flow. During anomaly detection, the key features extracted in real time are compared with these baseline features to determine whether the current traffic state is abnormal.

[0072] Step S133: Compare the key features of the traffic scene image of each road segment scene unit with the baseline features of normal traffic state, and extract feature difference information. The feature difference information includes the type, location and duration of the difference features.

[0073] Key features extracted from real-time traffic scene footage are compared dimension-by-dimensionally with baseline features of normal traffic conditions. If the value of a feature exceeds the normal range of the baseline feature, a feature difference is considered to exist. The type of the difference feature is recorded, such as abnormal vehicle shape, road obstacles, or traffic flow disorder; the location of occurrence, i.e., the specific coordinates of the difference feature in the video footage; and the duration, i.e., the length of time from the appearance of the difference feature to the present. This feature difference information is an important basis for determining whether a road segment scene unit is abnormal.

[0074] Step S134: Based on the feature difference information, select road segment scene units whose feature differences exceed the preset range, mark the road segment scene units as abnormal scene units, and collect all abnormal scene units to form an abnormal scene unit set.

[0075] Based on preset thresholds, feature difference information is filtered. If the degree of feature difference exceeds the set threshold, such as vehicle morphological integrity falling below a certain level, road obstacles persisting for a certain period of time, or traffic flow order parameters exceeding the normal range, the corresponding road segment scene unit is marked as an abnormal scene unit. All road segment scene units marked as abnormal are collected to form an abnormal scene unit set for subsequent event propagation path analysis.

[0076] Step S135: Extract the scene unit association relationship and association strength information of each abnormal scene unit in the abnormal scene unit set, and determine the directly associated road segment scene units and indirectly associated road segment scene units of each abnormal scene unit.

[0077] For each anomalous scene unit in the set of anomalous scene units, its association relationships and association strength information are queried from the video stream scene association network. Directly associated road segment scene units refer to road segment scene units that have a direct association relationship with the anomalous scene unit; the association strength between them is usually high. Indirectly associated road segment scene units refer to other road segment scene units that have an association relationship with the directly associated road segment scene unit. By identifying these associated units, the potential scope and direction of the anomalous event's spread can be understood.

[0078] Step S136: Analyze the transmission data of the feature difference information of the abnormal scene unit in the directly related road segment scene unit, and track the start time and diffusion process data of the feature difference information spreading from the abnormal scene unit to the directly related road segment scene unit.

[0079] Step S1361: Extract the start time from the feature difference information of the abnormal scene unit, and establish a time tracking coordinate system based on the start time.

[0080] The time of the first occurrence of feature differences is obtained from the feature difference information of the abnormal scene unit. This time is used as the origin of the time tracking coordinate system for subsequent tracking of the diffusion time of feature difference information.

[0081] Step S1362: Obtain the association strength information corresponding to the scene unit association relationship between the abnormal scene unit and each directly associated road segment scene unit, and sort the directly associated road segment scene units in descending order of association strength value.

[0082] Based on the correlation strength between anomalous scene units and directly related road segment scene units in the video stream scene association network, the directly related road segment scene units are ranked. Units with stronger correlation are more likely to be affected by anomalous events and should be prioritized for monitoring and analysis.

[0083] Step S1363: For each directly related road segment scene unit after sorting, retrieve the continuous video stream data of that road segment scene unit after the reference point, and analyze the key features of the traffic scene frame by frame in chronological order.

[0084] Based on the sorting results, video stream data for each directly related road segment scene unit is retrieved sequentially after the time when the feature differences of the abnormal scene unit occur. The video stream data is then analyzed frame by frame in chronological order to extract key features of the traffic scene, focusing on whether similar feature differences to those of the abnormal scene unit appear.

[0085] Step S1364: Identify the time point at which the key features of the traffic scene image of the directly related road segment scene unit first appear with the same type of feature difference information as the abnormal scene unit, and record this time point as the diffusion start time of the directly related road segment scene unit.

[0086] When analyzing video stream data of directly related road segment scene units, once a feature with the same type of difference as the abnormal scene unit is found in the key features of the traffic scene, the time point when the feature first appears is recorded. This time point is the diffusion start time of the directly related road segment scene unit, indicating that the abnormal event has begun to spread to this unit.

[0087] Step S1365: Starting from the diffusion start time, continuously track the change data of the corresponding type of difference features in the key features of the traffic scene image of the directly related road segment scene unit, and record the data of the expansion of the occurrence range of the difference features, the data of the intensity change, and the data of the superposition with other features.

[0088] Starting from the diffusion initiation time, continuously monitor the key features of traffic scene images of directly related road segment scene units. Record how the occurrence range of differential features expands, for example, from one lane to multiple lanes; how the intensity changes, for example, the size of obstacles increases or the degree of traffic flow congestion intensifies; and whether they appear in combination with other types of differential features, such as the simultaneous occurrence of obstacles and traffic flow turbulence.

[0089] Step S1366: Compare the feature difference information of the abnormal scene unit with the difference features tracked in the directly related road segment scene unit, analyze the consistency data of the two in terms of type, manifestation and change trend, and confirm the correlation data of feature transmission.

[0090] The differential features tracked in directly related road segment scene units are compared with the feature difference information of abnormal scene units to analyze whether they are consistent in type, similar in manifestation, and have the same trend of change. If these aspects show high consistency, the correlation of feature transmission can be confirmed, that is, the anomaly of the directly related road segment scene unit is caused by the feature difference transmission of the abnormal scene unit.

[0091] Step S1367: Based on the time tracking coordinate system, compare the diffusion start time of each directly associated road segment scene unit with the feature difference start time of the abnormal scene unit, and calculate the diffusion time difference, which includes the time interval data of feature transmission.

[0092] Using a time-tracking coordinate system, the time difference between the diffusion start time of each directly associated road segment scene unit and the start time of the feature difference of the abnormal scene unit is calculated. This time difference reflects the time interval required for the feature difference to be transmitted from the abnormal scene unit to the directly associated road segment scene unit.

[0093] Step S1368: Integrate the diffusion start time, difference feature change data and diffusion time difference of each directly related road segment scene unit to generate a feature transfer record of a single directly related road segment scene unit.

[0094] The diffusion start time, variation data of differential features, and diffusion time difference of each directly related road segment scene unit are integrated together to form the feature transmission record of that unit, which records in detail the transmission process of abnormal features in that unit.

[0095] Step S1369: Collect feature transfer records of all directly related road segment scene units, sort the results by association strength and diffusion time difference, and generate a feature transfer data summary.

[0096] The feature transmission records of all directly related road segment scene units are collected and sorted first according to the association strength from large to small. If the association strength is the same, they are then sorted according to the diffusion time difference from small to large to form a feature transmission data summary, so as to show the transmission of abnormal features in different directly related road segment scene units.

[0097] Step S13610: Based on the feature transfer data summary, output the start time and diffusion process data of feature difference information spreading from abnormal scene units to directly related road segment scene units.

[0098] Based on the feature transfer data, the starting time and specific diffusion process data of the diffusion of feature difference information from the abnormal scene unit to each directly related road segment scene unit are compiled, including the changes in the difference features and the diffusion speed.

[0099] Step S137: Based on the start time of the diffusion of feature difference information from abnormal scene units to each directly related road segment scene unit, as well as the spatial distance and traffic flow relationship between each directly related road segment scene unit and the abnormal scene unit, determine the diffusion priority of feature difference information among different related road segment scene units.

[0100] Taking into account factors such as the start time of diffusion of feature difference information, the spatial distance between related road segment scene units and abnormal scene units, and traffic flow relationships, the diffusion priority of abnormal features among different related road segment scene units is determined. Road segment scene units with earlier diffusion start time, shorter spatial distance, and smoother traffic flow have higher diffusion priority, and abnormal events are more likely to affect these units first.

[0101] Step S138: Record the road segment scene units through which the feature difference information diffuses according to the diffusion priority, and generate a transmission node sequence. Each node in the transmission node sequence corresponds to a road segment scene unit affected by the feature difference information.

[0102] Step S1381: Take the abnormal scene unit as the starting node of the transmission node sequence, and record the initial state of the road scene unit number, geographical location and feature difference information of the starting node.

[0103] The abnormal scene unit is used as the first node in the transmission node sequence. Its road segment scene unit number, central geographical coordinates, and initial state of feature difference information, such as the type and intensity of the difference features, are recorded.

[0104] Step S1382: Based on the diffusion priority sorting results, select the highest priority directly associated road segment scene unit as the first transmission node, and record the road segment scene unit number, geographical location, and feature difference information status transmitted from the starting node for this transmission node.

[0105] Based on the diffusion priority ranking results, the highest priority directly associated road segment scene unit is selected as the second node in the transmission node sequence. Its number, geographical location, and feature difference information status transmitted from the abnormal scene unit are recorded, including diffusion start time, type and intensity of difference features, etc.

[0106] Step S1383: Analyze the indirect associated road segment scene units of the first transmission node, and in combination with the diffusion priority ranking results, select the next highest priority associated road segment scene unit as the second transmission node, and record the relevant information and feature difference information of the transmission node.

[0107] For the first transmission node, analyze its indirectly related road segment scene units, i.e., other road segment scene units that are related to the first transmission node. Based on the diffusion priority ranking results, select the next highest priority indirectly related road segment scene unit as the third transmission node, and record its relevant information and the changes in its feature differences during the transmission process, such as whether the intensity weakens or the type changes.

[0108] Step S1384: Select subsequent transmission nodes in sequence. The selection of each transmission node is based on the associated road segment scene unit and diffusion priority sorting result of the previous transmission node, so that the transmission order conforms to the diffusion logic.

[0109] Following the method described above, subsequent propagation nodes are selected sequentially. The selection of each propagation node is based on the associated road segment scene unit of the previous propagation node and refers to the diffusion priority sorting results to ensure that the order of propagation nodes conforms to the diffusion logic of abnormal events and can accurately reflect the propagation path of events.

[0110] Step S1385: When recording the information of each transmission node, simultaneously record the time interval for the characteristic difference information to be transmitted from the previous transmission node to the current transmission node, the details of the changes during the transmission process, and the correlation strength information.

[0111] When recording information for each transmission node, detailed records are kept of the time interval taken for characteristic difference information to be transmitted from the previous transmission node to the current transmission node, details of changes in the difference characteristics during the transmission process, such as increase or decrease in intensity, expansion or contraction of range, and the correlation strength information between the two nodes. This information helps to analyze the speed and degree of impact of the event's spread.

[0112] Step S1386: For multiple associated road segment scene units with the same diffusion priority, select them as subsequent transmission nodes in order of their spatial distance from the current transmission node from near to far.

[0113] When multiple related road segment scene units have the same diffusion priority, they are selected in order of their spatial distance from the current transmission node from near to far. The units that are closer are given priority as subsequent transmission nodes, because the closer the distance, the greater the possibility of being affected and the faster the diffusion speed may be.

[0114] Step S1387: If a certain associated road segment scene unit has already been recorded as a transmission node, it will not be recorded again.

[0115] To avoid duplicate road segment scene units in the transmission node sequence, it is necessary to check whether the unit has already been recorded when selecting a transmission node. If it has already been recorded, it will not be used as a new transmission node to ensure the uniqueness and accuracy of the transmission node sequence.

[0116] Step S1388: During the construction of the transmission node sequence, the scene unit association relationship in the video stream scene association network is linked in real time, and the association validity between each transmission node and the previous node is checked.

[0117] During the construction of the transmission node sequence, it is necessary to continuously refer to the scene unit association relationships in the video stream scene association network to ensure that each transmission node has a valid association relationship with the previous node. If there is no association relationship between two nodes or the association strength is too low, it may indicate that the event will not propagate along the path, and it is necessary to select a new transmission node.

[0118] Step S1389: When the intensity of the feature difference information of the transmission node is lower than the preset threshold or there is no subsequent associated road segment scene unit, stop adding new transmission nodes and determine the termination node of the transmission node sequence.

[0119] When the intensity of the characteristic difference information of a certain transmission node decreases to below a preset threshold, it indicates that the impact of the abnormal event has weakened to a negligible level, or when the transmission node has no subsequent associated road segment scene units, the addition of new transmission nodes is stopped, and this node is the termination node of the transmission node sequence.

[0120] Step S13810: Organize all recorded transmission node information and arrange them in the transmission order to form a transmission node sequence. The transmission node sequence contains relevant data of all road segment scene units through which the feature difference information diffuses.

[0121] All recorded transmission nodes are arranged in the transmission order to form a complete transmission node sequence. This transmission node sequence contains relevant data such as the number, geographical location, diffusion time, and feature difference status of all road segment scene units traversed by the diffusion of feature difference information, demonstrating the diffusion path of abnormal events in the highway network.

[0122] Step S139: Extract the feature difference information, diffusion time and correlation strength information of each node in the transmission node sequence, and mark them between the corresponding nodes in the transmission node sequence to generate a transmission link containing time and strength dimensions.

[0123] The propagation node sequence extracts the characteristic difference information of each node, namely the diffusion time (the time when the node begins to be affected by the abnormal event) and the correlation strength information between the node and the previous node during diffusion. This information is then labeled between corresponding nodes in the propagation node sequence to form a propagation link. The propagation link not only includes the connection relationship between nodes but also information in the time dimension (diffusion time) and the strength dimension (correlation strength), allowing us to gain a more comprehensive understanding of the event's diffusion process.

[0124] Step S1310: Integrate the transmission node sequence and transmission link to construct the transmission path of the emergency event. The transmission path of the emergency event includes the diffusion path information and the order of influence information of the abnormal event between different road segment scene units.

[0125] By integrating the sequence of transmission nodes and the transmission links, a complete model of the transmission path of an emergency is constructed. This model demonstrates the propagation path of an abnormal event between different road segment scene units, as well as the order and timing of the impact on each road segment scene unit.

[0126] Step S140: Collect basic information on control resources across the entire highway area, dynamically adapt control resources and transmission paths based on the spatiotemporal attributes of emergency transmission paths, and generate a control resource adaptation plan.

[0127] To effectively handle emergencies on highways, it is necessary to understand the overall situation of control resources. The basic information collected on control resources includes: resource type, such as traffic police, breakdown vehicles, ambulances, and maintenance equipment; functional scope, i.e., the specific tasks that each resource can perform; deployment location, indicating the current location of the resource; availability status, such as whether it is on standby, busy, or under maintenance; and response initiation time, i.e., the preparation time from receiving the instruction to being able to depart.

[0128] After acquiring basic information about the controlled resources, dynamic adaptation is performed based on the spatiotemporal attributes of the emergency transmission path. First, the spatiotemporal attributes of the transmission path are analyzed, including the starting and ending road segment scene units, the direction of spread, the spread speed, and the duration of the impact at each transmission node. Then, based on the deployment geographical location and effective response radius of the controlled resources, the spatial distance from the resources to each transmission node is calculated, and candidate resources that can be reached within the effective timeframe are selected. The functional scope of the resources is considered to ensure it matches the type of abnormal impact at the transmission nodes, further narrowing down the range of candidate resources.

[0129] Based on the propagation speed and duration of the impact along the transmission path, a response time window for each candidate resource is calculated to ensure that the resource can function during the period when nodes are affected. Combining the propagation direction and resource deployment location, a reasonable scheduling route is generated to avoid resource waste and route conflicts. Based on the order of the propagation nodes and the duration of the impact, the response order of resources is determined, prioritizing resources that can quickly reach critical nodes. Simultaneously, the response order and routes are adjusted according to the availability of resources, providing supporting measures for partially available resources. Finally, considering that the event propagation speed may change, dynamic adjustment clauses are added to the plan to cope with unforeseen circumstances, generating the final resource management and adaptation plan.

[0130] Step S141: Collect basic information on management and control resources for the entire highway area. The basic information on management and control resources includes the type, functional scope, geographical location of deployment, availability status and response start time of the management and control resources.

[0131] Information on controlled resources across the entire highway management system is collected through its resource database. These resources are diverse, including but not limited to traffic police, tow trucks, medical rescue vehicles, road maintenance equipment, and emergency communication equipment. Each resource has a specific functional scope; for example, tow trucks are primarily responsible for towing disabled vehicles, traffic police are responsible for traffic control and on-site command, and medical rescue vehicles provide emergency medical assistance. Deployment geolocation is obtained through the Global Positioning System (GPS) or other positioning methods, accurate to specific latitude and longitude coordinates. Availability status is updated in real time, including standby, in operation, and maintenance status. Response start time is determined based on the resource type and current status; for example, a tow truck in standby status may have a shorter response start time, while equipment under maintenance may have a longer response start time. This information is organized into structured data and stored in the resource information database.

[0132] Step S142: Analyze the spatiotemporal attributes of the emergency transmission path, and extract the starting segment scene unit, ending segment scene unit, diffusion direction, diffusion speed, and the duration of influence corresponding to each transmission node of the emergency transmission path.

[0133] A thorough analysis of the constructed emergency transmission path is conducted to extract its spatiotemporal attributes. The initial road segment scene unit is the road segment where the abnormal event initially occurs, and the final road segment scene unit is the final impact range of the event's spread. The spread direction is determined by analyzing the changes in the geographical location of the transmission nodes, such as from west to east or from south to north. The spread speed is calculated based on the distance between transmission nodes and the spread time, reflecting the rate of event spread. The duration of impact corresponding to each transmission node is determined based on the duration of characteristic difference information at that node, i.e., the time interval from the start of the spread to the point where the intensity of the characteristic difference information decreases below a threshold. These spatiotemporal attributes are crucial for the rational allocation and control of resources, helping to determine the scheduling order and arrival time of resources.

[0134] Step S143: Based on the deployment geographic location, effective response radius, and transmission node geographic location of the control resource and the emergency transmission path, calculate the spatial distance from the control resource to each transmission node, filter out control resources whose spatial distance is less than or equal to the effective response radius of the control resource, and generate a candidate control resource set.

[0135] For each type of control resource, the spatial distance from the resource to each transmission node in the emergency transmission path is calculated based on its geographical location and preset effective response radius. The spatial distance can be calculated using distance calculation functions in a Geographic Information System (GIS), such as the Haversine formula based on latitude and longitude. Control resources with spatial distances less than or equal to the effective response radius are selected; these resources are considered capable of reaching the transmission nodes within a reasonable timeframe, forming a candidate control resource set. The effective response radius is determined based on the resource type and performance; for example, the effective response radius of a tow truck may be larger than that of a traffic police officer on foot patrol.

[0136] Step S144: Based on the spread speed of the emergency transmission path and the duration of the impact of each transmission node, calculate the response time window for each candidate control resource. The response time window contains the time range data of the control resource being able to arrive and take effect within the duration of the impact of the transmission node.

[0137] The response time window refers to the time range within which control resources can arrive and begin functioning when a transmission node is affected by an abnormal event. Calculating the response time requires considering the resource's response start-up time and travel time. The travel time is calculated based on the spatial distance from the resource to the transmission node and the resource's average travel speed. The start time of the response time window is the start time of the transmission node's impact, and the end time is the end time of the transmission node's impact minus the response time. If the start time of the response time window is earlier than the end time, the resource is effective within the response time window of that transmission node; otherwise, the resource cannot arrive during the period when the node is affected.

[0138] Step S145: Based on the functional scope of the control resources and the abnormal impact type of each transmission node, match the functional requirements of the candidate control resources with those of the transmission nodes, select control resources whose functional scope can cover the abnormal impact handling requirements of the transmission nodes, and update the candidate control resource set.

[0139] Step S1451: Extract the abnormal impact type of each transmission node. The abnormal impact types include vehicle malfunction lane occupation impact, road surface damage impact, traffic flow congestion impact, and foreign object intrusion impact.

[0140] Based on the characteristic differences of the transmission nodes, the type of abnormal impact for each node is determined. For example, if the characteristic difference information of a node shows abnormal vehicle morphology and traffic flow disorder, it may be due to a vehicle malfunction blocking the lane; if an obstacle is detected on the road surface and traffic flow is obstructed, it may be due to foreign object intrusion. Accurately extracting the type of abnormal impact is the foundation for achieving resource function matching.

[0141] Step S1452: Extract the functional scope related data for each candidate control resource. The functional scope related data includes the types of abnormal impacts that each candidate control resource can handle, the range of its processing capabilities, and the applicable road segment scenario conditions.

[0142] The functional scope data for each candidate resource is extracted from the basic information on managed resources. For example, the functional scope of a breakdown vehicle includes handling the impact of vehicle breakdowns blocking lanes, has a certain towing capacity, and is suitable for various road sections and scenarios; a road maintenance vehicle can handle the impact of road damage and is suitable for specific types of road structures. The handling capacity range and applicable conditions also need to be clearly defined to ensure that the resources can play an effective role.

[0143] Step S1453: Establish a matching table between the types of abnormal impacts and the functional scope of control resources. The table contains a description of the appropriate control resource functions for each type of abnormal impact.

[0144] Based on historical experience and resource characteristics, a matching table is established to map the types of abnormal impacts to the functional scope of controlled resources. For example, a vehicle breakdown blocking the road corresponds to the towing function of a breakdown vehicle and the traffic control function of traffic police; road damage corresponds to the repair function of road maintenance vehicles and engineering teams, etc.

[0145] Step S1454: For each transmission node, query the matching table according to its abnormal impact type to obtain the control resource function conditions required by the transmission node.

[0146] For each transmission node, the matching lookup table is consulted based on its abnormal impact type to determine which control resources and functions the node requires. For example, a node affected by traffic congestion needs resources with traffic diversion capabilities.

[0147] Step S1455: Check each control resource in the candidate control resource set one by one to determine whether its functional scope-related data fully includes the control resource functional conditions required by the transmission node.

[0148] For each resource in the candidate control resource set, check whether its functional scope can fully meet the functional requirements of the transmission node. If the functional scope of a resource includes all the functional conditions required by the node, then the resource is considered a match for the node.

[0149] Step S1456: Mark the control resources whose functional scope-related data can fully contain the required functional conditions as adaptive control resources and retain them in the candidate control resource set.

[0150] Reserve the control resources that are a perfect match in the candidate set; these resources are ideal for handling the impact of anomalies in the corresponding propagation nodes.

[0151] Step S1457: For control resources whose functional scope-related data contains the required functional conditions, further analyze whether their missing functions can be supplemented by other supporting resources. If they can be supplemented, mark the control resource and supporting resources as adaptive control resources and retain them in the candidate control resource set.

[0152] For partially matched resources, analyze whether the missing functions can be supplemented by other resources. For example, a tow truck has towing capabilities but lacks traffic management capabilities. It can be combined with traffic police resources to form a combined resource that meets the functional requirements. The combined resource is also marked as an adapted management resource.

[0153] Step S1458: Remove control resources from the candidate control resource set if the data related to the functional scope does not contain the required functional conditions and cannot be supplemented by supporting resources.

[0154] Resources that are completely mismatched and cannot be supplemented by combination will be removed from the candidate set to reduce unnecessary resource allocation considerations.

[0155] Step S1459: For each transmission node, summarize the filtered adaptation and control resources to generate a subset of exclusive adaptation and control resources corresponding to that transmission node.

[0156] All adapted management and control resources are aggregated for each transmission node to form a subset of adapted resources for that node, which facilitates subsequent resource scheduling and allocation.

[0157] Step S14510: Integrate the exclusive adaptation control resource subsets of all transmission nodes, remove duplicate control resources, and generate an updated candidate control resource set. The control resources in the updated candidate control resource set can meet the functional requirements of at least one transmission node.

[0158] The dedicated adaptation resource subsets of all propagation nodes are integrated, and duplicate resources are removed to obtain an updated candidate control resource set. All resources in this updated candidate control resource set can meet the functional requirements of at least one propagation node.

[0159] Step S146: Analyze the diffusion direction of the emergency transmission path, combine the deployment geographical location of each control resource in the updated candidate control resource set, incorporate the geographical range data of the road segment scene unit that has been affected by the abnormal event, and generate the scheduling route.

[0160] Based on the propagation direction of the emergency's transmission path, the direction of resource allocation for control is determined. Combining the geographical locations of candidate control resources, a scheduling route is planned from the current location of the resource to the target transmission node. When planning the route, it is necessary to avoid the geographical area of ​​road segments and scene units already affected by the abnormal event to ensure that resources can reach their destination quickly and safely. The path planning function of a Geographic Information System (GIS) can be used to generate the optimal scheduling route by comprehensively considering factors such as road conditions and traffic control.

[0161] Step S147: Determine the response order of each control resource based on the sequence of transmission nodes in the transmission path of the emergency and the duration of the impact of each transmission node. The response order of the control resources is correlated with the diffusion rhythm of the transmission path and the duration of the impact of each transmission node.

[0162] The response order of control resources should be determined based on the sequence of nodes in the transmission path and the duration of each node's impact. Typically, nodes affected first require priority response to control the spread of the event as quickly as possible. Simultaneously, the duration of each node's impact should be considered; if a node's impact is short-lived, resources need to arrive and handle it quickly. The determined response order should match the pace of the event's spread to ensure resources can function effectively and promptly.

[0163] Step S148: Adjust the response order and scheduling route based on the availability status of the managed resources, prioritize scheduling managed resources with good availability status, and supplement supporting measures for managed resources with partial availability status to ensure their normal functioning.

[0164] After determining the response order and scheduling route, adjustments need to be made based on the actual availability of the managed resources. Resources with good availability should be scheduled first; for resources with partial availability, their limiting factors should be analyzed, and necessary supporting measures should be provided, such as supplementing equipment and allocating personnel, to ensure that they can function normally. If a resource is currently busy, it is necessary to assess whether its task completion time is within the response time window. If so, it should be included in the scheduling plan; otherwise, the response order should be adjusted, and other available resources should be selected.

[0165] Step S149: Integrate the scheduling routes, response sequences, and supporting measures of the control resources to generate a preliminary control resource adaptation plan. The preliminary control resource adaptation plan includes a list of control resources and execution requirements for each transmission node.

[0166] By integrating information such as the scheduling routes, response sequences, and supporting measures for managed resources, a preliminary managed resource adaptation plan is formed. The plan clearly defines the managed resource list for each transmission node, including resource type, quantity, scheduling route, estimated arrival time, and resource execution requirements, such as task content and operating procedures.

[0167] Step S1410: Based on the changing trend of the spread speed of the emergency transmission path, add dynamic adjustment clauses to the preliminary control resource adaptation plan to generate the final control resource adaptation plan. The dynamic adjustment clauses include relevant operational specifications for adjusting the response order and scheduling route of control resources according to the changes in the spread speed.

[0168] For example, step S14101: Extract the diffusion speed data of the transmission path of the sudden event, arrange them in chronological order to form a diffusion speed sequence, and the diffusion speed sequence contains the diffusion speed values ​​corresponding to different time points.

[0169] Data on the spread rate is extracted from the emergency propagation path model and arranged chronologically to form a spread rate sequence. Analyzing this sequence allows us to understand how the event's spread rate changes, such as whether it is increasing, remaining stable, or decreasing.

[0170] Step S14102: Analyze the changing trend of the diffusion rate sequence, identify the rising phase, the stable phase and the falling phase of the diffusion rate, and determine the start time, duration and magnitude of the rate change for each phase.

[0171] Trend analysis was performed on the diffusion velocity sequence, and the rate of change of velocity over different time periods was calculated using methods such as sliding windows to identify the rising, stable, and declining phases of the diffusion velocity. The start time, duration, and magnitude of velocity change for each phase were determined.

[0172] Step S14103: Based on the change range of the diffusion rate increase phase and the statistical data of historical emergencies, set the first type of dynamic adjustment trigger condition. The first type of dynamic adjustment trigger condition is that the diffusion rate increase exceeds the increase rate threshold obtained from historical data statistics.

[0173] By referencing statistical data from historical emergencies, the normal range of variation during the escalation phase of the spread is determined. When the current increase in the spread rate exceeds the threshold set based on historical data, the first type of dynamic adjustment is triggered.

[0174] Step S14104: For the first type of dynamic adjustment triggering conditions, construct a corresponding adjustment strategy. The adjustment strategy includes operations related to advance control of resource response time, optimization of scheduling routes to shorten travel distance, and increase of control resources for key transmission nodes.

[0175] When the first type of dynamic adjustment condition is triggered, corresponding adjustment strategies are adopted. For example, the response time of resources is controlled in advance by reducing preparation time or choosing faster modes of transportation; scheduling routes are optimized to avoid congested sections and shorten travel distances; and the number of control resources at key transmission nodes is increased to enhance handling capabilities and prevent the incident from spreading further.

[0176] Step S14105: Based on the duration of the steady-state diffusion phase and the event type experience value, set a second type of dynamic adjustment trigger condition. The second type of dynamic adjustment trigger condition is that the duration of the steady-state phase exceeds the duration threshold determined based on the event type experience.

[0177] Based on empirical data from different types of emergencies, the normal duration of the stable phase of the spread rate is determined. When the duration of the stable phase exceeds a set threshold, a second type of dynamic adjustment is triggered.

[0178] Step S14106: For the second type of dynamic adjustment triggering conditions, construct a corresponding adjustment strategy. The adjustment strategy includes operations related to maintaining the current response order of control resources, optimizing the working coordination mode of control resources, and supplementing the control resource reserves of subsequent transmission nodes.

[0179] If the stable propagation phase lasts too long, it indicates that the event may be in a stalemate or requires more time to handle. In this case, maintain the current resource response sequence while optimizing resource coordination to improve efficiency. Furthermore, replenish resource reserves at subsequent propagation nodes to cope with potential re-propagation of the event.

[0180] Step S14107: Based on the change in the diffusion rate during the decline phase and the evaluation of the event handling effect, set a third type of dynamic adjustment trigger condition. The third type of dynamic adjustment trigger condition is that the diffusion rate decline exceeds the decline rate threshold determined according to the event handling effect evaluation.

[0181] Based on the assessment of the incident response effectiveness, a reasonable range of variation in the rate of decrease in the spread is determined. When the rate of decrease exceeds a set threshold, it indicates that the incident response has been effective, triggering the third type of dynamic adjustment.

[0182] Step S14108: For the third type of dynamic adjustment triggering conditions, construct a corresponding adjustment strategy. The adjustment strategy includes operations such as delaying the response time of some non-critical control resources, adjusting the scheduling route to avoid resource concentration areas, and reducing the investment of redundant control resources.

[0183] When the rate of incident spread slows significantly, the response time of some non-critical control resources can be appropriately delayed to avoid resource waste. Adjust scheduling routes to avoid areas where resources are already concentrated, reducing traffic congestion. Simultaneously, reduce redundant control resource deployment and allocate surplus resources to other areas where they are needed.

[0184] Step S14109: Organize the three types of dynamic adjustment trigger conditions and corresponding adjustment strategies into dynamic adjustment clauses, which include relevant data on applicable scenarios, triggering processes and execution requirements.

[0185] The three types of dynamic adjustment triggering conditions and corresponding adjustment strategies mentioned above are compiled into standardized dynamic adjustment clauses, clearly defining the applicable scenarios, triggering procedures, and execution requirements for each clause. For example, the applicable scenarios describe the conditions under which the clause is triggered, the triggering procedures specify the steps from condition detection to strategy execution, and the execution requirements clarify the specific content and standards of the adjustment operation.

[0186] Step S141010: Embed the dynamic adjustment clauses into the preliminary control resource adaptation plan, and organically combine them with the original scheduling routes, response sequences and supporting measures to generate the final control resource adaptation plan.

[0187] Integrating dynamic adjustment clauses into the initial control resource adaptation plan gives the plan flexibility to respond to changing events. During plan execution, the system can automatically or manually trigger corresponding dynamic adjustment clauses based on real-time monitoring of changes in the spread rate, adjusting the response sequence and scheduling routes of control resources to ensure the effectiveness and adaptability of the plan.

[0188] Step S150: Generate intelligent auxiliary decision-making instructions based on the control resource adaptation scheme. The intelligent auxiliary decision-making instructions include control resource scheduling instructions and road traffic control instructions, which are used to guide the on-site control equipment and the back-end control system to coordinate and execute emergency response operations.

[0189] Once the resource management and control adaptation plan is determined, it is converted into specific intelligent auxiliary decision-making instructions. The resource management and control scheduling instructions specify the destination, route, arrival time, and task content for each resource, such as "Dispatch tow truck A to road segment scenario unit 3, arrive before 10:30, and tow away the disabled vehicle." Road segment traffic control instructions then formulate traffic control measures for the affected road segment scenario units, such as "Close the overtaking lane of road segment scenario unit 2, with a speed limit of 60 km / h," and "Set up temporary traffic guidance signs in road segment scenario unit 5."

[0190] These instructions are transmitted via communication networks to on-site control equipment and the back-end control system. On-site control equipment, such as handheld terminals used by traffic police and onboard systems in tow trucks, receives the instructions and executes the corresponding operations. The back-end control system coordinates the work between various departments based on the instructions, monitors the progress of incident handling in real time, and adjusts decisions based on feedback. Through the issuance and execution of intelligent decision-making instructions, the collaborative work between on-site control equipment and the back-end control system is achieved, enabling efficient handling of highway emergencies and minimizing their impact.

[0191] In the above embodiments, the intelligent auxiliary decision-making system for highway emergency management, used to perform the above method embodiments, has at least one processor, a control module (chipset) coupled to at least one of the processors, a memory coupled to the control module, a non-volatile memory (NVM) / storage device coupled to the control module, at least one load to / output device coupled to the control module, and a network interface coupled to the control module.

[0192] This invention discloses a computer read storage medium that stores a computer program for electronic data interchange, wherein the computer program causes a computer to execute the steps in the intelligent auxiliary decision-making method for highway emergency management described in the foregoing embodiments.

[0193] Finally, it should be noted that the above-disclosed embodiments are merely preferred embodiments of the present invention and are only used to illustrate the technical solutions of the present invention, not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent auxiliary decision-making method for emergency management on highways, characterized in that, The method includes: Acquire continuous video stream data collected by monitoring equipment deployed throughout the highway, wherein the continuous video stream data includes real-time traffic scene recordings of each road segment at different times; A video stream scene association network is constructed based on the spatiotemporal distribution characteristics of continuous video stream data. The video stream scene association network includes road segment scene units, scene unit association relationships, and association strength information. The scene unit association relationships include traffic flow transmission information and event impact association information between different road segment scene units. Abnormal scene units are located by video stream scene association network, and the transmission path of sudden events is modeled based on association relationship and association strength information. The transmission path of sudden events includes the diffusion path information and the order of influence information of abnormal events among scene units in different road segments. Collect basic information on control resources across the entire highway network, dynamically adapt control resources and transmission paths based on the spatiotemporal attributes of emergency transmission paths, and generate control resource adaptation schemes. Intelligent auxiliary decision-making instructions are generated based on the resource adaptation scheme. These instructions include resource scheduling instructions and road traffic control instructions, which are used to guide on-site control equipment and the back-end control system to coordinate and execute emergency response operations.

2. The intelligent auxiliary decision-making method for highway emergency management according to claim 1, characterized in that, The construction of a video stream scene association network based on the spatiotemporal distribution features of continuous video stream data includes: According to the highway segment division standard, the monitoring coverage area corresponding to the continuous video stream data is divided into multiple independent segment scene units. Each segment scene unit corresponds to a fixed length of highway segment and the complete monitoring screen range of that highway segment. Extract the static association elements of each road segment scene unit. The static association elements include the number of lanes in the road segment corresponding to the road segment scene unit, the road surface structure type, the connection method of adjacent road segments, and the deployment density of fixed monitoring equipment. Extract the dynamic correlation elements of each road segment scene unit. The dynamic correlation elements include the traffic flow direction, traffic flow density change trend and traffic status change at different times within the road segment scene unit. The structural fit degree between different road segment scene units is calculated based on static correlation elements. The structural fit degree includes the connection and matching data of the road segment scene units on the road structure. A static correlation matrix is ​​generated, and the matrix elements of the static correlation matrix contain the structural fit degree related data of the corresponding two road segment scene units. Based on the dynamic correlation elements, the traffic flow transmission efficiency between different road segment scene units is calculated, and a dynamic correlation matrix is ​​generated. The traffic flow transmission efficiency includes relevant data on the smooth transmission of traffic flow between different road segment scene units, and the matrix elements of the dynamic correlation matrix contain relevant data on the traffic flow transmission efficiency of corresponding two road segment scene units. The static and dynamic correlation matrices are fused at the element level, and the fused correlation values ​​are calculated according to a preset weighting ratio to generate a fused correlation matrix. The matrix elements of the fused correlation matrix contain data related to the structural adaptation and traffic flow transmission between road segment scene units. Each road segment scene unit is used as a network node, and the road segment scene unit pairs corresponding to the non-zero elements in the fusion correlation matrix are used as the association objects to establish scene unit association relationships between road segment scene units. The association strength of the scene unit association relationship is determined by the corresponding element values ​​in the fusion correlation matrix. An initial network skeleton is constructed based on the relationship between network nodes and scene units. The initial network skeleton includes all road segment scene unit nodes and their corresponding relationships. The spatiotemporal attribute information of each road segment scene unit is embedded into the initial network skeleton. The spatiotemporal attribute information includes the geographical coordinate range of the road segment scene unit and the temporal coverage range of the video stream data, thereby strengthening the spatiotemporal identification of network nodes. By integrating the network skeleton with embedded spatiotemporal attribute information and the association relationship and association strength information of scene units, a video stream scene association network is generated. The video stream scene association network includes traffic flow transmission information and event impact association information between scene units of different road segments.

3. The intelligent auxiliary decision-making method for emergency management of highways according to claim 1, characterized in that, The step of locating abnormal scene units through a video stream scene association network and modeling the transmission path of sudden events based on association relationships and association strength information includes: Traverse each road segment scene unit in the video stream scene association network, and extract key features of traffic scene images from the continuous video stream data corresponding to each road segment scene unit. The key features of traffic scene images include vehicle shape integrity, road obstacle existence status and traffic flow order status. Retrieve the normal traffic state baseline features corresponding to each road segment scene unit. The normal traffic state baseline features are the statistical average of the key features of the traffic scene image of the road segment scene unit when no abnormal events occur. By comparing the key features of the traffic scene images of each road segment scene unit with the baseline features of normal traffic conditions, feature difference information is extracted. The feature difference information includes the type, location and duration of the difference features. Based on feature difference information, road segment scene units with feature differences exceeding a preset range are selected, and the road segment scene units are marked as abnormal scene units. All abnormal scene units are collected to form an abnormal scene unit set. Extract the scene unit association relationship and association strength information of each abnormal scene unit in the abnormal scene unit set, and determine the directly associated road segment scene units and indirectly associated road segment scene units of each abnormal scene unit; Analyze the transmission data of feature difference information of abnormal scene units in directly related road segment scene units, and track the start time and diffusion process data of feature difference information spreading from abnormal scene units to directly related road segment scene units; Based on the starting time of the diffusion of feature difference information from abnormal scene units to each directly related road segment scene unit, as well as the spatial distance and traffic flow relationship between each directly related road segment scene unit and the abnormal scene unit, the diffusion priority of feature difference information among different related road segment scene units is determined. Record the road segment scene units through which the feature difference information diffuses according to the diffusion priority, and generate a transmission node sequence. Each node in the transmission node sequence corresponds to a road segment scene unit affected by the feature difference information. Extract the feature difference information, diffusion time and correlation strength information of each node in the transmission node sequence, and label them between the corresponding nodes in the transmission node sequence to generate a transmission link containing time and strength dimensions. By integrating the transmission node sequence and transmission link, a sudden event transmission path is constructed. The sudden event transmission path includes the diffusion path information and impact sequence information of the abnormal event between different road segment scene units.

4. The intelligent auxiliary decision-making method for emergency management of highways according to claim 1, characterized in that, The process involves collecting basic information on control resources across the entire highway network, dynamically adapting control resources to the transmission paths of emergencies based on their spatiotemporal attributes, and generating a control resource adaptation scheme, including: Collect basic information on management and control resources across the entire highway area. This basic information includes the type, functional scope, geographical location of deployment, availability status, and response start time of the management and control resources. Analyze the spatiotemporal attributes of the transmission path of a sudden event, and extract the starting segment scene unit, ending segment scene unit, diffusion direction, diffusion speed, and the duration of influence corresponding to each transmission node of the sudden event transmission path; Based on the deployment geographic location, effective response radius, and geographic location of transmission nodes along the transmission path of emergencies of the managed resources, the spatial distance from the managed resources to each transmission node is calculated, and managed resources whose spatial distance is less than or equal to the effective response radius of the managed resource are selected to generate a set of candidate managed resources. Based on the spread speed of the emergency transmission path and the duration of the impact of each transmission node, the response time window of each candidate control resource is calculated. The response time window contains the time range data of the control resource that can arrive and play a role within the duration of the impact of the transmission node. Based on the functional scope of the control resources and the abnormal impact type of each transmission node, the functional requirements of the candidate control resources and the transmission nodes are matched, and the control resources whose functional scope can cover the abnormal impact handling requirements of the transmission nodes are selected, and the candidate control resource set is updated. Analyze the propagation direction of the emergency transmission path, combine the deployment geographical location of each control resource in the updated candidate control resource set, incorporate the geographical range data of the road segment scene unit that has been affected by the abnormal event, and generate the dispatch route; Based on the sequence of transmission nodes in the transmission path of an emergency and the duration of the impact of each transmission node, the response order of each control resource is determined, and the response order of the control resources is correlated with the diffusion rhythm of the transmission path and the duration of the impact of each transmission node. Adjust the response order and scheduling route based on the availability status of the managed resources, prioritize scheduling managed resources with good availability status, and supplement supporting measures for managed resources with partial availability status to ensure their normal functioning. The scheduling routes, response sequences, and supporting measures of integrated control resources are used to generate a preliminary control resource adaptation plan. The preliminary control resource adaptation plan includes a list of control resources and execution requirements for each transmission node. Based on the changing trend of the spread speed of the emergency transmission path, dynamic adjustment clauses are added to the preliminary control resource adaptation plan to generate the final control resource adaptation plan. The dynamic adjustment clauses include relevant operational specifications for adjusting the response order and scheduling route of control resources according to changes in the spread speed.

5. The intelligent auxiliary decision-making method for emergency management of highways according to claim 2, characterized in that, The calculation of structural fit between different road segment scene units based on static correlation elements, wherein the structural fit includes relevant data on the connection and matching of road segment scene units in the road structure, generates a static correlation matrix, including: The matching relationships of multiple structural elements between different road segment scene units are evaluated respectively. The matching relationships of multiple structural elements include lane number matching relationships, road surface structure type matching relationships, adjacent road segment connection method matching relationships, and fixed monitoring equipment deployment density matching relationships. For each pair of road segment scene units, the lane matching level is determined based on the matching relationship of their number of lanes, the road structure matching level is determined based on the matching relationship of their road structure type, the connection method matching level is determined based on the matching relationship of their adjacent road segment connection method, and the monitoring coverage matching level is determined based on the matching relationship of their fixed monitoring equipment deployment density. Based on the lane matching level, road surface structure matching level, connection method matching level, and monitoring coverage matching level, the structural adaptability between different road segment scene units is determined according to the preset comprehensive evaluation rules. Using the number of the road segment scene unit as the identifier of the matrix row and column, an initial empty matrix is ​​constructed. The number of rows and columns of the initial empty matrix are equal to the total number of road segment scene units. The calculated structural fit between each pair of road segment scene units is filled into the corresponding row and column positions in the initial empty matrix to generate an intermediate static correlation matrix. The matrix elements in the intermediate static correlation matrix are normalized so that the values ​​of all matrix elements are within the same preset range, thus eliminating the dimensional differences caused by different dimension coefficients. Check the matrix elements in the normalized intermediate static correlation matrix. If there are matrix elements with abnormal values, recalculate the structural fit of the corresponding road segment scene unit pair and replace the abnormal values. The final static association matrix is ​​generated based on the corrected intermediate static association matrix. The matrix elements of the static association matrix contain data related to the structural fit of the scene units corresponding to the two road segments.

6. The intelligent auxiliary decision-making method for highway emergency management according to claim 3, characterized in that, The analysis includes the transmission data of characteristic difference information of abnormal scene units to directly related road segment scene units, tracking the start time and diffusion process data of characteristic difference information spreading from abnormal scene units to directly related road segment scene units, including: Extract the start time from the feature difference information of the abnormal scene unit, and establish a time tracking coordinate system based on the start time. Obtain the association strength information corresponding to the association relationship between the abnormal scene unit and each directly associated road segment scene unit, and sort the directly associated road segment scene units in descending order of the association strength value; For each directly related road segment scene unit after sorting, retrieve the continuous video stream data of that road segment scene unit after the reference point, and analyze the key features of the traffic scene frame by frame in chronological order. The time point at which the key features of the traffic scene image of the directly related road segment scene unit first appear with the same type of feature difference information as the abnormal scene unit is identified, and this time point is recorded as the diffusion start time of the directly related road segment scene unit. Starting from the diffusion initiation time, continuously track the change data of the corresponding type of difference features in the key features of traffic scene images of directly related road segment scene units, and record the expansion of the occurrence range of difference features, the change of intensity, and the superposition data with other features. By comparing the feature difference information of the abnormal scene unit with the difference features tracked in the directly related road segment scene unit, we can analyze the consistency data of the two in terms of type, manifestation and change trend, and confirm the correlation data of feature transmission. Based on the time tracking coordinate system, the diffusion start time of each directly associated road segment scene unit is compared with the feature difference start time of the abnormal scene unit to calculate the diffusion time difference, which includes the time interval data of feature transmission. Integrate the diffusion start time, differential feature change data, and diffusion time difference of each directly related road segment scene unit to generate a feature transfer record for a single directly related road segment scene unit; Collect feature transfer records of all directly related road segment scene units, sort the results by association strength and diffusion time difference, and generate a feature transfer data summary. Based on the aggregation of feature transfer data, the starting time and diffusion process data of feature difference information spreading from abnormal scene units to directly related road segment scene units are output.

7. The intelligent auxiliary decision-making method for emergency management of highways according to claim 4, characterized in that, The process involves matching candidate control resources with the functional requirements of each transmission node based on the functional scope of the control resources and the anomaly impact type of each transmission node. Control resources whose functional scope can cover the anomaly impact handling requirements of the transmission nodes are then selected, and the candidate control resource set is updated. This includes: Extract the abnormal impact type for each transmission node. The abnormal impact types include vehicle breakdown lane occupation impact, road surface damage impact, traffic flow congestion impact, and foreign object intrusion impact. Extract the functional scope-related data for each candidate control resource. The functional scope-related data includes the types of abnormal impacts that each candidate control resource can handle, the range of its processing capabilities, and the applicable road segment scenario conditions. Establish a matching table between abnormal impact types and the functional scope of control resources. The table includes a description of the appropriate control resource functions for each type of abnormal impact. For each transmission node, query the matching table according to its abnormal impact type to obtain the control resource function conditions required for that transmission node; Each control resource in the candidate control resource set is checked one by one to determine whether its functional scope data fully includes the control resource functional conditions required by the transmission node. Control resources whose functional scope-related data can fully cover the required functional conditions are marked as suitable control resources and retained in the candidate control resource set; For control resources whose functional scope-related data includes the required functional conditions, further analysis is conducted to determine whether their missing functions can be supplemented by other supporting resources. If they can be supplemented, the control resource and supporting resources are combined and marked as an adaptive control resource, and retained in the candidate control resource set. Control resources whose functional scope data does not contain the required functional conditions and cannot be supplemented by supporting resources shall be removed from the candidate control resource set. For each transmission node, the filtered adaptation and control resources are aggregated to generate a subset of exclusive adaptation and control resources corresponding to that transmission node; Integrate the exclusive adaptive control resource subsets of all transmission nodes, remove duplicate control resources, and generate an updated candidate control resource set. The control resources in the updated candidate control resource set can meet the functional requirements of at least one transmission node.

8. The intelligent auxiliary decision-making method for emergency management of highways according to claim 2, characterized in that, The calculation of traffic flow transmission efficiency between different road segment scene units based on dynamic correlation elements, generating a dynamic correlation matrix, includes: The dynamic correlation elements between different road segment scene units are evaluated separately. These dynamic correlation elements include traffic direction relationship, traffic flow stability, traffic status smoothness, and traffic flow transmission status. For each pair of road segment scenario units, the directional association level is determined based on their traffic direction relationship, the stability level is determined based on the traffic flow stability state, the stability level is determined based on the traffic state smoothness, and the traffic flow transmission state level is evaluated based on traffic flow conversion data. Based on the aforementioned directional correlation level, stability level, smoothness level, and traffic flow transmission status level, the traffic flow transmission efficiency between different road segment scenario units is determined according to the preset comprehensive evaluation rules. Using the number of the road segment scene unit as the identifier of the matrix row and column, an initial empty dynamic matrix is ​​constructed. The number of rows and columns of the initial empty dynamic matrix are equal to the total number of road segment scene units. The calculated traffic flow transmission efficiency between each pair of road segment scene units is filled into the corresponding row and column positions in the initial empty dynamic matrix to generate an intermediate dynamic correlation matrix. The rationality of the matrix elements in the intermediate dynamic correlation matrix is ​​checked, and abnormal values ​​that deviate significantly from the normal range are removed by referring to historical traffic flow transmission data. The traffic flow transmission efficiency corresponding to the abnormal values ​​is then recalculated. The final dynamic correlation matrix is ​​generated based on the intermediate dynamic correlation matrix after verification and correction. The matrix elements of the dynamic correlation matrix contain traffic flow transmission efficiency data for the corresponding two road segment scene units.

9. The intelligent auxiliary decision-making method for emergency management of highways according to claim 3, characterized in that, The step of recording the scene units along the path of feature difference information diffusion according to diffusion priority to generate a transmission node sequence includes: The abnormal scene unit is used as the starting node of the transmission node sequence, and the initial state of the starting node, including the road segment scene unit number, geographical location, and feature difference information, is recorded. Based on the diffusion priority ranking results, the highest priority directly associated road segment scene unit is selected as the first transmission node, and the road segment scene unit number, geographical location, and feature difference information status transmitted from the starting node are recorded for this transmission node. Analyze the indirect associated road segment scene units of the first transmission node, and combine the diffusion priority ranking results to select the next highest priority associated road segment scene unit as the second transmission node. Record the relevant information and feature difference information of this transmission node. The subsequent transmission nodes are selected sequentially. The selection of each transmission node is based on the associated road segment scene unit and diffusion priority sorting result of the previous transmission node, so that the transmission order conforms to the diffusion logic. When recording information for each transmission node, the time interval between the transmission of feature difference information from the previous transmission node to the current transmission node, the details of changes during the transmission process, and the correlation strength information are recorded simultaneously. For multiple related road segment scene units with the same diffusion priority, they are selected as subsequent transmission nodes in order of their spatial distance from the current transmission node from closest to farthest. If a certain associated road segment scene unit has already been recorded as a transmission node, it will not be recorded again; During the construction of the transmission node sequence, the scene unit association relationship in the video stream scene association network is correlated in real time, and the association validity between each transmission node and the previous node is verified. When the intensity of the characteristic difference information of the transmission node is lower than the preset threshold or there are no subsequent related road segment scene units, stop adding new transmission nodes and determine the termination node of the transmission node sequence. All recorded transmission node information is organized and arranged in the transmission order to form a transmission node sequence. The transmission node sequence contains relevant data of all road segment scene units through which the feature difference information diffuses.

10. An intelligent auxiliary decision-making system for emergency management on highways, characterized in that, The invention includes a processor and a computer-readable storage medium storing machine-executable instructions, which, when executed by a computer, implement the intelligent auxiliary decision-making method for highway emergency management as described in any one of claims 1-9.

Images