Road monitoring data management system based on big data

Abstract

The present application relates to the technical field of traffic monitoring data management, in particular to a road monitoring data management system based on big data, which comprises a traffic flow monitoring input module, a path segment organization module, a path stability generation module, a behavior chain structure extraction module and an index path configuration module.In the present application, the identification records of vehicles at different monitoring points are classified to construct a data set with spatial and temporal characteristics, support trajectory modeling, combine time proximity and spatial coherence to complete path segment merging, enhance behavior description capability, analyze the frequency and change amplitude of path segments, extract stable path information for path sorting and classification, identify vehicle combinations with close positions and synchronized times, track their trajectories, depict the dynamic process of road aggregation behavior, construct a hierarchical structure based on path stability, improve data organization and retrieval efficiency, realize multi-source data fusion and behavior recognition, and enhance traffic state perception and response capability.

Description

Technical Field

[0001] This invention relates to the field of traffic monitoring data management technology, and in particular to a road monitoring data management system based on big data. Background Technology

[0002] The field of traffic monitoring data management technology involves key aspects such as the collection, storage, management, and analysis of road traffic operation status data. It primarily involves constructing a multi-source monitoring data acquisition system oriented towards the road traffic environment, including video images, license plate recognition, and radar detection. This system, combined with data modeling, format conversion, and semantic extraction, forms an information system oriented towards management objectives such as traffic events, traffic flow, and traffic efficiency. Methodologically, it builds management capabilities around steps such as data fusion, feature extraction, database design, and index construction. Traditional road monitoring data management systems involve deploying camera equipment at fixed monitoring points to collect traffic image information, then using a relational database to statically store the collected images and metadata. Historical data is retrieved using time-series retrieval methods, and traffic anomalies are identified through preset rules. This approach uses matching retrieval methods based on license plate recognition image fields or time-label-based data aggregation methods for data retrieval and organization. However, when faced with the continuous growth of complex, multi-source, and heterogeneous monitoring data, it has limitations in data organization and real-time processing.

[0003] Traditional road monitoring data management relies on image acquisition from fixed monitoring points and static storage in relational databases. This results in a lack of diverse data organization methods and an inability to effectively construct the continuity and spatial logical connections of vehicle trajectories. Time-series retrieval models can only achieve historical data backtracking along a time axis, failing to dynamically reconstruct the inherent relationships between multi-source identification information. This leads to fragmented trajectory information and disjointed behavioral processes, hindering the formation of effective path reconstruction and behavioral group extraction mechanisms. Aggregation retrieval methods based on image fields or time tags are limited by coarse retrieval granularity and missing path features, making it difficult to effectively identify high-frequency vehicle behaviors in complex road scenarios. For example, during peak traffic hours, multiple vehicles may pass through multiple monitoring points simultaneously, making it difficult for the system to identify their trajectory connections and thus failing to reflect the true trend of road traffic. Furthermore, static storage methods lack the ability to model real-time data evolution characteristics, leading to delayed responses to sudden aggregation phenomena and an inability to adjust strategies in a timely manner at the management level, resulting in decreased efficiency in traffic scheduling and event response. These problems become increasingly prominent against the backdrop of continuous data growth and increasingly complex scenarios, limiting the adaptability and accuracy of traditional systems in intelligent traffic management. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a road monitoring data management system based on big data.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a road monitoring data management system based on big data includes: The traffic flow monitoring input module extracts vehicle detection records within the monitoring area, integrates identification data of the same vehicle at different times, organizes and classifies vehicle time information, location coordinates and unique identifiers to form a dataset with spatial partition structure and time series features, and generates the original traffic flow dataset. The path segment organization module extracts the trajectory information of vehicles between different monitoring points based on the continuous recording of each vehicle in the original traffic flow dataset, merges temporally adjacent and spatially continuous path nodes, and generates a set of lane path segments. The path stability generation module extracts the frequency of occurrence of the same segment within different time periods from the lane path segment set, determines the magnitude of frequency change over time, and uses the change characteristics to measure the stability of the path segments, establishes sorted and distinguishable stability labeling information, and generates a stability-labeled path segment table. The behavior chain structure extraction module identifies combinations of vehicles that enter the monitoring area at the same time and are in close proximity based on the vehicle start position and time records in the stability labeled path segment table, marks the continuous aggregation phenomenon as the start node of the behavior chain, tracks the associated path, and generates the lane behavior chain structure.

[0006] As a further embodiment of the present invention, the original traffic flow dataset includes time information, location coordinates, and unique identifiers; the lane path segment set includes path node sequences, time interval features, and spatial continuity features; the stability-labeled path segment table includes the frequency of occurrence of path segments, the magnitude of frequency variation, and the stability level; and the lane behavior chain structure includes the starting position, the starting time, and the chain-like trajectory structure.

[0007] As a further aspect of the present invention, the traffic flow monitoring input module includes: The vehicle passage record extraction submodule extracts the vehicle passage detection records of the monitoring nodes in the monitoring area, calls the frame sequence of images captured by the nodes within a set time period, detects the vehicle body pixel shape and movement trajectory appearing in the images, judges them as independent vehicle passage events, records the occurrence time and node number, and obtains the initial number of vehicle passage records. The identification data integration submodule filters vehicle records that appear at different times based on the initial number of vehicle records, calculates the cosine distance between the license plate character fitting degree and the vehicle body color vector, selects records with a character fitting degree greater than the character fitting degree threshold and a color vector distance less than the threshold of the same type of color, labels them with a unified vehicle number, extracts the corresponding identification time point and location coordinates, arranges the records in chronological order, and obtains the vehicle trajectory identification sequence. The spatiotemporal feature generation submodule calls the vehicle trajectory recognition sequence to construct a continuous time period and spatial coordinate sequence for each vehicle, calculates the time difference, displacement difference and trajectory segment time span between adjacent nodes, obtains the spatiotemporal deviation value between nodes, maps the deviation value to feature weights, combines the time series length and coordinate aggregation partitioning, obtains the partition number and time window, and generates a spatial partition structure and time series feature dataset.

[0008] As a further aspect of the present invention, the path segment organization module includes: The path filtering submodule extracts the passage nodes, passage time and location coordinates based on the continuous records of each vehicle in the original traffic flow data, organizes the trajectory sequence in chronological order, removes records that do not meet the requirements of temporal and spatial continuity, and generates a continuous trajectory node set. The spatiotemporal merging submodule calls the continuous trajectory node set, filters trajectory nodes based on time interval and spatial distance conditions, merges node combinations that meet the continuity requirements, extracts the start and end information and trajectory features after merging, and obtains the node merging path index value. The path segment generation submodule identifies the changes in trajectory features in the merged path combination based on the node merging path index value, determines whether the path meets the segment division criteria, organizes the trajectory structure and direction information, and obtains a set of lane path segments.

[0009] As a further aspect of the present invention, the path stability generation module includes: The frequency extraction submodule extracts the occurrence records of each path segment in different time periods based on the set of lane path segments, counts the occurrence of path segments by time period, removes duplicate records, merges the occurrence of the same path number in different time intervals, and obtains the path segment occurrence frequency sequence. The change recognition submodule calls the frequency sequence of the path segments, analyzes the frequency change status in chronological order, constructs a time-frequency sequence curve for each segment, identifies the change amplitude and range between frequency values, obtains the degree of frequency change of the path segments, quantifies the fluctuation intensity of each path, and generates path frequency fluctuation values. The stability labeling submodule sets a stability interval threshold range based on the path frequency fluctuation value, determines the interval affiliation of the frequency fluctuation index corresponding to all path segments, marks the corresponding stability level label, sorts and classifies them into a callable format, and obtains a stability-labeled path segment table.

[0010] As a further aspect of the present invention, the behavior chain structure extraction module includes: The behavior initiation identification submodule extracts the first appearance time and corresponding position of the vehicle in the monitoring area based on the stability labeled path segment table, identifies vehicle combinations whose lateral coordinate difference is within a set distance threshold in adjacent time periods, determines whether vehicle groups that maintain a close state meet the continuous aggregation condition, marks the combined vehicles, and obtains the initiation node identification label set. The path aggregation analysis submodule calls the starting node to identify the tag set, obtains the path number sequence and time record of the marked vehicles, judges the consistency of path numbers and the synchronization of time sequences among vehicle combinations, filters vehicle combinations that meet the path aggregation characteristics, and obtains the path aggregation coupling value sequence. The chain trajectory generation submodule identifies vehicle combinations that meet the set requirements based on the path aggregation coupling value sequence, integrates path numbers and time series, constructs a continuous structure based on time sequence and path connection relationship, and obtains lane behavior chain structure.

[0011] As a further aspect of the present invention, the system further includes: The index path configuration module matches the corresponding stability information based on the path segments in the behavior chain structure, sorts all path segments according to stability order, constructs a hierarchical structure of trunk and auxiliary paths, and generates a path index configuration table. The path index configuration table includes path segment sorting results, main path, and auxiliary path; The monitoring area refers to the road space covered by multiple front-end traffic monitoring devices, and the boundary is defined by the coordinate range covered by the devices. The vehicle detection record refers to the information data collected by the monitoring equipment when a vehicle passes through the detection point, including the license plate number, collection time, vehicle coordinates, speed, and lane number fields; The difference time refers to the recorded time when the vehicle enters the difference monitoring point or the same monitoring point at a different clock time point, and the basis for the judgment is the inconsistency of the timestamp field value; The location coordinates are the geographical location information recorded when the vehicle passes through the monitoring point; The spatial partitioning structure refers to the logical space that divides the monitoring area according to roads, lanes, and coordinate grids.

[0012] The path node is a collection of location information formed by the vehicle during continuous monitoring, including basic elements such as timestamp, location coordinates, and lane number; The path segment refers to a vehicle trajectory segment composed of two or more adjacent path nodes, where the time interval and spatial displacement between nodes meet the continuity standard. The time period refers to the time window used to aggregate vehicle trajectory information, which can be defined by a fixed interval (five minutes) or an event-driven approach; The degree of stability refers to a value calculated by the magnitude of frequency changes; The set distance threshold is a value set based on the width of the road lanes; The aggregation phenomenon refers to the situation where the lateral coordinates of multiple vehicles are close and their spatial positions are continuously concentrated within a unit of time. After parallel judgment, it can be identified as aggregation. The behavior chain is an ordered structure formed by combining multiple vehicle trajectories that start from the same origin, have continuous paths, and have consistent behavioral characteristics. The main path is a set of path segments with a frequency stability index higher than the average and a large vehicle coverage in multiple time periods, and the auxiliary path is a path that does not meet the conditions. The path index configuration table is a data structure that records the path segment number, its level, and priority sorting fields.

[0013] As a further aspect of the present invention, the index path configuration module includes: The stability information matching submodule matches the stability information of the path in the stability labeling path segment table according to the path number in the behavior chain structure, establishes the correspondence between the path and the stability, and obtains the path stability matching list. The path level sorting submodule calls the data in the path stability matching list, divides the path into levels based on the stability information, and combines the occurrence of the path in the behavior chain to complete the sorting and stratification operation and obtain the path level stratification result. The structure index generation submodule organizes path numbers and establishes an index structure by level based on the hierarchical classification information in the path level hierarchical results, excludes overlapping numbers and overlapping levels, and generates a path index configuration table.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by classifying the identification records of vehicles at different monitoring points, a dataset with spatial and temporal characteristics is constructed, effectively supporting continuous modeling of traffic flow trajectories. By combining the temporal proximity and spatial coherence of path nodes, trajectory merging is completed, enabling path segments to have stable behavioral description capabilities. By analyzing the frequency and variation of path segments, highly stable path information is extracted, providing a basis for path classification and priority ranking. Behavioral chain tracking of vehicle combinations that are close in location and synchronized in time can reconstruct trajectory groups with consistent behavior, depicting the dynamic evolution process of road aggregation phenomena. Based on the stability of paths in the behavioral chain, a primary and secondary hierarchical structure is constructed to enhance the organization efficiency and retrieval performance of path data. Overall, the fusion processing of multi-source identification data and accurate identification of trajectory behavior are realized, improving the insight capability and management response efficiency in complex traffic environments. Attached Figure Description

[0015] Figure 1 This is a system flowchart of the present invention; Figure 2 This is a flowchart of the traffic flow monitoring input module of the present invention; Figure 3 This is a flowchart of the path segment organization module of the present invention; Figure 4 This is a flowchart of the path stability generation module of the present invention; Figure 5 This is a flowchart of the behavior chain structure extraction module of the present invention; Figure 6 This is a flowchart of the index path configuration module of the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0017] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0018] Please see Figure 1 The road monitoring data management system based on big data includes: The traffic flow monitoring input module extracts vehicle detection records within the monitoring area, integrates the identification data of the same vehicle at different times, organizes and classifies the vehicle's time information, location coordinates and unique identifiers, forms a dataset with spatial partition structure and time series features, and generates the original traffic flow dataset. The monitored area refers to the road space covered by multiple front-end traffic monitoring devices (video surveillance, geomagnetic sensors, millimeter-wave radar), and the boundary is defined by the coordinate range covered by the devices; Vehicle detection records refer to the information data collected by monitoring equipment when a vehicle passes through a detection point, including license plate number, collection time, vehicle coordinates, speed, and lane number fields; The difference time refers to the recorded time when a vehicle enters a different monitoring point or the same monitoring point at a different clock time. The basis for the distinction is the inconsistency of the timestamp field value. Location coordinates are the geographical location information recorded when the vehicle passes through the monitoring point; Spatial partitioning structure refers to the logical space that divides the monitoring area according to roads, lanes, and coordinate grids; The path segment organization module extracts the trajectory information of vehicles between different monitoring points based on the continuous records of each vehicle in the original traffic flow dataset, merges the path nodes that are temporally adjacent and spatially continuous, and generates a set of lane path segments. A path node is a collection of location information generated by a vehicle during continuous monitoring, including basic elements such as timestamp, location coordinates, and lane number. A path segment refers to a vehicle trajectory segment consisting of two or more adjacent path nodes, where the time interval and spatial displacement between nodes meet the continuity standard. The path stability generation module extracts the frequency of occurrence of the same segment in different time periods from the set of lane path segments, judges the magnitude of frequency change over time, and uses the change characteristics to measure the stability of the path segments, establishes stability labeling information for sorting and differentiation, and generates a stability-labeled path segment table. The time period refers to the time window used to aggregate vehicle trajectory information, which can be defined by a fixed interval (five minutes) or by an event-driven approach; The degree of stability refers to the value calculated from the magnitude of frequency changes; The behavior chain structure extraction module identifies vehicle combinations that enter the monitoring area simultaneously and whose lateral coordinate difference is within a set distance threshold based on the starting position and time records of vehicles in the stability label path segment table. It marks the continuous aggregation phenomenon as the starting node of the behavior chain, tracks the associated paths and integrates them into a chain trajectory structure to generate the lane behavior chain structure. The distance threshold is set as a value based on the road lane width; Aggregation refers to the situation where the lateral coordinates of multiple vehicles are close and their spatial positions are continuously concentrated within a unit of time. Aggregation can be identified after parallel judgment is satisfied. A behavior chain is an ordered structure formed by the combination of multiple vehicle trajectories that start from the same origin, follow continuous paths, and have consistent behavioral characteristics. The index path configuration module matches the corresponding stability information based on the path segments in the behavior chain structure, sorts all path segments according to stability order, constructs a hierarchical structure of trunk and auxiliary paths, and generates a path index configuration table. The main routes are a set of route segments with a frequency stability index higher than the average and a large vehicle coverage in multiple time periods, while auxiliary routes are routes that do not meet the conditions. The path index configuration table is a data structure that records the path segment number, its level, and priority sorting fields.

[0019] The original traffic flow dataset includes time information, location coordinates, and unique identifiers. The lane path segment set includes path node sequences, time interval features, and spatial continuity features. The stability-labeled path segment table includes the frequency of path segment occurrence, the magnitude of frequency variation, and the stability level. The lane behavior chain structure includes the starting position, starting time, and chain trajectory structure. The path index configuration table includes the path segment sorting results, main paths, and auxiliary paths.

[0020] Please see Figure 2 The traffic flow monitoring input module includes: The vehicle passage record extraction submodule extracts vehicle passage detection records from monitoring nodes within the monitoring area, calls the frame sequence of images captured by the nodes within a set time period, detects the vehicle body pixel shape and movement trajectory appearing in the images, determines them as independent vehicle passage events, records the occurrence time and node number, and obtains the initial number of vehicle passage records. To extract vehicle detection records within the monitored area, a monitoring time interval must first be selected at each deployed monitoring node. Generally, a time period with concentrated traffic flow is chosen as the target time axis, such as 07:00–09:00 during the morning rush hour. Video image frame sequences within this time period are extracted node by node, with a sampling frequency of 10 frames per second. For pixel clusters showing vehicle outlines in consecutive frames, inter-frame motion detection is performed to identify densely packed areas of inter-frame difference pixels, determining these as vehicle passage. By comparing the start and end points of frames where vehicle shapes stably appear in the frame difference image, the time range of vehicle passage is determined and recorded. and Further calculate the average time point at which it appears at this node. For example, if a vehicle appears and disappears at 08:21:10.6 and 08:21:12.0 respectively in a certain node, then its average time is 08:21:11.3, which is recorded as the timestamp of the vehicle at that node. The vehicle's center point position is obtained by combining the GPS calibration coordinates or camera position coordinates in that frame. For example, if the camera number is C021 and the coordinates are (128.2135, 35.5432), then the vehicle is recorded to have appeared at this point. The record structure is uniformly written in the following format: timestamp, vehicle position, node number, image sequence number, frame index, and image ID. Then, the above operation is repeated for multiple nodes, and the appearance information of each vehicle at each node is integrated into a group. The original event record set is formed by accumulating the numbers. Each data structure contains the fields ID, node number, timestamp, image number, coordinates, and frame segment start and end points. The number of records obtained is the initial number of vehicle passing records.

[0021] The identification data integration submodule filters vehicle records that appear at different times based on the initial number of vehicle records, calculates the cosine distance between the license plate character fitting degree and the vehicle body color vector, selects records with a character fitting degree greater than the character fitting degree threshold and a color vector distance less than the threshold of the same type of color, labels them with a unified vehicle number, extracts the corresponding identification time point and location coordinates, arranges the records in chronological order, and obtains the vehicle trajectory identification sequence. Based on the initial number of vehicle records, it is necessary to iterate through vehicle records at different time points one by one. By comparing the license plate characters and vehicle body color features in each record, it is determined whether they belong to the same vehicle. The license plate string and color vector are extracted from each record. During character matching, a position-by-character comparison method is used to construct character similarity, denoted as . For example, if the target license plate is "B362T8" and the current comparison item is "B36ZT8", and there are 5 similar characters and 4 exact matches, then the similarity is... For color features, extract the vector values ​​of the main color region of the image, such as (128, 140, 145), and calculate the cosine angle between it and the color vector of another record (130, 136, 150) to obtain the similarity value. Set the character similarity threshold to 0.7 and the color cosine similarity threshold to 0.98. When the condition is met... and If a match is deemed successful, the matching record group is uniformly labeled with a vehicle number, with the numbering rule being date + vehicle sequence number, such as "240501-C013". Then, the time point and coordinate information in all matching records are retrieved and sorted in ascending order of time. The record structure is: vehicle number, timestamp sequence, coordinate sequence, and node ID sequence, which yields the travel trajectory of each vehicle at multiple nodes, forming a vehicle trajectory recognition sequence.

[0022] The spatiotemporal feature generation submodule calls the vehicle trajectory recognition sequence to construct a continuous time period and spatial coordinate sequence for each vehicle, and calculates the time difference, displacement difference, and trajectory segment time span between adjacent nodes using the following formula: ; The spatiotemporal deviation values ​​between nodes are obtained through computation, and the deviation values ​​are mapped to feature weights. Combined with the time series length and coordinate aggregation partitioning, the partition number and time window are obtained to generate a spatial partitioning structure and time series feature dataset. in, Indicates vehicle In the The time interval between two nodes in the trajectory segment. This indicates the displacement distance. Indicates the segment of trajectory The time span it covers This represents the arc length of the curve representing the trajectory segment. This indicates that the vehicle is in all The average displacement distance in the segment, Indicates vehicle The spatiotemporal deviation value between nodes, Indicates vehicle The total number of trajectory segments, Number the trajectory segments; By calling the vehicle trajectory recognition sequence, the time and location coordinates of each vehicle between consecutive monitoring nodes are first extracted, and the shooting time of adjacent nodes is marked as... The corresponding spatial displacement distance is marked as For example, if vehicle A's shooting times between node 1 and node 2 are 09:01:12 and 09:01:14 respectively, then... The Euclidean distance is calculated by combining the positioning coordinates. Construct the length of the trajectory segment The driving trajectory is obtained by fitting a curve to the image of the vehicle, for example, the length of the fitted curve for a lane segment with a large curvature. The image frame range shows that the vehicle has a larger curvature in that section, and the trajectory segment took [time missing]. Extracted based on frame sequence duration; for example, if the segment spans 81 frames and the sampling frequency is 30fps, then... Number all trajectory segments The parameter values ​​are recorded in Table 2, and the difference between the velocity and curvature velocity between nodes is calculated. The total difference is obtained by summing the squares and taking the square root, combined with the displacement deviation. The sum of the absolute values ​​yields the spatiotemporal deviation value between nodes. The formula used is: ; The parameters are explained below: For the first Segment displacement, The time interval between shooting nodes. To fit the arc length, This is the time span of the arc segment. The summation sign is the average displacement of all segments of the vehicle. correspond arrive Squares, square roots, and absolute values ​​are used to enhance the expression of differences, and are calculated through the following example: Substitute the values ​​from Table 2 into the formula to perform multi-level calculations: Average displacement: ; The sum of squares of the velocity differences (Part 1): ; Displacement deviation term (Part 2): ; Calculation of overall deviation value: ; The results indicate that the vehicle's spatiotemporal deviation in the three trajectory segments is 0.988, which falls within the moderate deviation range. Based on the deviation level, the range is defined as follows: Determined to be a stable trajectory. The disturbance was determined to be minor. The change is determined to be drastic, therefore the current sample is on the critical edge of slight perturbation; As shown in Table 1, the parameters of the three trajectories have been entered in a standardized manner, supporting the full-process deduction and standardized application of formula calculation, and generating a spatial partition structure and time series feature dataset. The formula's operational logic is built upon a dual structure of "time-space coupling deviation" and "trajectory uniformity offset," and is divided into two parts: The first part is the summation and square root of the squared differences of the velocity offsets of trajectory segments. This measure the difference between the vehicle's linear velocity (displacement distance divided by the node time difference) between adjacent nodes and the actual fitted trajectory curvature velocity (arc length divided by the time taken for that segment). By squaring the differences of each segment, summing them, and taking the square root, the positive and negative cancellations are eliminated, and the deviation effect is strengthened. The square root operation reflects the overall deviation magnitude, ensuring that the result remains in the same order of magnitude as the velocity. The second part is the summation of the normalized deviations of the displacement values ​​in each trajectory segment relative to the average displacement. This reflects whether the vehicle exhibits significant acceleration, deceleration, or abnormal segment skipping behavior in continuous segments. The absolute value is used to ignore the deviation direction and retain only the offset magnitude. Normalization is used to eliminate the influence of inconsistent travel distance scales among different vehicles or zones. The two results are directly added because they represent the spatial motion trajectory stability and the uniformity of travel rhythm between segments, respectively, and have complementary evaluation significance. Together, they constitute the vehicle's comprehensive spatiotemporal deviation index. The spatiotemporal deviation value between nodes is used to measure the consistency of the trajectory behavior of the same vehicle during its passage between continuous monitoring nodes. It comprehensively reflects the vehicle's change characteristics in two dimensions: spatial path and temporal rhythm. Specifically, the larger the value, the greater the difference between the vehicle's driving speed and the fitted trajectory shape between adjacent nodes, which may indicate behaviors such as deviating from the predetermined route, abnormal lane changes, or temporary stops. It also indicates that the vehicle's inter-segment displacement changes are more volatile, with acceleration, deceleration, or skipping phenomena. This indicator can be used to filter unstable passage samples, identify vehicles with abnormal trajectories, or classify vehicle behavior patterns. In the structured processing of datasets, this value serves as an important spatiotemporal feature input, which helps in the training and judgment of subsequent clustering partitioning, behavior recognition, and dynamic prediction models.

[0023] Please see Figure 3 The path segment organization module includes: The path filtering submodule extracts passing nodes, passing times and location coordinates based on the continuous records of each vehicle in the original traffic flow data, organizes the trajectory sequence in chronological order, removes records that do not meet the requirements of temporal and spatial continuity, and generates a continuous trajectory node set. Based on continuous records of each vehicle in the original traffic flow data, the passage information of vehicles at different times is extracted sequentially, and the node number, passage time and corresponding spatial coordinates are identified. By constructing a trajectory list arranged in ascending order of time field, the continuous path behavior of each vehicle in the entire observation period is obtained. Adjacent trajectory nodes are combined in pairs in sequence, and their passage time difference and spatial displacement are calculated. A time threshold of 30 seconds and a spatial distance threshold of 1500 meters are set for each group of node records to determine whether the records meet the continuity requirements. When either of them exceeds the threshold, the group of records is removed. The remaining nodes are marked as "continuous" and retained as valid trajectory points. For a typical example, the judgment results of three vehicles in part of the path segment are selected for display, and the following table is generated: As shown in Table 2, vehicles V001 and V002 were determined to be continuous trajectories because their time and distance between nodes did not exceed the set thresholds. Vehicle V003 was determined to be a non-continuous trajectory because its time difference exceeded the threshold and its distance exceeded the upper limit. Continuous records were retained for subsequent trajectory merging processing, while data that did not meet the conditions were excluded. The data were organized by vehicle number to form a set of trajectory datasets with temporal and spatial continuity, and the continuous trajectory node set was obtained.

[0024] The spatiotemporal merging submodule calls the continuous trajectory node set, filters trajectory nodes based on time interval and spatial distance conditions, merges node combinations that meet the continuity requirements, extracts the start and end information and trajectory features after merging, and obtains the node merging path index value. The continuous trajectory node set is invoked, and the merging conditions between adjacent nodes of each vehicle in the sorted trajectory are analyzed in turn. The travel time field and node spatial coordinates are extracted respectively. By judging whether the travel time between nodes is within 15 seconds and whether the spatial distance is less than 300 meters, node pairs that meet the merging judgment rules are filtered. The trajectory pairs that meet the conditions are marked as mergeable trajectory segments. The travel time, position coordinates and node numbers of their start and end points are extracted to form a structured path segment record. The merged path identifier is constructed with vehicle number and path sequence number. The time difference, position span and speed index extracted for all merged paths constitute a path segment attribute set. The speed value is extracted by dividing the distance between two nodes by the time interval. Paths with small speed changes and consistent direction in continuous path segments are recorded as stable path combinations. Each group of merged path segments is accompanied by start and end position numbers and start and end timestamps. The overall path set is classified by vehicle number to form a trajectory structure attribute data set that can be used for segment judgment and obtain the node merged path index value.

[0025] The path segment generation submodule identifies the changes in trajectory features in the merged path combination based on the node merged path index value, determines whether the path meets the segment division criteria, organizes the trajectory structure and direction information, and obtains a set of lane path segments. Based on the node merging path index value, all merged path combinations for each vehicle are analyzed to determine the continuity, directional stability, speed fluctuation range, and trajectory point density of each path segment. Path segments are then categorized and identified, with a focus on the start and end information and the number of internal trajectory points. For path segments with sparse trajectory points and large speed fluctuations, segmentation is performed. If there are abrupt changes in direction between multiple consecutive nodes in a path segment, the segment is deemed to lack trajectory consistency and is split into multiple structurally clear sub-segments. All processed path segments are arranged in ascending order by node number and renumbered. Paths with relatively consistent speeds and controllable trajectory direction changes retain their original structure. Path numbers are constructed for the split path segments based on their start and end positions and added to the path segment set. An angle threshold is set for the directional change range of trajectory points; trajectory points exceeding this threshold are recorded as directional turning points. Paths with a directional turning point density greater than a preset ratio are determined to be structurally unstable segments. Using the merge number, vehicle number, and trajectory segment sequence number as a joint primary key, all path segments are summarized and organized to obtain the lane path segment set.

[0026] Please see Figure 4 The path stability generation module includes: The frequency extraction submodule is based on the set of lane path segments. It extracts the occurrence records of each path segment in different time periods, divides the path segments by time period to count the occurrence times, removes duplicate records, merges the occurrences of the same path number in different time intervals, and obtains the path segment occurrence frequency sequence. Based on the set of lane path segments, the records of each path segment in different time periods are extracted. Three elements need to be specified: the lane path segment number, the recording time, and the number of times the segment is passed. The number of passes is measured in "times" as the number of times the segment is accessed in all vehicle trajectories within that time period, with a granular time period of 10 minutes. For example, the segment with path number P001 is recorded 4, 5, and 6 times in three consecutive time periods, belonging to time period 1, time period 2, and time period 3 respectively. This statistical operation needs to be performed segment by segment, removing vehicles repeatedly counted within a certain time period. If a vehicle repeatedly passes through the same road segment within the same time period, it is only counted once. A frequency array is constructed for each path, arranged in order of time period, and bound to the segment number. Simultaneously, intermittent records across time periods are identified, and a null value handling strategy is set. If the frequency of two consecutive time periods is missing, the path is discarded. This process avoids discontinuous frequency calculations due to incomplete records, as shown in the table below: Table 3 lists the frequency records of the sample path numbers in three consecutive time periods. Among them, the frequency sequence of P001 is [4, 5, 6], and the record is complete. P002 is [7, 3, 4], and P003 is [5, 8, 6]. All path records are complete and can be used for subsequent fluctuation value calculation to obtain the frequency sequence of path segments.

[0027] The change recognition submodule calls the frequency sequence of path segments, analyzes the frequency changes in chronological order, constructs a time-frequency sequence curve for each segment, and identifies the relative magnitude and range of frequency values ​​using the following formula: ; The calculation obtains the degree of frequency change of path segments, quantifies the fluctuation intensity of each path, and generates path frequency fluctuation values. in, Representing a path The frequency variation stability value, in units of "times"; Representing a path In time period The frequency of occurrence within the range, expressed in "times"; Representing a path The average frequency value over the entire monitoring period, expressed in "times"; Representing a path The total number of time periods monitored, in units of "segments"; and Representing paths The maximum and minimum frequency values ​​across all time periods are both expressed in "times". The frequency sequence of path fragment occurrences is called, and the frequency array of each path number in each time period is processed sequentially. Let the path number be P002, and its frequency values ​​in the three time periods be... , , First, calculate the average frequency of the path using the following formula: ; Then calculate the average of the absolute differences: ; Next, the range term is calculated as follows: ; Finally, the variance root term is calculated as follows: ; Substituting into the corrected formula, we get: ; This value represents the total frequency fluctuation of path P002 within a time period, reflecting the strong instability of the path in time series, which is mainly caused by the large range term, generating the path frequency fluctuation value. The calculation logic of this formula aims to measure the degree of fluctuation in the frequency of path segments occurring in different time periods from multiple perspectives. First, This represents the average of the absolute values ​​of the differences between the frequency of each path in each time period and its average frequency. It measures the overall degree of deviation. This term combines all deviations positively through summation and absolute value calculations, avoiding the cancellation of positive and negative values, and reflects the overall dispersion of frequency changes. A larger value indicates stronger frequency fluctuations over most time periods. Secondly, This term describes the range between the maximum and minimum frequencies. It does not require normalization and retains its original value, reflecting the strongest fluctuation in path frequency across different time periods. It is suitable for identifying local extreme changes. Finally... The standard deviation of the frequency value is calculated by averaging the squared deviations and then taking the square root, which enhances the sensitivity to outlier fluctuations. This term amplifies the abnormal frequency that occurs in a particular period and is used to supplement the local blind spots of the mean absolute deviation. The three terms measure the intensity of frequency changes from the perspectives of average fluctuation, maximum fluctuation, and variance fluctuation, respectively. They are complementary in structure and converge into the total fluctuation value through addition. This allows the fluctuation dimensions that cannot be covered by a single indicator to be comprehensively reflected in this formula. The frequency variation of a route segment refers to the overall intensity of fluctuations in the number of times the same route segment is traversed by vehicles in different time periods. It reflects the traffic stability characteristics of the route in the time dimension. If the frequency of a route remains consistent or the variation is small in multiple time periods, it indicates that the traffic behavior of the route is relatively regular and the frequency variation is low. Conversely, if there are significant fluctuations or extreme differences in frequency between different time periods, it indicates that the traffic behavior of the route is unstable and the frequency variation is high. This indicator comprehensively considers the overall deviation of each time period from the average value, the range between the maximum and minimum values, and the dispersion of the frequency in all time periods. It can reflect the dynamic characteristics of the route in time series from three aspects: average fluctuation, local jumps, and global variance. It is the core quantitative basis for measuring route stability.

[0028] The stability labeling submodule sets the stability interval threshold range based on the path frequency fluctuation value, determines the interval assignment of the frequency fluctuation index corresponding to all path segments, marks the corresponding stability level label, outputs a structured information table, sorts and classifies it into a callable format, and obtains the stability labeled path segment table. Based on the frequency fluctuation values ​​of the paths, all paths are labeled and grouped. A stability interval threshold is set as follows: fluctuation values ​​less than or equal to 3 are classified as "stable segment," values ​​greater than 3 and less than or equal to 6 are classified as "transition segment," and values ​​greater than 6 are classified as "fluctuating segment." This interval is calculated based on the median of 4.2 and the standard deviation of 1.9 from 60 path samples. A reasonable boundary is constructed by adding or subtracting one standard deviation from the mean. Path number P002 has a fluctuation value of 7.26, which falls within the interval greater than 6, therefore it is classified as "fluctuating segment" and assigned a stability label of 3. This classification logic is repeated for path numbers P001 and P003, where P... The frequency sequence 001 is [4, 5, 6], with an average frequency of 5, an average absolute difference of 1, a range of 2, a standard deviation of 0.82, and a fluctuation of 1+2+0.82=3.82, falling between 3 and 6. It is assigned label 2. The frequency sequence P003 is [5, 8, 6], with an average of 6.33, an average absolute difference of 1.11, a range of 3, a standard deviation of 1.25, and a fluctuation of 5.36. It is assigned label 2. The path number, frequency fluctuation value, and assigned label are merged to form a stability mapping table. After sorting, the table is output in a structured format to obtain the stability-labeled path segment table. Additional commands can be added to generate this structured table.

[0029] Please see Figure 5 The behavior chain structure extraction module includes: The behavior initiation recognition submodule extracts the first appearance time and corresponding position of a vehicle in the monitoring area based on the stability labeled path segment table, identifies vehicle combinations whose lateral coordinate difference is within a set distance threshold in adjacent time periods, determines whether vehicle groups that maintain a close state meet the continuous aggregation condition, completes the labeling operation of combined vehicles, and obtains the initiation node recognition label set. Based on the stability-labeled path segment table, the time information and corresponding coordinates of the first appearance of all vehicles in the monitoring area are extracted to obtain the starting position point of each vehicle. This information is decomposed into three types of parameters: vehicle number, timestamp, and spatial coordinates. Within any time granularity, the time difference of all vehicles is obtained and arranged in chronological order. Then, the difference between the horizontal and vertical coordinates of each pair of vehicles in the spatial position parameters is detected to determine whether they are simultaneously below the spatial proximity threshold. Based on the judgment result, vehicle combinations with overlapping time periods and spatial aggregation characteristics are selected. If the vehicles in a combination maintain a spatial proximity relationship for at least three consecutive time periods, it is confirmed that it meets the continuous aggregation condition and the combination is identified as a valid starting combination. All vehicles in the combination are uniformly assigned an identification number, as shown in the table below: Table 4 lists the start time and location coordinates of vehicles V001 to V004. V001, V002, and V003 are identified as the initial aggregation group when the time difference is less than 10 seconds and the coordinate difference is less than 2 meters. V004 is not identified as an aggregation member because the time difference and location difference exceed the set threshold. Vehicles V001, V002, and V003 are identified as the same group of identification numbers, and the starting node identification label set is obtained.

[0030] The path aggregation analysis submodule calls the starting node to identify the tag set, obtains the path number sequence and time record of the marked vehicles, judges the consistency of path numbers and the synchronization of time sequences among vehicle combinations, filters vehicle combinations that meet the path aggregation characteristics, and obtains the path aggregation coupling value sequence. The system calls upon the identified aggregated vehicle IDs in the tag set of the starting node, extracts the path ID sequence and corresponding time record for each vehicle throughout the entire monitoring period, compares the vehicle path IDs side-by-side at the time granularity, and identifies the consistency of path IDs based on the degree of synchronization within each time period. It then compares the degree of path overlap for each vehicle combination across different time periods, selecting vehicle pairs with a path ID repetition rate greater than 60% and a time record difference less than a set threshold as valid combinations. Further analysis is conducted to determine whether these combinations consistently maintain the above characteristics across multiple consecutive time periods. If they meet the criteria, they are categorized into path synchronization aggregation groups, and each group is assigned a specific attribute. Vehicle consistency identifiers are used to record the correspondence between all vehicle pairs and their consistency levels. A table is established to correspond to combination numbers and consistency scores. In the combination with starting nodes V001, V002, and V003, the path number overlap comparison shows that the path consistency score between V001 and V002 is 0.68, V001 and V003 is 0.74, and V002 and V003 is 0.70. All of these scores are higher than the set aggregation threshold. The path time record difference is distributed between 3 and 8 seconds, which is lower than the standard value of 10 seconds for time synchronization error. All three paths meet the path synchronization aggregation conditions, and a path aggregation coupling value sequence is obtained.

[0031] The chain trajectory generation submodule identifies vehicle combinations that meet the set requirements based on the path aggregation coupling value sequence, integrates path numbers and time series, constructs a continuous structure based on time sequence and path connection relationship, and obtains lane behavior chain structure. Based on the path aggregation coupling value sequence, vehicle combinations with path consistency scores in the higher range are identified. The path number sequence of all vehicles in the combination is called, and the numbers are arranged according to the timestamp order. All consecutively occurring path numbers are connected, and breakpoints are set at the time points when the path numbers change. The combination is split into multiple time-segment structure segments. The segment structure between vehicle numbers in the same group is sequence fused to generate path structure data with chain connection logic. Taking the path number sequence V001, V002, V003 as an example, in the continuous path segment segment, there is a consistent connection P31→P32→P33→P35. Only P34 has a one-time jump segment that is not repeated by most vehicles, so it is removed from the chain structure. The output is a path number chain group with temporal and structural continuity, and the lane behavior chain structure is obtained.

[0032] Please see Figure 6 The index path configuration module includes: The stability information matching submodule matches the stability information of the path in the stability label path fragment table based on the path number in the behavior chain structure, filters out incomplete path items, establishes the correspondence between path and stability, and obtains the path stability matching list. Extract all records from the stability-labeled path segment table based on the path numbers in the behavior chain structure. Search the stability record values ​​one by one against the path number field to obtain initial pairing data between path numbers and stability values. Check for missing or duplicate path numbers in the pairing data; remove missing path records. For duplicate path records, filter by record time, selecting the most recent valid entry as the corresponding value. Perform a consistency check on the retained path numbers to confirm they match the identifier field in the stability-labeled path segment table. Establish a mapping dictionary structure for consistent fields. Check again for conflicts where multiple paths correspond to the same stability value. For conflicting entries, select the earliest appearing path number based on the path order in the behavior chain structure, remove conflicting paths to form a matching set, and construct a two-dimensional key-value structure with two fields: path number and stability value. Arrange the records in ascending order of path number to form structured records, as shown in Table 5. Table 5 lists the path numbers and stability record values ​​that were successfully matched in the behavior chain structure. Through the aforementioned steps, conflicting and non-valued items were eliminated, and the record order was ensured to be consistent, thus obtaining the path stability matching list.

[0033] The path level sorting submodule calls the data in the path stability matching list, classifies the path levels according to the stability information, and combines the occurrence of the path in the behavior chain to complete the sorting and stratification operation and obtain the path level stratification result. The system retrieves path numbers and stability record values ​​from the path stability matching list. All paths are then sorted in ascending order of stability value. Path numbers with identical stability values ​​are further filtered. The frequency of occurrence of the corresponding path in the behavior chain structure is used as the sorting reference field. The path frequency sorting standard is set as follows: first by stability, then by frequency in descending order. The sorted results are segmented, with the first half of the numbers designated as primary path candidates and the second half as secondary path candidates. The candidate path numbers are rearranged according to the behavior chain timestamp order to generate a main path set and a secondary path set. The system then checks for duplicate path numbers belonging to both groups. If duplicates exist, the path with the higher stability ranking is retained, the other path is removed, and the numbering intervals are redefined. A two-level path set structure is constructed, and a merged path and sorting identifier list is output, generating a path level hierarchy result.

[0034] The structure index generation submodule organizes path numbers and builds an index structure by level according to the hierarchical classification information in the path level hierarchical results, excludes overlapping numbers and overlapping levels, and generates a path index configuration table. Based on the trunk path numbers and auxiliary path numbers marked in the path hierarchy results, the number sequence is hierarchically classified and grouped with labels. The trunk path numbers are used as the first-level index and the auxiliary path numbers as the second-level index, and merged into a two-level list structure. The overlapping path numbers in the list are deduplicated. For trunk path numbers that intersect with auxiliary paths, a strategy of prioritizing trunk path numbers is applied. The path number format is standardized, and the number length and naming structure are unified. A complete hierarchical path set is generated, the number order is sorted, and a structured index structure form is generated. The path index configuration table is obtained.

[0035] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A road monitoring data management system based on big data, characterized in that, The system includes: The traffic flow monitoring input module extracts vehicle detection records within the monitoring area, integrates the identification data of the same vehicle at different times, organizes and classifies the vehicle time information, location coordinates and unique identifiers, forms a dataset with spatial partition structure and time series features, and generates the original traffic flow dataset. The path segment organization module extracts the trajectory information of vehicles between different monitoring points based on the continuous recording of each vehicle in the original traffic flow dataset, merges temporally adjacent and spatially continuous path nodes, and generates a set of lane path segments. The path stability generation module extracts the frequency of occurrence of the same segment within different time periods from the set of lane path segments, determines the magnitude of frequency change over time, and uses the change characteristics to measure the stability of the path segments, establishes sorted and distinguishable stability labeling information, and generates a stability-labeled path segment table. The behavior chain structure extraction module identifies combinations of vehicles that enter the monitoring area at the same time and are in close proximity based on the vehicle start position and time records in the stability labeled path segment table. It marks the continuous aggregation phenomenon as the starting node of the behavior chain, tracks the associated path, and generates the lane behavior chain structure.

2. The road monitoring data management system based on big data according to claim 1, characterized in that, The original traffic flow dataset includes time information, location coordinates, and unique identifiers. The lane path segment set includes path node sequences, time interval features, and spatial continuity features. The stability-labeled path segment table includes the frequency of path segment occurrence, the magnitude of frequency variation, and the stability level. The lane behavior chain structure includes the starting position, the starting time, and the chain-like trajectory structure.

3. The road monitoring data management system based on big data according to claim 1, characterized in that, The traffic flow monitoring input module includes: The vehicle passage record extraction submodule extracts the vehicle passage detection records of the monitoring nodes in the monitoring area, calls the frame sequence of images captured by the nodes within a set time period, detects the vehicle body pixel shape and movement trajectory appearing in the images, judges them as independent vehicle passage events, records the occurrence time and node number, and obtains the initial number of vehicle passage records. The identification data integration submodule filters vehicle records that appear at different times based on the initial number of vehicle records, calculates the cosine distance between the license plate character fitting degree and the vehicle body color vector, selects records with a character fitting degree greater than the character fitting degree threshold and a color vector distance less than the threshold of the same type of color, labels them with a unified vehicle number, extracts the corresponding identification time point and location coordinates, arranges the records in chronological order, and obtains the vehicle trajectory identification sequence. The spatiotemporal feature generation submodule calls the vehicle trajectory recognition sequence to construct a continuous time period and spatial coordinate sequence for each vehicle, generating a spatial partition structure and time series feature dataset.

4. The road monitoring data management system based on big data according to claim 3, characterized in that, The path segment organization module includes: The path filtering submodule extracts passing nodes, passing times and location coordinates based on the continuous records of each vehicle in the original traffic flow data, organizes the trajectory sequence in chronological order, and generates a continuous trajectory node set. The spatiotemporal merging submodule calls the continuous trajectory node set, filters trajectory nodes based on time interval and spatial distance conditions, merges node combinations that meet the continuity requirements, extracts the start and end information and trajectory features after merging, and obtains the node merging path index value. The path segment generation submodule identifies the changes in trajectory features in the merged path combination based on the node merging path index value, determines whether the path meets the segment division criteria, organizes the trajectory structure and direction information, and obtains a set of lane path segments.

5. The road monitoring data management system based on big data according to claim 4, characterized in that, The path stability generation module includes: The frequency extraction submodule extracts the occurrence records of each path segment in different time periods based on the set of lane path segments, counts the occurrence of path segments by time period, merges the occurrence of the same path number in different time intervals, and obtains the path segment occurrence frequency sequence. The change recognition submodule calls the frequency sequence of the path segments, analyzes the frequency change status in chronological order, constructs a time-frequency sequence curve for each segment, identifies the change amplitude and range between frequency values, obtains the degree of frequency change of the path segments, quantifies the fluctuation intensity of each path, and generates path frequency fluctuation values. The stability labeling submodule sets the stability interval threshold range based on the path frequency fluctuation value, determines the interval assignment of the frequency fluctuation index corresponding to all path segments, marks the corresponding stability level label, sorts and classifies them into a callable format, and obtains the stability labeling path segment table. Stability level labels refer to stability levels classified based on frequency fluctuation values.

6. The road monitoring data management system based on big data according to claim 5, characterized in that, The behavior chain structure extraction module includes: The behavior initiation identification submodule extracts the first appearance time and corresponding position of the vehicle in the monitoring area based on the stability labeled path segment table, identifies vehicle combinations whose lateral coordinate difference is within a set distance threshold in adjacent time periods, determines whether vehicle groups that maintain a close state meet the continuous aggregation condition, marks the combined vehicles, and obtains the initiation node identification label set. The path aggregation analysis submodule calls the starting node to identify the tag set, obtains the path number sequence and time record of the marked vehicles, judges the consistency of path numbers and the synchronization of time sequences among vehicle combinations, filters vehicle combinations that meet the path aggregation characteristics, and obtains the path aggregation coupling value sequence. The chain trajectory generation submodule identifies vehicle combinations that meet the set requirements based on the path aggregation coupling value sequence, integrates path numbers and time series, constructs a continuous structure based on time sequence and path connection relationship, and obtains lane behavior chain structure.

7. The road monitoring data management system based on big data according to claim 6, characterized in that, The system also includes: The index path configuration module matches the corresponding stability information based on the path segments in the behavior chain structure, sorts all path segments according to stability order, constructs a hierarchical structure of trunk and auxiliary paths, and generates a path index configuration table. The path index configuration table includes path segment sorting results, main path, and auxiliary path; The monitoring area refers to the road space covered by multiple front-end traffic monitoring devices, and the boundary is defined by the coordinate range covered by the devices. The vehicle detection record refers to the information data collected by the monitoring equipment when a vehicle passes through the detection point, including the license plate number, collection time, vehicle coordinates, speed, and lane number fields; The difference time refers to the recorded time when the vehicle enters the difference monitoring point or the same monitoring point at a different clock time point, and the basis for the judgment is the inconsistency of the timestamp field value; The location coordinates are the geographical location information recorded when the vehicle passes through the monitoring point; The spatial partitioning structure refers to the logical space that divides the monitoring area according to roads, lanes, and coordinate grids; The path node is a collection of location information formed by the vehicle during continuous monitoring, including basic elements such as timestamp, location coordinates, and lane number; The path segment refers to a vehicle trajectory segment composed of two or more adjacent path nodes, where the time interval and spatial displacement between nodes meet the continuity standard. The time period refers to the time window used to aggregate vehicle trajectory information, which can be defined by a fixed interval (five minutes) or an event-driven approach; The degree of stability refers to a value calculated by the magnitude of frequency changes; The set distance threshold is a value set based on the width of the road lanes; The aggregation phenomenon refers to the situation where the lateral coordinates of multiple vehicles are close and their spatial positions are continuously concentrated within a unit of time. After parallel judgment, it can be identified as aggregation. The behavior chain is an ordered structure formed by combining multiple vehicle trajectories that start from the same origin, have continuous paths, and have consistent behavioral characteristics. The main path is a set of path segments with a frequency stability index higher than the average and a large vehicle coverage in multiple time periods, and the auxiliary path is a path that does not meet the conditions. The path index configuration table is a data structure that records the path segment number, its level, and priority sorting fields.

8. The road monitoring data management system based on big data according to claim 7, characterized in that, The index path configuration module includes: The stability information matching submodule matches the stability information of the path in the stability labeling path segment table according to the path number in the behavior chain structure, establishes the correspondence between the path and the stability, and obtains the path stability matching list. The path level sorting submodule calls the data in the path stability matching list, divides the path into levels based on the stability information, and combines the occurrence of the path in the behavior chain to complete the sorting and stratification operation and obtain the path level stratification result. The structure index generation submodule organizes path numbers and establishes an index structure by level based on the hierarchical classification information in the path level hierarchical results, excludes overlapping numbers and overlapping levels, and generates a path index configuration table.

Images