A vehicle dynamic label generation method and system based on multi-dimensional behavior analysis

Abstract

The present application relates to the technical field of vehicle dynamic analysis, in particular to a vehicle dynamic label generation method and system based on multi-dimensional behavior analysis. The method is as follows: obtaining the vehicle pass database of the specified date of the camera, analyzing the pass record through multi-dimensional aggregation condition, generating the intermediate record containing the license plate number, label ID and other information and storing it in the intermediate statistical table; when generating the dynamic label, based on the research and judgment rules, the intermediate record of the target license plate in the research and judgment time window is queried, and the label is generated after the aggregation judgment meets the threshold. The present application pre-processes the calculation through the intermediate statistical table mechanism, compresses the data size and improves the efficiency; the multi-label independent analysis shares the intermediate table, avoiding repeated scanning of the original data; the effective data is screened to eliminate redundancy, ensuring the real-time performance and scalability of high concurrency tasks.

Description

Technical Field

[0001] This invention relates to the field of vehicle dynamic analysis technology, and in particular to a method and system for generating vehicle dynamic labels based on multi-dimensional behavior analysis. Background Technology

[0002] In the fields of intelligent transportation and public safety, the large-scale deployment of video surveillance and checkpoint systems generates massive amounts of vehicle traffic data daily. To support key vehicle management, abnormal behavior early warning, and traffic situation analysis, there is an urgent need to automatically extract semantically valuable vehicle behavior features from raw traffic records and express them in a structured form as "dynamic tags." For example, identifying complex behavioral patterns such as "active at night" or "frequent occurrence in specific areas" has become a key step in improving the efficiency of police analysis. However, directly calculating such multi-dimensional behavioral tags on raw data in real time faces significant performance challenges, thus requiring an efficient and scalable tag generation mechanism.

[0003] Current mainstream technical solutions typically employ static rule matching or simple statistical methods, directly performing label determination on the original vehicle traffic database. For example, an SQL query is used to count the number of times a license plate passes through a specific time period; if the number exceeds a threshold, a corresponding label is assigned. This approach has significant drawbacks: First, it is limited in its dimensionality and label description capabilities; existing technologies mostly rely on a single data dimension (such as location or time alone) for simple counting, failing to effectively integrate multiple dimensions such as time (e.g., day-night distribution), space (e.g., cross-regional movement), and frequency for collaborative analysis. Therefore, it is difficult to accurately characterize intertwined behavioral features such as "nocturnal activity" (requiring consideration of nighttime and all-day traffic ratios) and "resident location" (requiring consideration of location frequency and consecutive days). Second, the rules are rigid and lack adaptability; they rely on manually preset fixed thresholds and hard-coded logic, lacking flexible configuration and dynamic judgment capabilities. First, when business needs change (such as adding a "weekend active vehicle" tag), the code needs to be redeveloped and deployed, severely limiting the system's scalability. Second, the computational efficiency is low and the real-time performance is insufficient. Each tag generation requires traversing the entire number of traffic records, resulting in high computational overhead and response latency when dealing with hundreds of millions of data points. Furthermore, when multiple tags are generated in parallel, the same data is repeatedly scanned, leading to significant resource waste. Third, the system lacks an architecture that separates multidimensional statistics from analysis. The data processing flow is highly coupled, and no intermediate result reuse mechanism has been established. A large number of vehicle records that do not meet any behavioral characteristics still participate in the entire process, resulting in computational redundancy and storage expansion.

[0004] Therefore, this invention aims to solve the comprehensive technical problems in the prior art, such as single dimension, rigid rules, low computational efficiency and poor system scalability caused by directly calculating multi-dimensional behavior labels on the original traffic base database. It provides a label generation method that can accurately depict complex vehicle behavior and efficiently support dynamic updates under massive data. Summary of the Invention

[0005] To address one of the aforementioned technical problems, the present invention adopts the following technical solution:

[0006] According to one aspect of the present invention, a method for generating vehicle dynamic tags based on multi-dimensional behavior analysis is provided, the method comprising the following steps:

[0007] Retrieve the vehicle access database corresponding to a specific camera on a specified date. The vehicle access database contains all vehicle access event records recorded by the camera on that specified date, and each access event record includes at least the license plate number, access time, and the location identifier of the camera.

[0008] Using the multidimensional aggregation conditions corresponding to each vehicle behavior tag, an aggregation analysis is performed on all passage event records in the vehicle passage database to determine whether the vehicle corresponding to each license plate number meets the multidimensional aggregation conditions. The multidimensional aggregation conditions are used to characterize the behavioral features of the vehicle defined by the corresponding vehicle behavior tag in the time, space, or spatiotemporal combination dimensions.

[0009] If a passage event record corresponding to a certain license plate number meets the multidimensional aggregation conditions of a certain vehicle behavior tag, an intermediate record is generated and included in the intermediate statistics table. The intermediate record must at least contain the license plate number, the behavior tag ID of the vehicle behavior tag, the aggregation result value, and the specified date.

[0010] In response to the dynamic tag generation task, based on the behavior tag ID, dynamic judgment time window, and judgment threshold included in the judgment rules corresponding to the dynamic tag generation task, the system queries the intermediate statistics table for all intermediate records of the target license plate number under the behavior tag ID within the dynamic judgment time window. Vehicle behavior tags, multi-dimensional aggregation conditions, and judgment rules are all managed in a configurable manner, supporting dynamic addition, deletion, and modification without modifying the program code.

[0011] The intermediate records retrieved are aggregated and judged. If the judgment threshold is met, a dynamic tag corresponding to the target license plate number is generated.

[0012] According to a second aspect of the present invention, a vehicle dynamic label generation system based on multi-dimensional behavior analysis is provided, the system comprising:

[0013] The vehicle access database acquisition module is used to obtain the vehicle access database corresponding to a specific camera on a specified date. The vehicle access database contains all vehicle access event records recorded by the camera on that specified date, and each access event record includes at least the license plate number, access time, and the location identifier of the camera.

[0014] The behavior aggregation module is used to aggregate and analyze all passage event records in the vehicle passage database using the multidimensional aggregation conditions corresponding to each vehicle behavior tag, and to determine whether the vehicle corresponding to each license plate number meets the multidimensional aggregation conditions. The multidimensional aggregation conditions are used to characterize the behavioral features of the vehicle defined by the corresponding vehicle behavior tag in the time, space, or spatiotemporal combination dimensions.

[0015] The intermediate record generation module generates an intermediate record and adds it to the intermediate statistics table if the passage event record corresponding to a certain license plate number meets the multi-dimensional aggregation conditions of a certain vehicle behavior tag. The intermediate record must at least include the license plate number, the behavior tag ID of the vehicle behavior tag, the aggregation result value, and the specified date.

[0016] The dynamic analysis module responds to dynamic tag generation tasks. Based on the behavior tag ID, dynamic analysis time window, and analysis threshold included in the analysis rules corresponding to the dynamic tag generation task, it queries the intermediate statistics table for all intermediate records of the target license plate number under the behavior tag ID within the dynamic analysis time window. Vehicle behavior tags, multi-dimensional aggregation conditions, and analysis rules are all managed in a configurable manner, supporting dynamic addition, deletion, and modification without requiring modification of the program code.

[0017] The dynamic tag generation module is used to aggregate and judge the intermediate records retrieved. If the judgment threshold is met, a dynamic tag corresponding to the target license plate number is generated.

[0018] This invention has at least one of the following beneficial effects:

[0019] First, this invention effectively solves the problem of low computational efficiency caused by traversing the entire original traffic record for each tag generation in existing technologies by introducing an intermediate statistical table mechanism driven by vehicle behavior tags. Specifically, in this invention, an intermediate record containing the license plate number, behavior tag ID, aggregation result value, and date is generated and written into the intermediate statistical table only when the traffic event record of a certain license plate number on a specified date by a single camera meets the multidimensional aggregation conditions of the corresponding vehicle behavior tag. Subsequent dynamic tag generation tasks only need to query the historical intermediate records of the target license plate based on this intermediate statistical table, under the specified behavior tag ID and dynamic judgment time window, and perform lightweight aggregation judgment. Compared with the background technology's method of "directly scanning the hundreds of millions of original databases for each tag judgment", this solution forwards the complex multidimensional aggregation calculation into a daily batch processing task, and the intermediate statistical table only retains the valid data related to the tag, which significantly compresses the data scale in the tag judgment stage and greatly improves the tag generation efficiency.

[0020] Secondly, this invention overcomes the resource waste caused by repeatedly scanning the original data when generating multiple tags in parallel by independently performing aggregation analysis based on vehicle behavior tags and selectively generating intermediate records. In this invention, aggregation analysis is performed on the same vehicle traffic database for each vehicle behavior tag using its corresponding multi-dimensional aggregation conditions, and intermediate records are generated only for combinations of (license plate, tag ID) that meet the conditions. This means that the calculation logic of different tags is decoupled, and all tags share the same preprocessed intermediate statistics table, avoiding the redundant operation of "re-scanning the entire table every time a new tag is added" in the background technology. For example, when generating tags for "daytime and nighttime activity" and "permanent residence," the traditional solution requires traversing all traffic records twice, while this invention only requires a single database read to complete the generation of intermediate records for multiple tags, significantly reducing I / O load and CPU overhead, and improving the overall system throughput.

[0021] Furthermore, this invention effectively eliminates computational redundancy caused by a large number of invalid vehicle records participating in subsequent processing by implementing a data filtering mechanism based on multidimensional aggregation conditions during the intermediate statistics table construction stage. In this invention, the generation of intermediate records is strictly selective; only vehicles that meet the multidimensional aggregation conditions for specific behavioral tags on a given day (such as nighttime traffic exceeding a threshold) are written into the intermediate statistics table. Vehicle traffic event records that do not meet any tag conditions are directly filtered out and do not enter the intermediate layer. This contrasts sharply with the prior art's approach of "all vehicles, regardless of whether they possess behavioral characteristics, participate in tag determination." Since most vehicles in real-world scenarios do not conform to specific abnormal or key behavioral patterns, the amount of data in the intermediate statistics table is usually only a small portion of the original database. This significantly reduces unnecessary data reading and comparison operations in the subsequent tag analysis stage, saving storage and computing resources and further ensuring the real-time performance and scalability of high-concurrency tag generation tasks. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 A flowchart illustrating a method for generating dynamic vehicle tags based on multi-dimensional behavior analysis, provided in an embodiment of the present invention.

[0024] Figure 2 This is a structural framework diagram of a vehicle dynamic label generation system based on multi-dimensional behavior analysis, provided for an embodiment of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] As one possible embodiment of the present invention, such as Figure 1 As shown, a method for generating vehicle dynamic labels based on multi-dimensional behavior analysis is provided. The method includes the following steps:

[0027] S100: Obtain the vehicle access database corresponding to a specific camera on a specified date. The vehicle access database contains all vehicle access event records recorded by the camera on the specified date, and each access event record includes at least the license plate number, access time, and the location identifier of the camera.

[0028] S100 specifically includes:

[0029] S101: Retrieve the raw video stream or image sequence captured by a camera on a specified date.

[0030] S102: Perform license plate recognition and timestamp extraction on the original video stream or image sequence to obtain structured traffic event data.

[0031] S103: Collect structured traffic event data according to a specified date and the location identifier of a certain camera to form a vehicle traffic database.

[0032] In each passage event record, the passage time is generated based on the camera's system clock or timing signal, and the license plate number is the recognition result after confidence verification.

[0033] This step is the starting point for data input in the entire method. Specifically, the system first retrieves the raw video stream or image sequence captured by a specific camera (e.g., at checkpoint A at an intersection) on a specified date (e.g., February 20, 2025) from the video surveillance platform or checkpoint management platform (S101). Then, it processes this raw data using a mature license plate recognition (LPR) algorithm to accurately extract the license plate number from each frame and simultaneously record the precise passage timestamp (S102). To ensure data quality, the system performs confidence checks on the identified license plate numbers, retaining only high-confidence results. Finally, all processed structured data, including license plate numbers, passage times (generated based on the camera's own system clock or high-precision timing signals), and the camera's unique location identifier (e.g., checkpoint ID), is aggregated by "February 20, 2025" and "Checkpoint A" to form a complete, structured vehicle passage database (S103). This database serves as the foundational data source for all subsequent analyses. In the vehicle passage database corresponding to any camera, as long as the camera captures a vehicle passing through the monitored area, a record is formed for that vehicle. Since the number of vehicles passing through the monitored area in a day is huge, the amount of data in the vehicle passage database is also very huge, and most of the data is invalid.

[0034] S200: Using the multidimensional aggregation conditions corresponding to each vehicle behavior tag, perform aggregation analysis on all passage event records in the vehicle passage database to determine whether the vehicle corresponding to each license plate number meets the multidimensional aggregation conditions. The multidimensional aggregation conditions are used to characterize the behavioral features of the vehicle defined by the corresponding vehicle behavior tag in the time, space, or spatiotemporal combination dimensions.

[0035] Specifically, the multidimensional aggregation conditions include at least one of the following: the number of times a vehicle passes through a specified time period, the frequency of a vehicle's appearance in a specified geographical area, or the proportion of the number of times a vehicle passes through a specific time period and a specific area to its total number of passes through that camera on that day.

[0036] Vehicle behavior tags can include vehicles that operate during the day and are active at night, vehicles that reside at a fixed location, or vehicles that frequently move across regions. Each type of vehicle behavior tag corresponds to a unique multidimensional aggregation condition and judgment rule.

[0037] This step is the core step for the present invention to achieve accurate multi-dimensional behavior characterization. The system will traverse all preset vehicle behavior tags, and for each tag, apply its exclusive multi-dimensional aggregation conditions to independently and parallelly analyze the base library obtained in S100. Taking the "active at night" tag (behavior tag ID: 3) as an example, the corresponding multi-dimensional aggregation condition can be defined as: calculating the proportion of the traffic volume of the vehicle during the night period (e.g., from 21:00 to 05:00 the next day) to the total traffic volume within its complete 24-hour cycle (e.g., from 05:00 of the current day to 05:00 the next day). The system will traverse all records in the base library, first filter out the traffic events belonging to the target vehicle (such as license plate "Lu AXXXXX"), and then separately count the number of traffic occurrences during the above two time periods. Suppose on February 20, 2025, the vehicle passed through checkpoint A 15 times at night and a total of 20 times throughout the day, then its night proportion is 75%.

[0038] Taking the "permanent dwelling place" tag as another example, its multi-dimensional aggregation condition may be defined as: whether the total number of times the vehicle appears at the position of this camera on the current day exceeds a frequency threshold. This design enables the system to preliminarily judge the association strength between the vehicle and this location from the perspective of a single camera, thereby providing a basis for the recognition of complex behavior patterns.

[0039] To improve the flexibility and efficiency of the system, the present invention adopts a hierarchical processing architecture, clearly separating the vehicle dynamic tag generation process into a "multi-dimensional daily behavior statistics layer" and a "configurable tag judgment layer". The former is responsible for extracting core feature vectors from massive traffic data and storing them; the latter makes intelligent judgments based on the results of the former and external configuration rules. This decoupled design is the basis for the high efficiency and flexibility of this solution.

[0040] S300: If the traffic event record corresponding to a certain license plate number meets the multi-dimensional aggregation conditions of a certain vehicle behavior tag, then generate an intermediate record and record it in the intermediate statistical table. The intermediate record shall at least include the license plate number, the behavior tag ID of this vehicle behavior tag, the aggregation result value, and the specified date.

[0041] S301: If the traffic event record corresponding to a certain license plate number does not meet the multi-dimensional aggregation conditions corresponding to any vehicle behavior tag, then no intermediate record corresponding to the license plate number will be generated.

[0042] The intermediate statistical table is stored in partitions by date, and a joint index is established for the license plate number and the behavior tag ID to accelerate the query operation based on the behavior tag ID and the dynamic judgment time window.

[0043] Continuing with the above example of "active at night and dormant during the day", if the preset preliminary judgment threshold is 70%, and the calculated 75% > 70%, the system will generate an intermediate record for the license plate "Lu AXXXXX" and the "active at night and dormant during the day" tag. Taking the license plate "Lu AXXXXX" meeting the preliminary conditions of "active at night and dormant during the day" (behavior tag ID is 3) at checkpoint A on February 20, 2025 as an example, the intermediate record can be represented as the following structured data: {license_plate: Lu AXXXXX, tag_id: 3, date: 2025-02-20, night_pass_count: 15, total_pass_count: 20, night_ratio: 0.75}; where night_pass_count and total_pass_count are the original aggregated result values of the number of occurrences at night and throughout the day respectively, used to support possible complex research and judgment in the future, such as weighted calculation. night_ratio is the proportion of occurrences at night.

[0044] Crucially, if a vehicle does not meet the multi-dimensional aggregation conditions of any vehicle behavior tags on a certain day (for example, a private car that occasionally passes by checkpoint A only during the day), the system will not generate any intermediate records for it (S301). This "record only when hit" mechanism greatly compresses the data volume of the intermediate statistical table, only retaining the "candidate" data with business value, effectively solving the performance bottleneck caused by directly performing complex queries on a large amount of raw data. This intermediate statistical table is usually stored in a high-performance columnar database (such as ClickHouse), partitioned by date, and a joint index is established on the license plate number and behavior tag ID to optimize subsequent query performance.

[0045] Specifically, in a big data columnar database (such as ClickHouse), establishing a joint index means: on the intermediate statistical table, creating a composite index whose index key consists of two fields in a fixed order: for example, first is license_plate (license plate number), and second is tag_id (behavior tag ID).

[0046] The joint index allows the database engine to first quickly locate all record blocks of this license plate through license_plate = 'Lu AXXXXX', and then further filter through tag_id = 3 in this subset. Compared with scanning hundreds of millions of intermediate records without an index, this operation can reduce the I / O and CPU overhead by several orders of magnitude.

[0047] In addition, since the query conditions include both fields of the joint index (and in the same order), the database can make full use of the "leftmost prefix" feature of the index to achieve efficient point queries or small-range scans.

[0048] In actual business operations, the system needs to perform S400 queries in parallel on tens of thousands of license plates and dozens of tags every day. The composite index ensures that each query can be completed with O(log N) or even O(1) complexity, avoiding system avalanche caused by lock contention or resource contention.

[0049] As mentioned earlier, intermediate tables are typically partitioned by date. During queries, the system first limits the query to the nearest N partitions through partition pruning, and then uses a composite index within each partition for precise location, forming a dual filtering of "partition + index" that greatly improves throughput.

[0050] Furthermore, to ensure processing stability and throughput efficiency in massive data scenarios, this invention employs optimized strategies in both data reading and writing. Specifically, when performing S200 aggregation analysis, the system does not load all daily passage records from a single camera at once. Instead, it uses a pagination query mechanism (such as LIMIT / OFFSET based on license plate number or cursor pagination) to divide the millions of records into multiple small batches for sequential processing. After completing the multi-dimensional aggregation judgment for each batch, the intermediate records that meet the conditions are temporarily stored in a memory buffer. When the buffer reaches a preset capacity (e.g., 5000 records), the database's batch insert interface (such as executemany or ClickHouse's batch write API) is called to uniformly write the records to the intermediate statistics table. This approach effectively avoids memory overflow or database connection timeouts in a single operation, significantly improving the robustness and execution efficiency of daily feature extraction.

[0051] Similarly, during the dynamic tag generation phase from S400 to S500, the system also employs a pagination strategy to retrieve historical intermediate records in batches by license plate or tag ID when querying the intermediate statistics table, and submits the analysis results in batches. For example, when tags need to be generated for 100,000 vehicles per day, the system accumulates the final tag results into groups of thousands and writes them to the tag result table all at once using bulk_insert, instead of executing INSERT statements one by one. This collaborative mechanism of "pagination reading + batch writing" significantly reduces database I / O overhead and network round trips, enabling the present invention to achieve minute-level daily tag calculation in high-concurrency, large-data-volume real-world business environments, fully ensuring the real-time performance and scalability of the dynamic tag system.

[0052] S400: Responding to the dynamic tag generation task, based on the behavior tag ID, dynamic judgment time window, and judgment threshold included in the judgment rules corresponding to the dynamic tag generation task, it queries the intermediate statistics table for all intermediate records of the target license plate number under the behavior tag ID within the dynamic judgment time window. Vehicle behavior tags, multi-dimensional aggregation conditions, and judgment rules are all managed in a configurable manner, supporting dynamic addition, deletion, and modification without modifying the program code.

[0053] Specifically, the dynamic judgment time window is an N-day window that rolls based on the current time. The judgment threshold is M, indicating that there must be at least M intermediate records within the dynamic judgment time window. Both N and M are positive integers, and 1 ≤ M ≤ N.

[0054] By maintaining a label rule configuration table, the judgment logic (dependent statistical dimensions, time window, threshold conditions, special rules, etc.) of each dynamic label is parameterized. The judgment engine can dynamically combine different statistical dimensions and judgment logics by parsing the configuration, enabling the unified and flexible generation of diverse and complex vehicle behavior labels.

[0055] S500: Aggregate and judge the queried intermediate records. If the judgment threshold is met, generate the dynamic label corresponding to the target license plate number.

[0056] These two steps, S400 and S500, together constitute the flexible and configurable label judgment engine of the present invention. When the daily label generation task is triggered, the system reads the judgment rules of the "active at night and dormant during the day" label, which clearly specify the behavior label ID (3), the dynamic judgment time window (e.g., the last 30 days), and the judgment threshold (e.g., M = 15, which requires 15 days to meet the conditions within 30 days). The system then executes S400 to query all intermediate records with label ID 3 for the license plate "Lu AXXXXX" in the last 30 days from the intermediate statistical table. Since the intermediate table only stores records that "meet the preliminary conditions on a single day", the aggregate judgment in S500 becomes extremely simple: just count the number of queried records. If exactly 15 records are found, it indicates that the vehicle has met the feature of "nighttime occupancy > 70%" for 15 days in the past 30 days, meeting the judgment threshold M = 15. The system then generates a final dynamic label record ({license_plate: Lu AXXXXX, tag: zy3, match_count: 15}) and stores it in the label result table, where tag: zy3 is the dynamic label corresponding to "active at night and dormant during the day".

[0057] The judgment rule for the dynamic label of "cross-region movement": Based on the list of different regions where the vehicle appears in the daily statistics, directly count whether the number of regions reaches the threshold during judgment.

[0058] The method from S100 to S500 above is executed once a day for the daily vehicle passing bottom library of each camera, realizing the incremental update of the intermediate statistical table and enabling the dynamic label to be dynamically refreshed based on the latest dynamic judgment time window. The dynamic labels that no longer meet the conditions in history automatically become invalid.

[0059] As another possible embodiment of the present invention, after obtaining the vehicle passing bottom library corresponding to a certain camera on a specified date, the method further includes:

[0060] S110: Obtain the vehicle passage database from multiple cameras within a specified date, and globally aggregate it by license plate number to obtain the full passage event sequence for each vehicle on that specified date.

[0061] S120: For each vehicle's full passage event sequence, sort by passage time and cluster based on preset spatiotemporal proximity rules to merge consecutive passage events with time intervals less than a first threshold and spatial distances less than a second threshold into a single behavior segment.

[0062] S200: Using the multidimensional aggregation conditions corresponding to each vehicle behavior tag, perform aggregate analysis on all passage event records in the vehicle passage database, including:

[0063] S201: Based on behavioral fragments, determine whether the vehicle meets the multidimensional aggregation conditions.

[0064] The multidimensional aggregation condition is configured to depend on behavioral features across cameras, including:

[0065] The number of times a vehicle appears in at least K different geographical areas within a specified time period.

[0066] Does the average travel time of a vehicle from area A to area B fall within the preset range?

[0067] Does the number of cameras covering the vehicle's nighttime activity exceed the third threshold?

[0068] Building upon the basic embodiments of the methods defined in S100 to S500 above, this invention further provides more advanced analytical capabilities to address the limitations of single-point analysis. By executing S110 and S120, the system aggregates isolated passage events of a vehicle throughout the day into semantically meaningful "behavioral fragments." Based on this, the aggregation analysis in S201 can be judged based on these spatiotemporally coherent fragments. For example, multi-dimensional aggregation conditions can be configured to determine whether a vehicle's nighttime behavioral fragments cover at least three different administrative regions, thereby accurately identifying "nighttime roaming" vehicles; or whether its average travel time from residential areas to office areas stably falls within a specific range to identify "regular commuting" patterns; or the number of different cameras involved in its nighttime activities can be counted to assess the breadth of its activities. By introducing cross-camera spatiotemporal behavioral fragments, this embodiment elevates vehicle behavior analysis from isolated "points" to coherent "lines" and even "surfaces," greatly enriching the semantic connotation and business value of dynamic tags, and effectively solving the deficiency of existing technologies in being unable to identify complex movement patterns.

[0069] As another possible embodiment of the present invention, such as Figure 2 As shown, a vehicle dynamic label generation system based on multi-dimensional behavior analysis is provided. The system includes:

[0070] The vehicle access database acquisition module is used to obtain the vehicle access database corresponding to a specific camera on a specified date. The vehicle access database contains all vehicle access event records recorded by the camera on that specified date, and each access event record includes at least the license plate number, access time, and the location identifier of the camera.

[0071] The behavior aggregation module is used to aggregate and analyze all passage event records in the vehicle passage database using the multidimensional aggregation conditions corresponding to each vehicle behavior tag, and to determine whether the vehicle corresponding to each license plate number meets the multidimensional aggregation conditions. The multidimensional aggregation conditions are used to characterize the behavioral features of the vehicle defined by the corresponding vehicle behavior tag in the time, space, or spatiotemporal combination dimensions.

[0072] The intermediate record generation module generates an intermediate record and adds it to the intermediate statistics table if the passage event record corresponding to a certain license plate number meets the multi-dimensional aggregation conditions of a certain vehicle behavior tag. The intermediate record must at least include the license plate number, the behavior tag ID of the vehicle behavior tag, the aggregation result value, and the specified date.

[0073] The dynamic analysis module is used to respond to the dynamic tag generation task. Based on the behavior tag ID, dynamic analysis time window and analysis threshold included in the analysis rules corresponding to the dynamic tag generation task, it queries the intermediate statistics table for all intermediate records of the target license plate number under the behavior tag ID within the dynamic analysis time window.

[0074] The dynamic tag generation module is used to aggregate and judge the intermediate records retrieved. If the judgment threshold is met, a dynamic tag corresponding to the target license plate number is generated.

[0075] Furthermore, although the steps of the method in this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps.

[0076] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, mobile terminal, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0077] In an exemplary embodiment of this disclosure, an electronic device capable of implementing the above-described method is also provided.

[0078] Those skilled in the art will understand that various aspects of the present invention can be implemented as systems, methods, or program products. Therefore, various aspects of the present invention can be specifically implemented in the following forms: entirely in hardware, entirely in software (including firmware, microcode, etc.), or in a combination of hardware and software, collectively referred to herein as “circuit,” “module,” or “system.”

[0079] An electronic device according to this embodiment of the invention. The electronic device is merely an example and should not be construed as limiting the functionality or scope of the embodiments of the invention.

[0080] Electronic devices are manifested in the form of general-purpose computing devices. Components of an electronic device may include, but are not limited to: at least one processor, at least one memory, and buses connecting different system components (including memory and processor).

[0081] The memory stores program code that can be executed by a processor, causing the processor to perform the steps described in the "Exemplary Methods" section above, according to various exemplary embodiments of the present invention.

[0082] The storage may include readable media in the form of volatile storage, such as random access memory (RAM) and / or cache memory, and may further include read-only memory (ROM).

[0083] The storage may also include programs / utilities having a set (at least one) of program modules, including but not limited to: an operating system, one or more applications, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0084] A bus can represent one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus that uses any of the various bus architectures.

[0085] The electronic device can also communicate with one or more external devices (e.g., keyboards, pointing devices, Bluetooth devices, etc.), one or more devices that enable a user to interact with the electronic device, and / or any device that enables the electronic device to communicate with one or more other computing devices (e.g., routers, modems, etc.). This communication can be performed via input / output (I / O) interfaces. Furthermore, the electronic device can communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter. The network adapter communicates with other modules of the electronic device via a bus. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with the electronic device, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0086] In exemplary embodiments of this disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the methods described above is stored. In some possible embodiments, various aspects of the present invention may also be implemented as a program product comprising program code that, when the program product is run on a terminal device, causes the terminal device to perform the steps of the various exemplary embodiments of the present invention described in the "Exemplary Methods" section above.

[0087] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0088] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting programs for use by or in conjunction with an instruction execution system, apparatus, or device.

[0089] The program code contained on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0090] Program code for performing the operations of this invention can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0091] Furthermore, the accompanying drawings are merely illustrative of the processes included in the method according to exemplary embodiments of the present invention and are not intended to be limiting. It is readily understood that the processes shown in the above drawings do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0092] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0093] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for generating dynamic vehicle tags based on multi-dimensional behavior analysis, characterized in that, The method includes the following steps: Obtain the vehicle access database corresponding to a certain camera on a specified date; the vehicle access database contains all vehicle access event records recorded by the camera on the specified date, and each access event record includes at least the license plate number, access time and the location identifier of the camera; Using the multidimensional aggregation conditions corresponding to each vehicle behavior tag, perform aggregation analysis on all passage event records in the vehicle passage database to determine whether the vehicle corresponding to each license plate number meets the multidimensional aggregation conditions. The multidimensional aggregation conditions are used to characterize the behavioral features of the vehicle defined by the corresponding vehicle behavior label in the dimensions of time, space, or spatiotemporal combination. If a passage event record corresponding to a certain license plate number satisfies the multidimensional aggregation condition of a certain vehicle behavior tag, an intermediate record is generated and included in the intermediate statistics table; the intermediate record shall at least include the license plate number, the behavior tag ID of the vehicle behavior tag, the aggregation result value, and the specified date; In response to the dynamic tag generation task, based on the behavior tag ID, dynamic judgment time window and judgment threshold included in the judgment rule corresponding to the dynamic tag generation task, the intermediate statistics table is queried for all intermediate records of the target license plate number under the behavior tag ID within the dynamic judgment time window; vehicle behavior tags, multi-dimensional aggregation conditions and judgment rules are all managed in a configurable manner, supporting dynamic addition, deletion and modification without modifying the program code. The intermediate records retrieved are aggregated and judged. If the judgment threshold is met, a dynamic tag corresponding to the target license plate number is generated.

2. The method according to claim 1, characterized in that, After performing aggregate analysis on all passage event records in the vehicle passage database using the multidimensional aggregation conditions corresponding to each vehicle behavior tag, and determining whether the vehicle corresponding to each license plate number meets the multidimensional aggregation conditions, the method further includes: If the passage event record corresponding to a certain license plate number does not meet the multidimensional aggregation conditions corresponding to any vehicle behavior label, then no intermediate record corresponding to that license plate number will be generated.

3. The method according to claim 1, characterized in that, The multidimensional aggregation conditions include at least one of the following: the number of times a vehicle passes through a specified time period, the frequency of a vehicle's appearance in a specified geographical area, or the proportion of the number of times a vehicle passes through a specific time period and a specific area to its total number of passes through the camera on that day.

4. The method according to claim 1, characterized in that, The dynamic judgment time window is an N-day window that changes dynamically based on the current time; the judgment threshold is M, which means that there must be at least M days of intermediate records within the dynamic judgment time window, where N and M are both positive integers 1≤M≤N.

5. The method according to claim 1, characterized in that, The intermediate statistics table is stored in partitions by date, and a joint index is established for license plate number and behavior tag ID to accelerate query operations based on behavior tag ID and dynamic judgment time window.

6. The method according to claim 1, characterized in that, The vehicle behavior tags include vehicles that operate during the day and at night, vehicles that reside at a fixed location, or vehicles that frequently move across regions. Each type of vehicle behavior tag corresponds to a unique multidimensional aggregation condition and judgment rule.

7. The method according to claim 1, characterized in that, The method is executed once a day on the vehicle passage database of each camera, realizing incremental updates of the intermediate statistics table, and making the dynamic tags dynamically refreshed based on the latest dynamic judgment time window. Dynamic tags that no longer meet the conditions in the past are automatically invalidated.

8. The method according to claim 1, characterized in that, Retrieve the vehicle access data for a specific camera on a specified date, including: Retrieve the original video stream or image sequence captured by one of the cameras on the specified date; The original video stream or image sequence is processed by license plate recognition and timestamp extraction to obtain structured traffic event data; The structured traffic event data is collected according to the specified date and the location identifier of a certain camera to form the vehicle traffic database; In each passage event record, the passage time is generated based on the system clock or timing signal of the camera, and the license plate number is the recognition result after confidence verification.

9. The method as described in claim 1, characterized in that, After obtaining the vehicle access data from a specific camera on a specified date, the process also includes: Obtain the vehicle passage database from multiple cameras within the specified date, and globally aggregate it by license plate number to obtain the full passage event sequence for each vehicle on the specified date; For each vehicle's full passage event sequence, sort by passage time and cluster based on a preset spatiotemporal proximity rule to merge consecutive passage events with a time interval less than a first threshold and a spatial distance less than a second threshold into a single behavior segment. The method of using multidimensional aggregation conditions corresponding to each vehicle behavior tag to perform aggregation analysis on all passage event records in the vehicle passage database includes: Based on the behavioral fragment, determine whether the vehicle meets the multidimensional aggregation condition; The multidimensional aggregation conditions are configured to depend on behavioral features across cameras, including: The number of times a vehicle appears in at least K different geographical areas within a specified time period; Does the average travel time of a vehicle from area A to area B fall within the preset range? Does the number of cameras covering the vehicle's nighttime activity exceed the third threshold? 10. A vehicle dynamic label generation system based on multi-dimensional behavior analysis, characterized in that, The system includes: The vehicle access database acquisition module is used to acquire the vehicle access database corresponding to a certain camera on a specified date; the vehicle access database contains all vehicle access event records recorded by the camera on the specified date, and each access event record includes at least the license plate number, access time and the location identifier of the camera; The behavior aggregation module is used to perform aggregation analysis on all passage event records in the vehicle passage database using the multidimensional aggregation conditions corresponding to each vehicle behavior tag, and to determine whether the vehicle corresponding to each license plate number meets the multidimensional aggregation conditions; the multidimensional aggregation conditions are used to characterize the behavioral features of the vehicle defined by the corresponding vehicle behavior tag in the time, space or spatiotemporal combination dimensions. The intermediate record generation module is used to generate an intermediate record and add it to the intermediate statistics table if the passage event record corresponding to a certain license plate number meets the multidimensional aggregation condition of a certain vehicle behavior tag; the intermediate record includes at least the license plate number, the behavior tag ID of the vehicle behavior tag, the aggregation result value, and the specified date; The dynamic analysis module is used to respond to the dynamic tag generation task. Based on the behavior tag ID, dynamic analysis time window and analysis threshold included in the analysis rules corresponding to the dynamic tag generation task, it queries the intermediate statistics table for all intermediate records of the target license plate number under the behavior tag ID within the dynamic analysis time window. The vehicle behavior tags, multi-dimensional aggregation conditions and analysis rules are all managed in a configurable manner, supporting dynamic addition, deletion and modification without modifying the program code. The dynamic tag generation module is used to aggregate and judge the intermediate records found in the query. If the judgment threshold is met, a dynamic tag corresponding to the target license plate number is generated.

Images