Dynamic patrol-based street abnormal occupation detection method and system

Abstract

The application discloses a street abnormal occupation detection method and system based on dynamic patrol. The method comprises the following steps: acquiring a street video stream, GPS data and a timestamp; identifying a traffic area and municipal facilities in each frame of the street video stream, combining the timestamp and geographic information, performing semantic matching with a street rule library, combining with a street environment structure diagram, generating a dynamic compliance benchmark diagram of each frame, and determining an unidentified target; detecting and tracking the unidentified target, screening out a target in a stationary state, comparing the target with the dynamic compliance benchmark diagram, marking a target in violation, so as to obtain a suspected target; identifying attribute and behavior characteristics of the suspected target, judging compliance according to a business rule library, so as to determine a violation event. The method of the application realizes abnormal occupation detection with the fusion of mobile scene adaptability, dynamic business rule understanding and application ability.

Description

A Method and System for Detecting Abnormal Encroachment on Streets Based on Dynamic Patrol Technical Field

[0001] This invention relates to computer vision, and more specifically to a method and system for detecting abnormal road occupancy along streets based on dynamic patrol. Background Technology

[0002] In current urban management, enhancing the intelligent monitoring capabilities of public space order is crucial for achieving refined urban governance. However, existing technologies have several limitations and shortcomings in addressing this challenge.

[0003] First, traditional surveillance methods rely on fixed cameras or manual patrols. While fixed cameras can provide continuous monitoring in specific locations, their fixed viewing angles and numerous blind spots prevent them from achieving comprehensive, blind-spot-free coverage of urban streetscapes. On the other hand, while manual patrols can flexibly handle various complex situations, their low efficiency and high labor costs are limiting factors.

[0004] Secondly, although drone inspection, as an emerging mobile monitoring technology, has improved flexibility and coverage, it is limited by physical conditions such as short battery life and limited single inspection range. Coupled with the influence of weather factors, its stability and security face challenges, making it difficult to guarantee long-term effective monitoring services.

[0005] Furthermore, although modern image recognition technology can accurately detect a variety of target objects, it often encounters the dilemma of being easy to identify but difficult to judge in practical applications. This is especially true for various objects operating on the streets in cities, such as food trucks, goods stacks, and temporary structures. Because these objects are diverse and change rapidly, systems based on pure visual recognition struggle to effectively determine which behaviors comply with regulations and which require correction. There is a lack of a mechanism to deeply integrate information from the physical space with urban management rules to support the shift from simple identification to intelligent judgment.

[0006] Therefore, it is necessary to design a new method to achieve abnormal lane occupancy detection that integrates mobile scenario adaptability, dynamic business rule understanding and application capabilities, and solves the problems of logic and business rules being disconnected, low data processing efficiency and rigid rule configuration in existing technologies. Summary of the Invention

[0007] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and system for detecting abnormal road occupancy along the street based on dynamic inspection.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: a method for detecting abnormal road occupancy along streets based on dynamic patrol, comprising:

[0009] Acquire street view video streams, GPS data, and timestamps;

[0010] For each frame of the street view video stream, the traffic area and municipal facilities are identified, combined with timestamps and geographic information, semantically matched with the street view rule base, and combined with the street view environment structure map to generate a dynamic compliance benchmark map for each frame and determine unidentified targets.

[0011] The unidentified targets are detected and tracked, targets in a stationary state are filtered out, and compared with the dynamic compliance benchmark map to mark the non-compliant targets, so as to obtain the suspected targets;

[0012] The suspected targets are identified by attribute and behavioral characteristics, and compliance is judged based on the business rule base to determine the violation.

[0013] The further technical solution is as follows: For each frame of the street view video stream, road functional areas and municipal facilities are identified, and a dynamic compliance benchmark map for each frame is generated in conjunction with a rule base, and unidentified targets are determined, including:

[0014] Identify and classify the traffic areas and municipal facilities in each frame of the street view video stream, and combine GPS data to determine custom functional areas to obtain general area identification results, municipal facility identification results, and unidentified targets;

[0015] Based on the street view rule library, the general area identification results and municipal facility identification results are combined with timestamps and geographic information to dynamically generate a dynamic compliance benchmark map for each frame.

[0016] The further technical solution is as follows: based on the street view rule library, the general area identification results and municipal facility identification results are combined with timestamps and geographic information to dynamically generate a dynamic compliance benchmark map for each frame, including:

[0017] The general area identification results and municipal facility identification results are combined with timestamps and geographic information, and semantically matched with the content in the street view rule base to obtain the matching results; wherein, the street view rule base includes spatial range, time conditions, constraint content and priority;

[0018] The pairing results are combined with the street view environment structure map in the form of dynamic tags to obtain a dynamic compliance benchmark map for each frame.

[0019] The further technical solution is as follows: the dynamic compliance benchmark map is matched with the content of the street view rule library based on different traffic areas, municipal facilities, timestamps and geographic information.

[0020] The further technical solution is as follows: The detection and tracking of the unidentified targets, filtering out stationary targets, comparing them with the dynamic compliance benchmark map, and marking the non-compliant targets to obtain suspected targets includes:

[0021] Track the unidentified targets and analyze their motion state to filter out targets that are stationary, thus obtaining static targets;

[0022] Based on the dynamic compliance benchmark map, the static targets are spatially mapped and matched according to rules, and non-suspect targets and suspect targets are classified.

[0023] The further technical solution is as follows: tracking the unidentified target and analyzing the motion state of the unidentified target to filter out targets in a stationary state to obtain static targets includes:

[0024] The unidentified targets in the street view video stream are identified and roughly classified as motor vehicles, non-motor vehicles, people, or objects.

[0025] The system uses multi-target tracking and short-term trajectory analysis to determine whether the unidentified target is stationary, and marks the unidentified target that meets the criteria as a static target and records its basic information.

[0026] The further technical solution is as follows: The spatial location mapping and rule matching of the static target based on the dynamic compliance benchmark map, classifying non-suspect targets and suspect targets, includes:

[0027] Calculate the specific ground contact point of the image corresponding to the static target in the street view video stream, and map the specific ground contact point onto the dynamic compliance benchmark map to determine the functional area where it is located;

[0028] Matching is performed based on the functional area rules where the static target is located, excluding non-suspected targets located in unregulated areas or belonging to people, and marking suspected targets that violate regulations.

[0029] The further technical solution is as follows: The identification of the suspected target's attributes and behavioral characteristics, and the determination of compliance based on the business rule base, to identify the violation event, includes:

[0030] The suspected targets are meticulously classified, and their current behavior and state characteristics are analyzed to obtain attributes and behaviors.

[0031] The attributes and behaviors of the suspected target are compared with the business rule base. Based on the functional area and the nature of the activity, and combined with the time factor, the compliance of the suspected target is determined to identify the violation event.

[0032] The further technical solution is as follows: after identifying the attributes and behavioral characteristics of the suspected target, judging compliance based on the business rule base, and determining the violation event, it also includes:

[0033] A structured processing work order is generated for the violation event, and the structured processing work order is output; the structured processing work order includes target ID, geographic coordinates, functional area, attributes, behavior, timestamp, and confidence level.

[0034] This invention also provides a street-side abnormal encroachment detection system based on dynamic patrol, comprising:

[0035] The acquisition unit is used to acquire street view video streams, GPS data, and timestamps.

[0036] The generation unit is used to identify the passage area and municipal facilities in each frame of the street view video stream, combine them with timestamps and geographic information, perform semantic matching with the street view rule base, and combine them with the street view environment structure map to generate a dynamic compliance benchmark map for each frame and determine unidentified targets.

[0037] The suspected target determination unit is used to detect and track the unidentified targets, filter out targets in a stationary state, compare them with the dynamic compliance benchmark map, mark the non-compliant targets, and obtain suspected targets.

[0038] The violation event determination unit is used to identify the attributes and behavioral characteristics of the suspected target, and to determine compliance based on the business rule base in order to identify the violation event.

[0039] The advantages of this invention compared to existing technologies are as follows: By integrating mobile scenario adaptability, dynamic business rule understanding, and application capabilities, this invention solves the problems of disconnect between logic and business rules, low data processing efficiency, and rigid rule configuration in existing technologies. Specifically, it first acquires street view video streams, GPS data, and timestamps, and identifies the traffic areas and municipal facilities in each frame of the video stream. Semantic matching is then performed between geographic information and timestamps and the street view rule base to generate a dynamic compliance benchmark map for each frame and identify unidentified targets. Subsequently, these unidentified targets are detected and tracked, and stationary targets are selected and compared with the dynamic compliance benchmark map to mark violating targets as suspected targets. Finally, by identifying the attributes and behavioral characteristics of suspected targets, compliance is determined based on the business rule base, thereby accurately locating violation events. This process not only achieves real-time monitoring and analysis of the urban environment but also optimizes data processing efficiency through a progressive filtering mechanism. Simultaneously, it supports custom rule configuration, improving the system's flexibility and adaptability, and ensuring effective response to complex and ever-changing urban management needs.

[0040] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0041] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 is a flowchart illustrating the method for detecting abnormal road occupancy based on dynamic patrol provided in an embodiment of the present invention.

[0043] Figure 2 is a schematic diagram of an image provided in an embodiment of the present invention;

[0044] Figure 3 is a schematic diagram of the dynamic compliance benchmark diagram generation provided in an embodiment of the present invention;

[0045] Figure 4 is a schematic diagram of a target in a stationary state provided in an embodiment of the present invention;

[0046] Figure 5 is a schematic diagram of a suspected target provided in an embodiment of the present invention;

[0047] Figure 6 is a schematic block diagram of the street abnormality detection system based on dynamic patrol provided in an embodiment of the present invention;

[0048] Figure 7 is a schematic block diagram of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0049] 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, not all, of the embodiments of the present invention. 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.

[0050] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0051] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0052] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0053] Please refer to Figure 1, which is a flowchart illustrating the method for detecting abnormal street occupancy based on dynamic patrol according to an embodiment of the present invention. This method is applied to a server. By integrating street view video streams, GPS data, and timestamps, it identifies the traffic area and municipal facilities in each frame of the image and combines them with timestamps and geographic information to generate a dynamic compliance benchmark map, enabling the detection, tracking, and marking of unidentified targets and violations. The method further utilizes a business rule base to identify the attributes and behavioral characteristics of suspected targets to determine their compliance and identify violations. The entire process integrates mobile scenario adaptability, understanding and application of dynamic business rules, effectively solving the problems of disconnect between logic and business rules, low data processing efficiency, and rigid rule configuration in existing technologies. It achieves efficient and accurate identification and processing of abnormal street occupancy, improving the level of intelligent urban management. Furthermore, by generating structured processing work orders for violations, the efficiency and accuracy of subsequent processing are improved.

[0054] Figure 1 is a flowchart illustrating the method for detecting abnormal road occupancy based on dynamic patrol according to an embodiment of the present invention. As shown in Figure 1, the method includes the following steps S110 to S140.

[0055] S110: Acquire street view video stream, GPS data, and timestamps.

[0056] In this embodiment, the street view video stream refers to the real-time video of the street environment collected by the terminal equipment on the dynamic inspection vehicle. These video streams provide a continuous sequence of images for analyzing changes in the status of traffic areas and municipal facilities on the street, serving as the basic data source for detecting abnormal road occupancy. After the device is turned on, the video stream, GPS data, and timestamps are packaged and uploaded to the cloud server in real time through the vehicle communication unit, completing the initial data aggregation and storage.

[0057] GPS data refers to geographic information acquired through the Global Positioning System. For each frame of video imagery, the corresponding GPS data provides precise location information, enabling the system to associate identified road functional zones, municipal facilities, and unidentified targets with specific geographic locations. This is crucial for building dynamic compliance baseline maps, as it allows the system to apply appropriate management rules based on different geographic conditions.

[0058] Timestamp: Represents the point in time when each video frame or data record was generated. It provides the system with time-dimensional information, allowing dynamic management rules to be adjusted according to the requirements of different time periods (e.g., prohibiting parking during specific times). Furthermore, combining GPS data and timestamps allows for comparison of data from the same location at different times to monitor changes over time and ensure the traceability of all operations.

[0059] These three elements (street view video stream, GPS data, and timestamps) work together to provide the necessary input information for subsequent steps such as street view environment analysis, static target detection, preliminary screening of suspected targets, and feature behavior recognition. They not only support the deep integration of physical spatial information and dynamic management rules but also lay the foundation for moving from simple target recognition to rule-based intelligent judgment. Through this comprehensive processing approach, the system can more accurately identify illegal road occupation and improve urban management efficiency and response speed.

[0060] S120. For each frame of the street view video stream, identify the access area and municipal facilities, combine them with the timestamp and geographic information, perform semantic matching with the street view rule base, and combine them with the street view environment structure map to generate a dynamic compliance benchmark map for each frame, and determine the unidentified targets.

[0061] In this embodiment, the dynamic compliance baseline map refers to a base map dynamically generated based on information such as the passable area, municipal facility identification results, GPS data (geographical location), and timestamps of each frame of image, combined with the latest street view rule base content, for subsequent judgment of violations. It can reflect the current status of different functional areas and their corresponding management rules in real time, thereby supporting the system to make accurate judgments on whether the behavior of static or dynamic targets complies with regulations.

[0062] Unidentified targets: These refer to independent entities that, after initial identification by traffic areas and municipal facilities, cannot be clearly categorized into a target type (such as motor vehicles, non-motor vehicles, people, objects, etc.). These targets require further analysis to determine their attributes and potential violations.

[0063] In one embodiment, step S120 described above may include steps S121 to S122.

[0064] S121. Identify and classify the traffic areas and municipal facilities in each frame of the street view video stream, and combine GPS data to determine the custom functional area to obtain general area identification results, municipal facility identification results, and unidentified targets.

[0065] In this embodiment, the general region recognition result refers to the result obtained by performing pixel-level segmentation of the passable regions (such as "pedestrian walkways", "non-motorized vehicle lanes", "motorized vehicle lanes", etc.) in the image using a semantic segmentation model. These results help the system understand the functional partitions of the current scene.

[0066] Municipal facility identification results refer to common municipal facilities (such as parking spaces / lines, trash cans, etc.) identified in the same image frame by the object detection model. The location and status of these facilities are crucial for determining whether the target behavior is compliant.

[0067] This step aims to provide foundational information for subsequent steps, ensuring that all key elements are correctly identified and categorized.

[0068] S122. Based on the street view rule library, combine the general area identification results and municipal facility identification results with timestamps and geographic information to dynamically generate a dynamic compliance benchmark map for each frame.

[0069] In one embodiment, step S122 described above may include steps S1221 to S1222.

[0070] S1221. Combine the general area identification result and the municipal facility identification result with the timestamp and geographic information, and perform semantic pairing with the content in the street view rule base to obtain the pairing result; wherein, the street view rule base includes spatial range, time conditions, constraint content and priority.

[0071] In this embodiment, the matching result refers to the process of matching the identification results of the access area and municipal facilities with the corresponding entries (including spatial range, time conditions, constraints, and priorities) in the street view rule base. For example, a rule prohibiting parking in a certain location during a specific time period will be applied to the corresponding area identification result.

[0072] Street view rule bases are important tools in urban management for regulating and guiding various behaviors. They ensure the effective management and rational use of urban space by setting a set of clear, specific, and enforceable rules. The following are the main components of a street view rule base:

[0073] Spatial scope refers to the specific geographical location or area to which the rules apply. This can include, but is not limited to:

[0074] Functional zoning: such as urban zoning for different purposes, such as commercial areas, residential areas, and industrial areas.

[0075] Specific locations: such as parks, squares, pedestrian streets and other public places.

[0076] Geographic boundaries: Define which geographic coordinate ranges are subject to this rule.

[0077] By clearly defining the spatial scope, it is possible to ensure that various rules are customized for specific locations, thereby improving the accuracy and relevance of urban management.

[0078] The time condition refers to the periodic time or periodic characteristic during which the rule takes effect. This includes:

[0079] All-day rule: Applicable at any time of day.

[0080] Specific time period rules: These rules are only valid within a specified time period, such as business hours restrictions between 7 a.m. and 9 p.m.

[0081] Seasonal or holiday rules: Management measures are adjusted according to seasonal changes or special holidays, such as extending outdoor seating time in summer and relaxing some traffic control during the Spring Festival.

[0082] Time constraints allow for flexible adjustments to the rules based on actual circumstances to adapt to changing needs at different times.

[0083] Constraints are the core of the rule base, clearly defining permitted or prohibited behaviors and their associated requirements. Common constraints may include:

[0084] Code of conduct: such as prohibiting littering and not occupying public spaces without permission.

[0085] Regulations on the use of facilities: Regulations on how to properly use urban facilities, such as bicycle lanes are for cyclists only and motor vehicles are not allowed to enter.

[0086] Environmental protection requirements include regulations on noise control and green space protection.

[0087] These restrictions aim to maintain the city's order, safety, and aesthetics, and to improve the quality of life for residents.

[0088] Priority determines which rule should be followed when multiple rules apply simultaneously. Typically, rule priority can be set based on the following factors:

[0089] Urgency level: such as fire safety regulations, which usually have the highest priority.

[0090] Temporary regulations in special circumstances: During large-scale events, there may be temporary special regulations that go beyond the regular rules.

[0091] Clearly defining priorities helps resolve rule conflicts and ensures the consistency and authority of urban management decisions.

[0092] In conclusion, the Street View rule base provides scientific and systematic methodological support for urban management by specifying in detail the spatial scope, temporal conditions, constraints, and priorities, thus promoting the harmonious development of the urban environment.

[0093] S1222. The pairing results are combined with the street view environment structure map in the form of dynamic labels to obtain a dynamic compliance benchmark map for each frame.

[0094] The dynamic compliance benchmark map is matched with the content of the street view rule base based on different traffic areas, municipal facilities, timestamps, and geographic information.

[0095] The dynamic compliance baseline map is dynamically generated based on different access areas, municipal facilities, timestamps, and geographic information, combined with the content of the street view rule base. This enables the system to accurately determine the types of behaviors allowed or prohibited in each area at any given time, thereby realizing the transformation from simple target recognition to rule-based intelligent judgment.

[0096] In this embodiment, as shown in Figures 2 and 3, each frame of the acquired street view image is analyzed in detail to identify road functional areas and municipal management facilities, and a dynamic compliance benchmark atlas is generated by combining the latest street view rule base. This process involves not only the identification of static elements (such as traffic areas and municipal facilities), but also the dynamic adjustment of these elements as they change with time and geographical location.

[0097] A semantic segmentation model is used to perform pixel-level segmentation on each image input in real time. Based on the color and shape distribution characteristics of each lane, the road is divided into key functional areas such as "pedestrian walkway", "non-motorized vehicle lane" and "motorized vehicle lane".

[0098] This process requires handling a large amount of visual information to ensure accurate differentiation of different types of passageways, even under varying lighting conditions or weather conditions. Therefore, it is highly adaptable and dynamic, capable of handling a variety of complex scenarios.

[0099] Municipal facility recognition: The same frame of image is simultaneously fed into the target detection model to detect common municipal facilities, such as parking spaces / lines, trash cans, etc.

[0100] The target detection model must not only be able to quickly and accurately locate these facilities, but also be able to adapt to new facility types, maintaining the system's flexibility and scalability.

[0101] Custom function area: By comparing the GPS location of the image screen area with the GPS of the dynamically updated custom function area sent from the cloud, the custom function management area is identified.

[0102] The existence of a customizable function area allows the system to flexibly configure the monitoring scope and rules according to specific needs, further enhancing the system's dynamic response capabilities.

[0103] The results of general area identification and municipal facility identification are overlaid to form a preliminary street view environmental structure map. The core of this step is to build a compliance judgment benchmark that is dynamically generated and updated frame by frame during the inspection process.

[0104] Street View Rule Base: Allows users to define rules for each area through text descriptions or custom bounding boxes. The minimum structured data in the rule base includes: specific spatial range (i.e., the effective geographical range), time conditions (i.e., the effective time of the rule), constraint content (i.e., the combination of prohibited or allowed objects), and priority (used to determine when there is a rule conflict).

[0105] The rule base is designed to take into account the changing factors of time and space, so that the rules can be dynamically adjusted as the actual situation changes, thereby ensuring the timeliness and accuracy of the system.

[0106] Based on the general area identification results, municipal facility identification results, GPS, and timestamps of the street view environment structure map, semantic pairing is performed with the content in the street view rule base, and the results are combined with the street view environment structure map in the form of dynamic tags to form a unique compliance benchmark map for each frame.

[0107] The dynamic nature of the system is evident in the fact that the functional area boundaries, municipal facilities, location, and time may differ for each image, resulting in variations in the corresponding street view rule set. This means the system can provide the most suitable compliance judgment criteria for each frame, ensuring accurate detection and timely response to violations.

[0108] In summary, through the aforementioned mechanisms, the street view environment analysis and compliance benchmark construction module not only effectively identifies static elements but also dynamically adjusts its working mode and judgment criteria based on changes in time and geographical location, greatly improving the system's adaptability and efficiency. This dynamic adjustment capability is the key difference between this system and traditional methods, providing strong technical support for urban management.

[0109] Through the above steps, the system can not only accurately identify and classify various targets, but also dynamically adjust its judgment criteria according to the actual situation, thereby improving the accuracy and efficiency of road occupancy detection, while reducing the false judgment rate and enhancing the system's adaptability and flexibility.

[0110] S130. Detect and track the unidentified targets, filter out the targets in a stationary state, compare them with the dynamic compliance benchmark map, mark the targets that violate the rules, and obtain the suspected targets.

[0111] In this embodiment, suspected targets refer to those targets that, after preliminary analysis, are identified as potentially violating regulations. These targets require further refined analysis to confirm whether they have indeed violated relevant rules. Specifically, suspected targets refer to static targets located in specific functional areas (such as pedestrian walkways, vehicle lanes, etc.) whose existence or behavior may not conform to the regulations of that area. For example, a motor vehicle parked on a pedestrian walkway where parking is not permitted is considered a suspected target.

[0112] In one embodiment, step S130 described above may include steps S131 to S132.

[0113] S131. Track the unidentified target and analyze the motion state of the unidentified target to filter out targets that are stationary, so as to obtain static targets.

[0114] In this embodiment, a static target refers to all targets that have not been identified by previous modules in the video stream, and targets that are stationary are filtered out by analyzing their motion state.

[0115] In one embodiment, step S131 described above may include steps S1311 to S1312.

[0116] S1311. Identify and roughly classify the unidentified targets in the street view video stream as motor vehicles, non-motor vehicles, people, or objects.

[0117] First, a lightweight, general-purpose object detection model is used to roughly classify unidentified targets in the street view video stream. This step categorizes unidentified targets into four main types: motor vehicles, non-motor vehicles, people, or objects.

[0118] S1312. Determine whether the unidentified target is stationary through multi-target tracking and short-term trajectory analysis, and mark the unidentified target that meets the conditions as a static target and record its basic information.

[0119] Next, a multi-target tracking algorithm is used to assign a unique ID to each detected target, and short-term trajectory analysis is used to determine whether these targets are stationary. If a target remains relatively stationary for N consecutive frames, it is marked as a static target, and basic information including ID, image slice, location, and general category is recorded.

[0120] S132. Based on the dynamic compliance benchmark map, perform spatial location mapping and rule matching on the static target, and classify non-suspect targets and suspect targets.

[0121] In one embodiment, step S132 described above may include steps S1321 to S1322.

[0122] S1321. Calculate the specific ground contact point of the image corresponding to the static target in the street view video stream, and map the specific ground contact point onto the dynamic compliance benchmark map to determine the functional area where it is located.

[0123] The specific ground contact point of each static target is calculated and mapped onto the dynamic compliance baseline map corresponding to the current frame to determine the specific functional area where the target is located. For example, it determines whether the target is located on a pedestrian walkway, a driveway, a parking space, or a custom functional area.

[0124] S1322. Match according to the functional area rules where the static target is located, exclude non-suspected targets located in unregulated areas or belonging to people, and mark suspected targets that violate regulations.

[0125] A rapid matching process is performed based on the functional area where the static target is located and its underlying prohibition rules. If the target is located in an unregulated area or belongs to the "person" category, it is directly excluded as a non-suspect target. Conversely, if the target is located in an area where its existence or activity is prohibited, it is marked as a suspect target and prepared for the next stage of refined analysis.

[0126] In this embodiment, as shown in Figures 4 and 5, a lightweight general object detection model is used to coarsely classify all remaining unidentified independent objects in the video frame. This model can quickly select these objects and categorize them into a limited set of general basic categories such as "motor vehicle," "non-motor vehicle," "person," or "object." This method not only improves processing speed but also ensures the effectiveness of subsequent analysis.

[0127] Once the initial classification of all targets is complete, the next step is to analyze the motion state of each target using a multi-target tracking algorithm. Specifically, the system assigns a unique ID to each detected target and tracks its position changes across consecutive frames, forming short-term motion trajectories. Based on these trajectories, the system uses motion state analysis to determine whether a target is in a relatively stationary state. If a target remains relatively stationary for N consecutive frames, it is marked as a static target. Subsequently, the system outputs relevant information about the static target, including its ID, image slice, location, and general category, providing necessary data support for the next stage of initial screening.

[0128] After identifying the static target, the next step is to use the center point of its ground bounding box bottom edge as the ground contact point location and map it onto the dynamic compliance baseline map generated in the current frame. This process is crucial for accurately determining the specific functional area where the target is located (such as pedestrian walkways, vehicle lanes, parking spaces, custom functional areas, etc.). The dynamic compliance baseline map not only reflects the physical spatial structure but also integrates the latest management business rules, thus it can automatically update based on the time and geographical location of each frame to adapt to constantly changing environmental conditions and management requirements.

[0129] Based on the above spatial location mapping results, the system will quickly match and sort the static targets with the basic prohibition rules of each functional area in the dynamic compliance benchmark diagram:

[0130] Scenario A (Non-Suspicious Target): If the target is a "person," or is located in an area explicitly not regulated by the system's rule base, the system will directly exclude it without further processing. This is because some areas may not require monitoring of specific types of targets due to special regulations.

[0131] Scenario B (Suspected Target): If a target is located in an area that contains a "prohibited" judgment corresponding to its basic category, such as a "motor vehicle" or "object" stationary on the sidewalk, this is considered a violation and requires further attention.

[0132] On the other hand, if the target is located within the permitted area corresponding to its basic category, such as a "motor vehicle" parked in a parking space, it is considered compliant.

[0133] In this way, the role of the dynamic compliance benchmark map is fully realized. It not only helps the system accurately locate the position of each static target and its corresponding functional area, but also determines whether these targets have engaged in any violations based on real-time updated management rules. This intelligent analysis method, which combines physical spatial information and business rules, greatly improves the system's accuracy and intelligence level, while effectively reducing the consumption of computing resources and achieving efficient operation of urban management.

[0134] Through steps S130-S132 described above, the system can effectively filter out truly suspicious targets from a large number of unidentified targets. This method not only improves the system's processing efficiency and reduces unnecessary computational resource consumption, but also enhances the accuracy and intelligence of the judgment. It represents a significant shift from simple target identification to rule-based intelligent judgment, ensuring the effective implementation and timely response of urban management measures.

[0135] S140. Identify the attributes and behavioral characteristics of the suspected target, and determine compliance based on the business rule base to identify the violation event.

[0136] In this embodiment, a violation event refers to the system's detailed analysis of the suspected target, including its attributes (such as motor vehicle type, non-motor vehicle type, or object type) and its behavior (such as parking, loading and unloading goods, etc.), combined with the functional area rules and time factors in the dynamic compliance benchmark map generated in the current frame, to ultimately determine whether the target has violated the preset management rules.

[0137] In one embodiment, step S140 described above may include steps S141 to S142.

[0138] S141. The suspected targets are classified in detail, and the current behavior and state characteristics of the suspected targets are analyzed to obtain attributes and behaviors.

[0139] In this embodiment, the attribute refers to further refining the specific category of the suspected target. For example, "motor vehicle" can be further subdivided into taxis, private cars, freight trucks, etc.; "non-motor vehicle" can be further subdivided into bicycles, tricycles, etc.; and "object" can be a pile of goods, a billboard, etc.

[0140] Behavior refers to the specific activities or states of a suspected target identified through real-time monitoring and analysis. This may include, but is not limited to, parking, loading and unloading goods, and commercial operations. To accurately capture these behaviors, the system employs high-precision target and behavior feature recognition technology to ensure that the behavioral description of each suspected target is as accurate as possible.

[0141] In this step, the system first meticulously categorizes suspected targets, identifying their specific attributes. Next, the system analyzes the current behavior and state characteristics of the suspected targets to gain a comprehensive understanding of the nature of their activities. This is done to provide detailed information support for subsequent compliance assessments.

[0142] S142. Compare the attributes and behaviors of the suspected target with the business rule base, determine the compliance of the suspected target based on the functional area and the nature of the activity, and combine the time factor, so as to identify the violation event.

[0143] At this stage, the system compares the attribute and behavioral information of the suspected target obtained in S141 with the regulations in the business rule base. The business rule base not only contains basic prohibition rules for each functional area (such as prohibiting motor vehicles from parking on pedestrian walkways), but also takes into account changes in rules over different time periods (such as allowing temporary parking in certain areas during specific times). In addition, the business rule base also supports custom rule configuration, which can flexibly adapt to the temporary and regional control needs in urban management.

[0144] In practice, the system first identifies the functional area (such as a pedestrian walkway or parking space) where the suspected target is located. Then, it combines the target's attributes and behavior with the current time factor to determine whether the target complies with the relevant business rules. If, after comparison, it is found that the behavior of a suspected target violates the regulations of its functional area, it will be marked as a violation event, and a detailed report will be generated, including the target ID, geographical coordinates, functional area, attributes, behavior, timestamp, and confidence level. This information will then be pushed to the relevant business systems for subsequent processing measures.

[0145] Referring to Figure 5, when processing scenario B (suspect targets), the system first needs to perform high-precision ontological attribute and behavioral feature identification on these targets. This process provides the basic data for subsequent business rule verification.

[0146] Refined identification of ontological attributes:

[0147] Motor vehicles: further subdivided into different types such as taxis, private cars, freight trucks, mobile food trucks, and construction vehicles.

[0148] Non-motorized vehicles: including bicycles, tricycles, and mobile food trucks, etc.

[0149] Objects: encompassing various types such as cargo piles, construction waste, billboards, and construction facilities.

[0150] Behavioral and State Feature Analysis: The system monitors and analyzes the target's instantaneous activities in real time, such as parking, loading and unloading goods, or commercial operations. In this way, the system can not only understand the target's physical form but also grasp its specific behavioral patterns.

[0151] After completing the above detailed classification and behavioral feature analysis, the system will automatically judge the "attribute + behavior" combination of each suspected target based on the latest business rule base to determine whether it meets the regulations.

[0152] For example, if a target identified as a "mobile food truck" is engaged in "operational" activities, the system needs to determine whether this behavior is permitted based on the functional area where the target is located and the corresponding time conditions. If such behavior is prohibited in the current functional area and time period, the target will be marked as violating regulations; otherwise, it will be considered compliant.

[0153] If the identified target is a "taxi" and it is performing a "parking" action, the system needs to check whether the functional area to which the parking location belongs allows taxi parking. If the area explicitly prohibits taxi parking, it is ultimately determined to be a violation; if it allows, it is considered a compliant operation.

[0154] Through this process design, the system can not only efficiently identify various targets and their behavioral characteristics, but also ensure that all decisions strictly adhere to existing business rules and management requirements, thereby achieving intelligent and standardized urban management. This meticulous approach greatly improves the system's accuracy and response speed, contributing to the construction of a safer and more orderly urban environment.

[0155] Through the close coordination of the two steps described above, the system can automate the entire process from initial screening to precise judgment, thereby effectively improving urban management efficiency, reducing the misjudgment rate, and enhancing the system's business adaptability and flexibility. This method not only improves the accuracy and intelligence of the judgment but also optimizes the overall operational efficiency of the system, enabling urban managers to respond to various violations more quickly and accurately.

[0156] Furthermore, in another embodiment, the above method further includes:

[0157] A structured processing work order is generated for the violation event, and the structured processing work order is output; the structured processing work order includes target ID, geographic coordinates, functional area, attributes, behavior, timestamp, and confidence level.

[0158] After the comprehensive analysis of the above steps, the system will generate detailed structured processing orders. These orders contain key data points such as target ID, geographic coordinates, functional area, attribute description, behavior details, timestamp information, and confidence score. This information will then be pushed to relevant business systems to execute necessary follow-up processing measures, such as notifying relevant departments to take action or recording it in the management system for historical data reference.

[0159] Once an event is confirmed as a violation, the system automatically generates a structured action order. This action order is a detailed report created based on the system's analysis results, containing the following key data points:

[0160] Target ID: Used to uniquely identify each monitored object.

[0161] Geographic coordinates: Provides the specific geographical location information of the target.

[0162] Functional area: describes the urban management or functional zoning where the target is located, such as commercial area, residential area, etc.

[0163] Attributes: Record the physical type of the target, such as motor vehicle, non-motor vehicle, or object.

[0164] Behavior: The behavioral characteristics of the target are described in detail, such as parking and operation.

[0165] Timestamp: Marks the exact time when an event occurred.

[0166] Confidence score: indicates the degree of certainty by which the system believes the event constitutes a violation.

[0167] A structured processing work order is generated for the aforementioned violation event, and the structured processing work order is output. This process emphasizes the importance of presenting the violation event information in a standardized format to facilitate subsequent operations and management.

[0168] The generated structured processing orders are automatically pushed to the relevant business systems. This step aims to achieve closed-loop processing, forming a complete process chain from the occurrence of an event to its final resolution. Specifically, this information can be used for:

[0169] Notify relevant departments to take action: Based on the information in the work order, relevant departments can respond quickly and take appropriate measures.

[0170] Recorded in the management system as historical data reference: All violations that have been processed will be recorded and become valuable historical data for future decision-making reference.

[0171] This detailed process design not only improves the speed and efficiency of response to urban violations but also ensures that every step is based on evidence, thereby enhancing the overall level of urban management. Furthermore, this approach helps build a more transparent and efficient management system, promoting harmonious urban development.

[0172] This embodiment's method elevates the identification of road occupancy behavior from simple visual recognition to rule-based intelligent judgment by deeply integrating physical spatial information and dynamic management rules. This combination not only solves the disconnect between judgment logic and actual business rules in traditional methods but also ensures an accurate understanding and intelligent response to urban space occupancy. To improve system processing efficiency, a progressive screening mechanism is introduced. This mechanism first excludes non-static and explicitly compliant targets, thereby significantly reducing the amount of data requiring high-precision analysis. This approach not only reduces the system's computational load but also significantly improves overall processing efficiency, enabling the system to react more quickly to environmental changes. To better adapt to temporary and regional needs in urban management, a system that supports flexible configuration of custom functional areas and their exclusive business rules is provided. This allows managers to quickly adjust rule settings according to actual conditions, effectively addressing various special situations and specific regional needs, enhancing the system's flexibility and practicality.

[0173] Specifically, this embodiment digitally integrates the street view physical spatial structure with management business rules to form a dynamic compliance benchmark map. This benchmark map provides the system with the core basis for judging whether a target is compliant, realizing a shift from simple spatial recognition to rule-based intelligent decision-making. A multi-level process of "environmental understanding - static initial screening - behavioral fine-tuning" is adopted to achieve accurate screening and efficient analysis of suspected targets. This method not only improves the accuracy of analysis but also effectively reduces system resource consumption and speeds up response. Users can set exclusive business rules for specific functional areas, and a built-in priority mechanism for handling rule conflicts is included. This design greatly enhances the system's adaptability and flexibility, enabling it to dynamically adjust strategies according to actual needs.

[0174] This approach, by combining rules with spatial information, enables the system to more accurately distinguish between compliant parking and illegal parking, significantly reducing the false positive rate and improving the accuracy and intelligence of the judgment. The progressive filtering mechanism avoids high-precision analysis of all targets, reducing unnecessary waste of computing resources and improving the system's response speed and overall performance. Configurable custom rules allow the management system to quickly adjust to changes in actual needs, increasing the system's practical value and application scope.

[0175] The "Dynamic Compliance Benchmark Map" and "Progressive Judgment Paradigm" proposed in this embodiment aim to fill the gap in existing technologies for accurate and automated input of violation events. By dynamically integrating and matching environmental structure, target attributes, behavioral characteristics, and business rules, it achieves a leap from "seeing the target" to "understanding the violation," providing strong support for downstream systems and forming an indispensable link in the complete technology chain.

[0176] The aforementioned method for detecting abnormal road occupancy based on dynamic patrol, by integrating mobile scene adaptability, dynamic business rule understanding and application capabilities, solves the problems of disconnect between logic and business rules, low data processing efficiency, and rigid rule configuration in existing technologies. Specifically, it first acquires street view video streams, GPS data, and timestamps, and identifies the traffic areas and municipal facilities in each frame of the video stream. Semantic matching is then performed between geographic information and timestamps and the street view rule base to generate a dynamic compliance benchmark map for each frame and identify unidentified targets. Subsequently, these unidentified targets are detected and tracked, and stationary targets are filtered out and compared with the dynamic compliance benchmark map, marking the violating targets as suspected targets. Finally, by identifying the attributes and behavioral characteristics of suspected targets, compliance is determined based on the business rule base, thereby accurately locating the violation. This process not only achieves real-time monitoring and analysis of the urban environment but also optimizes data processing efficiency through a progressive filtering mechanism. Simultaneously, it supports custom rule configuration, improving the system's flexibility and adaptability, and ensuring effective response to complex and ever-changing urban management needs.

[0177] Figure 6 is a schematic block diagram of a street-side abnormal road occupancy detection system 300 based on dynamic patrol provided by an embodiment of the present invention. As shown in Figure 6, corresponding to the above-described street-side abnormal road occupancy detection method based on dynamic patrol, the present invention also provides a street-side abnormal road occupancy detection system 300 based on dynamic patrol. This street-side abnormal road occupancy detection system 300 based on dynamic patrol includes units for executing the above-described street-side abnormal road occupancy detection method based on dynamic patrol, and the system can be configured in a server. Specifically, referring to Figure 6, the street-side abnormal road occupancy detection system 300 based on dynamic patrol includes an acquisition unit 301, a generation unit 302, a suspected target determination unit 303, and a violation event determination unit 304.

[0178] The acquisition unit 301 is used to acquire street view video stream, GPS data, and timestamps; the generation unit 302 is used to identify the traffic area and municipal facilities in each frame of the street view video stream, combine them with the timestamp and geographic information, perform semantic matching with the street view rule base, and combine them with the street view environment structure map to generate a dynamic compliance benchmark map for each frame and identify unidentified targets; the suspected target determination unit 303 is used to detect and track the unidentified targets, filter out targets in a stationary state, compare them with the dynamic compliance benchmark map, and mark the targets that violate regulations to obtain suspected targets; the violation event determination unit 304 is used to identify the attributes and behavioral characteristics of the suspected targets, judge compliance based on the business rule base, and determine the violation event.

[0179] In one embodiment, the generation unit 302 includes:

[0180] The identification and classification subunit is used to identify and classify the passable areas and municipal facilities in each frame of the street view video stream, and combine GPS data to determine the custom functional area to obtain the general area identification result, the municipal facility identification result, and the unidentified target; the benchmark map generation subunit is used to combine the general area identification result and the municipal facility identification result with the timestamp and geographic information according to the street view rule library to dynamically generate a dynamic compliance benchmark map for each frame.

[0181] In one embodiment, the reference map generation subunit includes:

[0182] The semantic pairing module is used to combine the general area identification results and municipal facility identification results with timestamps and geographic information, and perform semantic pairing with the content in the street view rule base to obtain pairing results; wherein, the street view rule base includes spatial range, time conditions, constraint content and priority; the combination module is used to combine the pairing results with the street view environment structure map in the form of dynamic tags to obtain a dynamic compliance benchmark map for each frame.

[0183] In one embodiment, the suspect target determination unit 303 includes:

[0184] The initial screening subunit is used to track the unidentified targets and analyze their motion state to filter out targets that are stationary, thus obtaining static targets. The classification subunit is used to perform spatial location mapping and rule matching on the static targets based on the dynamic compliance benchmark map, and classify them into non-suspect targets and suspect targets.

[0185] In one embodiment, the primary screening unit includes:

[0186] The identification module is used to identify and roughly classify the unidentified targets in the street view video stream as motor vehicles, non-motor vehicles, people or objects; the recording module is used to determine whether the unidentified targets are stationary through multi-target tracking and short-term trajectory analysis, and to mark the unidentified targets that meet the conditions as static targets and record basic information.

[0187] In one embodiment, the classification subunit includes:

[0188] The calculation module is used to calculate the specific ground contact point of the image corresponding to the static target in the street view video stream, and map the specific ground contact point onto the dynamic compliance benchmark map to determine the functional area where it is located; the exclusion module is used to match according to the rules of the functional area where the static target is located, exclude non-suspected targets located in unregulated areas or belonging to people, and mark suspected targets that violate regulations.

[0189] In one embodiment, the violation event determination unit 304 includes:

[0190] The analysis subunit is used to classify the suspected targets in detail and analyze their current behavior and status characteristics to obtain attributes and behaviors. The event determination subunit is used to compare the attributes and behaviors of the suspected targets with the business rule base, determine the compliance of the suspected targets based on functional areas and activity nature, and combine time factors to identify violation events.

[0191] In one embodiment, the system described above further includes:

[0192] The output unit is used to generate a structured processing work order for the violation event and output the structured processing work order; the structured processing work order includes target ID, geographic coordinates, functional area, attributes, behavior, timestamp, and confidence level.

[0193] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned dynamic patrol-based abnormal road occupancy detection system 300 and its various units can be found in the corresponding descriptions in the aforementioned method embodiments. For the sake of convenience and brevity, these details will not be repeated here.

[0194] The aforementioned abnormal road occupancy detection system 300 based on dynamic patrol can be implemented as a computer program, which can run on the computer device shown in Figure 7.

[0195] Please refer to Figure 7, which is a schematic block diagram of a computer device provided in an embodiment of this application. The computer device 500 can be a server, wherein the server can be a standalone server or a server cluster composed of multiple servers.

[0196] Referring to Figure 7, the computer device 500 includes a processor 502, a memory, and a network interface 505 connected via a system bus 501. The memory may include a non-volatile storage medium 503 and internal memory 504.

[0197] The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions that, when executed, cause the processor 502 to perform a method for detecting abnormal road occupancy along the street based on dynamic inspection.

[0198] The processor 502 provides computing and control capabilities to support the operation of the entire computer device 500.

[0199] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute a method for detecting abnormal road occupancy along the street based on dynamic inspection.

[0200] The network interface 505 is used for network communication with other devices. Those skilled in the art will understand that the structure shown in FIG. 7 is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device 500 to which the present application is applied. A specific computer device 500 may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0201] The processor 502 is used to run the computer program 5032 stored in the memory to implement all the steps of the dynamic patrol-based abnormal road occupancy detection method.

[0202] It should be understood that in the embodiments of this application, the processor 502 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0203] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.

[0204] Therefore, the present invention also provides a storage medium. This storage medium may be a computer-readable storage medium. The storage medium stores a computer program, wherein when executed by a processor, the computer program causes the processor to perform all the steps of the dynamic patrol-based abnormal road occupancy detection method.

[0205] The storage medium can be any computer-readable storage medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), magnetic disk, or optical disk.

[0206] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0207] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of each unit is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0208] The steps in the method of this invention can be adjusted, merged, or reduced in order according to actual needs. The units in the system of this invention can be merged, divided, or reduced according to actual needs. Furthermore, the functional units in the various embodiments of this invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0209] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0210] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered 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 detecting abnormal road occupancy along streets based on dynamic patrol, characterized in that, include: Acquire street view video streams, GPS data, and timestamps; For each frame of the street view video stream, the system identifies the passable areas and municipal facilities, combines them with timestamps and geographic information, performs semantic matching with the street view rule base, and combines them with the street view environment structure map to generate a dynamic compliance benchmark map for each frame. It also identifies unidentified targets, including: identifying and classifying the passable areas and municipal facilities in each frame of the street view video stream, and determining custom functional areas using GPS data to obtain general area identification results, municipal facility identification results, and unidentified targets; combining the general area identification results and municipal facility identification results with timestamps and geographic information, and comparing them with the street view rule base... The content is semantically paired to obtain pairing results; the street view rule base is a set of rules used in urban management to regulate and guide various behaviors, including spatial scope, time conditions, constraints, and priorities; the pairing results are combined with the street view environment structure map in the form of dynamic tags to generate a dynamic compliance benchmark map for each frame; the unidentified targets are detected and tracked, static targets are filtered out and compared with the dynamic compliance benchmark map, and the non-compliant targets are marked to obtain suspected targets; the suspected targets are identified by attribute and behavioral characteristics, and compliance is judged according to the business rule base to determine the violation event.

2. The method for detecting abnormal road occupancy based on dynamic patrol according to claim 1, characterized in that, The dynamic compliance benchmark map is matched with the content of the street view rule base based on different traffic areas, municipal facilities, timestamps, and geographic information.

3. The method for detecting abnormal road occupancy based on dynamic patrol according to claim 1, characterized in that, The process of detecting and tracking the unidentified targets, filtering out stationary targets, comparing them with the dynamic compliance benchmark map, and marking non-compliant targets to obtain suspected targets includes: tracking the unidentified targets, analyzing the motion state of the unidentified targets to filter out stationary targets to obtain static targets; and performing spatial location mapping and rule matching on the static targets based on the dynamic compliance benchmark map to classify non-suspected targets and suspected targets.

4. The method for detecting abnormal road occupancy based on dynamic patrol according to claim 3, characterized in that, The process of tracking the unidentified targets and analyzing their motion states to filter out stationary targets and obtain static targets includes: identifying and roughly classifying the unidentified targets in the street view video stream as motor vehicles, non-motor vehicles, people, or objects; determining whether the unidentified targets are stationary through multi-target tracking and short-term trajectory analysis; marking the unidentified targets that meet the conditions as static targets; and recording basic information.

5. The method for detecting abnormal road occupancy based on dynamic patrol according to claim 3, characterized in that, The step of spatially mapping and matching the static targets based on the dynamic compliance benchmark map, and classifying non-suspect targets and suspect targets, includes: calculating the specific ground contact point of the image corresponding to the static target in the street view video stream, and mapping the specific ground contact point onto the dynamic compliance benchmark map to determine the functional area where it is located; matching according to the rules of the functional area where the static target is located, excluding non-suspect targets located in unregulated areas or belonging to people, and marking suspected targets that violate regulations.

6. The method for detecting abnormal road occupancy based on dynamic patrol according to claim 1, characterized in that, The process of identifying the attributes and behavioral characteristics of the suspected targets and determining compliance based on the business rule base to identify violations includes: meticulously classifying the suspected targets and analyzing their current behavior and status characteristics to obtain attributes and behaviors; comparing the attributes and behaviors of the suspected targets with the business rule base, determining the compliance of the suspected targets based on functional areas and activity nature, and combining time factors to identify violations.

7. The method for detecting abnormal road occupancy based on dynamic patrol according to claim 1, characterized in that, After identifying the suspected target's attributes and behavioral characteristics, and determining compliance based on the business rule base to identify the violation event, the process further includes: generating a structured processing work order for the violation event and outputting the structured processing work order; the structured processing work order includes target ID, geographical coordinates, functional area, attributes, behavior, timestamp, and confidence level.

8. A street-side abnormal encroachment detection system based on dynamic patrol, employing the method described in any one of claims 1-7, characterized in that, include: The acquisition unit is used to acquire street view video streams, GPS data, and timestamps. The generation unit is used to identify the passage area and municipal facilities in each frame of the street view video stream, combine them with timestamps and geographic information, perform semantic matching with the street view rule base, and combine them with the street view environment structure map to generate a dynamic compliance benchmark map for each frame and determine unidentified targets. The suspected target determination unit is used to detect and track the unidentified targets, filter out targets in a stationary state, compare them with the dynamic compliance benchmark map, mark the non-compliant targets, and obtain suspected targets. The violation event determination unit is used to identify the attributes and behavioral characteristics of the suspected target, and to determine compliance based on the business rule base in order to identify the violation event.

Images