A management method and system for urban road encroachment control

Abstract

This invention relates to the field of data governance technology, specifically to a management method and system for urban street occupation management, comprising: obtaining gridded street monitoring images based on images acquired by roadside cameras; analyzing the street monitoring images to obtain the location and range of street monitoring blind spots; obtaining entrance and exit area information based on the range of the street monitoring blind spots, and establishing a monitoring blind spot associated image dataset based on the images obtained from monitoring the entrance and exit area information; obtaining the characteristics of street vendors occupying the road based on the associated image dataset and identifying them; recording the time interval between street vendors entering and exiting the monitoring blind spots based on the identification tags to determine the individual street vendors; comparing and excluding individuals with abnormal lingering and the shop information of the monitoring blind spots to obtain mobile street occupation data. This allows for the analysis of the business situation in the obstructed parts of the road using mobile data, thereby improving management efficiency.

Description

Technical Field

[0001] This invention relates to the field of data governance technology, and in particular to a management method and system for managing urban road encroachment. Background Technology

[0002] Urban road occupation management refers to measures to regulate and manage the unauthorized occupation of roads, sidewalks, and other public spaces in cities. With the acceleration of urbanization, the increase in the number of vehicles, and the prosperity of commercial activities, road occupation has become increasingly common, not only affecting traffic flow but also posing a threat to pedestrian safety.

[0003] Current methods for managing street encroachment typically involve using cameras to capture street images and employing image recognition technology to identify instances of encroachment and notify relevant personnel for intervention. However, trees are usually situated along streets, which can obstruct the view of cameras, making it impossible to observe the obstructed portions of the encroachment and reducing management efficiency. Summary of the Invention

[0004] The purpose of this invention is to provide a management method and system for urban road occupancy management, which aims to improve management efficiency by combining traffic data analysis of the operation of road obstruction.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a management method for urban road occupation control, comprising obtaining gridded street monitoring images based on images acquired by roadside cameras;

[0006] Analyze street monitoring images to determine the location and extent of blind spots in street monitoring;

[0007] Information on import and export areas is obtained based on the blind spots of street monitoring, and an image dataset associated with the blind spots is established based on the images obtained from monitoring the import and export areas.

[0008] Features of objects illegally occupying roads are obtained from associated image datasets and their identities are labeled.

[0009] Based on the time interval between the entry and exit of objects illegally occupying the roadside in the monitoring blind zone, the individual illegally occupying the roadside can be identified;

[0010] By comparing and excluding individuals with abnormal lingering behavior and information about shops in monitoring blind spots, data on mobile street occupancy is obtained.

[0011] The method further includes, after comparing and excluding abnormally staying individuals and the store information where the monitoring blind spots are located to obtain mobile road occupation data, extracting image features from the mobile road occupation data and matching and comparing the image features throughout the monitoring network to predict the mobile road occupation status of individuals.

[0012] The specific steps for obtaining gridded street monitoring images based on images acquired from roadside cameras include:

[0013] Set the image acquisition frequency to capture image information on the street;

[0014] Use filtering algorithms to remove noise from image information;

[0015] Cropping the image to focus on the street area;

[0016] The street area is divided into several grid units of the same size;

[0017] By using image stitching technology, grid cells generated from monitoring images of all areas are combined into a complete image to obtain street monitoring image information.

[0018] The specific steps for analyzing street monitoring images to obtain the location and extent of street monitoring blind spots include:

[0019] Acquire street monitoring image information;

[0020] The method of occlusion recognition is used to identify stationary obstacles in road monitoring images and obtain obstacle information.

[0021] By comparing the locations of road areas and obstacles marked on street monitoring images, the location and extent of street monitoring blind spots can be determined.

[0022] The specific steps for identifying stationary obstacles in road monitoring images using the occlusion recognition method to obtain obstacle information include:

[0023] Collect features of common obstacles in the target area;

[0024] The target obstacle is obtained by feature matching based on the characteristics of common obstacles in road monitoring image information;

[0025] Obstacle information is extracted based on the target obstacle, and the obstacle information includes the obstacle type and location.

[0026] The specific steps for obtaining entry and exit area information based on the blind spot range of street monitoring and establishing a monitoring blind spot associated image dataset based on the images obtained from monitoring the entry and exit area information include:

[0027] Obtain map data showing the location of the blind spot;

[0028] Mark key road points on the map data;

[0029] Obtain the corresponding camera IDs based on key road points;

[0030] Based on the camera ID, images from related cameras are combined to form a dataset of associated images for monitoring blind spots.

[0031] The specific steps for obtaining features of street vendors based on associated image datasets and identifying them include:

[0032] The Canny algorithm was used to detect the outlines of individual objects in an associated image dataset.

[0033] The object detection algorithm is applied to identify all object types in the object outline and define their positions.

[0034] Identify and label objects.

[0035] The specific steps for determining the individual street vendor based on the time interval between their entry and exit from the monitoring blind zone recorded by their identity tags include:

[0036] Motion detection algorithms are used to capture events of each identified object entering or leaving the monitoring blind zone;

[0037] Record an entry timestamp for each detected entry event and a departure timestamp for each detected departure event;

[0038] For each entry / exit event, calculate the time interval during which the object remains within the monitoring blind zone;

[0039] The Z-score algorithm is used to identify operating objects with time intervals exceeding the normal range, thus identifying individuals with abnormal dwell times.

[0040] The specific steps for comparing and excluding individuals exhibiting abnormal lingering behavior and store information located in monitoring blind spots to obtain data on mobile street occupancy include:

[0041] Obtain store data within the monitored area;

[0042] Based on store data, determine whether an individual with abnormal lingering behavior belongs to a store; if not, it is considered mobile obstruction of the road.

[0043] Secondly, the present invention also provides a management system for urban street occupation management, including a monitoring image acquisition module, a monitoring blind spot detection module, an entrance and exit area acquisition module, an identity marking module, a street occupation judgment module, and an exclusion module;

[0044] The monitoring image acquisition module is used to obtain gridded street monitoring images based on images acquired by roadside cameras;

[0045] The blind spot detection module is used to analyze street monitoring images to obtain the location and range of street monitoring blind spots;

[0046] The import / export area acquisition module is used to acquire import / export area information based on the blind spot range of street monitoring, and to establish a monitoring blind spot associated image dataset based on the images obtained from the monitoring of import / export area information.

[0047] The identity labeling module is used to obtain the features of objects illegally occupying the road based on the associated image dataset and to label their identities.

[0048] The street vendor determination module is used to determine the individual street vendor based on the time interval between the entry and exit of the street vendor object into the monitoring blind zone recorded by the identity tag.

[0049] The exclusion module is used to compare and exclude individuals with abnormal lingering status and store information located in monitoring blind spots to obtain data on mobile street occupation.

[0050] This invention discloses a management method and system for urban street vendor management. It utilizes roadside cameras to capture real-time street conditions, covering a wide area. The acquired images are then gridded to create a series of street monitoring images. The system performs in-depth analysis on these images to identify blind spots—areas difficult to observe directly due to limited viewing angles or obstructions. To better understand and manage these blind spots, the system further collects information about entrances and exits. By combining image data from these locations, an image dataset related to the blind spots is created. Using this correlated image dataset, the system can identify the characteristics of street vendors, such as the shape, color, and size of the stalls, and label them. This labeling process is automated, allowing the system to quickly and accurately identify specific street vendors, even if they have moved to different locations. The time interval between these labeled street vendors entering and exiting the blind spots is then recorded to determine which are individual vendors. This helps distinguish between long-term occupation of public resources and temporary stays, thus enabling more precise targeting of intervention needs. For individuals exhibiting unusual activity, the system compares and excludes them against information from surrounding shops. This process confirms the existence of legitimate business activities or determines whether such behavior constitutes illegal street vending. By combining mobile data analysis with the operational status of areas obstructed by pedestrian traffic, management efficiency can be improved. The system can filter out genuine cases of mobile street vending, supporting urban management decisions while protecting the interests of legitimate businesses. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the 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.

[0052] Figure 1 This is a flowchart of a management method for urban road occupancy control according to the present invention.

[0053] Figure 2 This is a flowchart of the present invention for obtaining gridded street monitoring images based on images acquired by roadside cameras.

[0054] Figure 3 This is a flowchart of the present invention for analyzing street monitoring images to obtain the location and range of street monitoring blind spots.

[0055] Figure 4 This invention is a flowchart illustrating how a method for identifying stationary obstacles in road monitoring images is used to obtain obstacle information.

[0056] Figure 5 This is a flowchart of the present invention for obtaining import and export area information based on the blind spot range of street monitoring, and establishing a monitoring blind spot associated image dataset based on the images obtained from the monitoring of import and export area information.

[0057] Figure 6 This is a flowchart of the present invention for obtaining features of street vendors based on associated image datasets and performing identity labeling.

[0058] Figure 7 This is a flowchart of the present invention that uses the time interval between the entry and exit of objects occupying the road for business to enter and exit the monitoring blind zone based on identity tags to determine the individual occupying the road for business.

[0059] Figure 8 This invention is a flowchart for comparing and excluding information on individuals with abnormal lingering and shops located in monitoring blind spots to obtain data on mobile street occupation. Detailed Implementation

[0060] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0061] First Embodiment

[0062] Please see Figures 1 to 8This invention provides a management method for urban road encroachment control, comprising:

[0063] S101 generates gridded street monitoring images based on images acquired by roadside cameras;

[0064] The specific steps include:

[0065] S201 sets the image acquisition frequency to capture image information on the street;

[0066] Setting the image acquisition frequency is crucial for capturing dynamic changes on the street. This step determines how quickly the system can respond to real-time events. The choice of acquisition frequency must balance data volume and real-time requirements; acquisitions that are too frequent may overload data processing, while acquisitions that are too infrequent may result in missing critical events.

[0067] S202 uses a filtering algorithm to remove noise from image information;

[0068] Next, filtering algorithms are used to remove noise from the image. Due to environmental factors such as weather conditions, lighting changes, and limitations of the camera itself, images often contain various forms of noise. This noise can interfere with subsequent image analysis. Employing advanced filtering techniques, such as adaptive filters or deep learning models, can effectively reduce unnecessary interference and ensure image quality.

[0069] S203 crops the image to focus on the street area;

[0070] Next, the image is cropped to focus on the street area. To improve processing efficiency and focus on the region of interest, namely the actual street portion, the original image needs to be cropped. This process may involve computer vision techniques such as edge detection to accurately locate street boundaries and exclude irrelevant background elements (such as the sky, buildings, etc.), making the processing more efficient and targeted.

[0071] S204 divides the street area into several grid units of the same size;

[0072] The street area is divided into several uniformly sized grid cells. This is done to enable more granular data management and analysis. Each grid cell represents a small segment of the street, and by monitoring and analyzing each grid cell separately, more detailed and accurate information about the street conditions can be obtained. This division method also facilitates subsequent spatial data analysis and pattern recognition.

[0073] S205 uses image stitching technology to combine the grid cells generated from the monitoring images in all areas into a complete image, thus obtaining street monitoring image information.

[0074] Image stitching technology is used to combine grid cells generated from monitoring images across all areas into a single, complete image. This step involves image processing algorithms, such as feature point matching and perspective transformation, to ensure seamless connectivity between different grid cells. The final output is a high-resolution image comprehensively reflecting the entire street condition. This image is composed of multiple individual grid cell images, yet it appears as a coherent whole. The street monitoring image information obtained in this way not only helps traffic management departments with effective urban planning and management but also provides the public with a more intuitive and detailed display of street conditions.

[0075] S102 analyzes street monitoring images to obtain the location and extent of street monitoring blind spots;

[0076] The specific steps include:

[0077] S301 acquires street monitoring image information;

[0078] This step relies on previously captured and processed gridded street images from cameras. These images provide the visual data foundation for the streets and their surroundings, which is a prerequisite for further analysis.

[0079] S302 uses an obstruction recognition method to identify stationary obstacles in road monitoring images and obtain obstacle information.

[0080] To identify objects that may affect the effectiveness of surveillance, such as parked vehicles, temporary structures, trees, or other factors that obstruct the view.

[0081] The specific steps include:

[0082] S401 collects features of common obstacles in the target area;

[0083] First, a series of data samples of typical obstacles are systematically collected within the target area. This step aims to build a comprehensive and representative database covering all object types that may affect road monitoring effectiveness. Data acquisition can be accomplished in various ways, such as capturing high-resolution images or video sequences using high-definition cameras, obtaining 3D point cloud data using LiDAR, or combining this with drone photography. For each sample, its appearance characteristics, such as shape, size, color, and texture, should be recorded, and its location and category should be labeled.

[0084] S402 performs feature matching on road monitoring image information based on the characteristics of common obstacles to obtain the target obstacle;

[0085] Machine learning or deep learning algorithms are applied to real-time road monitoring images and compared with a pre-established obstacle feature database. This process utilizes advanced pattern recognition algorithms such as Convolutional Neural Networks (CNNs) and Support Vector Machines (SVMs) to accurately locate potential obstacles in the images. By comparing the features of newly acquired images with known features in the database, objects that exist in the road environment but are not considered normal traffic elements can be effectively detected.

[0086] S403 extracts obstacle information based on the target obstacle, the obstacle information including obstacle type and location.

[0087] Once the presence of suspected obstacles is confirmed, the next step is to extract detailed information about them. This involves not only identifying their category (e.g., vehicle, building, or naturally growing vegetation) but also precisely locating their spatial coordinates. To do this, computer vision techniques such as geometric transformations and perspective projection are typically used to map obstacles from two-dimensional images back into the three-dimensional physical world, thereby obtaining a more accurate description of their location.

[0088] S303 compares the locations of road areas and obstacles marked on street monitoring images to determine the location and extent of blind spots in street monitoring.

[0089] In this step, the known road layout is matched with the detected obstacle information to determine which areas are difficult to monitor effectively due to the presence of obstacles. This comparison can be done using a Geographic Information System (GIS) or other spatial analysis tools, with the ultimate goal of creating a map showing all potential blind spots. This not only helps evaluate the effectiveness of the current surveillance system but also guides decisions on future camera deployment or obstacle removal, thereby improving the safety and management efficiency of the entire street.

[0090] S103 obtains information on entrance and exit areas based on the blind zone range of street monitoring, and establishes a monitoring blind zone associated image dataset based on the images obtained from monitoring the entrance and exit area information;

[0091] The specific steps include:

[0092] S501 obtains map data showing the location of the blind spot;

[0093] This step relies on the blind spot locations and extents determined by previous analysis (as described in S102). Map data includes not only basic information about the road network but may also cover detailed geographic features such as buildings and greenbelts. To accurately reflect reality, the latest high-resolution map data should be used.

[0094] S502 marks key road points on map data;

[0095] These key points typically refer to strategically important locations on a road, such as intersections, turns, bridge entrances, or tunnel exits. For blind spots, it is particularly important to mark the entrances and exits—the points where vehicles or pedestrians enter and leave the blind spot. These key points will serve as reference coordinates in subsequent steps to correlate camera positions and fields of view.

[0096] S503 obtains the corresponding camera number based on key road points;

[0097] Each camera installed near a key point has a unique identification number used to identify and manage a large number of surveillance devices. Through the integration of GIS technology and a camera database, the nearest camera can be automatically matched based on the location of the key point. If there is no suitable camera coverage near a key point, it is necessary to consider redeploying or adjusting the angles of existing cameras to ensure effective coverage.

[0098] S504 uses the camera ID to create a dataset of associated images for monitoring blind spots.

[0099] This step involves synchronizing images from different sources according to their timestamps to ensure temporal consistency between the images. Simultaneously, image preprocessing is required, such as correcting lens distortion and enhancing contrast, to improve image quality. The resulting image dataset should comprehensively reflect the monitoring blind spots and their surrounding environment, providing a solid data foundation for subsequent analysis.

[0100] S104 obtains the features of objects illegally occupying the road based on the associated image dataset and performs identity labeling;

[0101] The specific steps include:

[0102] S601 uses the Canny algorithm to detect the outlines of various objects in an associated image dataset;

[0103] To extract object boundary information from complex backgrounds, the Canny edge detection algorithm is used to process each frame of the associated image dataset. The Canny algorithm is a multi-level edge detection method that effectively suppresses noise while preserving realistic edge information. The specific steps include: applying Gaussian blur to the original image to reduce noise; calculating the intensity gradient magnitude and direction of each pixel in the image to determine the edge location; retaining only gradients with local maxima as edges and removing other non-edge pixels; using two thresholds (high and low) to track strong and weak edges; and forming a continuous contour by connecting strong edges and parts of the weak edges.

[0104] After these steps, a clear object outline can be obtained, which provides a foundation for subsequent target recognition.

[0105] S602 uses an object detection algorithm to identify all object types in an object outline and define their positions;

[0106] Next, advanced object detection algorithms (such as YOLO, SSD, or Faster R-CNN) are applied to identify all object types within the object's outline and accurately pinpoint their locations. These object detection algorithms can not only distinguish between different object categories (e.g., vehicles, pedestrians, stalls, etc.) but also precisely mark the location of each object in the image.

[0107] In addition, additional features such as color and texture can be combined to enhance the recognition effect, ensuring that even objects with similar appearances can be correctly classified.

[0108] S603 identifies objects.

[0109] The final step is to identify the detected objects. This step is crucial for long-term monitoring and management because it helps distinguish between legitimate business activities and illegal street vending. Identification can be achieved by comparing detected objects with existing business registration information to confirm whether they belong to legitimate operators. It can also be done by analyzing the frequency and duration of object occurrences in long-term image sequences of the same location to determine if illegal street vending exists. For cases that are difficult to determine automatically, a manual review mechanism is introduced, with staff making the final judgment based on the actual situation. A unique identification tag is generated for each detected object, including but not limited to object ID, type, location coordinates, first appearance time, and status (legal / illegal). These tags can be stored in a database for subsequent querying and management.

[0110] S105 identifies individuals operating illegally on the street by recording the time interval between their entry and exit from the monitoring blind zone based on identification tags.

[0111] The specific steps include:

[0112] The S701 uses a motion detection algorithm to capture each event of an identified object entering or leaving a blind spot in the monitoring area;

[0113] Advanced motion detection algorithms (such as background subtraction, optical flow, or deep learning models) are used to capture each instance of an identified object entering or leaving the detection blind zone. These motion detection algorithms can distinguish between static backgrounds and moving objects, ensuring accurate capture of every entry and exit moment. The key to this step is:

[0114] Establish a stable background model to distinguish moving objects in the background and foreground. Set an appropriate motion detection threshold based on the specific application scenario to avoid false alarms. Detect object movement by comparing differences between consecutive frames. If the monitoring blind zone is covered by multiple cameras, ensure time synchronization between different cameras to guarantee consistent event capture.

[0115] S702 records an entry timestamp for each detected entry event and a departure timestamp for each detected departure event;

[0116] Once an object enters or exits the monitoring blind zone, a corresponding timestamp is immediately recorded for each event. This involves ensuring that all cameras and processing units use the same time source to achieve millisecond-level timestamp recording. Detailed log entries are created for each entry and exit event, including the object ID, type, location coordinates, and precise timestamp. Considering the potentially large amount of timestamp data generated, efficient data structures and compression techniques are required to optimize storage.

[0117] For each entry / exit event, S703 calculates the time interval during which the object remains in the monitoring blind zone;

[0118] For each entry / exit event, calculate the time interval during which the object remains within the monitoring blind zone. This step can be accomplished with simple arithmetic: subtract the entry timestamp from the exit timestamp. To improve accuracy, the following factors should also be considered: correcting for minor errors caused by network latency or device response time. For incomplete events caused by device malfunction or external interference (such as entry without exit), a reasonable upper limit can be set as the default exit time.

[0119] S704 uses the Z-score algorithm to identify operating objects with time intervals exceeding the normal range, thus identifying individuals with abnormal dwell times.

[0120] The Z-score algorithm from statistics is used to identify business entities with time intervals that deviate from the normal range. The Z-score is a method that measures how much a data point deviates from the mean, and it helps identify outliers.

[0121] The specific steps are as follows:

[0122] Data collection and preprocessing: Collect data on the time intervals of all object dwell times over a period of time, and remove obvious outliers or erroneous records.

[0123] Calculate the mean and standard deviation: Calculate the mean (μ) and standard deviation (σ) of all time interval data.

[0124] Calculate the Z-score: For each time interval t, calculate its Z-score: Z = σ(t - μ)

[0125] Based on urban management and policy needs, a reasonable Z-score threshold is set, such as |Z|>3, to determine whether an activity constitutes abnormal lingering. All objects whose Z-score exceeds the set threshold for any given time interval are identified as individuals exhibiting abnormal lingering, and a report is generated for further review.

[0126] S106 compares and excludes individuals with abnormal lingering behavior and information about shops in monitoring blind spots to obtain data on mobile street occupation.

[0127] The specific steps include:

[0128] S801 acquires store data within the monitored area;

[0129] This requires acquiring data on all legitimate stores within the monitored area. This step involves several aspects of the work:

[0130] Database integration: Obtain the latest store registration information from official channels such as the Administration for Industry and Commerce and the Urban Planning Department, and integrate it into a unified database.

[0131] Geographic Information System (GIS) Integration: Ensure that store data includes detailed geographic location information, such as latitude and longitude coordinates, street addresses, etc., to facilitate subsequent spatial analysis.

[0132] Regular updates: Considering the opening, closing, or changes of stores, a regular update mechanism will be established to ensure the timeliness and accuracy of the data.

[0133] Multi-source data fusion: In addition to official data, you can also combine publicly available resources such as social media and business review websites to supplement relevant information about the store.

[0134] S802 determines whether individuals with abnormal lingering behavior belong to a store based on store data; if not, it is considered mobile obstruction of the road.

[0135] Based on the collected store data, it is determined whether individuals exhibiting unusual activity belong to a legitimate store. This involves the following specific operations: using GIS technology to spatially match the location of individuals exhibiting unusual activity with the location of stores to determine if they are within the legitimate operating area of ​​a particular store; and using image processing technology to identify features such as signs and logos on objects and comparing them with image data in the store registration information.

[0136] For situations difficult to determine automatically, a manual review mechanism is introduced, with staff making the final judgment based on the actual circumstances. Once it is confirmed that an individual lingering abnormally does not belong to any legal shop, they are marked as a mobile street vendor. This data reflects unauthorized street vending activities, a key focus of urban management.

[0137] S107 extracts image features from mobile road occupation data and matches and compares these features across the entire monitoring network to predict individual mobile road occupation situations.

[0138] By extracting image features from mobile street vendor data and matching and comparing them across the entire monitoring network, the behavior of individual street vendors can be predicted. This process aims to leverage the similarity of image features to track and predict the behavioral patterns of individuals operating street vending businesses.

[0139] Specifically, image features that effectively characterize individuals operating illegally on the street can be selected, such as shape, color, and texture. A pre-trained deep learning model (such as a convolutional neural network, CNN) is used to automatically extract high-level feature representations from the image. The extracted features are then converted into fixed-length feature vectors for easier subsequent comparison and matching.

[0140] The system searches the city's surveillance network for images that match known characteristics of mobile street occupancy data, expanding the tracking scope. Within specific areas (such as the same block or adjacent streets), it focuses on finding images with similar characteristics to improve matching efficiency. By combining temporal and spatial dimensions, it analyzes flow patterns between different locations and predicts new destinations individuals might visit. As new data is continuously added, the matching algorithm is continuously updated and optimized to ensure the system's adaptability and accuracy.

[0141] Based on historical movement patterns, the project predicts the possible future routes and locations of individual street vendors. It identifies areas with high frequency of mobile street vending and proactively deploys management measures. Long-term data analysis helps understand the changing trends of mobile street vending, providing a basis for policy formulation.

[0142] Second Embodiment

[0143] This invention also provides a management system for urban street occupation management, including a monitoring image acquisition module, a monitoring blind spot detection module, an entrance / exit area acquisition module, an identification marking module, a street occupation judgment module, and an exclusion module. The monitoring image acquisition module is used to obtain gridded street monitoring images based on images acquired by roadside cameras. The monitoring blind spot detection module is used to analyze the street monitoring images to obtain the location and range of street monitoring blind spots. The entrance / exit area acquisition module is used to obtain entrance / exit area information based on the range of the street monitoring blind spots, and establish a monitoring blind spot associated image dataset based on the images obtained from monitoring the entrance / exit area information. The identification marking module is used to obtain the characteristics of street occupation objects based on the associated image dataset and perform identification marking. The street occupation judgment module is used to record the time interval between the entry and exit of street occupation objects into the monitoring blind spot based on the identification markings, and judge individual street occupation operators. The exclusion module is used to compare and exclude individuals with abnormal lingering and the shop information where the monitoring blind spot is located, to obtain mobile street occupation data.

[0144] In this embodiment, the monitoring image acquisition module collects images from cameras installed on the street and converts these images into a gridded street monitoring image. An appropriate image acquisition frequency is set to capture dynamic changes on the street; a filtering algorithm is used to remove image noise and improve image quality.

[0145] The blind spot detection module, based on acquired street surveillance images, identifies areas with insufficient monitoring and their impact range caused by obstacles or other factors through a series of image processing steps. It constructs a stable background model using long-term image sequences, compares differences between consecutive frames to detect stationary obstacles, and applies edge detection algorithms to find non-background contours.

[0146] The import / export area acquisition module obtains import / export area information based on the monitoring blind zone range and establishes a monitoring blind zone associated image dataset. Specifically, it acquires map data containing the blind zone locations and marks key points such as intersections and turns on the map. Based on the key point locations, it obtains the corresponding camera numbers to ensure effective coverage of the blind zone and its surroundings.

[0147] The identification and labeling module automatically detects and classifies street vendors obstructing traffic based on associated image datasets, and labels them accordingly. It extracts object contours using the Canny algorithm or other edge detection methods. Deep learning models (such as YOLO and SSD) are applied to identify object types and define their locations. A unique label is generated for each detected object, including object ID, type, location coordinates, first appearance time, and status.

[0148] The street vendor enforcement module records the time intervals between objects entering and exiting the monitoring blind zone, and identifies individuals exhibiting abnormal lingering through time management and statistical analysis. Motion detection algorithms are used to capture object entry and exit events and record precise timestamps. The time intervals during which objects remain within the monitoring blind zone are calculated. The Z-score algorithm is used to identify time intervals exceeding the normal range, thus identifying individuals exhibiting abnormal lingering.

[0149] The exclusion module compares individuals with abnormal lingering with information from legitimate shops to distinguish between legal operations and illegal street vending, ultimately obtaining data on mobile street vending.

[0150] In conclusion, this intelligent management system, through the collaborative work of its various modules, achieves refined management of street vending activities in the city. It not only helps maintain urban cleanliness and safety but also provides strong technical support to relevant law enforcement departments, improving the efficiency and intelligence level of urban management.

[0151] The above description discloses only one preferred embodiment of the present invention, and should not be construed as limiting the scope of the present invention. Those skilled in the art will understand that all or part of the processes of the above embodiments can be implemented, and equivalent changes made in accordance with the claims of the present invention are still within the scope of the invention.

Claims

1. A management method for addressing urban road encroachment. Its features are, This includes: obtaining gridded street monitoring images based on images acquired from roadside cameras; Analyze street monitoring images to determine the location and extent of blind spots in street monitoring; Information on import and export areas is obtained based on the blind spots of street monitoring, and an image dataset associated with the blind spots is established based on the images obtained from monitoring the import and export areas. Features of objects illegally occupying roads are obtained from associated image datasets and their identities are labeled. Based on the time interval between the entry and exit of objects illegally occupying the roadside in the monitoring blind zone, the individual illegally occupying the roadside can be identified; By comparing and excluding individuals with abnormal lingering behavior and information about shops in monitoring blind spots, data on mobile street occupancy is obtained.

2. The management method for urban road encroachment control as described in claim 1, characterized in that, After comparing and excluding abnormally staying individuals and store information in monitoring blind spots to obtain mobile road occupation data, the method further includes extracting image features from the mobile road occupation data and matching and comparing the image features throughout the monitoring network to predict the mobile road occupation status of individuals.

3. The management method for urban road encroachment control as described in claim 2, characterized in that, The specific steps for obtaining gridded street monitoring images based on images acquired from roadside cameras include: Set the image acquisition frequency to capture image information on the street; Use filtering algorithms to remove noise from image information; Cropping the image to focus on the street area; The street area is divided into several grid units of the same size; By using image stitching technology, grid cells generated from monitoring images of all areas are combined into a complete image to obtain street monitoring image information.

4. The management method for urban road encroachment control as described in claim 3, characterized in that, The specific steps for analyzing street monitoring images to obtain the location and extent of street monitoring blind spots include: Acquire street monitoring image information; The method of occlusion recognition is used to identify stationary obstacles in road monitoring images and obtain obstacle information. By comparing the locations of road areas and obstacles marked on street monitoring images, the location and extent of street monitoring blind spots can be determined.

5. A management method for urban road encroachment control as described in claim 4, characterized in that, The specific steps for identifying stationary obstacles in road monitoring images using the occlusion recognition method to obtain obstacle information include: Collect features of common obstacles in the target area; The target obstacle is obtained by feature matching based on the characteristics of common obstacles in road monitoring image information; Obstacle information is extracted based on the target obstacle, and the obstacle information includes the obstacle type and location.

6. A management method for urban road encroachment control as described in claim 5, characterized in that, The specific steps for obtaining entry and exit area information based on the blind spot range of street monitoring, and establishing a monitoring blind spot associated image dataset based on the images obtained from monitoring the entry and exit area information, include: Obtain map data showing the location of the blind spot; Mark key road points on the map data; Obtain the corresponding camera IDs based on key road points; Based on the camera ID, images from related cameras are combined to form a dataset of associated images for monitoring blind spots.

7. A management method for urban road encroachment control as described in claim 6, characterized in that, The specific steps for obtaining features of street vendors based on associated image datasets and identifying them include: The Canny algorithm was used to detect the outlines of individual objects in an associated image dataset. The object detection algorithm is applied to identify all object types in the object outline and define their positions. Identify and label objects.

8. A management method for urban road encroachment control as described in claim 7, characterized in that, The specific steps for determining the individual street vendor based on the time interval between the entry and exit of the monitoring blind zone recorded by the identity tag include: Motion detection algorithms are used to capture events of each identified object entering or leaving the monitoring blind zone; Record an entry timestamp for each detected entry event and a departure timestamp for each detected departure event; For each entry / exit event, calculate the time interval during which the object remains within the monitoring blind zone; The Z-score algorithm is used to identify operating objects with time intervals exceeding the normal range, thus identifying individuals with abnormal dwell times.

9. A management method for urban road encroachment control as described in claim 8, characterized in that, The specific steps for comparing and excluding individuals exhibiting abnormal lingering behavior and store information located in monitoring blind spots to obtain data on mobile street occupancy include: Obtain store data within the monitored area; Based on store data, determine whether an individual with abnormal lingering behavior belongs to a store; if not, it is considered mobile obstruction of the road.

10. A management system for urban road encroachment management, applied to the management method for urban road encroachment management as described in any one of claims 1 to 9, characterized in that, It includes a monitoring image acquisition module, a monitoring blind spot detection module, an entrance and exit area acquisition module, an identification marking module, a street vendor judgment module, and an exclusion module; The monitoring image acquisition module is used to obtain gridded street monitoring images based on images acquired by roadside cameras; The blind spot detection module is used to analyze street monitoring images to obtain the location and range of street monitoring blind spots; The import / export area acquisition module is used to acquire import / export area information based on the blind spot range of street monitoring, and to establish a monitoring blind spot associated image dataset based on the images obtained from the monitoring of the import / export area information. The identity labeling module is used to obtain the features of objects illegally occupying the road based on the associated image dataset and to label their identities. The street vendor determination module is used to determine the individual street vendor based on the time interval between the entry and exit of the street vendor object into the monitoring blind zone recorded by the identity tag. The exclusion module is used to compare and exclude individuals with abnormal lingering status and store information located in monitoring blind spots to obtain data on mobile street occupation.

Images