A high-position camera inter-vehicle license plate data complementary method, device, equipment and medium

Abstract

This application provides a method, apparatus, device, and medium for complementing license plate data among high-position cameras, comprising: after the deployment of multiple high-position cameras, pre-dividing the parking space areas corresponding to each camera based on the actual installation position and overhead view of each camera, and constructing topological information between the cameras; determining the associated camera set of the current camera based on the topological information, and uploading the parking space coordinates and parking space number corresponding to each parking space in the inspection images captured periodically by the current camera and the associated camera set to the platform; determining an extended recognition area in the inspection images based on the coordinates of each parking space; and performing license plate recognition on vehicles in the extended recognition area based on the platform database and parking space location information rules to determine new parking order records. This significantly reduces the parking order false negative rate and improves the order generation accuracy and business compliance.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a method, apparatus, device and medium for complementing license plate data between high-position cameras. Background Technology

[0002] Current smart parking systems in cities commonly employ multiple high-position cameras (installed on poles or gantry tops) for overhead monitoring of parking lots. Each camera, after deployment, is pre-defined as managing an independent parking area. The platform only performs vehicle detection and license plate recognition within the image area of ​​its assigned parking space and generates parking orders accordingly. However, this deployment model has inherent structural flaws: due to the height gradient of the camera installation positions and the linear arrangement of parking spaces, license plates of vehicles in farther parking spaces are often too small, severely distorted, and easily obscured by vehicles in front. Meanwhile, license plates of vehicles in closer parking spaces are difficult to capture completely due to vehicle shape obstructions or license plate angle issues. This results in vehicles with obscured license plates not having their entry and exit times recorded, preventing parking fee calculations and causing losses for the parking lot. Summary of the Invention

[0003] In view of this, the purpose of this application is to provide a method, device, equipment and medium for complementing license plate data between high-position cameras, and to construct a full-chain technical solution of "topology association modeling → extended identification area generation → parking space allocation decision", which significantly reduces the parking order underreporting rate and improves the order generation accuracy and business compliance.

[0004] This application provides a method for complementing license plate data between high-position cameras, the complementation method comprising: After the deployment of multiple high-position cameras is completed, based on the actual installation position and topological angle of each camera, the berth area corresponding to each camera is pre-divided, and the topological information between the cameras is constructed. Based on the topology information, determine the associated camera set of the current camera, and upload the berth coordinates and berth number corresponding to each berth in the inspection images captured by the current camera and the associated camera set at regular intervals to the platform. Based on the coordinates of each berth, an extended recognition area is determined in the inspection image; Based on the platform database and parking space location information rules, license plate recognition is performed on vehicles in the extended recognition area to determine new parking order records.

[0005] In one possible implementation, the step of pre-dividing the berth area corresponding to each camera based on the actual installation position and topological angle of each camera, and constructing the topological information between the cameras, includes: For any given camera, based on its actual installation location, pitch angle, focal length, and the spatial distribution of the preset berth area, identify at least two associated cameras that have spatial overlap with its field of view and whose number of managed berths and viewing angle characteristics are complementary. Generate a topology table for each camera, labeling the associated camera with its identity and type. The topology table is bound to the berth coordinates of the berth area corresponding to the camera to generate topology information between cameras.

[0006] In one possible implementation, determining the extended recognition region in the inspection image based on each berth coordinate includes: In the inspection image, a managed berth identification area and an extended identification area are determined; wherein, the managed berth identification area is the managed berth area of ​​the management camera that acquired the inspection image; Based on the berth coordinates of any berth in the inspection image and the management berth identification area, the rear edge line corresponding to the berth with the farthest physical location outside the management berth identification area in the inspection image and the front edge line corresponding to the berth with the closest physical location are identified. Using the front edge line as a reference, the front edge line is extended upward along the side edge line of the berth where it is located until it intersects with the upper boundary of the inspection image. The strip-shaped area enclosed by the resulting line segment and the upper boundary of the inspection image is defined as the upper edge recognition area. Using the rear edge line as a reference, the rear edge line is extended downward along the side edge of the berth where it is located until it intersects with the lower boundary of the inspection image. The strip-shaped area enclosed by the resulting line segment and the lower boundary of the inspection image is defined as the lower edge recognition area. The extended recognition region is constructed based on the upper edge recognition region and the lower edge recognition region; wherein, the shape of the extended recognition region is jointly determined by the geometric properties of the reference and the image boundary constraints.

[0007] In one possible implementation, the associated camera set includes at least one adjacent camera located in the same mounting area as the current camera but with a different overhead view, and at least one far / near camera located in the front and rear mounting areas and having a field of view that overlaps with the current camera.

[0008] In one possible implementation, the step of performing license plate recognition on vehicles in the extended recognition area based on the platform database and parking space location information rules to determine newly added parking order records includes: The license plate character information, corresponding parking space number, and collection timestamp in the extended recognition area are matched with the parking records stored in the platform database in a multi-dimensional manner. If no record with the same parking space number and license plate character information is found within the preset time tolerance, then based on the parking space location information rules and the spatial attribution of the license plate character information in the inspection image, it is determined to be the corresponding target management camera, and a new parking order record is generated.

[0009] In one possible implementation, the step of determining the corresponding target management camera based on the parking space location information rules and the spatial attribution of the license plate character information in the inspection image, and generating a new parking order record, includes: The target management camera is determined based on the berth number and the topology information between the cameras; The target management camera, parking space number, license plate character information, and timestamp information are bound together to determine the new parking order record.

[0010] In one possible implementation, after the new parking order record is determined, the complementary method further includes: The new parking order record is encapsulated as a task to be reviewed, and the task to be reviewed is distributed to the human agent workstation through the message queue service; The system receives an operation instruction from a human agent for the pending task. If the operation instruction is approved, the status of the new parking order record is updated to "effective," and the billing module is started simultaneously.

[0011] This application embodiment also provides a complementary device for license plate data between high-position cameras, the complementary device comprising: The topology building module is used to pre-divide the berth areas corresponding to each camera and build the topology information between the cameras after multiple high-position cameras have been deployed, based on the actual installation position and top view of each camera. The information acquisition module is used to determine the associated camera set of the current camera based on the topology information, and upload the berth coordinates and berth number corresponding to each berth in the inspection images captured by the current camera and the associated camera set at regular intervals to the platform. The area recognition module is used to determine the extended recognition area in the inspection image based on the coordinates of each berth. The multi-dimensional detection module is used to perform license plate recognition on vehicles in the extended recognition area based on the platform database and parking space location information rules, and to determine new parking order records.

[0012] This application embodiment also provides an electronic device, including: a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. When the machine-readable instructions are executed by the processor, the steps of the complementary method for license plate data between high-position cameras as described above are performed.

[0013] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the above-described method for complementing license plate data between high-position cameras.

[0014] This application provides a method, apparatus, device, and medium for complementing license plate data among high-position cameras. The complementation method includes: after multiple high-position cameras are deployed, pre-dividing the parking space areas corresponding to each camera based on the actual installation position and overhead view of each camera, and constructing topological information between the cameras; determining the associated camera set of the current camera based on the topological information, and uploading the parking space coordinates and parking space number corresponding to each parking space in the inspection images captured periodically by the current camera and the associated camera set to the platform; determining an extended recognition area in the inspection images based on the coordinates of each parking space; and performing license plate recognition on vehicles in the extended recognition area based on the platform database and parking space location information rules to determine new parking order records. This constructs a full-chain technical solution of "topological association modeling → extended recognition area → parking space ownership decision," significantly reducing the parking order false negative rate and improving the order generation accuracy and business compliance.

[0015] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating a method for complementing license plate data between high-position cameras provided in this application embodiment; Figure 2 This is one of the structural schematic diagrams of a complementary device for license plate data between high-position cameras provided in an embodiment of this application; Figure 3 A second schematic diagram of a complementary device for license plate data between high-position cameras provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. Based on the embodiments of this application, every other embodiment obtained by those skilled in the art without inventive effort falls within the scope of protection of this application.

[0019] First, the applicable scenarios for this application will be introduced. This application can be applied to the field of image processing technology.

[0020] Based on this, this application provides a method for complementing license plate data between high-position cameras, constructing a full-chain technical solution of "topological association modeling → expanding the identification area → parking space allocation decision", which significantly reduces the parking order underreporting rate and improves the order generation accuracy and business compliance.

[0021] Please see Figure 1 , Figure 1 A flowchart illustrating a method for complementing license plate data between high-position cameras, provided as an embodiment of this application. Figure 1 As shown in the embodiments of this application, the complementary method includes: S101: After the deployment of multiple high-position cameras, based on the actual installation position and topological angle of each camera, the berth area corresponding to each camera is pre-divided, and the topological information between the cameras is constructed.

[0022] In this step, after the deployment of multiple high-position cameras is completed, the berth areas corresponding to each camera are pre-divided according to the actual installation position and topological angle of each camera, and the topological information between the cameras is constructed.

[0023] For example, in some implementation scenarios, there are physical adjacencies and overlapping fields of view between near-end cameras and far-end cameras in the same parking area. In this case, the topological information between the cameras represents the physical adjacency and overlapping fields of view between the near-end cameras and the far-end cameras. When only high-position cameras are installed, there are also physical adjacencies and overlapping fields of view between the front and rear high-position cameras. In this case, the topological information between the cameras represents the physical adjacency and overlapping fields of view between the high-position cameras.

[0024] It should be noted that if the current camera is a close-up camera... A(For example, managing only 3 berths), the associated cameras are remote camera B currently installed at the same location (for example, managing only 5 berths), and remote camera C installed on the high mast behind (for example, managing only 5 berths). These two remote cameras are associated with each other. B And C can compensate for the current close-range cameras A The issue of missed identification due to obstruction or excessively large downward angles arises because in the parking areas managed by the camera, parking spaces closer to the front and rear are prone to having their vehicles obscured by neighboring vehicles due to their large field of view, preventing the high-position camera from recognizing the vehicles' valid identity information. It is understood that the solution proposed in this application is applicable not only to scenarios with near-end and far-end cameras installed in the same installation area, but also to scenarios with only one high-position recognition camera installed in the same area.

[0025] In one possible implementation, the step of pre-dividing the berth area corresponding to each camera based on the actual installation position and topological angle of each camera, and constructing the topological information between the cameras, includes: A: For any given camera, based on its actual installation location, pitch angle, focal length, and the spatial distribution of the preset berth area, identify at least two associated cameras that have spatial overlap with its field of view and whose number of managed berths and viewing angle characteristics are complementary.

[0026] It should be noted that the complementary relationship is as follows: when the camera is a near-end camera, its associated camera is a far-end camera located at a higher position on the same mounting pole, and a far-end camera located on an adjacent pole at a higher position behind it with a larger downward angle coverage; when the camera is a far-end high-position camera, its associated camera is a near-end high-position camera located at a lower position on the same mounting pole, and a near-end high-position camera located on an adjacent pole at a higher position in front of it with a smaller downward angle and higher image resolution.

[0027] B: Generate a topology table for each camera, labeling the associated camera's identity and type; bind the topology table to the berth coordinates of the berth area corresponding to the camera to generate topology information between cameras.

[0028] Here, the topology information is stored in the platform database in the form of a directed graph. The nodes in the graph represent high-position cameras, the edges represent one-way complementary recognition support relationships, and the physical constraint parameters corresponding to each edge are labeled, including relative height difference, horizontal distance, maximum field of view overlap threshold, and viewing angle difference.

[0029] In this application, through the aforementioned structured topology table generation and berth-level binding mechanism, the platform no longer relies on manual configuration of camera association relationships. Instead, it automatically constructs a complementary network with clear spatial semantics based on physically measurable parameters (angle, height, coordinates, overlap rate). Crucially, by shifting the "associated camera" from the device dimension to the individual berth dimension, subsequent steps such as "edge area extension" and "license plate collision" can be precisely anchored to the weak points in the identification of specific berths, thereby fundamentally solving the problem of missed order reporting in cross-area areas.

[0030] S102: Determine the associated camera set of the current camera based on the topology information, and upload the berth coordinates and berth number corresponding to each berth in the inspection images captured by the current camera and the associated camera set at regular intervals to the platform.

[0031] In this step, the set of associated cameras of the current camera is determined from the topology information. Using the current camera and the set of associated cameras, the captured inspection images, along with the berth coordinates and berth number corresponding to each berth in the inspection images, are uploaded to the platform according to the timed capture logic.

[0032] It should be noted that the connection node of the current camera is determined in the topology information, and the camera corresponding to the connection node is used as the set of associated cameras of the current camera.

[0033] S103: Based on the coordinates of each berth, determine the extended recognition area in the inspection image.

[0034] In this step, the algorithm deployed on the platform divides the inspection image into the management berth area corresponding to the camera based on the berth coordinates. Outside the management berth area, the upper and lower edges of the recognition area are extended to the upper and lower edges of the image, thus generating an extended recognition area that includes the upper and lower edge areas.

[0035] It should be noted that the extended recognition area is not limited to the berths at the top and bottom edges, but can also be the recognition area corresponding to any berth within the field of vision outside the managed berth.

[0036] In one possible implementation, determining the extended recognition region in the inspection image based on each berth coordinate includes: a: Determine the management berth identification area and the extended identification area in the inspection image; wherein, the management berth identification area is the management berth area of ​​the management camera that acquired the inspection image.

[0037] b: Based on the berth coordinates of any berth in the inspection image and the management berth identification area, identify the rear edge line corresponding to the berth with the farthest physical location outside the management berth identification area in the inspection image, and the front edge line corresponding to the berth with the closest physical location.

[0038] Here, in the same image, the smaller the (y) coordinate value of a parking space, the closer its physical location (because nearby vehicles are imaged at the top of the image); the larger the (y) coordinate value of a parking space, the farther its physical location (because distant vehicles are imaged at the bottom of the image). Therefore, for all marked parking spaces in the current image, we search for the front edge line with the smallest (y) coordinate among all parking spaces, corresponding to the parking space with the closest physical location; and we search for the rear edge line with the largest (y) coordinate among all parking spaces, corresponding to the parking space with the farthest physical location.

[0039] It should be noted that "nearest / farthest" here refers to the extreme value of Euclidean distance in the depth direction of the camera's field of view, rather than the horizontal distance, which meets the actual management needs of parking scenarios.

[0040] For any berth received by the platform, the pixel coordinates of its four uploaded vertices are parsed. A least-squares linear fit is performed on this set of four points to calculate its upper and lower edges. The upper edge (L_{text{upper}}) is defined as the straight line segment connecting the two vertices with the smallest y-coordinates among the four points; the lower edge (L_{text{lower}}) is defined as the straight line segment connecting the two vertices with the largest y-coordinates among the four points. c: Using the front edge line as a reference, extend the front edge line upward along the side edge line of the berth where it is located until it intersects with the upper boundary of the inspection image. The strip-shaped area enclosed by the resulting line segment and the upper boundary of the inspection image is defined as the upper edge recognition area.

[0041] Here, the upper edge recognition region is defined as follows: Extend the front edge line (L_upper}} upwards along its own direction, and find its intersection point (Q_top}) with the upper boundary of the image (i.e., the line (y=0)). Simultaneously, find the intersection points (Q_left} and (Q_right}) of this line within the image with the left boundary (x=0) and the right boundary (x=W). The upper edge recognition region is then a trapezoidal or triangular strip-shaped area enclosed by the intersection points (Q_left, Q_right) and the upper boundary of the image (y=0).

[0042] d: Using the rear edge line as a reference, extend the rear edge line downwards along the side edge line of the berth where it is located until it intersects with the lower boundary of the inspection image. The strip-shaped area enclosed by the resulting line segment and the lower boundary of the inspection image is defined as the lower edge recognition area.

[0043] Here, the lower edge recognition region is defined as follows: Extend the rear edge line downwards along its own direction and find its intersection point (Q_{text{bottom}}) with the lower boundary of the image. Similarly, find its intersection points (Q'_{text{left}}) and (Q'_{text{right}}) with the left and right boundaries. Then, the lower edge recognition region is a trapezoidal or triangular strip-shaped region enclosed by the intersection points (Q'_{text{left}}Q'_{text{right}}}) and the lower boundary of the image.

[0044] e: The extended recognition region is constructed based on the upper edge recognition region and the lower edge recognition region; wherein, the shape of the extended recognition region is determined by the geometric properties of the reference and the image boundary constraints.

[0045] Here, the upper edge recognition region and the lower edge recognition region are combined in the image plane, and the resulting composite region is the extended recognition region defined in this invention. The shape of this region is determined by the following two types of constraints: the geometric properties of the baseline: including the slope reflecting the berth orientation, the intercept (reflecting the vertical position of the berth in the image), and the relative distance between the two lines (reflecting the depth difference between near and far berths); and image boundary constraints: i.e., hard cropping of the image boundary to ensure that it is always within the effective image range and to avoid invalid calculations.

[0046] In this application, the design does not simply expand the detection box, but rather follows the imaging geometry of a high-position camera—vehicles in near-end parking spaces appear larger in the image and are mostly located in the upper half, making their license plates easily captured by the upper edge region; vehicles in far-end parking spaces appear smaller due to perspective compression and are mostly located in the lower half of the image, requiring the lower edge region to cover their potentially complete license plate area. Furthermore, this dynamic expansion strategy improves the license plate recall rate compared to fixed-ratio expansion.

[0047] It should be noted that near-end cameras, limited by their low installation height and small downward angle, struggle to clearly capture license plates of distant parking spaces; far-end cameras, on the other hand, suffer from low recognition rates for near-distance parking spaces due to their excessive downward angle, severe image distortion, and susceptibility to occlusion. This application does not rely on increasing the resolution of a single camera or adding supplementary lighting equipment. Instead, it constructs a near-end-far-end camera topology relationship, enabling each camera to both utilize its own field-of-view redundancy to fill the recognition blind spots in the "vertical distance dimension" (e.g., near-end cameras use their upper edge area to fill distant parking spaces) and utilize the redundancy of associated cameras to fill the occlusion blind spots in the "lateral spatial dimension" (e.g., far-end cameras use their lower edge area to fill near parking spaces obscured by the vehicle in front), achieving a leap in system-level perception capabilities with zero hardware increments.

[0048] S104: Based on the platform database and parking space location information rules, perform license plate recognition on vehicles in the extended recognition area to determine new parking order records.

[0049] In this step, the platform database and parking space location information rules are used to perform license plate recognition on vehicles in the extended recognition area. Only when all preset business rules are met will a new parking order record with legal effect be generated.

[0050] In one possible implementation, the step of performing license plate recognition on vehicles in the extended recognition area based on the platform database and parking space location information rules to determine newly added parking order records includes: (1): The license plate character information, corresponding parking space number and collection timestamp in the extended recognition area are matched with the parking records stored in the platform database in a multidimensional way.

[0051] Here, the platform database performs a status comparison of license plate character information, corresponding parking space number, and collection timestamp. Only when all dimensions meet the preset consistency conditions is the recognition result determined to be "a valid record already exists"; otherwise, it is considered a legitimate data source for generating a new order. This method fundamentally avoids systemic risks such as cross-parking space misjudgment, time drift omissions, and equipment ownership confusion caused by single-dimensional matching (such as only checking license plates).

[0052] (2): If no record with the same parking space number and the same license plate character information is matched within the preset time tolerance, the target management camera corresponding to the license plate character information is determined based on the parking space location information rules and the spatial ownership of the license plate character information in the inspection image, and a new parking order record is generated.

[0053] Here, if no record with the same parking space number and license plate character information is found within the preset time tolerance, the corresponding target management camera is determined based on the parking space location information rules and the spatial assignment of the license plate character information in the inspection image, and a new parking order record is generated. If a match is found, the license plate character information is not processed.

[0054] In one possible implementation, the step of determining the corresponding target management camera based on the parking space location information rules and the spatial attribution of the license plate character information in the inspection image, and generating a new parking order record, includes: The target management camera is determined based on the parking space number and the topology information between cameras; the target management camera, parking space number, license plate character information, and timestamp information are bound together to determine a new parking order record. Here, the target management camera is determined based on the parking space number and the topology information between cameras, and the target management camera, parking space number, license plate character information, and timestamp information are bound together to determine a new parking order record.

[0055] In one possible implementation, after the new parking order record is determined, the complementary method further includes: The new parking order record is encapsulated as a task to be reviewed, and the task to be reviewed is distributed to the human agent's workbench through the message queue service; the operation instructions of the human agent for the task to be reviewed are received, and if the operation instructions are approved, the status of the new parking order record is updated to effective, and the billing module is started simultaneously.

[0056] Here, new parking order records are encapsulated as pending review tasks and distributed to human agent workstations via a message queue service. Upon receiving a pending review task, the human agent can view multimodal auxiliary information (including the original image cropping frame, enlarged license plate image, parking space geolocation diagram, and historical records of vehicles with the same license plate entering and exiting), and execute the following two types of operation instructions: Approve: The agent clicks the "Confirm Validity" button, and the workstation frontend sends a `POST / v1 / review / {task_id} / approve` request to the platform API gateway, carrying a JWT authentication token and the operator ID; Reject: The agent selects the rejection reason (such as "license plate blurred and unrecognizable," "parking space ownership questionable," "suspected duplicate entry"), and can add voice / text notes. After receiving the `Approve` instruction, the platform executes an atomic state update transaction: ① Using `task_id` as the key, and under the guarantee of a distributed transaction coordinator (such as SeataAT mode), synchronously update the `status` field of the corresponding parking order record in the database from `"pending_review"` to `"active"`; ② Publish the domain event `OrderActivatedEvent` to the event bus, whose payload includes `order_id`, `activation_time`, and `approver_id`; ③ The billing module, as a subscriber to this event, listens to `OrderActivatedEvent` in real time, parses the parking space number and start time, calls the rate engine (supporting multi-dimensional billing strategies for time periods / vehicle types / membership levels) to generate an initial bill, and triggers the payment gateway to implement the deduction process.

[0057] In a specific embodiment, (1) after the equipment is installed and debugged, the respective camera management berth areas are divided out; (2) the front-end equipment can capture the inspection map according to the timed capture logic, and simultaneously report the berth coordinates and berth number information; (3) the platform establishes the topology information of the equipment and obtains the front and rear cameras of each camera; the camera relationship is as follows: if the current camera is the near-end camera A (for example, only managing 3 berths), the associated camera is the far-end camera B installed at the same location (for example, only managing 5 berths), and the far-end camera C installed on the high pole behind (for example, only managing 5 berths). The two associated far-end cameras can make up for the problem of missed identification caused by the current near-end camera due to occlusion or too large downward angle. Similarly, if the current camera is remote camera 1 (e.g., managing only 5 berths), the associated cameras are near-end camera 1 set up at the same location (e.g., managing only 3 berths) and near-end camera 3 installed on the high pole in front (e.g., managing only 3 berths). These two associated near-end cameras can make up for the missed identification problem caused by the current remote camera due to occlusion or too large downward angle. (4) The images captured by the camera during regular inspections are uploaded to the platform along with the images, parking space coordinates, and parking space number information. The algorithm deployed on the platform will form a large recognition area by using the parking space coordinate information to identify the farthest and nearest parking spaces in the image. The upper and lower edges of the recognition area will be extended to the upper and lower edges of the image, thus generating an upper edge area and a lower edge area. (5) At this time, vehicle detection and license plate recognition will be performed on the edge area. If a license plate is identified in the edge area, the license plate information will be compared with the license plate in the platform database. If the license plate already exists in the platform database, it will be discarded. Otherwise, a new parking record will be generated according to the parking space location information rules, including the license plate number, parking start time, and parking space number. This information can also be pushed to manual confirmation. Through the above operations, combined with platform data collision, the problem of missed recognition by near and far cameras can be reduced, the problem of missed parking order reporting can be reduced, the overall order accuracy can be improved, and the management of operators can be facilitated.

[0058] This application provides a method for complementing license plate data among high-position cameras. The method includes: after deploying multiple high-position cameras, pre-dividing parking space areas corresponding to each camera based on the actual installation position and overhead view of each camera, and constructing topological information between the cameras; determining the associated camera set of the current camera based on the topological information, and uploading the parking space coordinates and parking space number corresponding to each parking space in the periodically captured inspection images of the current camera and the associated camera set to the platform; determining an extended recognition area in the inspection images based on the coordinates of each parking space; and performing license plate recognition on vehicles in the extended recognition area based on the platform database and parking space location information rules to determine new parking order records. This complete chain of technical solutions—"topology association modeling → extended recognition area → parking space ownership decision"—significantly reduces the parking order false alarm rate and improves order generation accuracy and business compliance.

[0059] Please see Figure 2 , Figure 3 , Figure 2 This is one of the structural schematic diagrams of a complementary device for license plate data between high-position cameras provided in an embodiment of this application; Figure 3 This is a second schematic diagram of a device for complementing license plate data between high-position cameras, provided as an embodiment of this application. Figure 2 As shown, the complementary device 200 for license plate data between high-position cameras includes: The topology building module 210 is used to pre-divide the berth area corresponding to each camera based on the actual installation position and top view of each camera after the deployment of multiple high-position cameras; Information acquisition module 220 is used to determine the associated camera set of the current camera based on the topology information, and upload the berth coordinates and berth number corresponding to each berth in the inspection images captured by the current camera and the associated camera set at regular intervals to the platform. The area recognition module 230 is used to determine the extended recognition area in the inspection image based on the coordinates of each berth. The multidimensional detection module 240 is used to perform license plate recognition on vehicles in the extended recognition area based on the platform database and parking space location information rules, and to determine new parking order records.

[0060] Furthermore, the topology construction module 210 is used to pre-divide the berth areas corresponding to each camera based on the actual installation position and topological angle of each camera, and construct the topology information between the cameras: For any given camera, based on its actual installation location, pitch angle, focal length, and the spatial distribution of the preset berth area, identify at least two associated cameras that have spatial overlap with its field of view and whose number of managed berths and viewing angle characteristics are complementary. Generate a topology table for each camera, labeling the associated camera with its identity and type. The topology table is bound to the berth coordinates of the berth area corresponding to the camera to generate topology information between cameras.

[0061] Furthermore, the area recognition module 230 is used to determine an extended recognition area in the inspection image based on the coordinates of each berth: In the inspection image, a managed berth identification area and an extended identification area are determined; wherein, the managed berth identification area is the managed berth area of ​​the management camera that acquired the inspection image; Based on the berth coordinates of any berth in the inspection image and the management berth identification area, the rear edge line corresponding to the berth with the farthest physical location outside the management berth identification area in the inspection image and the front edge line corresponding to the berth with the closest physical location are identified. Using the front edge line as a reference, the front edge line is extended upward along the side edge line of the berth where it is located until it intersects with the upper boundary of the inspection image. The strip-shaped area enclosed by the resulting line segment and the upper boundary of the inspection image is defined as the upper edge recognition area. Using the rear edge line as a reference, the rear edge line is extended downward along the side edge line of the berth where it is located until it intersects with the lower boundary of the inspection image. The strip-shaped area enclosed by the resulting line segment and the lower boundary of the inspection image is defined as the lower edge recognition area. The extended recognition region is constructed based on the upper edge recognition region and the lower edge recognition region; wherein, the shape of the extended recognition region is jointly determined by the geometric properties of the reference and the image boundary constraints.

[0062] Furthermore, the multi-dimensional detection module 240 is used to perform license plate recognition on vehicles in the extended recognition area based on the platform database and parking space location information rules, thereby identifying newly added parking order records. The license plate character information, corresponding parking space number, and collection timestamp in the extended recognition area are matched with the parking records stored in the platform database in a multi-dimensional manner. If no record with the same parking space number and license plate character information is found within the preset time tolerance, the corresponding target management camera is determined based on the parking space location information rules and the spatial attribution of the license plate character information in the inspection image, and a new parking order record is generated.

[0063] Furthermore, the multi-dimensional detection module 240 is used to determine the corresponding target management camera based on the parking space location information rules and the spatial attribution of the license plate character information in the inspection image, and generate a new parking order record: The target management camera is determined based on the berth number and the topology information between the cameras; By binding the target management camera, parking space number, license plate character information, and timestamp information, a new parking order record is determined. Further, such as... Figure 3 As shown, the complementary device 300 for license plate data between high-position cameras also includes a confirmation module 250, which is used for: The new parking order record is encapsulated as a task to be reviewed, and the task to be reviewed is distributed to the human agent workstation through the message queue service; The system receives an operation instruction from a human agent for the pending task. If the operation instruction is approved, the status of the new parking order record is updated to "effective," and the billing module is started simultaneously.

[0064] This application provides a complementary device for license plate data between high-position cameras. The complementary device includes: a topology construction module, used to pre-divide the parking space areas corresponding to each camera based on the actual installation position and overhead view of each camera after the deployment of multiple high-position cameras, and construct the topology information between the cameras; an information acquisition module, used to determine the associated camera set of the current camera according to the topology information, and upload the parking space coordinates and parking space number corresponding to each parking space in the inspection images captured by the current camera and the associated camera set at regular intervals to the platform; an area recognition module, used to determine the extended recognition area in the inspection image based on the coordinates of each parking space; and a multi-dimensional detection module, used to perform license plate recognition on vehicles in the extended recognition area based on the platform database and parking space location information rules, and determine new parking order records. This constructs a full-chain technical solution of "topology association modeling → extended recognition area → parking space ownership decision", significantly reducing the parking order false negative rate and improving the order generation accuracy and business compliance.

[0065] Please see Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 4 As shown, the electronic device 400 includes a processor 410, a memory 420, and a bus 430.

[0066] The memory 420 stores machine-readable instructions executable by the processor 410. When the electronic device 400 is running, the processor 410 communicates with the memory 420 via the bus 430. When the machine-readable instructions are executed by the processor 410, they can perform the operations described above. Figure 1 The steps of the method for complementing license plate data between high-position cameras in the illustrated method embodiment can be found in the method embodiment for specific implementation methods, which will not be repeated here.

[0067] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can perform the above-described actions. Figure 1 The steps of the method for complementing license plate data between high-position cameras in the illustrated method embodiment can be found in the method embodiment for specific implementation methods, which will not be repeated here.

[0068] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0069] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the shown or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0070] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0071] In addition, the functional units in the various embodiments of this application 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.

[0072] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion 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, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0073] Finally, it should be noted that the above-described embodiments are merely specific implementations of this application, used to illustrate the technical solutions of this application, and not to limit them. The scope of protection of this application is not limited thereto. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features, within the scope of the technology disclosed in this application. Such modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for complementing license plate data between high-position cameras, characterized in that, When applied to the platform side, the complementary method includes: After the deployment of multiple high-position cameras is completed, based on the actual installation position and topological angle of each camera, the berth area corresponding to each camera is pre-divided, and the topological information between the cameras is constructed. Based on the topology information, determine the associated camera set of the current camera, and upload the berth coordinates and berth number corresponding to each berth in the inspection images captured by the current camera and the associated camera set at regular intervals to the platform. Based on the coordinates of each berth, an extended recognition area is determined in the inspection image; Based on the platform database and parking space location information rules, license plate recognition is performed on vehicles in the extended recognition area to determine new parking order records.

2. The complementary method according to claim 1, characterized in that, Based on the actual installation position and topographic view of each camera, the berth area corresponding to each camera is pre-divided, and the topological information between the cameras is constructed, including: For any given camera, based on its actual installation location, pitch angle, focal length, and the spatial distribution of the preset berth area, identify at least two associated cameras that have spatial overlap with its field of view and whose number of managed berths and viewing angle characteristics are complementary. Generate a topology table for each camera, labeling the associated camera with its identity and type. The topology table is bound to the berth coordinates of the berth area corresponding to the camera to generate topology information between cameras.

3. The complementary method according to claim 1, characterized in that, The step of determining the extended recognition region in the inspection image based on the coordinates of each berth includes: In the inspection image, a managed berth identification area and an extended identification area are determined; wherein, the managed berth identification area is the managed berth area of ​​the management camera that acquired the inspection image; Based on the berth coordinates of any berth in the inspection image and the management berth identification area, the rear edge line corresponding to the berth with the farthest physical location outside the management berth identification area in the inspection image and the front edge line corresponding to the berth with the closest physical location are identified. Using the front edge line as a reference, the front edge line is extended upward along the side edge line of the berth where it is located until it intersects with the upper boundary of the inspection image. The strip-shaped area enclosed by the resulting line segment and the upper boundary of the inspection image is defined as the upper edge recognition area. Using the rear edge line as a reference, the rear edge line is extended downward along the side edge line of the berth where it is located until it intersects with the lower boundary of the inspection image. The strip-shaped area enclosed by the resulting line segment and the lower boundary of the inspection image is defined as the lower edge recognition area. The extended recognition region is constructed based on the upper edge recognition region and the lower edge recognition region; wherein, the shape of the extended recognition region is jointly determined by the geometric properties of the reference and the image boundary constraints.

4. The complementary method according to claim 1, characterized in that, The associated camera set includes at least one adjacent camera that is in the same installation area as the current camera but has a different overhead view, and at least one far / near camera that is located in the front and rear installation areas and has an overlapping field of view with the current camera.

5. The complementary method according to claim 1, characterized in that, The process of performing license plate recognition on vehicles in the extended recognition area based on the platform database and parking space location information rules to determine newly added parking order records includes: The license plate character information, corresponding parking space number, and collection timestamp in the extended recognition area are matched with the parking records stored in the platform database in a multi-dimensional manner. If no record with the same parking space number and license plate character information is found within the preset time tolerance, the corresponding target management camera is determined based on the parking space location information rules and the spatial attribution of the license plate character information in the inspection image, and a new parking order record is generated.

6. The complementary method according to claim 5, characterized in that, The process of determining the corresponding target management camera based on the parking space location information rules and the spatial attribution of the license plate character information in the inspection image, and generating a new parking order record, includes: The target management camera is determined based on the berth number and the topology information between the cameras; The target management camera, parking space number, license plate character information, and timestamp information are bound together to determine the new parking order record.

7. The complementary method according to claim 1, characterized in that, After the new parking order record is determined, the complementary method further includes: The new parking order record is encapsulated as a task to be reviewed, and the task to be reviewed is distributed to the human agent workstation through the message queue service; The system receives an operation instruction from a human agent for the pending task. If the operation instruction is approved, the status of the new parking order record is updated to "effective," and the billing module is started simultaneously.

8. A device for complementing license plate data between high-position cameras, characterized in that, The complementary device includes: The topology building module is used to pre-divide the berth areas corresponding to each camera and build the topology information between the cameras after multiple high-position cameras have been deployed, based on the actual installation position and top view of each camera. The information acquisition module is used to determine the associated camera set of the current camera based on the topology information, and upload the berth coordinates and berth number corresponding to each berth in the inspection images captured by the current camera and the associated camera set at regular intervals to the platform. The area recognition module is used to determine the extended recognition area in the inspection image based on the coordinates of each berth. The multi-dimensional detection module is used to perform license plate recognition on vehicles in the extended recognition area based on the platform database and parking space location information rules, and to determine new parking order records.

9. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus. The machine-readable instructions are executed by the processor to perform the steps of the complementary method for license plate data between high-position cameras as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the method for complementing license plate data between high-position cameras as described in any one of claims 1 to 7.

Images