Dynamic path planning-based berth inspection method and system

Abstract

The invention discloses a parking space inspection method and system based on dynamic path planning, and relates to the technical field of intelligent inspection. The method comprises the following steps: acquiring a panoramic layout and real-time environment data of a parking area, and dividing into a plurality of planned plots; calculating the occupation index of each land based on the static and dynamic occupation evaluation models; in combination with a preset path priority rule and the dynamically corrected threshold value, screening a pre-passing land parcel and determining a target land parcel, and generating a real-time moving path; and periodically updating the indexes and dynamically adjusting the path in the inspection process. The system comprises an inspection device, a background management system, a data storage module and a communication module, wherein the inspection device is integrated with multiple sensors and a man-machine cooperation auxiliary module. According to the method, the path dynamically adapts to the complex environment, the inspection efficiency and safety are improved, the inspection comprehensiveness is guaranteed, and the method is suitable for various berth inspection scenes.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically, to a berth inspection method based on dynamic path planning. Background Technology

[0002] With the acceleration of urbanization and the continuous growth of motor vehicle ownership, intelligent parking management systems are increasingly widely used in urban traffic management. Parking space inspection, as a core component ensuring the stable operation of the system, directly impacts the quality of urban parking resource utilization through its efficiency and safety. Currently, parking space inspection mainly covers key tasks such as vehicle information photography and data entry, equipment battery replacement, and fault diagnosis. However, traditional manual inspection methods have many unavoidable limitations: Firstly, manual inspection is easily affected by severe weather. Under extreme conditions such as heavy rain, blizzards, and high temperatures, inspection efficiency drops significantly, and may even lead to interruptions, compromising the continuity of inspections. Secondly, inspection areas are often densely populated with public vehicles. When carrying spare parts for equipment maintenance, staff face personal safety risks from passing vehicles, and the inconvenience of carrying spare parts also reduces the ease of maintenance work. Furthermore, manual inspection is labor-intensive, with a high proportion of repetitive tasks, making it prone to missed or incorrect inspections, and also hindering real-time synchronization and efficient management of inspection data.

[0003] To address the pain points of manual inspections, autonomous driving and robotic inspection technologies are increasingly being introduced into parking space and parking lot management. While parking lot inspection robots have, to some extent, compensated for the shortcomings of traditional manual patrols and fixed monitoring systems, enabling the monitoring of vehicle parking and infrastructure status, existing technologies still have significant limitations. Traditional parking lot inspection robot path planning methods are mostly based on fixed routes or simple rules, failing to adequately adapt to the complex and ever-changing environment of parking lots and parking spaces—an area containing numerous stationary objects (such as parked vehicles, pillars, and obstacles) and moving objects (such as vehicles and pedestrians), which can easily obstruct robot movement. Existing path planning technologies lack precise assessment mechanisms for static and dynamic occupancy, and cannot dynamically adjust paths according to real-time environmental changes. This leads to frequent obstacle encounters during robot inspections, resulting in low inspection efficiency and difficulty in achieving both comprehensiveness and specificity in the inspection process.

[0004] Therefore, in view of the current safety and efficiency issues of manual inspection, as well as the shortcomings of existing inspection robots in terms of path planning flexibility, functional adaptability, and intelligence, there is an urgent need for an AI-powered unmanned vehicle application solution that can adapt to berth inspection scenarios, has dynamic path planning capabilities, and integrates multiple practical functions, so as to improve the intelligence level, operational efficiency, and safety assurance capabilities of berth inspection. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a berth inspection method based on dynamic path planning, comprising the following steps: acquiring panoramic layout information and real-time environmental data of the berth area, wherein the panoramic layout information includes berth distribution, passage boundaries, and fixed obstacle positions, and the real-time environmental data includes information on stationary and moving objects along the inspection path, wherein the stationary object information includes the size and position of parked vehicles and temporary obstacles, and the moving object information includes the real-time position and movement trajectory of moving vehicles and pedestrians; dividing the berth area into several planned plots based on the panoramic layout information; calculating the static occupancy index of each planned plot based on a preset static occupancy evaluation model, combined with the maximum width of the passage, the width of stationary object occupancy, and the width of the inspection device itself; and collecting inspection data in real time based on a preset dynamic occupancy evaluation model. The device calculates the dynamic occupancy index of each planned plot by using the current distance between the device and moving objects within each planned plot, the distance change data over historical time periods, and the movement speed of moving objects. It then sets path priority rules based on the berth inspection task requirements, including the priority of uninspected berths and the inspection weight of the intelligent berth manager. Threshold filtering is applied to the static and dynamic occupancy indicators of each planned plot, retaining plots that meet both criteria for passage. Priority scores are calculated for these plots according to the path priority rules, and the plot with the highest priority is selected as the target passage plot. A real-time movement path is generated for the inspection device from its current location to the target passage plot. As the inspection device travels along the real-time movement path, the above steps are repeated at preset intervals, updating the static and dynamic occupancy indicators in real time and dynamically adjusting the movement path.

[0006] Furthermore, the specific process for obtaining panoramic layout information and real-time environmental data of the berth area is as follows: Panoramic layout information of the berth area is obtained through a combination of multi-sensor fusion acquisition and background map calibration. The multi-sensors include LiDAR and cameras. LiDAR acquires the spatial coordinates of area boundaries, passageway directions, and fixed obstacles. Cameras acquire images of berth markers and passageway boundary lines. Coordinate calibration is performed with the pre-stored map in the background to determine the specific locations of berth distribution, passageway boundaries, and fixed obstacles. Environmental data along the inspection path is collected in real-time by multi-sensor devices mounted on the inspection device, including LiDAR, cameras, and infrared sensors. LiDAR detects the outline size and spatial coordinates of stationary objects to obtain their size and position information. Cameras and infrared sensors collaboratively track the motion trajectory of moving objects, calculating real-time positions and fitting movement trajectories through image frame analysis, and simultaneously acquiring the moving speed of the moving objects.

[0007] Furthermore, the specific process of dividing the berth area into several planning plots based on the panoramic layout information is as follows: Using the passageway direction as a reference, obtain the panoramic layout coordinate data of the berth area; calculate the number of berths per unit area to obtain the berth distribution density coefficient; preset multiple density threshold intervals, each corresponding to a preset number of adjacent berth combinations; match the calculated berth distribution density coefficient with the preset density threshold intervals to determine the corresponding number of adjacent berth combinations; for continuously distributed berth areas, according to the matched number of adjacent berth combinations, divide the corresponding number of adjacent berths and the passageway range of the area into individual planning plots. The process involves: dividing the berth into plots; obtaining the straight-line distance between a single berth and its nearest neighbor, setting a distance threshold, and classifying berths with a straight-line distance greater than the threshold as independent and scattered berths; setting maintenance operation space parameters and passage reserved space parameters, and calculating passage area range parameters based on these two parameters; dividing independent and scattered berths into individual planning plots based on the berth and the passage area corresponding to the calculated passage area range parameters; performing gridding on the panoramic layout coordinate data of the berth area, calculating the coordinates of the diagonal vertices of each planning plot using a coordinate division algorithm to form closed area boundary data, and storing the boundary data in the plot information database.

[0008] Furthermore, the specific description of the preset static occupancy assessment model is as follows: Static occupancy index = (total occupancy width of stationary objects + width of the inspection device itself) ÷ maximum width of the inspection path. Wherein, the total occupancy width of stationary objects is the sum of the projected widths of all stationary objects on the inspection path within the corresponding planned plot in the direction of the inspection path, and the maximum width of the inspection path is the maximum straight-line distance between the two boundaries of the inspection path within the corresponding planned plot. During the calculation process, the total occupancy width of stationary objects is extracted from the contour data of stationary objects collected by sensors, and combined with the preset width of the inspection device itself and the maximum width of the inspection path, the static occupancy index of each planned plot is obtained.

[0009] Furthermore, the specific description of the preset dynamic occupancy assessment model is as follows: Dynamic occupancy index = (historical distance ÷ current distance) × [1 ÷ (1 + exponential function ((current distance - historical distance) ÷ preset sampling interval))] where the historical distance is the distance between the inspection device and the active object in the corresponding planned plot at a historical time, the current distance is the distance between the inspection device and the active object at the current time, the time interval between the historical time and the current time is the preset sampling interval, and the preset sampling interval is adapted to the moving speed of the active object and the response requirements of the inspection device; during the calculation process, the current distance is collected in real time by the distance sensor of the inspection device, the historical distance at the preset sampling interval time is retrieved from the local cache, and the dynamic occupancy index of each planned plot is calculated.

[0010] Furthermore, the specific process for setting path priority rules and threshold filtering based on berth inspection task requirements is as follows: Path priority rule setting: Path priority score = Uninspected berth weight × Adaptive weight coefficient + Smart berth manager weight × Adaptive weight coefficient; Preset uninspected berth weight values ​​and inspected berth weight values, with the preset weight value corresponding to uninspected berths being higher than that corresponding to inspected berths; Preset core area smart berth manager weight values ​​and ordinary area smart berth manager weight values, with the preset weight value corresponding to the core area being greater than that corresponding to the ordinary area; Threshold filtering process: Set a first initial threshold and a second initial threshold, preset a busy period correction factor and an unbusy period correction factor, and set the value of the busy period correction factor... The value of the correction factor for non-busy periods is greater than or equal to 1; the current time period is determined to be either a busy or non-busy period, and the correction factor for the corresponding period is retrieved; the first initial threshold and the second initial threshold are corrected respectively using the correction formula: preset threshold = initial threshold × corresponding period correction factor, to generate the first preset threshold and the second preset threshold; wherein, the first preset threshold generated during busy periods is less than the first initial threshold, and the second preset threshold is less than the second initial threshold, while the first preset threshold generated during non-busy periods is greater than or equal to the first initial threshold, and the second preset threshold is greater than or equal to the second initial threshold; planned plots with static occupancy indicators less than the first preset threshold and dynamic occupancy indicators less than the second preset threshold are selected as pre-access plots.

[0011] Furthermore, the specific process of calculating priority scores for pre-passage plots and generating real-time movement paths is as follows: Calculate the priority score of each pre-passage plot according to the path priority rule, and determine the plot with the highest score as the target passage plot; if there are multiple plots with the same highest priority score, select the plot closest to the current plot as the target passage plot; use a path search algorithm, taking the center point of the current plot as the starting point and the center point of the target passage plot as the ending point, and combine the plot boundary coordinates, channel width, and occupancy index data to plan a collision-free optimal movement path. The path data includes the coordinates of the passing points, the driving direction, and the driving speed parameters, and is sent to the execution unit of the inspection device.

[0012] Furthermore, the specific process of the inspection device continuously updating indicators and dynamically adjusting the movement path is as follows: When the inspection device travels along the real-time movement path, it repeatedly executes the steps of panoramic layout and environmental data acquisition, occupancy indicator calculation, priority screening and path generation according to a preset cycle, re-collects environmental data and updates static and dynamic occupancy indicators; if the updated indicators show that the plot corresponding to the current travel path still meets the passage requirements, the original path is maintained and the device continues to travel; if the indicators exceed the preset threshold or the priority of the target passage plot is surpassed by other plots, the priority score is recalculated and a new target passage plot is determined, and a new real-time movement path is generated.

[0013] A berth inspection system based on dynamic path planning, implementing the above method, includes: an inspection device integrating a multi-sensor fusion module, a positioning module, a data processing module, a path planning module, and an execution module, used to collect environmental and layout data, calculate berth occupancy indicators, receive path instructions, and execute movement operations; a back-end management system used to store panoramic layout data of the berth area, receive real-time data uploaded by the inspection device, generate inspection tasks, calculate peak correction factors and preset thresholds, and send them to the inspection device; a data storage module used to store the boundary coordinates of the planned plots, historical inspection data, berth occupancy indicator calculation results, and path planning data, providing data support for threshold correction and path optimization; and a communication module used to realize real-time data transmission between the inspection device and the back-end management system, ensuring the stability of instruction issuance and data upload.

[0014] Furthermore, the inspection device also integrates a human-machine collaborative assistance module, including a storage unit, a flashing warning unit, and a remote interaction unit. The storage unit is used to store spare batteries and special tools required for the maintenance of the intelligent berth manager, and is stored in partitions according to spare parts type. The flashing warning unit is used to activate the warning function when the inspector performs maintenance work, forming a warning range suitable for the safety of working in public areas. The remote interaction unit is used to establish a communication connection between the inspector and the back-end technical platform, supporting fault image uploading, real-time voice interaction, and maintenance manual push functions.

[0015] The positive and progressive effects of this invention are as follows: This invention achieves real-time dynamic adjustment of berth inspection paths by dividing planned areas, constructing a dual-occupancy assessment model, and implementing a dynamic threshold correction mechanism. This effectively adapts to the complex environment where stationary and moving objects coexist in berth areas, solving the problems of rigid traditional inspection path planning and frequent obstacle encounters. Simultaneously, by combining preset weight rules and priority filtering, it ensures the targeted and comprehensive nature of inspections, overcomes the limitations of inclement weather on inspection operations, and reduces reliance on manual labor. Furthermore, with the human-machine collaborative assistance module, it simultaneously addresses pain points such as inconvenient spare parts carrying and high on-site operational safety risks, significantly improving the efficiency, safety, and stability of berth inspections and providing an efficient and reliable technical solution for intelligent berth inspection. Attached Figure Description

[0016] Figure 1 This is a flowchart of the steps of the present invention. Detailed Implementation

[0017] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0018] Reference Figure 1A berth inspection method based on dynamic path planning includes the following steps: acquiring panoramic layout information and real-time environmental data of the berth area, wherein the panoramic layout information includes berth distribution, passage boundaries, and fixed obstacle positions, and the real-time environmental data includes information on stationary objects and moving objects on the inspection path, wherein the stationary object information includes the size and position of parked vehicles and temporary obstacles, and the moving object information includes the real-time position and movement trajectory of moving vehicles and pedestrians; dividing the berth area into several planning plots according to the panoramic layout information; calculating the static occupancy index of each planning plot based on a preset static occupancy evaluation model, combined with the maximum width of the passage, the width of stationary object occupancy, and the width of the inspection device itself within each planning plot; and collecting real-time data on the relationship between the inspection device and each planning plot based on a preset dynamic occupancy evaluation model. The system calculates the dynamic occupancy index of each planned plot based on the current distance of moving objects, historical distance changes over a given time period, and the moving speed of the moving objects. It then sets path priority rules based on the berth inspection task requirements, including the priority of uninspected berths and the inspection weight of the intelligent berth manager. Threshold-based screening is applied to the static and dynamic occupancy indices of each planned plot, retaining plots that meet both criteria for passage. Priority scores are calculated for each plot according to the path priority rules, and the plot with the highest priority is selected as the target plot. A real-time movement path is generated for the inspection device from its current location to the target plot. As the inspection device travels along the real-time movement path, the above steps are repeated at a preset cycle, updating the static and dynamic occupancy indices in real time and dynamically adjusting the movement path.

[0019] Furthermore, the specific process for obtaining panoramic layout information and real-time environmental data of the berth area is as follows: Panoramic layout information of the berth area is obtained through a combination of multi-sensor fusion acquisition and background map calibration. In one example, it should be further explained that the multi-sensors include LiDAR and cameras. First, the LiDAR performs a 3D scan of the berth area, acquiring the three-dimensional spatial coordinates of the area boundaries, passageway directions, and fixed obstacles to generate an initial point cloud model. Then, the camera captures image data of berth markers, passageway boundaries, and fixed obstacles along a preset acquisition route, extracting the coordinates of feature points in the images. The initial point cloud model is fused and aligned with the image feature point coordinates, and then registered with the pre-stored base map of the berth area in the background coordinate system to eliminate system inconsistencies during the scanning and capturing process. By eliminating system errors and environmental interference, the precise berth distribution locations, channel boundary ranges, and specific coordinates of fixed obstacles are ultimately determined. Furthermore, environmental data along the inspection path is collected in real time by multi-sensor devices mounted on the inspection unit, including lidar, cameras, and infrared sensors. The lidar scans the inspection path in real time, detecting the outline dimensions and spatial coordinates of stationary objects, and calculates the size and position information of the stationary objects through coordinate analysis. The camera and infrared sensor work together; the camera continuously captures images of moving objects, and the infrared sensor captures the thermal imaging trajectory of the moving objects. The real-time position of the moving objects is calculated using the image frame difference method, and the complete movement trajectory is obtained by fitting multiple consecutive frames of data. Simultaneously, the movement speed of the moving objects is obtained based on the trajectory change rate.

[0020] Furthermore, the specific process of dividing the berth area into several planning plots based on the panoramic layout information is as follows: Using the passageway direction as a reference, obtain the panoramic layout coordinate data of the berth area; calculate the number of berths per unit area to obtain the berth distribution density coefficient; preset multiple density threshold intervals, each corresponding to a preset number of adjacent berth combinations; match the calculated berth distribution density coefficient with the preset density threshold intervals to determine the corresponding number of adjacent berth combinations; for continuously distributed berth areas, according to the matched number of adjacent berth combinations, divide the corresponding number of adjacent berths and the passageway range of the area into individual planning plots. The process involves: dividing the berth into plots; obtaining the straight-line distance between a single berth and its nearest neighbor, setting a distance threshold, and classifying berths with a straight-line distance greater than the threshold as independent and scattered berths; setting maintenance operation space parameters and passage reserved space parameters, and calculating passage area range parameters based on these two parameters; dividing independent and scattered berths into individual planning plots based on the berth and the passage area corresponding to the calculated passage area range parameters; performing gridding on the panoramic layout coordinate data of the berth area, calculating the coordinates of the diagonal vertices of each planning plot using a coordinate division algorithm to form closed area boundary data, and storing the boundary data in the plot information database.

[0021] Furthermore, the specific description of the preset static occupancy assessment model is as follows: Static occupancy index = (total occupancy width of stationary objects + width of the inspection device itself) ÷ maximum width of the inspection path, where the total occupancy width of stationary objects is the sum of the projected widths of all stationary objects on the inspection path within the corresponding planned plot in the direction of the inspection path, and the maximum width of the inspection path is the maximum straight-line distance between the two boundaries of the inspection path within the corresponding planned plot; during the calculation process, the total occupancy width of stationary objects is extracted by the contour data of stationary objects collected by sensors, and combined with the preset width of the inspection device itself and the maximum width of the inspection path, the static occupancy index of each planned plot is obtained.

[0022] Furthermore, the specific description of the preset dynamic occupancy assessment model is as follows: Dynamic occupancy index = (historical distance ÷ current distance) × [1 ÷ (1 + exponential function ((current distance - historical distance) ÷ preset sampling interval))], where the historical distance is the distance between the inspection device and the active object in the corresponding planned plot at a historical time, the current distance is the distance between the inspection device and the active object at the current time, the time interval between the historical time and the current time is the preset sampling interval, and the preset sampling interval is adapted to the moving speed of the active object and the response requirements of the inspection device; during the calculation process, the current distance is collected in real time by the distance sensor of the inspection device, the historical distance at the preset sampling interval time is retrieved from the local cache, and the dynamic occupancy index of each planned plot is calculated.

[0023] Furthermore, the specific process for setting path priority rules and threshold filtering based on berth inspection task requirements is as follows: Path priority rule setting: Path priority score = Uninspected berth weight × Adaptive weight coefficient + Smart berth manager weight × Adaptive weight coefficient; Preset uninspected berth weight values ​​and inspected berth weight values, with the preset weight value corresponding to uninspected berths being higher than that corresponding to inspected berths; Preset core area smart berth manager weight values ​​and ordinary area smart berth manager weight values, with the preset weight value corresponding to the core area being greater than that corresponding to the ordinary area; Threshold filtering process: Set a first initial threshold and a second initial threshold, preset a busy period correction factor and an unbusy period correction factor, and set the value of the busy period correction factor... The value of the correction factor for non-busy periods is greater than or equal to 1; the current time period is determined to be either a busy or non-busy period, and the correction factor for the corresponding period is retrieved; the first initial threshold and the second initial threshold are corrected respectively using the correction formula: preset threshold = initial threshold × corresponding period correction factor, to generate the first preset threshold and the second preset threshold; wherein, the first preset threshold generated during busy periods is less than the first initial threshold, and the second preset threshold is less than the second initial threshold, while the first preset threshold generated during non-busy periods is greater than or equal to the first initial threshold, and the second preset threshold is greater than or equal to the second initial threshold; planned plots with static occupancy indicators less than the first preset threshold and dynamic occupancy indicators less than the second preset threshold are selected as pre-access plots.

[0024] Furthermore, the specific process of calculating priority scores for pre-passage plots and generating real-time movement paths is as follows: Calculate the priority score of each pre-passage plot according to the path priority rule, and determine the plot with the highest score as the target passage plot; if there are multiple plots with the same highest priority score, select the plot closest to the current plot as the target passage plot; use a path search algorithm, taking the center point of the current plot as the starting point and the center point of the target passage plot as the ending point, and combine the plot boundary coordinates, channel width, and occupancy index data to plan a collision-free optimal movement path. The path data includes the coordinates of the passing points, the driving direction, and the driving speed parameters, and is sent to the execution unit of the inspection device.

[0025] Furthermore, the specific process of the inspection device continuously updating indicators and dynamically adjusting the movement path is as follows: When the inspection device travels along the real-time movement path, it repeatedly executes the steps of panoramic layout and environmental data acquisition, occupancy indicator calculation, priority screening and path generation according to a preset cycle, re-collects environmental data and updates static and dynamic occupancy indicators; if the updated indicators show that the plot corresponding to the current travel path still meets the passage requirements, the original path is maintained and the device continues to travel; if the indicators exceed the preset threshold or the priority of the target passage plot is surpassed by other plots, the priority score is recalculated and a new target passage plot is determined, and a new real-time movement path is generated.

[0026] A berth inspection system based on dynamic path planning, implementing the above method, includes: an inspection device integrating a multi-sensor fusion module, a positioning module, a data processing module, a path planning module, and an execution module, used to collect environmental and layout data, calculate berth occupancy indicators, receive path instructions, and execute movement operations; a back-end management system used to store panoramic layout data of the berth area, receive real-time data uploaded by the inspection device, generate inspection tasks, calculate peak correction factors and preset thresholds, and send them to the inspection device; a data storage module used to store the boundary coordinates of the planned plots, historical inspection data, berth occupancy indicator calculation results, and path planning data, providing data support for threshold correction and path optimization; and a communication module used to realize real-time data transmission between the inspection device and the back-end management system, ensuring the stability of instruction issuance and data upload.

[0027] Furthermore, the inspection device also integrates a human-machine collaborative assistance module, including a storage unit, a flashing warning unit, and a remote interaction unit. The storage unit is used to store spare batteries and special tools required for the maintenance of the intelligent berth manager, and is stored in partitions according to spare parts type. The flashing warning unit is used to activate the warning function when the inspector performs maintenance work, forming a warning range suitable for the safety of working in public areas. The remote interaction unit is used to establish a communication connection between the inspector and the back-end technical platform, supporting fault image uploading, real-time voice interaction, and maintenance manual push functions. In one example, it should be further explained that the inspection device can take the form of an autonomous vehicle or drone. The autonomous vehicle has real-time network connectivity and can complete inspection data collection and precise navigation through AI image recognition and accurate location positioning. It also has storage functions to store spare parts and tools, and uses a flashing warning function to remind surrounding vehicles to slow down. The autonomous vehicle integrates battery charging function and pre-positions the required proportion of spare batteries for the corresponding inspection section. It receives the number and location information of batteries that need maintenance and replacement from the system every day. In routine work, it completes license plate recording and certificate collection in sequence according to the preset inspection section, unaffected by inclement weather, reducing the intensity of manual labor. The inspector and the autonomous vehicle will arrive at the location of the equipment to be maintained together according to system instructions. The autonomous vehicle provides maintenance tools and parts, and the inspector performs equipment maintenance or battery replacement operations. The replaced used batteries can be directly charged and recycled inside the vehicle. During the maintenance process, the autonomous vehicle acts as a warning device to ensure the safety of the inspector, and can also provide the inspector with AI remote technical services to help quickly resolve complex faults.

[0028] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included 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.

[0029] Further refinement was carried out, with each paragraph of the specific implementation explained in as much detail as possible, while still using the format shown in the example. Detailed Implementation

[0030] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0031] Reference Figure 1 A berth inspection method based on dynamic path planning includes the following steps: acquiring panoramic layout information and real-time environmental data of the berth area, wherein the panoramic layout information includes berth distribution, passage boundaries, and fixed obstacle positions, and the real-time environmental data includes information on stationary objects and moving objects on the inspection path, wherein the stationary object information includes the size and position of parked vehicles and temporary obstacles, and the moving object information includes the real-time position and movement trajectory of moving vehicles and pedestrians; dividing the berth area into several planning plots according to the panoramic layout information; calculating the static occupancy index of each planning plot based on a preset static occupancy evaluation model, combined with the maximum width of the passage, the width of stationary object occupancy, and the width of the inspection device itself within each planning plot; and collecting real-time data on the relationship between the inspection device and each planning plot based on a preset dynamic occupancy evaluation model. The system calculates the dynamic occupancy index of each planned plot based on the current distance of moving objects, historical distance changes over a given time period, and the moving speed of the moving objects. It then sets path priority rules based on the berth inspection task requirements, including the priority of uninspected berths and the inspection weight of the intelligent berth manager. Threshold-based screening is applied to the static and dynamic occupancy indices of each planned plot, retaining plots that meet both criteria for passage. Priority scores are calculated for each plot according to the path priority rules, and the plot with the highest priority is selected as the target plot. A real-time movement path is generated for the inspection device from its current location to the target plot. As the inspection device travels along the real-time movement path, the above steps are repeated at a preset cycle, updating the static and dynamic occupancy indices in real time and dynamically adjusting the movement path. In one example, it is worth noting that the execution logic of the entire method follows a closed-loop process of "data collection - analysis and processing - decision generation - dynamic optimization". The steps are coordinated through real-time data flow. For example, the collected environmental data is directly used for occupancy index calculation, and the index results support path selection in real time, ensuring the continuity and efficiency of the inspection process. In addition, all preset parameters in the method (such as threshold, weight, and period) can be customized and adjusted by the back-end management system according to the actual parking scenario (such as open-air parking spaces and underground parking garage parking spaces) to adapt to the inspection needs of different scenarios.

[0032] Furthermore, the specific process for obtaining panoramic layout information and real-time environmental data of the berth area is as follows: Panoramic layout information of the berth area is obtained through a combination of multi-sensor fusion acquisition and background map calibration. In one example, it should be further explained that the multi-sensors include LiDAR and cameras. The LiDAR uses a 16-line or 32-line LiDAR, performing a 360-degree omnidirectional scan of the berth area at a frequency of 0.1 seconds / frame, collecting the three-dimensional spatial coordinates of the area boundaries, passageway directions, and fixed obstacles (such as pillars, walls, and fixed parking space locks), generating an initial point cloud model containing coordinate information. The camera uses a high-definition industrial camera (replaced here with "high-resolution industrial camera"), capturing image data of berth markers (such as parking space number plates), passageway boundary lines (such as white dividing lines), and fixed obstacles at a frequency of 2 frames / second along a preset acquisition route. Feature points (such as marker corners and boundary line endpoints) in the image are extracted using image recognition algorithms and converted into pixel coordinates. Then, the pixel coordinates are converted into world coordinates using camera calibration parameters. The initial point cloud model is fused and aligned with the world coordinates of the image feature points, and the ICP (Iterative Closest Point) algorithm is used to eliminate the two types of data. Spatial deviations are then addressed by registering the coordinates with a pre-stored base map of the berth area (including design coordinates and berth attributes). Gaussian filtering removes environmental noise, ultimately determining the precise berth distribution locations, passageway boundaries, and the specific coordinates of fixed obstacles. Furthermore, the inspection device utilizes multi-sensor equipment, including LiDAR, cameras, and infrared sensors, to collect real-time environmental data along the inspection path. The LiDAR scans the inspection path at a frequency of 0.05 seconds per frame, using a contour extraction algorithm to detect the contour dimensions of stationary objects (parked vehicles, temporary storage items). Combined with the scanned coordinates, the three-dimensional dimensions and spatial location information of the stationary objects are calculated. The camera and infrared sensor work in tandem. The camera continuously captures images of moving objects at a frequency of 3 frames per second, while the infrared sensor simultaneously captures the thermal imaging trajectory of the moving objects. By comparing pixel changes in adjacent frames using the frame difference method, the real-time two-dimensional coordinates of the moving objects are calculated. Combining data from more than 10 consecutive frames, the least squares method is used to fit the complete movement trajectory. Simultaneously, the moving speed of the moving objects is calculated based on the time interval between adjacent coordinate points on the trajectory, with speed data retained to two decimal places to ensure accuracy.

[0033] Furthermore, the specific process of dividing the berth area into several planning plots based on the panoramic layout information is as follows: Using the passageway direction as a reference, obtain the panoramic layout coordinate data of the berth area; calculate the number of berths per unit area to obtain the berth distribution density coefficient, calculated as "berth distribution density coefficient = total number of berths in the area ÷ total area of ​​the area"; preset multiple density threshold intervals (such as low-density intervals, medium-density intervals, and high-density intervals), each density threshold interval corresponding to a preset number of adjacent berth combinations (such as 2 adjacent berths for low-density intervals, 3-4 adjacent berths for medium-density intervals, and 5 adjacent berths for high-density intervals); match the calculated berth distribution density coefficient with the preset density threshold intervals to determine the corresponding number of adjacent berth combinations; for continuously distributed berth areas, according to the matched number of adjacent berth combinations, divide the corresponding number of adjacent berths and the passageway range of the area into individual planning plots; obtain the information of a single berth and its nearest surrounding berths. The system calculates the straight-line distance and sets a preset distance threshold. Berths with a straight-line distance greater than the preset threshold are classified as independent, dispersed berths. Preset maintenance operation space parameters and passage space parameters are also used. Maintenance operation space parameters include the inspector's working radius and tool placement space width. Passage space parameters include the minimum width required for surrounding vehicle passage. The passage area range parameters are calculated using the formula: "Passage area range parameter = Maintenance operation space parameter + Passage space parameter". For independent, dispersed berths, each berth and the corresponding passage area based on the calculated passage area range parameters are divided into individual planning plots. The panoramic layout coordinate data of the berth area is then gridded. Coordinate partitioning algorithms (such as rectangle partitioning algorithms) are used to calculate the diagonal vertex coordinates (starting point and ending point coordinates) of each planning plot, forming closed area boundary data. This boundary data is stored in the plot information database in the format "Plot number - Starting point X - Starting point Y - Ending point X - Ending point Y". In one example, it should be further explained that the preset density threshold range is based on the coverage and mobility of the single data collection of the inspection device. For example, the setting of the high-density range is intended to avoid data processing delays caused by too many berths in a single plot. In addition, the preset distance threshold needs to be combined with the normal spacing of berths, which is usually set to 1.5 times the normal berth spacing to ensure accurate determination of independent and scattered berths and avoid overly fragmented plot division.

[0034] Furthermore, the specific description of the preset static occupancy assessment model is as follows: Static occupancy index = (total occupancy width of stationary objects + width of the inspection device itself) ÷ maximum width of the inspection path, where the total occupancy width of stationary objects is the sum of the projected widths of all stationary objects on the inspection path within the corresponding planned plot in the direction of the inspection path, and the maximum width of the inspection path is the maximum straight-line distance between the two boundaries of the inspection path within the corresponding planned plot; during the calculation process, the total occupancy width of stationary objects is extracted by the contour data of stationary objects collected by sensors, and combined with the preset width of the inspection device itself and the maximum width of the inspection path, the static occupancy index of each planned plot is obtained. In one example, it is necessary to further explain that the extraction process of the total occupied width of stationary objects is as follows: the contour data of each stationary object is projected and transformed, and the three-dimensional contour is projected onto the extension direction of the inspection path to obtain a one-dimensional projection line segment. The length of the line segment is the occupied width of a single stationary object. The occupied widths of all stationary objects are added together to obtain the total occupied width. In addition, the width of the inspection device itself is a preset fixed value, which is entered into the system when the device leaves the factory. The maximum width of the inspection path is calculated through the channel boundary coordinates in the panoramic layout data, that is, the distance between the two sides of the channel in the direction perpendicular to the path.

[0035] Furthermore, the specific description of the preset dynamic occupancy assessment model is as follows: Dynamic occupancy index = (historical distance ÷ current distance) × [1 ÷ (1 + exponential function ((current distance - historical distance) ÷ preset sampling interval))], where the historical distance is the distance between the inspection device and the active object in the corresponding planned plot at a historical time, the current distance is the distance between the inspection device and the active object at the current time, the time interval between the historical time and the current time is the preset sampling interval, and the preset sampling interval is adapted to the moving speed of the active object and the response requirements of the inspection device; during the calculation process, the current distance is collected in real time by the distance sensor of the inspection device, the historical distance at the preset sampling interval time is retrieved from the local cache, and the dynamic occupancy index of each planned plot is calculated. In one example, it is worth noting that the distance sensor employs a fusion scheme of ultrasonic and laser rangefinders. The ultrasonic sensor is used for short-range (0-5 meters) distance acquisition, while the laser rangefinder is used for medium- to long-range (5-50 meters) distance acquisition. The data from both sensors complement each other to ensure distance accuracy. Furthermore, the exponential function adjusts the sensitivity of distance changes to the dynamic occupancy indicator. When a moving object approaches rapidly (the current distance is much smaller than the historical distance), the exponential function value increases, causing the dynamic occupancy indicator to rise significantly, thus highlighting the risk warning effect.

[0036] Furthermore, the specific process for setting path priority rules and threshold filtering based on berth inspection task requirements is as follows: Path priority rule setting: Path priority score = Uninspected berth weight × Adaptation weight coefficient + Smart berth manager weight × Adaptation weight coefficient; Preset uninspected berth weight values ​​and inspected berth weight values ​​(e.g., uninspected berth weight value set to 1.0, inspected berth weight value set to 0.3), the preset weight value corresponding to uninspected berths is higher than the preset weight value corresponding to inspected berths; Preset core area smart berth manager weight values ​​and ordinary area smart berth manager weight values ​​(e.g., core area weight value set to 1.2, ordinary area weight value set to 1.0), the preset weight value corresponding to the core area is greater than the preset weight value corresponding to the ordinary area; Threshold filtering process: Set a first initial threshold and a second initial threshold (e.g., both set to 1.0), preset a busy period correction factor. For both busy and off-peak periods, a correction factor is applied. The value of the correction factor for busy periods is less than 1 (e.g., 0.6-0.8), while the value of the correction factor for off-peak periods is greater than or equal to 1 (e.g., 1.0-1.1). Based on the current time point, the system determines whether the current period is busy or off-peak and retrieves the corresponding correction factor. Using the correction formula "preset threshold = initial threshold × corresponding period correction factor," the first and second initial thresholds are corrected to generate the first and second preset thresholds. Specifically, the first preset threshold generated during busy periods is less than the first initial threshold, and the second preset threshold is less than the second initial threshold; the first preset threshold generated during off-peak periods is greater than or equal to the first initial threshold, and the second preset threshold is greater than or equal to the second initial threshold. Planned plots with static occupancy indicators less than the first preset threshold and dynamic occupancy indicators less than the second preset threshold are selected as pre-access plots. In one example, it should be further explained that the adaptation weight coefficient is preset to 0.5 to ensure that the weight of uninspected parking spaces and smart parking managers has a balanced impact on the score. In addition, the correction factor is determined based on historical traffic flow data statistics. By analyzing the vehicle traffic volume at different times in the past 3 months, the time period with the highest traffic volume is identified as the busy time period. The corresponding correction factor is set inversely proportional to the traffic volume percentage. The higher the traffic volume, the smaller the correction factor and the stricter the screening conditions.

[0037] Furthermore, the specific process of calculating priority scores for pre-passage plots and generating real-time movement paths is as follows: Calculate the priority score of each pre-passage plot according to the path priority rule, and determine the plot with the highest score as the target passage plot; if there are multiple plots with the same highest priority score, select the plot closest to the current plot as the target passage plot; use a path search algorithm (such as the A* algorithm), taking the center point of the current plot as the starting point and the center point of the target passage plot as the ending point, and combine the plot boundary coordinates, channel width, and occupancy index data to plan a collision-free optimal movement path. The path data includes the coordinates of the passing points, the driving direction, and the driving speed parameters, and is sent to the execution unit of the inspection device. In one example, it should be further explained that the interval between the coordinate points along the route is set to 1.5 times the length of the vehicle body of the inspection device to ensure a smooth and complete path. The driving speed parameter is dynamically adjusted according to the occupancy index. The lower the static occupancy index and the lower the dynamic occupancy index, the higher the driving speed, with a maximum of 5 km / h and a minimum of 1 km / h. In addition, the route planning should avoid areas where the static occupancy index exceeds the standard, and a safe distance (such as not less than 1 meter) should be reserved with moving objects.

[0038] Furthermore, the specific process of the inspection device continuously updating indicators and dynamically adjusting its movement path is as follows: When the inspection device travels along the real-time movement path, it repeatedly executes the steps of panoramic layout and environmental data acquisition, occupancy indicator calculation, priority filtering, and path generation at a preset cycle. It re-collects environmental data and updates static and dynamic occupancy indicators. If the updated indicators show that the plot corresponding to the current travel path still meets the passage requirements, it continues to travel along the original path. If the indicators exceed the preset threshold or the priority of the target passable plot is surpassed by other plots, the priority score is recalculated, a new target passable plot is determined, and a new real-time movement path is generated. In one example, it should be further explained that the preset cycle is set according to the environmental complexity. The cycle is set to 0.5 seconds for complex environments (such as dense traffic and many obstacles) and 1 second for simple environments to ensure timely response to environmental changes. In addition, a smooth transition algorithm is used during path adjustment to avoid the inspection device from accelerating or turning sharply, ensuring movement stability.

[0039] A berth inspection system based on dynamic path planning, implementing the above method, includes: an inspection device integrating a multi-sensor fusion module, a positioning module, a data processing module, a path planning module, and an execution module, used to collect environmental and layout data, calculate berth occupancy indicators, receive path instructions, and execute movement operations; a back-end management system used to store panoramic layout data of the berth area, receive real-time data uploaded by the inspection device, generate inspection tasks, calculate peak correction factors and preset thresholds, and send them to the inspection device; a data storage module used to store the boundary coordinates of the planned plots, historical inspection data, berth occupancy indicator calculation results, and path planning data, providing data support for threshold correction and path optimization; and a communication module used to realize real-time data transmission between the inspection device and the back-end management system, ensuring the stability of instruction issuance and data upload. In one example, it is worth further explaining that the multi-sensor fusion module adopts a "data-level fusion" scheme, which converts the raw data collected by each sensor into a unified format before fusion processing to eliminate data redundancy and conflicts; the positioning module adopts a fusion positioning method of GPS + LiDAR + Bluetooth beacon, with GPS providing global positioning and LiDAR and Bluetooth beacon used for local positioning correction, and the positioning accuracy error not exceeding 0.3 meters; the data processing module uses an embedded processor to support parallel computing, ensuring the real-time performance of occupancy index calculation and path planning; in addition, the communication module adopts 5G and WiFi dual-mode communication, with 5G used for long-distance, high-bandwidth data transmission and WiFi used for short-distance (such as underground parking garage) blind spot filling to ensure uninterrupted communication.

[0040] Furthermore, the inspection device also integrates a human-machine collaborative assistance module, including a storage unit, a flashing warning unit, and a remote interaction unit. The storage unit is used to store spare batteries and special tools required for the maintenance of the intelligent berth manager, and is stored in partitions according to spare parts type. The flashing warning unit is used to activate the warning function when the inspector performs maintenance work, forming a warning range suitable for the safety of working in public areas. The remote interaction unit is used to establish a communication connection between the inspector and the back-end technical platform, supporting fault image uploading, real-time voice interaction, and maintenance manual push functions. In one example, it should be further explained that the inspection device can take the form of an unmanned vehicle or drone. The unmanned vehicle has real-time network connectivity, communicating bidirectionally with the backend system via 4G / 5G networks. It can identify license plates and equipment fault characteristics using AI image recognition (based on convolutional neural network algorithms), and combine this with precise location positioning to complete inspection data collection and accurate navigation. The unmanned vehicle's storage unit is designed with "battery area - tool area - recycling area." The battery area is equipped with a buffer and fixing structure to prevent battery collisions, and the tool area has slot-type storage spaces designed according to tool sizes. The strobe warning unit uses a red LED strobe light with a warning range of 10-15 meters in radius centered on the vehicle, supporting both automatic (triggered by the inspector starting maintenance mode) and manual (controlled by a handheld terminal) activation modes. Furthermore, the unmanned vehicle integrates a battery charging function, using a fast-charging module that can charge the spent battery within 30 minutes. Once the old battery is charged to 80%, each autonomous vehicle will have a spare battery of the required proportion for the corresponding inspection route pre-loaded. The system receives information daily on the number and location of batteries requiring maintenance or replacement via a backend system. During routine operations, the vehicle sequentially completes license plate registration and certification according to pre-set inspection routes, unaffected by severe weather such as heavy rain, blizzards, or high temperatures, significantly reducing the workload of manual labor. Inspectors and the autonomous vehicle will arrive at the location of the equipment to be maintained according to system instructions. The autonomous vehicle provides maintenance tools and parts, and the inspector performs equipment maintenance or battery replacement. Replaced batteries can be directly placed inside the vehicle for charging and recycling. During maintenance, the autonomous vehicle continuously activates a flashing warning function to provide safety protection for the inspector. Simultaneously, it provides AI remote technical services to the inspector through a remote interaction unit. Backend technicians can provide real-time guidance based on uploaded fault images, and the system automatically pushes relevant maintenance manuals to help quickly resolve complex faults.

[0041] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included 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 berth inspection method based on dynamic path planning, characterized in that, Includes the following steps: The system acquires panoramic layout information and real-time environmental data of the berth area. The panoramic layout information includes berth distribution, passage boundaries, and fixed obstacle positions. The real-time environmental data includes information on stationary objects and moving objects on the inspection path. The information on stationary objects includes the size and position of parked vehicles and temporary obstacles. The information on moving objects includes the real-time position and movement trajectory of moving vehicles and pedestrians. Based on the panoramic layout information, the berth area is divided into several planned plots; Based on the preset static occupancy assessment model, the static occupancy index of each planned plot is calculated by combining the maximum width of the passage, the width of the stationary object, and the width of the inspection device itself within each planned plot. Based on a preset dynamic occupancy assessment model, the system collects real-time data on the current distance between the inspection device and moving objects within each planned plot, the distance changes over historical time periods, and the moving speed of moving objects to calculate the dynamic occupancy index of each planned plot. Path priority rules are set in conjunction with the berth inspection task requirements, which include the priority of uninspected berths and the inspection weight of the intelligent berth manager. Threshold filtering is performed on the static and dynamic occupancy indicators of each planned plot, retaining pre-passage plots where both indicators meet the passage requirements. Calculate priority scores for the pre-passage plots according to the path priority rule, select the plot with the highest priority as the target passage plot, and generate a real-time movement path for the inspection device from the current plot to the target passage plot; As the inspection device travels along the real-time moving path, it repeats the above steps at a preset cycle, updating the static and dynamic occupancy indicators in real time and dynamically adjusting the moving path.

2. The berth inspection method based on dynamic path planning according to claim 1, characterized in that, The specific process for obtaining panoramic layout information and real-time environmental data of the berth area is as follows: Panoramic layout information of the berth area is obtained through a combination of multi-sensor fusion acquisition and background map calibration. The multi-sensors include LiDAR and cameras. LiDAR acquires the spatial coordinates of area boundaries, passageway directions, and fixed obstacles. Cameras acquire images of berth markers and passageway boundary lines. Coordinate calibration is performed with the pre-stored map in the background to determine the specific locations of berth distribution, passageway boundaries, and fixed obstacles. Environmental data along the inspection path is collected in real-time by multi-sensor devices mounted on the inspection device, including LiDAR, cameras, and infrared sensors. LiDAR detects the outline size and spatial coordinates of stationary objects, obtaining their size and position information. Cameras and infrared sensors collaboratively track the motion trajectory of moving objects, calculating real-time positions and fitting movement trajectories through image frame analysis, and simultaneously acquiring the moving speed of the moving objects.

3. The berth inspection method based on dynamic path planning according to claim 1, characterized in that, The specific process of dividing the berth area into several planned plots based on the panoramic layout information is as follows: Using the passageway direction as a reference, obtain the panoramic layout coordinate data of the berth area, calculate the number of berths per unit area to obtain the berth distribution density coefficient; preset multiple density threshold intervals, each corresponding to a preset number of adjacent berth combinations; match the calculated berth distribution density coefficient with the preset density threshold intervals to determine the corresponding number of adjacent berth combinations; for continuously distributed berth areas, divide the corresponding number of adjacent berths and the passageway range of the area into individual planned plots according to the matched number of adjacent berth combinations; obtain the straight-line distance between a single berth and the nearest surrounding berth, preset a distance threshold, and determine berths with a straight-line distance greater than the preset distance threshold as independent and scattered berths; preset maintenance operation space parameters and passageway reserved space parameters, and calculate the passageway range parameters based on these two parameters. For independent and scattered berths, each berth and the channel area corresponding to the calculated channel area range parameters are divided into individual planning plots. The panoramic layout coordinate data of the berth area is processed into a grid, and the coordinates of the diagonal vertices of each planning plot are calculated through a coordinate division algorithm to form closed area boundary data. The boundary data is stored in the plot information database.

4. The berth inspection method based on dynamic path planning according to claim 1, characterized in that, The specific description of the preset static occupancy assessment model is as follows: Static occupancy index = (total occupancy width of stationary objects + width of the inspection device itself) ÷ maximum width of the inspection path. Wherein, the total occupancy width of stationary objects is the sum of the projected widths of all stationary objects on the inspection path within the corresponding planned plot in the direction of the inspection path, and the maximum width of the inspection path is the maximum straight-line distance between the two boundaries of the inspection path within the corresponding planned plot. During the calculation process, the total occupancy width of stationary objects is extracted from the contour data of stationary objects collected by sensors, and combined with the preset width of the inspection device itself and the maximum width of the inspection path, the static occupancy index of each planned plot is obtained.

5. The berth inspection method based on dynamic path planning according to claim 1, characterized in that, The specific description of the preset dynamic occupancy assessment model is as follows: Dynamic occupancy index = (historical distance ÷ current distance) × [1 ÷ (1 + exponential function ((current distance - historical distance) ÷ preset sampling interval))] where the historical distance is the distance between the inspection device and the active object in the corresponding planned plot at a historical time, the current distance is the distance between the inspection device and the active object at the current time, and the time interval between the historical time and the current time is the preset sampling interval. The preset sampling interval is adapted to the moving speed of the active object and the response requirements of the inspection device. During the calculation process, the current distance is collected in real time by the distance sensor of the inspection device, and the historical distance at the preset sampling interval time is retrieved from the local cache to calculate the dynamic occupancy index of each planned plot.

6. The berth inspection method based on dynamic path planning according to claim 1, characterized in that, The specific process of setting path priority rules and threshold filtering based on berth inspection task requirements is as follows: The path priority rule is set as follows: Path priority score = weight of uninspected berth × adaptation weight coefficient + weight of intelligent berth manager × adaptation weight coefficient; The preset weight values ​​for uninspected berths and inspected berths are set, with the preset weight value for uninspected berths being higher than that for inspected berths. The preset weight values ​​for the core area intelligent parking space manager and the ordinary area intelligent parking space manager are set, with the preset weight value for the core area being greater than that for the ordinary area. The threshold screening process is as follows: set a first initial threshold and a second initial threshold, preset a busy period correction factor and an unbusy period correction factor, the value of the busy period correction factor is less than 1, and the value of the unbusy period correction factor is greater than or equal to 1. Determine whether the current time period is a busy or non-busy period based on the current time point, and retrieve the corresponding correction factor for the time period. By modifying the formula preset threshold = initial threshold × corresponding time period correction factor, the first initial threshold and the second initial threshold are modified respectively to generate the first preset threshold and the second preset threshold; wherein, the first preset threshold generated during busy periods is less than the first initial threshold and the second preset threshold is less than the second initial threshold, and the first preset threshold generated during non-busy periods is greater than or equal to the first initial threshold and the second preset threshold is greater than or equal to the second initial threshold. Planned land parcels with static occupancy indicators less than a first preset threshold and dynamic occupancy indicators less than a second preset threshold are selected as pre-access land parcels.

7. The berth inspection method based on dynamic path planning according to claim 1, characterized in that, The specific process of calculating priority scores for pre-passage plots and generating real-time movement paths is as follows: Calculate the priority score of each pre-passage plot according to the path priority rule, and determine the plot with the highest score as the target passage plot; if there are multiple plots with the same highest priority score, select the plot closest to the current plot as the target passage plot; use a path search algorithm, taking the center point of the current plot as the starting point and the center point of the target passage plot as the ending point, and combine the plot boundary coordinates, channel width, and occupancy index data to plan a collision-free optimal movement path. The path data includes the coordinates of the passing points, the driving direction, and the driving speed parameters, and is sent to the execution unit of the inspection device.

8. The berth inspection method based on dynamic path planning according to claim 1, characterized in that, The specific process of the inspection device continuously updating indicators and dynamically adjusting the movement path is as follows: When the inspection device travels along the real-time movement path, it repeatedly executes the steps of panoramic layout and environmental data acquisition, occupancy indicator calculation, priority screening and path generation according to a preset cycle, re-collects environmental data and updates static and dynamic occupancy indicators; if the updated indicators show that the plot corresponding to the current travel path still meets the passage requirements, the original path is maintained and the device continues to travel; if the indicators exceed the preset threshold or the priority of the target passage plot is surpassed by other plots, the priority score is recalculated and a new target passage plot is determined, and a new real-time movement path is generated.

9. A berth inspection system based on dynamic path planning that implements the method of any one of claims 1-8, characterized in that, include: The inspection device integrates a multi-sensor fusion module, a positioning module, a data processing module, a path planning module, and an execution module. It is used to collect environmental and layout data, calculate occupancy indicators, receive path instructions, and execute movement operations. The background management system is used to store panoramic layout data of the berth area, receive real-time data uploaded by the inspection device, generate inspection tasks, calculate peak correction factors and preset thresholds, and send them to the inspection device. The data storage module stores the boundary coordinates of the planned plots, historical inspection data, occupancy index calculation results, and path planning data, providing data support for threshold correction and path optimization; the communication module enables real-time data transmission between the inspection device and the back-end management system, ensuring the stability of command issuance and data upload.

10. The berth inspection system based on dynamic path planning according to claim 9, characterized in that, The inspection device also integrates a human-machine collaborative assistance module, including a storage unit, a flashing warning unit, and a remote interaction unit. The storage unit is used to store spare batteries and special tools required for the maintenance of the intelligent berth manager, and is stored in partitions according to spare parts type. The flashing warning unit is used to activate the warning function when the inspector performs maintenance work, forming a warning range suitable for the safety of working in public areas. The remote interaction unit is used to establish a communication connection between the inspector and the back-end technical platform, supporting fault image uploading, real-time voice interaction, and maintenance manual push functions.

Images