Low-altitude unmanned aerial vehicle inspection conflict detection and resolution method based on digital twinning

Abstract

The application discloses a low-altitude unmanned aerial vehicle inspection conflict detection and elimination method based on digital twinning, relates to the technical field of port low-altitude unmanned aerial vehicle cluster cooperative control, and comprises the following steps: acquiring port BIM model data, GIS data and Internet of Things real-time data flow, and establishing a digital twinning operation instance containing a three-dimensional space-time grid coordinate mapping relationship; analyzing target equipment ID and target component categories according to an inspection task work order, generating an initial flight intention hyperedge carrying an inspection observation semantic label, and mapping the initial flight intention hyperedge to the three-dimensional space-time grid coordinate mapping relationship; performing real-time comparison on the initial flight intention hyperedge of the unmanned aerial vehicle based on the three-dimensional space-time grid coordinate mapping relationship, and obtaining the inspection observation semantic label. The application realizes unmanned aerial vehicle inspection conflict detection, cooperative elimination and dynamic closed-loop updating for a complex port operation environment.

Description

Technical Field

[0001] This invention relates to the field of collaborative control technology for low-altitude unmanned aerial vehicle (UAV) swarms in ports, and in particular to a method for detecting and resolving conflicts during low-altitude UAV inspections based on digital twins. Background Technology

[0002] With the continuous improvement of port intelligence and automation, port equipment inspection technology based on digital twins, 3D modeling and multi-source perception fusion has gradually become an important technical direction for equipment status monitoring and operation management. Existing technologies usually acquire port BIM model data, GIS data and IoT real-time data streams to build port equipment spatial information models, and combine them with UAV inspection technology to realize remote observation and status collection of port operation equipment. By using digital twin operation examples to manage the spatial location, operation status and inspection task information of equipment, the level of informatization of port inspection process is improved.

[0003] Existing port drone inspection technologies based on digital twins typically focus on flight path planning, equipment status acquisition, or spatial location matching in multi-drone collaborative inspection scenarios. They lack a unified description of the dynamic relationship between the observed target, observation perspective, and the operating status of port equipment during drone inspection tasks. This makes it difficult to accurately identify the actual conflict relationship and coordinate the influence between drone inspection behavior and the operating status of port equipment when different drones have overlapping lines of sight to the same equipment component. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a method for detecting and resolving conflicts in low-altitude UAV inspections based on digital twins, which solves the problems of difficulty in uniformly describing the correlation between observation targets, observation perspectives and the operating status of port operation equipment during multi-UAV collaborative inspections, as well as the difficulty in identifying real conflicts.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides a method for conflict detection and resolution of low-altitude UAV inspection based on digital twins, which includes acquiring port BIM model data, GIS data and IoT real-time data streams, and establishing a digital twin operation instance containing three-dimensional spatiotemporal grid coordinate mapping relationships;

[0008] Based on the inspection task work order, the target equipment ID and target component category are parsed, and an initial flight intention hyperedge carrying inspection observation semantic tags is generated. The initial flight intention hyperedge is then mapped to the three-dimensional spatiotemporal grid coordinate mapping relationship.

[0009] Based on the three-dimensional spatiotemporal grid coordinate mapping relationship, the initial flight intention hyperedge of the UAV is compared in real time, and inspection observation semantic labels are obtained.

[0010] The conflict grid data is identified based on the target equipment ID, target component category, and observation azimuth angle to obtain true conflict cluster information. Based on the true conflict cluster information, the observation priority and the estimated cost of inspection interruption failure are determined, and the port operation equipment associated with the conflict grid is identified to construct a heterogeneous cost function.

[0011] Based on the calculation results of the heterogeneous cost function, a composite control package is generated. Real-time feedback data generated by the UAV and port operation equipment after executing the composite control package is collected, and the operation status and occupancy status in the three-dimensional spatiotemporal grid coordinate mapping relationship in the digital twin operation instance are updated.

[0012] As a preferred embodiment of the low-altitude UAV inspection conflict detection and resolution method based on digital twins described in this invention, the establishment of a digital twin operational instance containing a three-dimensional spatiotemporal grid coordinate mapping relationship includes,

[0013] By acquiring port BIM model data, GIS data, and IoT real-time data streams, and completing spatial benchmark unification and time synchronization processing respectively, spatial base plate data and real-time status data sources are formed.

[0014] Based on the spatial base plate data, a three-dimensional spatial mesh is divided, and a structural skeleton of the three-dimensional spatiotemporal mesh coordinate mapping relationship composed of spatial raster index and time slice index is established in combination with the time dimension.

[0015] The static information from the spatial base plate data and the dynamic information from the real-time status data source are written into the corresponding grid cells of the structural skeleton of the three-dimensional spatiotemporal grid coordinate mapping relationship to form the three-dimensional spatiotemporal grid coordinate mapping relationship. The three-dimensional spatiotemporal grid coordinate mapping relationship is then registered to the digital twin running instance.

[0016] As a preferred embodiment of the low-altitude UAV inspection conflict detection and resolution method based on digital twins described in this invention, wherein: the generation of the initial flight intent hyperedge carrying inspection observation semantic labels includes,

[0017] Perform field reading processing on the inspection task work order, read the target equipment number field and the target component category field in the inspection task work order, and obtain the target equipment ID and the target component category;

[0018] Based on the target device ID, query the spatial bounding box data of the target device in the digital twin operation instance, and calculate the observation azimuth angle based on the target component category. Combine the target device ID, target component category and observation azimuth angle to form the inspection observation semantic tag.

[0019] Based on the start and end location information in the inspection task work order and the spatial bounding box data corresponding to the target equipment ID, an initial waypoint sequence is generated. The initial waypoint sequence is used as the geometric trajectory data of the initial flight intention hyperedge, and the inspection observation semantic label is associated with the initial flight intention hyperedge.

[0020] As a preferred embodiment of the low-altitude UAV inspection conflict detection and resolution method based on digital twins described in this invention, the step of mapping the initial flight intention hyperedge to the three-dimensional spatiotemporal grid coordinate mapping relationship includes,

[0021] The time parameter discretization process is performed on the initial flight intention hyperedge carrying the inspection observation semantic label. The expected flight time of the initial flight intention hyperedge is divided into a continuous time slice index sequence according to the preset time slice interval. The continuous spatial coordinate points in the geometric trajectory data of the initial flight intention hyperedge are converted into the corresponding spatial raster index according to the preset spatial raster side length.

[0022] The spatial raster indexes corresponding to the same timestamp are combined into a spatiotemporal occupancy tuple to obtain a spatiotemporal occupancy sequence consisting of multiple spatiotemporal occupancy tuples. The spatiotemporal occupancy sequence is used as the data source for writing the three-dimensional spatiotemporal grid coordinate mapping relationship.

[0023] Write the inspection observation semantic tags in the spatiotemporal occupancy sequence into the grid cells of the corresponding time slice index and spatial raster index, and update the occupancy status record in the three-dimensional spatiotemporal grid coordinate mapping relationship.

[0024] As a preferred embodiment of the low-altitude UAV inspection conflict detection and resolution method based on digital twins described in this invention, wherein: obtaining the inspection observation semantic tags includes,

[0025] The updated 3D spatiotemporal grid coordinate mapping relationship is jointly traversed by time slice index and spatial grid index. Based on the source identifier of the initial flight intention hyperedge in the occupancy record corresponding to the spatial grid index, the initial flight intention hyperedge corresponding to different UAVs is filtered, and the inspection observation semantic label associated with the initial flight intention hyperedge is read.

[0026] Based on the target device ID and target component category in the inspection observation semantic label, establish the association relationship of the observation target, and collect the initial flight intention hyperedges corresponding to different UAVs with the same target device ID and target component category into a set of overlapping records of the same observation target;

[0027] Based on the spatial raster index and time slice index in the record set of overlapping observation targets, the relationship between the observation rays from the physical UAV to the target component is constructed, and the degree of line-of-sight overlap is calculated according to the projection relationship of the observation rays corresponding to different physical UAVs on the surface of the target component.

[0028] Based on the surface geometric features of the target component, the line-of-sight overlap judgment parameters are adjusted. The overlapping records of the same observation target that meet the preset judgment conditions are identified as observation line-of-sight conflict records, and the inspection observation semantic tags corresponding to the observation line-of-sight conflict records are summarized.

[0029] As a preferred embodiment of the low-altitude UAV inspection conflict detection and resolution method based on digital twins described in this invention, the true conflict cluster information includes:

[0030] Based on the azimuth angle of the observation view in the semantic tags of different inspection observations, a spatial correspondence between the corresponding observation rays and the observation area on the surface of the target component is established. The overlap of the observation range is judged according to the coverage area relationship of the target component surface corresponding to different observation rays. The observation line conflict records with overlapping observation ranges are collected to obtain the observation line overlap conflict record set.

[0031] Based on the inspection observation semantic tags in the set of overlapping conflict records of observation lines, combined with the amount of effective data collected, the completeness of target component observation, and the data discrimination ability of observation angle azimuth, the observation priority parameters are determined and associated with the corresponding observation line conflict records to obtain the set of priority-labeled observation conflict records.

[0032] Based on the observation continuity information in the priority-labeled observation conflict record set, the impact of different inspection observation semantic tags on the data validity after the interruption of the observation task is calculated, the estimated cost of inspection interruption failure is determined, and it is associated with the corresponding observation line conflict record to obtain the basic data of cost-related conflict clusters.

[0033] Based on the observation priority parameters and the estimated cost of inspection interruption failure in the basic data of cost-related conflict clusters within the same spatiotemporal grid cell, representative conflict records are determined, and the basic data of the associated cost-related conflict clusters are aggregated to form true conflict clusters, thus obtaining true conflict cluster information with the attribute of representative conflict records.

[0034] As a preferred embodiment of the low-altitude UAV inspection conflict detection and resolution method based on digital twins described in this invention, the construction of the heterogeneous cost function includes:

[0035] Based on the true conflict cluster information with attached conflict association representative record attributes, the corresponding spatiotemporal grid cell is located in the three-dimensional spatiotemporal grid coordinate mapping relationship, the port operation equipment operation information associated with the spatiotemporal grid cell is read, and the conflict association relationship between the UAV and the port operation equipment is established according to the spatial occupancy relationship within the spatiotemporal grid cell, thus obtaining the conflict-associated port operation equipment set.

[0036] Based on the observation priority parameters, the estimated cost of inspection interruption failure, and the azimuth angle of the observation view in the true conflict cluster information with attached conflict association representative record attributes, the degree of change in observation continuity after the UAV inspection task is adjusted is calculated, and the impact characterization of UAV inspection continuity is established based on the degree of change in observation continuity.

[0037] Based on the set of conflict-related port operation equipment and corresponding spatiotemporal grid units, the operation connection relationship between port operation equipment is calculated, the state changes of conflict-related port operation equipment are propagated along the operation connection relationship, the port operation impact propagation relationship is established, and the port operation continuity impact characterization is formed.

[0038] Based on the characterization of the continuity impact of UAV inspections and port operations, a cross-subject collaborative constraint relationship is constructed. This cross-subject collaborative constraint relationship is then mapped to true conflict cluster information, forming a heterogeneous cost function jointly constrained by the continuity impact terms of UAV inspections and port operations.

[0039] As a preferred embodiment of the low-altitude UAV inspection conflict detection and resolution method based on digital twins described in this invention, the generation of the composite control package includes:

[0040] Based on the continuous impact terms of UAV inspection, port operation, and observation priority parameters in the heterogeneous cost function, the impact relationships of UAV path adjustment, observation angle adjustment, and port operation equipment coordination adjustment are calculated to obtain the heterogeneous cost function calculation results.

[0041] Based on the results of the heterogeneous cost function calculation, including the impact of UAV path adjustment, the impact of observation perspective adjustment, and the impact of port operation equipment collaborative adjustment, the corresponding adjustment strategy is determined according to the observation priority parameter. The adjustment strategy is then converted into UAV control commands and port equipment collaborative commands, generating a composite control package containing both UAV control commands and port equipment collaborative commands.

[0042] As a preferred embodiment of the low-altitude UAV inspection conflict detection and resolution method based on digital twins described in this invention, the real-time transmitted data includes:

[0043] By combining the UAV control commands and port equipment coordination commands in the composite control package, the UAV and port operation equipment are controlled to perform corresponding adjustment operations, and the pose status information, equipment operation status information and task execution status information of the UAV and port operation equipment after executing the composite control package are collected simultaneously.

[0044] The pose status information, equipment operation status information, and task execution status information of the UAV and port operation equipment after executing the composite control package are associated with the time slice index and spatial grid index to form real-time back transmission data.

[0045] As a preferred embodiment of the low-altitude UAV inspection conflict detection and resolution method based on digital twins described in this invention, the step of updating the running status and occupancy status in the three-dimensional spatiotemporal grid coordinate mapping relationship of the digital twin running instance includes...

[0046] Based on the pose status information, device operation status information and task execution status information in the real-time transmitted data, the corresponding device status, task status and operation parameters in the digital twin operation instance are synchronously updated to obtain the updated operation status in the digital twin operation instance.

[0047] Based on the running status of the updated digital twin instance, the spatial location change information corresponding to the real-time transmitted data is associated with and updated with the time slice index and spatial raster index, and the occupancy status of the corresponding grid cell in the three-dimensional spatiotemporal grid coordinate mapping relationship is adjusted.

[0048] The beneficial effects of this invention are as follows: By integrating port BIM model data, GIS data, and real-time IoT data streams, a digital twin operational instance containing a three-dimensional spatiotemporal grid coordinate mapping relationship is constructed. Combined with the target equipment ID and target component category parsed from the inspection task work order, an initial flight intent hyperedge carrying inspection observation semantic tags is generated, realizing the dynamic mapping of UAV inspection tasks in the three-dimensional spatiotemporal grid coordinate mapping relationship. Based on the three-dimensional spatiotemporal grid coordinate mapping relationship, spatiotemporal correlation analysis is performed on the initial flight intent hyperedges of different UAVs. An observation ray relationship is established by combining the target equipment ID, target component category, and observation azimuth angle, identifying real observation conflicts and forming... The system generates true conflict cluster information; based on this information, it determines observation priority parameters and estimated costs of inspection interruption failure, associates the port operation equipment status corresponding to the conflict grid, and constructs a heterogeneous cost function that integrates the continuity impact of UAV inspection and the continuity impact of port operations; based on the calculation results of the heterogeneous cost function, it generates a composite control package to coordinate the UAV inspection path, observation perspective, and port operation equipment adjustment, and collects real-time feedback data after execution to update the operation status in the digital twin operation instance and the occupancy status in the three-dimensional spatiotemporal grid coordinate mapping relationship, thereby realizing UAV inspection conflict detection, collaborative resolution, and dynamic closed-loop update for complex port operation environments. Attached Figure Description

[0049] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a flowchart of a low-altitude UAV inspection conflict detection and resolution method based on digital twins. Detailed Implementation

[0051] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0052] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0053] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0054] Reference Figure 1 As one embodiment of the present invention, this embodiment provides a method for conflict detection and resolution of low-altitude unmanned aerial vehicle (UAV) inspections based on digital twins, comprising the following steps:

[0055] S1. Acquire port BIM model data, GIS data, and real-time IoT data streams to establish a digital twin operational instance containing three-dimensional spatiotemporal grid coordinate mapping relationships.

[0056] S1.1. By acquiring port BIM model data, GIS data, and IoT real-time data streams, and completing spatial benchmark unification and time synchronization processing respectively, spatial base plate data and real-time status data sources are formed.

[0057] Furthermore, port BIM model data is acquired through a port 3D modeling platform; GIS data is acquired through a geographic information acquisition platform; and real-time IoT data streams are acquired through IoT sensing devices deployed on port operating equipment. These real-time IoT data streams are obtained through equipment status acquisition interfaces, including equipment operating status information, equipment location change information, and equipment operating parameter information. Addressing the issue of inconsistent spatial reference systems among port BIM model data, GIS data, and real-time IoT data streams, the spatial coordinates of equipment in the port BIM model data, the geographic coordinates in the GIS data, and the equipment positioning information in the real-time IoT data streams are read separately. Spatial location mapping relationships are established between different data sources through spatial coordinate transformation relationships, unifying the equipment entity locations in the port BIM model data, the geographic spatial locations in the GIS data, and the equipment status locations in the real-time IoT data streams to the same spatial reference frame. This forms spatial baseboard data containing port equipment spatial structure information and regional spatial relationship information. By reading the time stamp information in the real-time IoT data streams and performing time calibration on the equipment status change process based on a unified time benchmark, a correlation is established between the equipment status change information and the corresponding time nodes, forming a real-time status data source.

[0058] S1.2. Based on the spatial base plate data, perform three-dimensional spatial mesh division, and combine the time dimension to establish a structural skeleton of three-dimensional spatiotemporal mesh coordinate mapping relationship composed of spatial raster index and time slice index.

[0059] Furthermore, the three-dimensional spatial division boundary is determined based on the spatial coordinate range in the spatial base plate data. The continuous spatial region is divided into multiple spatial raster units with independent spatial location identifiers. A spatial raster index is generated based on the spatial coordinate positions corresponding to the spatial raster units. The spatial raster index is obtained through the mapping relationship between spatial location coordinates and spatial raster division rules. It is used to identify the spatial location of equipment entities, UAV flight paths, and occupied areas within the port area. A time dimension division relationship is established based on the time identifier information in the real-time status data source. The continuous equipment operation status change process is divided into multiple time slices, and a time slice index is established for each time slice. The spatial raster index and the time slice index are associated so that the same spatial location corresponds to different time slice indices under different time states. The spatial raster index describes the spatial location relationship, and the time slice index describes the status change process, forming a structural skeleton of three-dimensional spatiotemporal grid coordinate mapping relationship composed of the spatial raster index and the time slice index.

[0060] S1.3. Write the static information in the spatial base plate data and the dynamic information in the real-time status data source into the structural skeleton corresponding to the grid unit of the three-dimensional spatiotemporal grid coordinate mapping relationship to form the three-dimensional spatiotemporal grid coordinate mapping relationship, and register the three-dimensional spatiotemporal grid coordinate mapping relationship to the digital twin running instance.

[0061] Furthermore, based on the mapping relationship between the spatial coordinates of the equipment entity and the spatial grid index, the spatial location, geometric structure, spatial relationship of equipment components, and regional spatial boundary information of the port operation equipment are stored in the grid cells corresponding to the spatial grid index. Dynamic information from the real-time status data source is written into the corresponding grid cells according to the association relationship between the time slice index and the spatial grid index. Based on the association relationship between the equipment hardware serial number and the equipment entity identifier, the equipment operating status information, equipment location change information, and equipment operating parameter information are mapped to the spatial grid index under the corresponding time slice index, realizing the corresponding storage of dynamic status information and spatial entity information. An index association relationship is established for the grid cells after writing static and dynamic information, so that the spatial grid index, time slice index, equipment entity information, and equipment status information form a unified mapping relationship, forming a three-dimensional spatiotemporal grid coordinate mapping relationship, and the three-dimensional spatiotemporal grid coordinate mapping relationship is registered in the digital twin operation instance.

[0062] S2. Based on the inspection task work order, parse the target equipment ID and target component category, generate an initial flight intention hyperedge carrying inspection observation semantic tags, and map the initial flight intention hyperedge to the three-dimensional spatiotemporal grid coordinate mapping relationship.

[0063] S2.1 Perform field reading processing on the inspection task work order, read the target equipment number field and the target component category field in the inspection task work order, and obtain the target equipment ID and target component category.

[0064] Furthermore, the inspection task work order is generated by the port inspection task management platform according to the inspection plan. The inspection task work order includes at least the target equipment number field, the target component category field, the inspection start position field, the inspection end position field, and the task identifier field. The data structure of the inspection task work order is parsed according to the preset field mapping rules. The field positioning relationship is established by the field name and the field type to locate the data position corresponding to the target equipment number field and the target component category field. The field mapping rules can be implemented by field name matching. For example, the field name is matched item by item with the predefined field dictionary. When the field name is consistent with the target equipment number field in the field dictionary, the corresponding field value is read as the target equipment ID; when the field name is consistent with the target component category field in the field dictionary, the corresponding field value is read as the target component category.

[0065] The system searches the device index directory maintained by the digital twin runtime instance to see if the target device ID has a corresponding device entity. If a corresponding device entity is found, the target device ID is output as a valid device identifier. If no corresponding device entity is found, the task is marked as abnormal and returned to the task management platform for re-verification. The system performs standardized mapping processing on the target component category. Based on the component category mapping table, the component descriptions in the inspection task work order are uniformly converted into standard component category identifiers defined within the digital twin runtime instance, so that a unique association is established between the target device ID and the target component category.

[0066] S2.2. Based on the target device ID, query the spatial bounding box data of the target device in the digital twin operation instance, and calculate the observation azimuth angle based on the target component category. Combine the target device ID, target component category and observation azimuth angle to form the inspection observation semantic label.

[0067] Furthermore, the target device ID is input into the digital twin runtime instance, and the target device spatial query function is called to retrieve the 3D model object corresponding to the target device from the device spatial index maintained by the digital twin runtime instance, and read the spatial bounding box data corresponding to the device. The target device spatial query function establishes a device entity index according to the target device ID, and then reads the boundary information of the 3D model associated with the device entity. By extracting the minimum bounding volume of the 3D model, the spatial bounding box data corresponding to the target device is obtained. The spatial bounding box data includes the bounding box boundary coordinates, the spatial center position, and the boundary range in each direction.

[0068] The positional relationship of the corresponding component within the spatial bounding box is determined based on the target component category. This is achieved using a component position mapping table, which pre-establishes a correspondence between the target component category and the component number of the equipment model. The spatial region where the corresponding component is located is located based on the target component category. The azimuth angle of the observation view is calculated by combining the spatial bounding box data. The observation direction calculation function uses the outward normal direction of the spatial region where the target component is located as the observation reference direction. The observation direction is calculated based on the positional relationship between the center position of the target component and the center position of the spatial bounding box, generating the azimuth angle of the observation view. The target equipment ID, target component category, and azimuth angle of the observation view are combined according to the preset field order to form an inspection observation semantic tag, and an association relationship is established between the target equipment ID and the inspection observation semantic tag.

[0069] The spatial bounding box data expression corresponding to the target device ID is:

[0070] ;

[0071] in, This refers to the spatial bounding box data corresponding to the target device ID. For the target device ID, This is a function to query the target device space in a digital twin runtime instance.

[0072] The expression for the azimuth angle of the observation viewpoint is:

[0073] ;

[0074] in, For the azimuth angle of the observation viewpoint, For the target component category, This is a function for calculating the observation direction based on the target component category and spatial bounding box data.

[0075] S2.3. Based on the start and end location information in the inspection task work order and the spatial bounding box data corresponding to the target equipment ID, generate an initial waypoint sequence, use the initial waypoint sequence as the geometric trajectory data of the initial flight intention hyperedge, and associate the inspection observation semantic label with the initial flight intention hyperedge.

[0076] Furthermore, the inspection start position field and inspection end position field in the inspection task work order are read to obtain the coordinates of the inspection start position and the inspection end position; at the same time, the spatial bounding box data corresponding to the target equipment is read, the spatial bounding box center position and observation position constraint information are extracted, the track connection relationship is established according to the inspection start position, the target equipment spatial bounding box and the inspection end position, and the initial waypoint sequence is generated according to the principle of flight continuity. The initial waypoint sequence can be generated by key waypoint interpolation. Specifically, the starting position of the inspection is used as the starting waypoint, the observation position corresponding to the spatial bounding box of the target equipment is used as the intermediate waypoint, and the ending position of the inspection is used as the ending waypoint. Then, a set of continuous waypoints is generated based on the spatial connection between adjacent waypoints. The set of continuous waypoints forms a complete initial waypoint sequence. The initial waypoint sequence is used as the geometric trajectory data of the initial flight intention hyperedge. At the same time, the inspection observation semantic label is written into the attribute set of the initial flight intention hyperedge, so that the initial flight intention hyperedge contains both geometric trajectory information and inspection observation semantic information. The association method can adopt a unique identifier association method, that is, establish a correspondence between the initial flight intention hyperedge identifier and the inspection observation semantic label identifier.

[0077] S2.4. Perform time parameter discretization processing on the initial flight intention hyperedge carrying inspection observation semantic tags. Divide the expected flight duration of the initial flight intention hyperedge into a continuous time slice index sequence according to the preset time slice interval. Convert the continuous spatial coordinate points in the geometric trajectory data of the initial flight intention hyperedge into the corresponding spatial raster index according to the preset spatial raster side length.

[0078] Furthermore, the geometric trajectory data and estimated flight duration are read from the initial flight intent hyperedge. The estimated flight duration is obtained from the flight distance between adjacent waypoints in the initial waypoint sequence and the UAV's cruising speed. Based on pre-established time-slice division rules, the estimated flight duration is discretized, and the continuous flight time is sequentially divided according to preset time-slice intervals. A corresponding time-slice index is created for each time-slice, forming a continuous sequence of time-slice indices. Using the flight start time as the time base, the time-slice intervals are sequentially accumulated to generate corresponding time nodes, and a correspondence is established between time nodes and time-slice indices, mapping the entire flight process to a continuous sequence of time-slice indices.

[0079] Spatial discretization is performed on the geometric trajectory data in the initial flight intent hyperedge. Based on the spatial grid division rules established in the 3D spatiotemporal grid coordinate mapping relationship, the spatial grid index mapping relationship is invoked to map each continuous spatial coordinate point to the corresponding spatial grid cell, and the spatial grid index corresponding to that spatial grid cell is read. When continuous spatial coordinate points are located near the boundary of two spatial grids, the principle of proximity to the center of the spatial grid to which the coordinate point belongs can be used for determination. That is, the distance from the continuous spatial coordinate point to the center point of the adjacent spatial grid is calculated, and the spatial grid index corresponding to the smallest distance is used as the spatial grid index corresponding to that spatial coordinate point.

[0080] S2.5 Combine the spatial raster indices corresponding to the same timestamp into a spatiotemporal occupancy tuple to obtain a spatiotemporal occupancy sequence composed of multiple spatiotemporal occupancy tuples. Use the spatiotemporal occupancy sequence as the data source for writing the three-dimensional spatiotemporal grid coordinate mapping relationship.

[0081] Furthermore, based on the time slice index, all spatial raster indexes corresponding to each time slice index are retrieved to establish a correspondence between the time slice index and the spatial raster index. Each spatial raster index originates from the spatial discretization result of the initial flight intention hyperedge geometric trajectory data, and each time slice index originates from the temporal discretization result of the expected flight duration. The association is completed according to the corresponding order of the two in the initial flight intention hyperedge. The time information, spatial raster index, and corresponding initial flight intention hyperedge identifier belonging to the same time slice index are combined to form a spatiotemporal occupancy tuple. The spatiotemporal occupancy tuple includes at least the time slice index, the spatial raster index, and the flight intention identifier. When the same time slice index corresponds to multiple spatial raster indices, multiple spatiotemporal occupancy tuples are created sequentially according to the track passage order in the geometric trajectory data. When the same time slice index corresponds to only one spatial raster index, a single spatiotemporal occupancy tuple is created directly. Based on the time sequence of the time slice indexes, all spatiotemporal occupancy tuples are sorted to generate a continuous spatiotemporal occupancy sequence. The temporal correlation between adjacent spatiotemporal occupancy tuples in the sequence is maintained, and the spatiotemporal occupancy sequence is registered as data of three-dimensional spatiotemporal grid coordinate mapping relationship and written into the queue.

[0082] S2.6 Write the inspection observation semantic tags in the spatiotemporal occupancy sequence into the grid cells of the corresponding time slice index and spatial raster index, and update the occupancy status record in the three-dimensional spatiotemporal grid coordinate mapping relationship.

[0083] Furthermore, using the time slice index in the spatiotemporal occupancy tuple as the temporal positioning basis and the spatial raster index as the spatial positioning basis, the corresponding grid cell is located in the three-dimensional spatiotemporal grid coordinate mapping relationship, and the currently saved occupancy status record of that grid cell is read. If the current grid cell does not have an occupancy status record, a new occupancy status record is created and written into the inspection observation semantic tag; if the current grid cell already has an occupancy status record, the current inspection observation semantic tag is added to the existing occupancy status record, and an association index is established between the current flight intention identifier and the existing flight intention identifier, so that multiple inspection observation semantic tags can correspond to the same grid cell. The occupancy status record in the grid cell is updated, and a unified association relationship is established between the current time slice index, spatial raster index, flight intention identifier, and inspection observation semantic tag, and written into the occupancy status record of the corresponding grid cell.

[0084] S3. Based on the three-dimensional spatiotemporal grid coordinate mapping relationship, the initial flight intention hyperedge of the UAV is compared in real time, and inspection observation semantic labels are obtained.

[0085] S3.1 Perform a joint traversal of the time slice index and spatial grid index on the updated three-dimensional spatiotemporal grid coordinate mapping relationship. Based on the source identifier of the initial flight intention hyperedge in the occupancy record corresponding to the spatial grid index, filter the initial flight intention hyperedge corresponding to different UAVs, and read the inspection observation semantic tags associated with the initial flight intention hyperedge.

[0086] Furthermore, using the time slice index as the first traversal object, all spatial raster indices within the corresponding time layer are read sequentially according to the order of the time slice indices. After completing the reading of all spatial raster indices under the current time slice index, the next time slice index is traversed, completing the joint traversal process combining the time dimension and the spatial dimension. When traversing each spatial raster index, the occupancy status record in the corresponding grid cell of the spatial raster index is read, and all initial flight intention hyperedge source identifiers in the occupancy status record are parsed. The initial flight intention hyperedge source identifier is a unique identifier that is synchronously established when the initial flight intention hyperedge is generated, including the association between the UAV identifier and the flight intention identifier. Grouping processing is performed according to the UAV identifier in the source identifier, and initial flight intention hyperedges from the same UAV are grouped into the same flight intention set, while initial flight intention hyperedges from different UAVs are divided into different flight intention sets.

[0087] The system counts the types of UAV identifiers corresponding to the current spatial grid index. If the number of UAV identifier types exceeds a preset number, it is determined that the current spatial grid index is shared by different UAVs, and the initial flight intention hyperedges corresponding to different UAVs are retained. If the UAV identifier type corresponds to only one UAV, the subsequent analysis of the current spatial grid index is terminated, and the system continues to traverse the next spatial grid index. Using the initial flight intention hyperedge identifier as the retrieval index, the system queries the corresponding inspection observation semantic tag in the flight intention association table and reads the target device ID, target component category, and observation azimuth angle. The system then establishes the association between the initial flight intention hyperedges corresponding to different UAVs under the current spatial grid index and the inspection observation semantic tags.

[0088] S3.2. Based on the target device ID and target component category in the inspection observation semantic label, establish the relationship between the observation target and the target device ID. Then, group the initial flight intention hyperedges corresponding to different UAVs with the same target device ID and target component category into a set of records that overlap with the same observation target.

[0089] Furthermore, based on the initial flight intention hyperedge and its corresponding inspection observation semantic label, the target device ID, target component category, and flight intention identifier in the inspection observation semantic label are parsed. The target device ID and target component category are both derived from the inspection observation semantic label, eliminating the need to parse the inspection task work order again. Subsequently, using the target device ID as the first association key, device classification processing is performed on all inspection observation semantic labels. Initial flight intention hyperedges with the same target device ID are grouped into the same device association set. Taking each device association set as the processing object, initial flight intention hyperedges with completely identical target component categories are grouped into the same component association subset. For initial flight intention hyperedges with the same target device ID but different target component categories, independent component association subsets are established respectively.

[0090] The system counts the types of UAV identifiers in the current component association subset. If there are two or more different UAV identifiers, the current component association subset is determined to meet the condition of the same observation target. All initial flight intention hyperedges in the component association subset are then aggregated to form a set of records overlapping with the same observation target. If all initial flight intention hyperedges in the current component association subset originate from the same UAV, the aggregation process of the component association subset is terminated. When generating the set of records overlapping with the same observation target, a set index is simultaneously established, and an association record is established for each initial flight intention hyperedge in the set, along with the corresponding flight intention identifier, UAV identifier, target device ID, target component category, spatial raster index, and time slice index.

[0091] S3.3 Based on the spatial raster index and time slice index in the overlapping record set of the same observation target, construct the observation ray relationship from the physical UAV to the target component, and calculate the degree of line-of-sight overlap according to the projection correlation of the observation rays corresponding to different physical UAVs on the surface of the target component.

[0092] Furthermore, the spatial location of the UAV within the corresponding time slice is queried based on the flight intent identifier. The UAV's spatial location is obtained through the center position of the spatial grid corresponding to the spatial grid index. The spatial model of the target device in the digital twin runtime instance is queried based on the target device ID, and the three-dimensional geometric region corresponding to the target component is read according to the target component category. The geometric center position of the target component is calculated. The observation ray relationship from the physical UAV to the target component is established, with the UAV's spatial location as the starting point and the target component's geometric center position as the ending point. When there are multiple different UAVs in the overlapping record set of the same observation target, an independent observation ray relationship is established for each UAV, and the association between the UAV identifier, target component identifier, and time slice index corresponding to each observation ray is saved.

[0093] A projection surface for the target component is established based on the three-dimensional geometric surface corresponding to the target component. Each observation ray is projected onto the surface of the target component along its ray direction to obtain the corresponding projection area. Projection correlation analysis is performed on the projection areas corresponding to multiple UAVs. The projection area overlap calculation method is adopted, that is, the common coverage area between the projection areas corresponding to each observation ray is calculated separately, and the line-of-sight overlap degree is calculated based on the positional correspondence between the common coverage area and each projection area. Then, the corresponding line-of-sight overlap degree is generated based on the correspondence of the projection area occupied by the common coverage area. If there is no common coverage area, it is determined that there is no line-of-sight overlap relationship between the corresponding observation rays.

[0094] Physical UAV spatial positioning calculation

[0095] ;

[0096] in, For the first A physical drone in time slice index The spatial coordinates below, For the first Spatial raster index corresponding to each physical drone For time slice index, This is a function for spatial coordinate transformation based on the three-dimensional spatiotemporal grid coordinate mapping relationship.

[0097] Construction of observation ray relationships:

[0098] ;

[0099] in, For the first The observation ray corresponding to each physical drone For the physical spatial coordinates of the UAV, The coordinates are the geometric center coordinates of the target component.

[0100] Observation ray direction vector:

[0101] ;

[0102] in, For the first The observation ray direction vector corresponding to each physical UAV.

[0103] The projection area of ​​the observation ray on the surface of the target component is calculated by projecting the observation ray onto the surface of the target component to obtain the corresponding projection area:

[0104] ;

[0105] in, For the first A physical drone observes the projection area of ​​rays on the surface of the target component. The target component surface area.

[0106] Calculation of line-of-sight overlap:

[0107] ;

[0108] in, For the first The physical drone and the first The degree of line-of-sight overlap between the observation rays corresponding to each physical drone. For the first Each physical drone corresponds to the area of ​​the projection of the observed ray on the surface of the target component. For the first Each physical drone corresponds to the area of ​​the projection of the observed ray on the surface of the target component. The first physical UAV to participate in the set of records overlapping with the observed target is designated as its number. The number of the second physical drone participating in the set of records overlapping with the observed target.

[0109] S3.4 Adjust the line-of-sight overlap judgment parameters based on the surface geometric features of the target component, determine the overlapping records of the same observation target that meet the preset judgment conditions as observation line-of-sight conflict records, and summarize the inspection observation semantic tags corresponding to the observation line-of-sight conflict records.

[0110] Furthermore, based on the set of records overlapping with the observed target, the target device ID, target component category, line-of-sight overlap degree, and three-dimensional geometric model of the target component are obtained for each record. The three-dimensional geometric model of the target component is obtained by querying the target device ID and target component category in the digital twin running instance without rebuilding the device model. The surface geometric features of the target component are extracted from the three-dimensional geometric model of the target component. By traversing the surface mesh data of the three-dimensional model of the target component, the relationship of the normal direction change of the target component surface and the continuous area of ​​the surface are statistically analyzed to form the surface geometric feature data of the target component. The surface geometric features of the target component include at least the distribution information of the continuous area of ​​the surface and the surface direction change information.

[0111] Target components are classified based on the distribution of continuous areas on their surfaces. When a target component's surface is a single continuous surface, the first type of line-of-sight overlap judgment parameter is used. When the target component's surface contains multiple directional change areas, the second type of line-of-sight overlap judgment parameter is used. The overlap judgment relationship between the corresponding observation ray projection areas is corrected based on the surface direction change information. When the degree of line-of-sight overlap meets the preset judgment condition corresponding to the current target component, the current overlap record with the observed target is marked as an observation line-of-sight conflict record. When the degree of line-of-sight overlap does not meet the corresponding judgment condition, the current overlap record with the observed target is retained but does not participate in subsequent conflict processing. The flight intention identifier in the observation line-of-sight conflict record is used as a retrieval index. The corresponding inspection observation semantic tag is read from the flight intention association relationship, and a summary set is established according to the relationship of the observation line-of-sight conflict records. The summary set stores the association relationship between the corresponding target device ID, target component category, observation angle azimuth, and flight intention identifier.

[0112] S4. Identify the conflict grid data based on the target equipment ID, target component category, and observation azimuth angle to obtain true conflict cluster information. Determine the observation priority and the estimated cost of inspection interruption failure based on the true conflict cluster information, identify the port operation equipment associated with the conflict grid, and construct a heterogeneous cost function.

[0113] S4.1 Based on the azimuth angle of the observation view in the semantic tags of different inspection observations, establish the spatial correspondence between the corresponding observation rays and the observation area on the surface of the target component. Determine the overlap of the observation range according to the coverage area relationship of the target component surface corresponding to different observation rays. Collect the observation line conflict records where the observation range has overlapping coverage to obtain the observation line overlap conflict record set.

[0114] Furthermore, the 3D model and surface mesh data of the target component are queried in the digital twin operation instance. The surface mesh data originates from the surface triangular mesh of the 3D model of the target component. By reading all surface mesh elements and their corresponding normal information, a surface region representation model of the target component is established. Using the azimuth angle of the observation viewpoint as the observation direction constraint, and combining the observation ray relationship from the physical UAV to the target component, surface mapping of the target component is performed for each observation ray. When the observation ray can form a visible spatial relationship with the surface mesh of the target component, the corresponding surface mesh is marked as the observation area of ​​the current observation ray. All marked surface meshes are combined to form the target component surface observation area corresponding to the observation ray, and a correspondence is established between the observation ray identifier and the observation area identifier, thereby establishing a spatial correspondence between the observation ray and the target component surface observation area.

[0115] Coverage area relationship analysis is performed on different observation areas corresponding to the same target component. The surface grid identifier sets corresponding to different observation areas are read, and intersection calculations are performed on two surface grid identifier sets. If the surface grid identifier sets corresponding to the two observation areas have a common surface grid, it is determined that the corresponding observation areas have overlapping coverage; if the two surface grid identifier sets do not have a common surface grid, it is determined that the corresponding observation areas do not have overlapping coverage. The above-mentioned overlapping coverage relationship is determined by the surface grid identifier overlap method. The target device ID, target component category, and overlapping coverage area identifier are used as the aggregation basis. Corresponding observation line-of-sight conflict records are merged into the same observation line-of-sight overlapping conflict record set, and a set index is established. The flight intention identifier, UAV identifier, observation area identifier, and observation angle azimuth angle corresponding to each observation line-of-sight conflict record in the set are associated to form an observation line-of-sight overlapping conflict record set.

[0116] S4.2. Based on the inspection observation semantic tags in the set of overlapping conflict records of observation lines, combined with the amount of effective data collected, the completeness of target component observation, and the data discrimination ability of observation angle azimuth, determine the observation priority parameters and associate them with the corresponding observation line conflict records to obtain the priority-labeled observation conflict record set.

[0117] Furthermore, using the set of overlapping and conflicting observation line records as the processing object, a statistical relationship of observation data is established for the inspection process corresponding to each inspection observation semantic tag within the set. The amount of valid data already collected is not re-collected, but rather statistically analyzed based on the continuously recorded inspection data cache in the digital twin operation instance. Using the flight intention identifier as an index, the data frames already formed in the current inspection process are quality-filtered, retaining only data frames that meet the requirements of image clarity, complete inclusion of the target component, and attitude stability. The number of data frames that meet the conditions is counted, forming the amount of valid data already collected for the corresponding inspection observation semantic tag. Using the three-dimensional surface mesh of the target component as the evaluation benchmark, the observation area corresponding to the valid data already collected is mapped to the surface mesh of the target component, and the set of surface meshes that have been covered is counted. Through the correspondence between the set of covered surface meshes and the set of all surface meshes of the target component, the observation completeness of the target component is formed. Whether the surface mesh of the target component has completed observation coverage is used as the evaluation object, so that the observation completeness can directly reflect the actual observation range of the target component.

[0118] An angular distribution relationship is established between the current observation azimuth and all existing observation azimuths of the same target component. A difference mapping relationship is then established based on the newly added coverage area on the target component surface corresponding to each observation azimuth. When the current observation azimuth covers a target component surface area not covered by existing observation azimuths, the corresponding newly added coverage area is included in the data discrimination capability; conversely, when the coverage area corresponding to the current observation azimuth is completely contained within the existing coverage area, its data discrimination capability is judged to be low. The entire process is evaluated through the newly added coverage relationship of the target component surface grid, establishing a multi-factor evaluation relationship based on the amount of effective data collected, the completeness of target component observation, and the data discrimination capability of the observation azimuth. Specifically, a hierarchical evaluation method can be adopted: firstly, a first-level ranking is performed based on the completeness of target component observation; for records with consistent observation completeness, a second-level ranking is performed based on data discrimination capability; if the same situation still exists, a third-level ranking is performed based on the amount of effective data collected. Observation priority parameters are generated based on the ranking results, and a correlation is established between the observation priority parameters and the corresponding observation line-of-sight conflict records, forming a priority-labeled observation conflict record set.

[0119] S4.3 Based on the observation continuity information in the priority-labeled observation conflict record set, calculate the impact of different inspection observation semantic tags on the data validity after the interruption of the observation task, determine the estimated cost of inspection interruption failure, and associate it with the corresponding observation line-of-sight conflict record to obtain the basic data of cost-related conflict clusters.

[0120] Furthermore, for each observation line-of-sight conflict record in the priority-marked observation conflict record set, a corresponding observation continuity description relationship is established. The observation continuity information is generated based on the flight intention execution process and consists of the observation states corresponding to the continuous time slice index. Using the time slice index as the sequence index, the continuous observation states corresponding to the same flight intention are connected to form continuous observation segments. Based on the continuous observation segments, the correspondence between the completed and incomplete observation parts of the current inspection task is determined. The current observation state is taken as the interruption position. The effective observation area formed before the interruption position is taken as the retained data, and the surface area of ​​the target component that has not been observed after the interruption position is taken as the missing data. Based on the target component surface grid, the correspondence between the coverage areas before and after the interruption is established, forming the data retention state and data missing state after the observation task is interrupted.

[0121] Data areas that can still form complete inspection interpretations after an interruption are designated as valid data areas, while data areas that cannot form complete observations due to interruption are designated as invalid data areas. A proportional relationship between valid and invalid data areas is established. Simultaneously, observation priority parameters are used to correct invalid data areas. When an invalid area corresponds to high-priority observation content, the corresponding impact level is increased; when an invalid area corresponds to low-priority observation content, the corresponding impact level is decreased, forming the impact level of inspection data validity. Using flight intention identifiers as association indexes, the estimated cost of inspection interruption failure is written into the corresponding observation line-of-sight conflict record. At the same time, the association relationship between inspection observation semantic tags, observation priority parameters, and estimated cost of inspection interruption failure is established, forming the basic data of cost association conflict clusters.

[0122] Calculation of the impact of data continuity:

[0123] ;

[0124] in, For the first Each inspection observation semantic label corresponds to the degree of impact on the data validity of the observation task. For the first Each inspection observation semantic tag corresponds to the amount of valid data collected before the observation task was interrupted. For the first Each inspection observation semantic label corresponds to the total amount of data required for the target component to complete the inspection. For the first Each inspection observation semantic label corresponds to an observation continuity weight. Semantic tags for inspection and observation.

[0125] Calculation of estimated cost of inspection interruption failure:

[0126] ;

[0127] in, For the first The estimated cost of inspection interruption failure corresponding to each inspection observation semantic label. For the first Each inspection observation semantic label corresponds to the degree of impact on the data validity of the observation task. For the first Each inspection observation semantic label corresponds to the parameters required for recovering continuous data sequences.

[0128] Correlation between estimated cost of inspection interruption failure and the following:

[0129] };

[0130] in, For the first Basic data of cost-related conflict clusters For the first One record of conflicting lines of sight. For the first The observation priority parameter corresponding to each observation line-of-sight conflict record. To record the conflict of sightlines.

[0131] Cost-related conflict cluster basic data formation:

[0132] ;

[0133] in, To associate conflict clusters with basic data at a cost, To observe the record of line-of-sight conflicts, This is the observation priority parameter.

[0134] S4.4. Based on the observation priority parameters and the estimated cost of inspection interruption failure in the basic data of cost-related conflict clusters within the same spatiotemporal grid cell, determine the representative record of conflict association, aggregate the basic data of the associated cost-related conflict clusters to form a true conflict cluster, and obtain the true conflict cluster information with the attribute of the representative record of conflict association.

[0135] Furthermore, a spatiotemporal grid positioning relationship is established based on the time slice index and spatial raster index stored in the cost-related conflict cluster basic data. All records with the same time slice index and corresponding spatial raster index are grouped into the same aggregation set, forming a conflict candidate record set corresponding to the same spatiotemporal grid unit. A two-parameter sorting relationship is established for each record in the conflict candidate record set. The first level of sorting is performed based on the observation priority parameter, placing records with higher observation priority parameters at the front. For records with the same observation priority parameter, a second level of sorting is performed based on the estimated cost of inspection interruption failure, placing records with higher estimated cost of inspection interruption failure at the front. Records are prioritized; if there are still cases where the sorting results are consistent, a fixed sorting order is established according to the flight intent identifier to ensure the uniqueness of the representative record selection process. The record that ranks first in the sorting results is identified as the conflict-related representative record. All records belonging to the same spatiotemporal grid unit and having a target device ID association with the conflict-related representative record are uniformly classified into the same true conflict cluster, and a unique identifier for the true conflict cluster is established. The observation priority parameters, inspection interruption failure cost estimates, flight intent identifiers, and inspection observation semantic tags in the conflict-related representative record are synchronously written into the true conflict cluster attribute set to form true conflict cluster information with the attributes of the conflict-related representative record.

[0136] S4.5. Based on the true conflict cluster information with attached conflict association representative record attributes, locate the corresponding spatiotemporal grid cell in the three-dimensional spatiotemporal grid coordinate mapping relationship, read the port operation equipment operation information associated with the spatiotemporal grid cell, and establish the conflict association relationship between the UAV and the port operation equipment according to the spatial occupancy relationship within the spatiotemporal grid cell to obtain the conflict-associated port operation equipment set.

[0137] Furthermore, based on the time slice index and spatial raster index in the true conflict cluster information, a corresponding spatiotemporal grid positioning relationship is established in the three-dimensional spatiotemporal grid coordinate mapping relationship to determine the target spatiotemporal grid cell where the true conflict cluster is located. The associated port operation equipment operation records are retrieved from the target spatiotemporal grid cell. These records originate from real-time status data continuously maintained by the digital twin operation instance, including equipment entity identification, equipment operating status, and current space occupancy information, and are already associated with the corresponding spatiotemporal grid cell. The spatial occupancy area corresponding to the port operation equipment is mapped to the current spatial raster index. Simultaneously, the spatial occupancy area corresponding to the UAV in the true conflict cluster is extracted, and a spatial positional correspondence between the two within the same spatiotemporal grid cell is established. When the UAV's spatial occupancy area and the port operation equipment's spatial occupancy area share a common spatial raster, or when the UAV's observation ray passes through the port operation equipment's spatial occupancy area, a conflict association relationship is determined between the UAV and the port operation equipment. Port operation equipment satisfying the conflict association relationship is uniformly grouped into a conflict-associated port operation equipment set, and an association relationship is established between the true conflict cluster identifier and the port operation equipment identifier, ensuring that each true conflict cluster corresponds to a conflict-associated port operation equipment set.

[0138] S4.6 Based on the set of conflict-related port operation equipment and corresponding spatiotemporal grid cells, calculate the operation connection relationship between port operation equipment, propagate the state changes of conflict-related port operation equipment along the operation connection relationship, establish the port operation impact propagation relationship, and form a representation of the port operation continuity impact.

[0139] Furthermore, based on the observation priority parameters, the estimated cost of inspection interruption failure, and the azimuth angle of the observation view in the true conflict cluster information with attached conflict association representative record attributes, the degree of change in observation continuity after the UAV inspection task is adjusted is calculated. Based on the degree of change in observation continuity, a characterization of the impact of UAV inspection continuity is established. Taking the conflict association representative record in the true conflict cluster as the analysis object, the description relationship of the inspection status before adjustment is established according to the observation priority parameters and the estimated cost of inspection interruption failure. The observation priority parameters are used to characterize the importance of the current observation task, and the estimated cost of inspection interruption failure is used to characterize the degree of data loss after the current inspection task is interrupted.

[0140] A set of permissible observation directions is established around the target component. The adjusted observation azimuth angle is mapped onto the surface of the target component. The surface coverage state of the target component is re-established based on the adjusted observation area and compared with the surface coverage state before adjustment. The correspondence between newly added covered areas, overlapping covered areas, and uncovered areas before and after adjustment is obtained. A correlation analysis relationship is established between the adjusted surface coverage state of the target component and the estimated cost of inspection interruption failure. When the adjusted observation area can maintain the original continuous coverage relationship, the original inspection continuity is maintained. When the adjustment causes the continuous coverage relationship to be interrupted, or an uncovered area is added, the degree of change in observation continuity after adjustment is calculated based on the grid range of the target component surface corresponding to the newly uncovered area, combined with the estimated cost of inspection interruption failure. The degree of change in observation continuity is associated with the true conflict cluster according to the flight intention identifier. A correlation relationship is established between the observation priority parameter, the estimated cost of inspection interruption failure, the degree of change in observation continuity, and the flight intention identifier, forming a characterization of the impact of UAV inspection continuity.

[0141] Calculation of the degree of continuous change observed:

[0142] ;

[0143] in, For the first Each inspection observation semantic label corresponds to the degree of change in the continuity of observation after the adjustment of the UAV inspection task. For the first The observation priority parameter corresponding to each inspection observation semantic label. For the first The estimated cost of inspection interruption failure corresponding to each inspection observation semantic label. For the first Each inspection observation semantic label corresponds to the continuous observation influence weight of the observation azimuth angle.

[0144] Calculation of the influence weight of the azimuth angle of the observation view:

[0145] ;

[0146] in, To adjust the previous number Each inspection observation semantic tag corresponds to the observation view azimuth angle. For the adjusted number Each inspection observation semantic tag corresponds to the observation view azimuth angle. This refers to the range of changes in the azimuth angle of the observation viewpoint.

[0147] Establishment of a characterization of the continuity impact of drone inspections:

[0148] ;

[0149] in, To characterize the impact of drone inspection continuity, For the first Each inspection observation semantic label corresponds to the degree of change in the continuity of the observation. The number of inspection observation semantic tags contained in the true conflict cluster information with attached conflict association representative record attributes.

[0150] S4.7 Based on the set of conflict-related port operation equipment and corresponding spatiotemporal grid cells, calculate the operation connection relationship between port operation equipment, propagate the state changes of conflict-related port operation equipment along the operation connection relationship, establish the port operation impact propagation relationship, and form a representation of the port operation continuity impact.

[0151] Furthermore, based on the port operation process configuration data maintained in the digital twin operation instance, the corresponding operation input objects and operation output objects for each port operation equipment are extracted. Operational connection relationships between equipment are established based on the correspondence between the operation output object and the next equipment's operation input object. The operation process configuration data can be derived from the equipment process flow definition imported from the port production scheduling system, including equipment identification, process number, and process dependencies. For the same cargo flow object, when the output object of one equipment matches the input object of another equipment, an operational connection relationship is established between the two equipment. If no correspondence exists, no operational connection relationship is established, and the current operating state of the conflicting port operation equipment is used as the initial state, propagating sequentially along the established operational connection relationships. During propagation, for each associated equipment, it is determined whether its operation input depends on the output result of the previous equipment. If a dependency exists, the state change of the previous equipment is synchronously mapped to the current equipment, and propagation continues to the next level of equipment. If the current equipment does not have an input dependency, the propagation branch is terminated.

[0152] All devices along the propagation path are arranged into a directed association sequence according to the propagation order, and the source of state change, propagation level, and propagation termination position of each device are recorded. When multiple propagation paths eventually converge on the same device, the propagation sources are merged and identified, so that the device retains all propagation source relationships, forming a complete port operation impact propagation relationship. The propagation path coverage area is used as the continuity evaluation object, and the set of devices affected by propagation is counted. Impact association relationships are established according to the propagation level, and the port operation impact propagation relationship is mapped to a port operation continuity impact representation. At the same time, the association relationship between the true conflict cluster identifier, the propagation path identifier, and the port operation continuity impact representation is established.

[0153] Establishing operational coordination relationships for port equipment:

[0154] ;

[0155] in: For the operational coordination of port equipment; For conflict-related port operation equipment sets, For the first A conflict-related port operations equipment This is a set of operational connections between port equipment.

[0156] The operational coordination relationship between the two port operation equipment is represented as follows:

[0157] ;

[0158] in, For the first The port operation equipment and the first The operational coordination relationships between various port operation equipment. This is the time slice index in the spatiotemporal grid cell corresponding to the u-th port operation equipment. For the first Each port operation equipment corresponds to a time slice index in the spatiotemporal grid cell. This is the spatial raster index corresponding to the u-th port operation equipment. For the first Spatial raster index corresponding to each port operation equipment.

[0159] The expression for the degree of state change is:

[0160] ;

[0161] in, For the first The degree of change in the status of port operation equipment after being affected by related factors. For the first Changes in the state of port operating equipment after a conflict. For the uth port operation equipment and the The operational coordination relationships between various port operation equipment. The number of port operation equipment in the conflict-related port operation equipment set.

[0162] Establishing the propagation relationship of port operations impacts:

[0163] ;

[0164] in, The impact of port operations on transmission relationships, For the first state change Port operation equipment For the affected Port operation equipment For the first The degree of change in the state of port operation equipment after being affected by the propagation.

[0165] Characterization of the impact of port operation continuity:

[0166] ;

[0167] in, This characterizes the impact on the continuity of port operations. For the first The degree of change in the status of port operation equipment. For the first The correlation weight of each port operation equipment in the port operation process. The number of port operation equipment in the conflict-related port operation equipment set. To affect the port operation equipment number on the initiating side, The number of port operation equipment that affects the receiving side.

[0168] S4.8 Based on the characterization of the continuity impact of UAV inspection and port operation, a cross-subject collaborative constraint relationship is constructed. This cross-subject collaborative constraint relationship is mapped to true conflict cluster information, forming a heterogeneous cost function jointly constrained by the continuity impact terms of UAV inspection and port operation.

[0169] Furthermore, using the continuous impact representations of UAV inspections and port operations as input data, continuous impact items for UAV inspections and port operations are constructed respectively. The continuous impact item for UAV inspections consists of flight intention identifiers, degree of change in observation continuity, observation priority parameters, and estimated cost of inspection interruption failure. The continuous impact item for port operations consists of port operation equipment identifiers, operation impact propagation relationships, propagation levels, and a description of port operation continuous impacts. Using true conflict cluster identifiers as association indexes, two types of impact items belonging to the same true conflict cluster are established to establish a correspondence. Based on the time slice index and spatial grid index corresponding to the same true conflict cluster, a cross-subject collaborative constraint relationship is established to determine whether the UAV activity area corresponding to the continuous impact item for UAV inspections and the equipment impact area corresponding to the continuous impact item for port operations belong to the same spatiotemporal grid unit, or whether their impact propagation ranges share a common spatiotemporal grid unit. If either condition is met, a corresponding collaborative constraint relationship is established, and the two types of impact items are connected by association edges to form a cross-subject collaborative constraint network.

[0170] Using the continuity impact of UAV inspections and the continuity impact of port operations as two types of constraint nodes, bidirectional constraint propagation is performed along the associated edges. When the state of any constraint node changes, the changed state is propagated to the associated nodes, and the corresponding constraint state is updated. This propagation is repeated until the states of all constraint nodes remain stable, forming the collaborative constraint results of the corresponding true conflict clusters. Based on the collaborative constraint results, a heterogeneous cost function is constructed, using the continuity impact of UAV inspections as the first cost term, the continuity impact of port operations as the second cost term, and the cross-entity collaborative constraint relationship as the coupling constraint term. The association calculation relationship between the two types of cost terms is established based on the coupling constraint term, generating the heterogeneous cost function.

[0171] Construction of cross-entity collaborative constraint relationships:

[0172] ;

[0173] in, This represents a cross-entity collaborative constraint relationship. To characterize the impact of drone inspection continuity, This characterizes the impact on the continuity of port operations. The relationship between the continuous impact characterization of UAV inspections and the continuous impact characterization of port operations is defined.

[0174] Cross-subject impact association mapping:

[0175] ;

[0176] in, For the cross-subject impact correlation corresponding to true conflict cluster information, This represents a cross-entity collaborative constraint relationship. For the true conflict cluster information that includes the associated conflict association representative record attributes, This represents the mapping relationship between cross-subject collaborative constraint relationships and true conflict cluster information.

[0177] The expression for the heterogeneous cost function is:

[0178] ;

[0179] For heterogeneous cost functions, To characterize the impact of drone inspection continuity, This characterizes the impact on the continuity of port operations. This establishes the cross-stakeholder collaborative constraint relationship between the characterization of the continuity impact of UAV inspections and the characterization of the continuity impact of port operations. The weights corresponding to the continuity impact items of drone inspections. The weights corresponding to the port operation continuity impact item. The weights correspond to the cross-entity collaborative constraint relationships.

[0180] S5. Generate a composite control package based on the calculation results of the heterogeneous cost function, collect the real-time feedback data generated by the UAV and port operation equipment after executing the composite control package, and update the running status and occupancy status in the three-dimensional spatiotemporal grid coordinate mapping relationship in the digital twin running instance.

[0181] S5.1 Based on the continuous impact terms of UAV inspection, port operation, and observation priority parameters in the heterogeneous cost function, the impact relationships of UAV path adjustment, observation angle adjustment, and port operation equipment coordination adjustment are calculated to obtain the heterogeneous cost function calculation results.

[0182] Furthermore, based on the initial flight intent hyperedge corresponding to the conflict, while keeping the starting waypoint and target waypoint unchanged, a detour search is performed on the conflict space grid to generate candidate adjustment paths. The candidate adjustment paths are then remapped to the three-dimensional spatiotemporal grid coordinate mapping relationship. The spatiotemporal occupancy sequence corresponding to the adjustment is calculated, and the degree of change in observation continuity after adjustment is calculated based on the UAV inspection continuity impact term, forming the impact relationship corresponding to the UAV adjustment path. Using the allowable observation area of ​​the target component as a constraint, the observation ray relationship is re-established based on the surface normal direction of the target component and the azimuth angle of the current observation viewpoint, and the impact relationship corresponding to the observation viewpoint adjustment method is formed based on the updated observation continuity change information. The impact relationship corresponding to the port operation equipment collaborative adjustment method is formed based on the changes in propagation level and propagation range. The impact relationships corresponding to the UAV adjustment path, the observation viewpoint adjustment method, and the port operation equipment collaborative adjustment method are respectively used as components in the heterogeneous cost function, and joint calculation is performed in combination with the observation priority parameter and the weight of the port operation equipment collaborative adjustment method impact relationship to obtain the heterogeneous cost function calculation result.

[0183] The expression for the heterogeneous cost function calculation result is:

[0184] ;

[0185] in, The result of the heterogeneous cost function calculation. For observation priority parameters, Adjusting the path of the drone to reflect the impact relationships. The influence relationship corresponding to the method of adjusting the observation perspective. The impact relationships corresponding to the coordinated adjustment methods of port operation equipment. The influence weight of the coordinated adjustment method of port operation equipment.

[0186] S5.2 Based on the results of the heterogeneous cost function calculation, the impact of UAV path adjustment, the impact of observation perspective adjustment, and the impact of port operation equipment collaborative adjustment, the corresponding adjustment strategy is determined according to the observation priority parameter, and the adjustment strategy is converted into UAV control commands and port equipment collaborative commands to generate a composite control package containing UAV control commands and port equipment collaborative commands.

[0187] Furthermore, adjustment strategies are selected based on observation priority parameters. When the observation priority parameter corresponds to a high-priority inspection task, observation continuity is prioritized, and the observation perspective adjustment method is selected. If the observation perspective adjustment cannot eliminate the conflict, the UAV path adjustment is then selected. If neither of the above two adjustment methods can meet the conflict resolution conditions, the port operation equipment collaborative adjustment method is introduced to participate in the joint adjustment. For ordinary priority inspection tasks, the adjustment method with the smaller total cost corresponding to the heterogeneous cost function calculation result is selected first, and combined to form the final adjustment strategy. The adjustment strategy is converted into equipment executable control commands. The UAV path adjustment is converted into waypoint update commands, flight speed adjustment commands, and flight altitude adjustment commands. The observation perspective adjustment method is converted into gimbal attitude control commands and camera observation direction control commands. The port operation equipment collaborative adjustment method is converted into equipment start / stop control commands, equipment waiting control commands, or equipment operation sequence adjustment commands. UAV control command sets and port equipment collaborative command sets are generated according to the control object, and are uniformly encapsulated according to the execution time sequence, true conflict cluster identifier, time slice index, and spatial grid index to form a composite control package.

[0188] S5.3. By using the UAV control commands and port equipment coordination commands in the composite control package, control the UAV and port operation equipment to perform corresponding adjustment operations, and simultaneously collect the pose status information, equipment operation status information and task execution status information of the UAV and port operation equipment after executing the composite control package.

[0189] Furthermore, the composite control package is parsed according to the controlled object. The UAV control commands are sent to the UAV flight control interface, and the port equipment coordination commands are sent to the corresponding port operation equipment control interface. Various adjustment operations are executed synchronously according to the execution order in the composite control package. The current position, flight attitude, and flight speed of the UAV are obtained through the navigation and positioning results output by the UAV flight controller. The pose status information is established according to the execution time of the control commands. The current operating status of the equipment is obtained through the port operation equipment control interface, including the equipment operating mode, equipment execution status, and equipment space occupancy status, forming equipment operating status information. Based on the inspection task execution record and the port operation task execution record, the current task execution progress and task completion status are determined, forming task execution status information.

[0190] S5.4. The pose status information, equipment operation status information and task execution status information of the UAV and port operation equipment after executing the composite control package are associated and processed according to the time slice index and spatial grid index to form real-time back transmission data.

[0191] Furthermore, based on the acquisition time corresponding to each status information, the corresponding time slice index is matched to establish a unified time reference for data acquired within the same time slice. The corresponding spatial grid index is redefined based on the spatial location in the pose status information and the spatial occupancy area in the device operation status information. A mapping relationship is established between the spatial location and the spatial grid index. Using the time slice index and the spatial grid index as a joint association key, pose status information, device operation status information, and task execution status information within the same spatiotemporal grid cell are combined to form a unified status record. When multiple status information corresponds to the same time slice index and spatial grid index, they are merged into a single status record; when they correspond to different time slices or different spatial grids, independent status records are established for each.

[0192] S5.5 Based on the pose status information, device operation status information and task execution status information in the real-time transmitted data, the corresponding device status, task status and operation parameters in the digital twin operation instance are synchronously updated to obtain the updated operation status in the digital twin operation instance.

[0193] Furthermore, based on the data type identifiers in the real-time transmitted data, the pose status information, equipment operation status information, and task execution status information are respectively divided into pose update datasets, equipment status update datasets, and task status update datasets. The pose status information originates from the UAV navigation and positioning results, the equipment operation status information originates from the control feedback of port operation equipment, and the task execution status information originates from the execution feedback of UAV inspection tasks and port operation tasks. Corresponding entity objects are retrieved in the digital twin operation instances, and update mapping relationships are established. For the pose update dataset, the corresponding UAV digital twin object is located based on the UAV identifier, and the spatial position, flight attitude, and motion status in the pose status information are overwritten and updated to the corresponding UAV object. For the equipment status update dataset, the corresponding equipment digital twin object is located based on the port operation equipment identifier, and the equipment operation status, equipment occupancy status, and equipment execution status are updated to the corresponding equipment object. For the task status update dataset, the corresponding inspection task object and port operation task object are located based on the task identifier, and the task execution progress, task completion status, and task interruption status are updated to the corresponding task object.

[0194] S5.6 Based on the running status in the updated digital twin running instance, the spatial location change information corresponding to the real-time transmitted data is associated with and updated with the time slice index and spatial grid index, and the occupancy status of the corresponding grid cell in the three-dimensional spatiotemporal grid coordinate mapping relationship is adjusted.

[0195] Furthermore, entity objects that have undergone state changes are extracted from the updated digital twin operational instance, and spatial location change information is read from the corresponding real-time transmitted data. The spatial location change information is obtained by establishing a correspondence between the spatial coordinates before and after the update, including changes in the spatial location of UAVs and changes in the space occupancy of port operation equipment. The updated spatial coordinates are remapped to the three-dimensional spatiotemporal grid coordinate mapping relationship. The new spatial grid index of the entity object is determined by referring to the spatial grid division rules, and the time slice index is re-determined according to the acquisition time corresponding to the real-time transmitted data. Thus, a correspondence between the updated time slice index and the spatial grid index is established. The grid cell corresponding to the original spatial grid index is used as the release object, and the occupancy record of the corresponding entity object is removed. Then, the grid cell corresponding to the updated spatial grid index is used as the target object, and the occupancy record of the corresponding entity object is written into the new grid cell. When an entity object moves from one spatial grid to another, a grid migration record corresponding to the movement trajectory is established simultaneously. When the operating status of port operation equipment changes but its spatial location remains unchanged, only the equipment status attributes in the corresponding grid cell are updated, while the spatial occupancy relationship remains unchanged. Using the time slice index as the order, multiple spatial grid indices corresponding to the same entity object within consecutive time slices are established into a continuous association relationship to form a new spatiotemporal occupancy sequence, and the occupancy status record in the three-dimensional spatiotemporal grid coordinate mapping relationship is updated synchronously.

[0196] In summary, this invention integrates port BIM model data, GIS data, and real-time IoT data streams to construct a digital twin operational instance containing a three-dimensional spatiotemporal grid coordinate mapping relationship. It also combines the target equipment ID and target component category parsing of inspection task work orders to generate initial flight intent hyperedges carrying inspection observation semantic tags, achieving dynamic mapping of UAV inspection tasks within the three-dimensional spatiotemporal grid coordinate mapping relationship. Based on the three-dimensional spatiotemporal grid coordinate mapping relationship, it performs spatiotemporal correlation analysis on the initial flight intent hyperedges of different UAVs, establishing observation ray relationships by combining the target equipment ID, target component category, and observation viewpoint azimuth angle, identifying real observation conflicts, and forming... True conflict cluster information; based on the true conflict cluster information, determine the observation priority parameters and the estimated cost of inspection interruption failure, associate the port operation equipment status corresponding to the conflict grid, and construct a heterogeneous cost function that integrates the continuous impact of UAV inspection and the continuous impact of port operations; generate a composite control package based on the calculation results of the heterogeneous cost function, coordinate the UAV inspection path, observation perspective and port operation equipment collaborative adjustment, and collect real-time feedback data after execution to update the operation status in the digital twin operation instance and the occupancy status in the three-dimensional spatiotemporal grid coordinate mapping relationship, so as to realize UAV inspection conflict detection, collaborative resolution and dynamic closed-loop update for the complex port operation environment.

[0197] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for conflict detection and resolution of low-altitude UAV inspections based on digital twins, characterized in that: include, Acquire port BIM model data, GIS data, and real-time IoT data streams to establish a digital twin operational instance containing three-dimensional spatiotemporal grid coordinate mapping relationships; Based on the inspection task work order, the target equipment ID and target component category are parsed, and an initial flight intention hyperedge carrying inspection observation semantic tags is generated. The initial flight intention hyperedge is then mapped to the three-dimensional spatiotemporal grid coordinate mapping relationship. Based on the three-dimensional spatiotemporal grid coordinate mapping relationship, the initial flight intention hyperedge of the UAV is compared in real time, and inspection observation semantic labels are obtained. The conflict grid data is identified based on the target equipment ID, target component category, and observation azimuth angle to obtain true conflict cluster information. Based on the true conflict cluster information, the observation priority and the estimated cost of inspection interruption failure are determined, and the port operation equipment associated with the conflict grid is identified to construct a heterogeneous cost function. Based on the calculation results of the heterogeneous cost function, a composite control package is generated. Real-time feedback data generated by the UAV and port operation equipment after executing the composite control package is collected, and the operation status and occupancy status in the three-dimensional spatiotemporal grid coordinate mapping relationship in the digital twin operation instance are updated.

2. The method for conflict detection and resolution of low-altitude UAV inspection based on digital twin as described in claim 1, characterized in that: The establishment of a digital twin runtime instance containing a three-dimensional spatiotemporal grid coordinate mapping relationship includes... By acquiring port BIM model data, GIS data, and IoT real-time data streams, and completing spatial benchmark unification and time synchronization processing respectively, spatial base plate data and real-time status data sources are formed. Based on the spatial base plate data, a three-dimensional spatial mesh is divided, and a structural skeleton of the three-dimensional spatiotemporal mesh coordinate mapping relationship composed of spatial raster index and time slice index is established in combination with the time dimension. The static information from the spatial base plate data and the dynamic information from the real-time status data source are written into the corresponding grid cells of the structural skeleton of the three-dimensional spatiotemporal grid coordinate mapping relationship to form the three-dimensional spatiotemporal grid coordinate mapping relationship. The three-dimensional spatiotemporal grid coordinate mapping relationship is then registered to the digital twin running instance.

3. The method for conflict detection and resolution of low-altitude UAV inspection based on digital twin as described in claim 2, characterized in that: The generation of the initial flight intent hyperedge carrying inspection observation semantic labels includes, Perform field reading processing on the inspection task work order, read the target equipment number field and the target component category field in the inspection task work order, and obtain the target equipment ID and the target component category; Based on the target device ID, query the spatial bounding box data of the target device in the digital twin operation instance, and calculate the observation azimuth angle based on the target component category. Combine the target device ID, target component category and observation azimuth angle to form the inspection observation semantic tag. Based on the start and end location information in the inspection task work order and the spatial bounding box data corresponding to the target equipment ID, an initial waypoint sequence is generated. The initial waypoint sequence is used as the geometric trajectory data of the initial flight intention hyperedge, and the inspection observation semantic label is associated with the initial flight intention hyperedge.

4. The method for conflict detection and resolution of low-altitude UAV inspection based on digital twin as described in claim 3, characterized in that: The mapping relationship between the initial flight intent hyperedge and the three-dimensional spatiotemporal grid coordinates includes, The time parameter discretization process is performed on the initial flight intention hyperedge carrying the inspection observation semantic label. The expected flight time of the initial flight intention hyperedge is divided into a continuous time slice index sequence according to the preset time slice interval. The continuous spatial coordinate points in the geometric trajectory data of the initial flight intention hyperedge are converted into the corresponding spatial raster index according to the preset spatial raster side length. The spatial raster indexes corresponding to the same timestamp are combined into a spatiotemporal occupancy tuple to obtain a spatiotemporal occupancy sequence consisting of multiple spatiotemporal occupancy tuples. The spatiotemporal occupancy sequence is used as the data source for writing the three-dimensional spatiotemporal grid coordinate mapping relationship. Write the inspection observation semantic tags in the spatiotemporal occupancy sequence into the grid cells of the corresponding time slice index and spatial raster index, and update the occupancy status record in the three-dimensional spatiotemporal grid coordinate mapping relationship.

5. The method for conflict detection and resolution of low-altitude UAV inspection based on digital twin as described in claim 4, characterized in that: The obtained inspection observation semantic tags include, The updated 3D spatiotemporal grid coordinate mapping relationship is jointly traversed by time slice index and spatial grid index. Based on the source identifier of the initial flight intention hyperedge in the occupancy record corresponding to the spatial grid index, the initial flight intention hyperedge corresponding to different UAVs is filtered, and the inspection observation semantic label associated with the initial flight intention hyperedge is read. Based on the target device ID and target component category in the inspection observation semantic label, establish the association relationship of the observation target, and collect the initial flight intention hyperedges corresponding to different UAVs with the same target device ID and target component category into a set of overlapping records of the same observation target; Based on the spatial raster index and time slice index in the record set of overlapping observation targets, the relationship between the observation rays from the physical UAV to the target component is constructed, and the degree of line-of-sight overlap is calculated according to the projection relationship of the observation rays corresponding to different physical UAVs on the surface of the target component. Based on the surface geometric features of the target component, the line-of-sight overlap judgment parameters are adjusted. The overlapping records of the same observation target that meet the preset judgment conditions are identified as observation line-of-sight conflict records, and the inspection observation semantic tags corresponding to the observation line-of-sight conflict records are summarized.

6. The method for conflict detection and resolution of low-altitude UAV inspection based on digital twin as described in claim 5, characterized in that: The true conflict cluster information includes, Based on the azimuth angle of the observation view in the semantic tags of different inspection observations, a spatial correspondence between the corresponding observation rays and the observation area on the surface of the target component is established. The overlap of the observation range is judged according to the coverage area relationship of the target component surface corresponding to different observation rays. The observation line conflict records with overlapping observation ranges are collected to obtain the observation line overlap conflict record set. Based on the inspection observation semantic tags in the set of overlapping conflict records of observation lines, combined with the amount of effective data collected, the completeness of target component observation, and the data discrimination ability of observation angle azimuth, the observation priority parameters are determined and associated with the corresponding observation line conflict records to obtain the set of priority-labeled observation conflict records. Based on the observation continuity information in the priority-labeled observation conflict record set, the impact of different inspection observation semantic tags on the data validity after the interruption of the observation task is calculated, the estimated cost of inspection interruption failure is determined, and it is associated with the corresponding observation line conflict record to obtain the basic data of cost-related conflict clusters. Based on the observation priority parameters and the estimated cost of inspection interruption failure in the basic data of cost-related conflict clusters within the same spatiotemporal grid cell, representative conflict records are determined, and the basic data of the associated cost-related conflict clusters are aggregated to form true conflict clusters, thus obtaining true conflict cluster information with the attribute of representative conflict records.

7. The method for conflict detection and resolution of low-altitude UAV inspection based on digital twin as described in claim 6, characterized in that: The construction of the heterogeneous cost function includes Based on the true conflict cluster information with attached conflict association representative record attributes, the corresponding spatiotemporal grid cell is located in the three-dimensional spatiotemporal grid coordinate mapping relationship, the port operation equipment operation information associated with the spatiotemporal grid cell is read, and the conflict association relationship between the UAV and the port operation equipment is established according to the spatial occupancy relationship within the spatiotemporal grid cell, thus obtaining the conflict-associated port operation equipment set. Based on the observation priority parameters, the estimated cost of inspection interruption failure, and the azimuth angle of the observation view in the true conflict cluster information with attached conflict association representative record attributes, the degree of change in observation continuity after the UAV inspection task is adjusted is calculated, and the impact characterization of UAV inspection continuity is established based on the degree of change in observation continuity. Based on the set of conflict-related port operation equipment and corresponding spatiotemporal grid units, the operation connection relationship between port operation equipment is calculated, the state changes of conflict-related port operation equipment are propagated along the operation connection relationship, the port operation impact propagation relationship is established, and the port operation continuity impact characterization is formed. Based on the characterization of the continuity impact of UAV inspections and port operations, a cross-subject collaborative constraint relationship is constructed. This cross-subject collaborative constraint relationship is then mapped to true conflict cluster information, forming a heterogeneous cost function jointly constrained by the continuity impact terms of UAV inspections and port operations.

8. The method for conflict detection and resolution of low-altitude UAV inspection based on digital twin as described in claim 7, characterized in that: The generated composite control package includes, Based on the continuous impact terms of UAV inspection, port operation, and observation priority parameters in the heterogeneous cost function, the impact relationships of UAV path adjustment, observation angle adjustment, and port operation equipment coordination adjustment are calculated to obtain the heterogeneous cost function calculation results. Based on the results of the heterogeneous cost function calculation, including the impact of UAV path adjustment, the impact of observation perspective adjustment, and the impact of port operation equipment collaborative adjustment, the corresponding adjustment strategy is determined according to the observation priority parameter. The adjustment strategy is then converted into UAV control commands and port equipment collaborative commands, generating a composite control package containing both UAV control commands and port equipment collaborative commands.

9. The method for conflict detection and resolution of low-altitude UAV inspection based on digital twin as described in claim 8, characterized in that: The real-time feedback data includes, By combining the UAV control commands and port equipment coordination commands in the composite control package, the UAV and port operation equipment are controlled to perform corresponding adjustment operations, and the pose status information, equipment operation status information and task execution status information of the UAV and port operation equipment after executing the composite control package are collected simultaneously. The pose status information, equipment operation status information, and task execution status information of the UAV and port operation equipment after executing the composite control package are associated with the time slice index and spatial grid index to form real-time back transmission data.

10. The method for conflict detection and resolution of low-altitude UAV inspection based on digital twin as described in claim 9, characterized in that: The updated running status of the digital twin instance and the occupancy status in the three-dimensional spatiotemporal grid coordinate mapping relationship include: Based on the pose status information, device operation status information and task execution status information in the real-time transmitted data, the corresponding device status, task status and operation parameters in the digital twin operation instance are synchronously updated to obtain the updated operation status in the digital twin operation instance. Based on the running status of the updated digital twin instance, the spatial location change information corresponding to the real-time transmitted data is associated with and updated with the time slice index and spatial raster index, and the occupancy status of the corresponding grid cell in the three-dimensional spatiotemporal grid coordinate mapping relationship is adjusted.

Images