A structure defect intelligent detection method based on BIM and unmanned aerial vehicle cooperation

Abstract

This invention provides an intelligent structural defect detection method based on BIM and UAV collaboration, belonging to the field of intelligent building inspection technology. The method includes: acquiring and analyzing the BIM model of the structure to be inspected to obtain BIM data; dividing the BIM model into a 3D mesh and determining the component inspection priority; marking key inspection areas and setting safe flight parameters for the UAV; calculating an initial inspection path; optimizing the initial inspection path using a multi-objective optimization algorithm to generate a 3D optimized path; dynamically adjusting the 3D optimized flight path to obtain the optimal UAV inspection path; acquiring inspection data; associating the inspection data with the BIM model; and generating a structured inspection report. This invention solves the problems of insufficient deep integration between traditional inspection methods and BIM models, disconnect between inspection planning and building information, difficulty in using inspection results to update the model, low efficiency, poor accuracy, and high risk.

Description

Technical Field

[0001] This invention belongs to the field of building intelligent inspection technology, and in particular relates to an intelligent structural defect detection method based on BIM and UAV collaboration. Background Technology

[0002] my country has made remarkable achievements in building and bridge construction, with complex structures such as super high-rise buildings and long-span bridges constantly pushing the technological limits. However, structural inspection still faces severe challenges during the operation and maintenance phase: traditional manual inspection is not only inefficient and costly, but also poses significant safety hazards due to high-altitude operations; while the introduction of drone technology has partially solved the accessibility problem, the fixed-route inspection mode cannot adapt to complex building forms, resulting in frequent blind spots; and manual flight inspection relying on pilot experience suffers from unstable data quality and poor repeatability, making it difficult to meet the standardized requirements of engineering inspection.

[0003] Meanwhile, with the large-scale promotion and application of BIM technology in my country's engineering construction field, digital management of the entire building life cycle has become an industry trend. However, the deep integration of existing inspection methods with BIM models is still insufficient, and the problems of inspection planning being disconnected from building information and inspection results being difficult to feed back into model updates are becoming increasingly prominent. Summary of the Invention

[0004] To address the aforementioned shortcomings in existing technologies, this invention provides an intelligent structural defect detection method based on BIM and UAV collaboration, which solves the problems of insufficient deep integration between traditional detection methods and BIM models, disconnect between detection planning and building information, difficulty in using detection results to update models, low efficiency, poor accuracy, and high risk.

[0005] To achieve the above objectives, the technical solution adopted by this invention is: an intelligent structural defect detection method based on BIM and UAV collaboration, comprising the following steps:

[0006] S1. Obtain the BIM model of the structure to be inspected and parse it to obtain BIM data. Divide the BIM model into three-dimensional meshes to obtain three-dimensional spatial data. Based on the component type, determine the component inspection priority, mark the key inspection areas, and set the safe flight parameters of the UAV.

[0007] S2. Based on the 3D spatial data of the BIM model, key inspection areas, and component inspection priorities, the initial inspection path is calculated using the improved A* algorithm.

[0008] S3. The initial detection path is optimized by a multi-objective optimization algorithm. Combined with the safe flight parameters of the UAV, a three-dimensional optimized path is generated. The three-dimensional optimized flight path is dynamically adjusted by combining BIM data to obtain the optimal detection path of the UAV.

[0009] S4. Based on the optimal inspection path, obtain inspection data, associate the inspection data with the BIM model, mark the defect locations, obtain the inspected BIM model, and generate a structured inspection report.

[0010] The beneficial effects of this invention are as follows: This invention extracts BIM data by parsing the BIM model, automatically marks the detection priority and generates the initial flight path, and uses a multi-objective optimization algorithm to optimize the flight path, thereby realizing the automation and intelligence of the detection process. It also dynamically adjusts the flight parameters according to the real-time environment and automatically associates the detection data with the BIM model, which significantly improves the efficiency and quality of structural inspection and significantly reduces the risk of high-altitude operations.

[0011] It achieves full and deep integration of structural defect detection methods with BIM models, close connection between detection planning and building information, and feedback of detection results to update the model.

[0012] Further, S1 includes the following steps:

[0013] S101. Obtain the BIM model of the structure to be inspected, parse the BIM model, and obtain BIM data containing geometric data, component attribute information, and historical inspection records.

[0014] S102. Divide the BIM model into three-dimensional meshes, establish a detection space coordinate system, and obtain three-dimensional space data by generating a three-dimensional space map;

[0015] S103. Based on component type, set the priority of load-bearing components and connection nodes to the highest level to determine component inspection priority;

[0016] S104. Based on historical defect data in historical inspection records, mark key inspection areas in the BIM model and set safe flight parameters for UAVs.

[0017] Furthermore, the calculation expression for the safe distance parameter of the UAV is as follows:

[0018] ;

[0019] in, This indicates the minimum distance between the drone and the structure's surface. This indicates the vertical distance from the drone to the sampling point. Indicates the pixel size of the camera sensor. Indicates the camera's focal length. Indicates the camera's field of view. This represents the material correction factor.

[0020] The beneficial effects of the above-mentioned further solutions are as follows: the present invention effectively avoids the risk of collision in complex structural environments through BIM-driven intelligent path planning, and improves the accuracy and reliability of path planning data; and by setting safe flight parameters for UAVs, it improves the safety of intelligent detection of structural defects, and the accurate obstacle avoidance algorithm in the safe flight parameters reduces the risk of collision.

[0021] Furthermore, the cost function of the improved A* algorithm is as follows:

[0022] ;

[0023] ;

[0024] ;

[0025] ;

[0026] in, Represents a node n Overall priority Indicates the distance from the starting point to the node. n The actual cost, Represents a node n Heuristic cost estimation to the destination Indicates the priority item for BIM components. Represents a node n Priority of BIM component inspection within radius R This indicates the shooting quality constraints. This represents the drone camera orientation vector. Represents the surface normal vector of the component. Indicates the angle of incidence during shooting. Indicates the minimum angle of incidence for shooting. Indicates the angle penalty coefficient. All of these represent weighting coefficients.

[0027] The beneficial effects of the above-mentioned further solutions are as follows: By using the improved A* algorithm for initial path planning and a detection strategy based on component priority, the present invention improves the defect detection rate of key parts and reduces invalid flight paths.

[0028] By adding a term that includes BIM component priority, image quality constraints, and flight status to the cost function of the A* algorithm, the detection efficiency and the generation of effective detection points were improved.

[0029] Furthermore, step S3 includes the following steps:

[0030] S301. The initial detection path is optimized by a multi-objective optimization algorithm. The distance and path of each waypoint in the initial detection path to the structure surface are optimized to generate a three-dimensional optimized path.

[0031] S302. Based on 3D optimized path, use UAVs for flight detection;

[0032] The methods of drone flight detection mentioned include: single-drone flight detection and multi-drone collaborative detection;

[0033] S303 responds to the UAV's flight process by using the UAV's onboard sensors to monitor environmental changes in real time and dynamically adjusts the 3D optimized path in combination with BIM data to obtain the optimal detection path for the UAV.

[0034] Furthermore, S301 specifically includes:

[0035] A multi-objective optimization model is established by using a multi-objective optimization algorithm to minimize flight time and maximize detection coverage. The distance and path of each detection point on the structure surface in the initial detection path are optimized. Based on the shooting quality requirements and combined with the safe flight parameters of the UAV, the optimal detection scheme is solved by a genetic algorithm to generate a three-dimensional optimized path.

[0036] The beneficial effects of the above-mentioned further solutions are as follows: This invention optimizes the detection path through a multi-objective optimization algorithm, generates a three-dimensional optimized path that takes into account both detection efficiency and safety, and improves the stability and reliability of detection data quality through a dynamic adjustment mechanism. It also includes automatic compliance checks to avoid illegal flights, and outputs the optimal detection route of the UAV that complies with safety regulations and enables the UAV to accurately cover all surfaces to be inspected, while ensuring that the shooting angle and distance meet the requirements for defect identification.

[0037] Furthermore, step S4 includes the following steps:

[0038] S401. Based on the optimal detection route, use a drone to collect data and obtain detection data including image and point cloud data.

[0039] S402. Associate the inspection data with the components in the BIM model to obtain the associated BIM model;

[0040] S403. Based on the associated BIM model, use image recognition algorithms to mark the defect locations, obtain the inspected BIM model, and generate a structured inspection report.

[0041] The beneficial effects of the above-mentioned further solutions are as follows: By automatically associating the detection data with the BIM model, the present invention saves data processing time, compares and analyzes historical data, improves the accuracy of defect prediction, achieves accurate spatial matching between the detection data and the BIM model, provides a reliable basis for subsequent structural health assessment, and is compatible with various building structure types, including bridges, high-rise buildings and industrial facilities, and is compatible with mainstream BIM software platforms, thus reducing implementation costs. Attached Figure Description

[0042] Figure 1 This is a flowchart of the method of the present invention.

[0043] Figure 2 This is a schematic diagram illustrating the process of associating detection data with the BIM model in this embodiment.

[0044] Figure 3 This is a system architecture diagram for this embodiment.

[0045] Figure 4 This is a flowchart of the route planning process in this embodiment.

[0046] Figure 5 This is a schematic diagram of the route planning for a typical arch bridge in this embodiment. Detailed Implementation

[0047] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0048] Before describing this embodiment, the following terms will be explained:

[0049] BIM model: Building Information Model;

[0050] A* algorithm: A* algorithm;

[0051] Pareto optimal solution;

[0052] YOLOv5 model: A real-time object detection model based on deep learning;

[0053] SIFT algorithm: Scale-invariant feature transformation algorithm;

[0054] ID: Identification number;

[0055] JSON format: A lightweight data exchange format;

[0056] Octree algorithm: an octree algorithm;

[0057] NSGA-II Algorithm: Non-dominated sorting genetic algorithm II;

[0058] DJI Waypoint 2.0 Protocol: DJI drone waypoint mission control protocol;

[0059] RTK positioning: Real-time dynamic positioning;

[0060] MAVLink protocol: A lightweight, open-source communication protocol for drones.

[0061] Example 1

[0062] In this embodiment, the problems of insufficient deep integration between traditional inspection methods and BIM models, disconnect between inspection planning and building information, difficulty in using inspection results to update models, low efficiency, poor accuracy, and high risk are addressed by innovatively combining BIM model data with UAV intelligent inspection technology. This fully leverages the technical advantages of BIM models, such as high geometric accuracy and comprehensive information dimensions, to achieve intelligent planning and dynamic optimization of inspection routes, effectively solving the industry pain points of low efficiency, poor accuracy, and high risk associated with traditional inspection methods.

[0063] like Figure 1 As shown, this invention provides a method for intelligent detection of structural defects based on BIM and UAV collaboration, the implementation of which is as follows:

[0064] S1. Obtain and parse the BIM model of the structure to be inspected to obtain BIM data. Divide the BIM model into a 3D mesh to obtain 3D spatial data. Based on the component type, determine the component inspection priority, mark key inspection areas, and set the safe flight parameters for the UAV. The specific steps are as follows:

[0065] S101. Obtain the BIM model of the structure to be inspected, parse the BIM model, and obtain BIM data containing geometric data, component attribute information, and historical inspection records.

[0066] S102. Divide the BIM model into three-dimensional meshes, establish a detection space coordinate system, and obtain three-dimensional space data by generating a three-dimensional space map;

[0067] S103. Based on the component type, set the priority of load-bearing components and connection nodes to the highest level to determine the component inspection priority.

[0068] In this embodiment, BIM data preprocessing is performed to obtain the BIM model of the structure to be inspected. In order to extract geometric data and semantic information from the model, the geometric data, component attribute information and historical inspection records in the model are parsed to obtain BIM data; the component attribute information includes the three-dimensional coordinates and dimension information of the component.

[0069] The BIM model is divided into three-dimensional meshes, a detection spatial coordinate system is established, a three-dimensional spatial map that can be used for path planning is generated, and three-dimensional spatial data is obtained.

[0070] The inspection priority is determined by marking the component type, with critical load-bearing components having a higher priority than general components, and load-bearing components and connection nodes having the highest priority.

[0071] S104. Based on historical defect data in historical inspection records, mark key inspection areas in the BIM model and set safe flight parameters for UAVs.

[0072] In this embodiment, a detection requirement analysis is performed, and detection parameters are determined based on the detection target, including shooting resolution, coverage density, and detection angle. Combined with historical defect data in historical detection records, key detection areas are marked in the BIM model, and safe flight parameters for the UAV are set, including flight altitude, shooting angle, and safe distance.

[0073] The calculation expression for the safe distance parameter of the UAV is as follows:

[0074] ;

[0075] in, This indicates the minimum distance between the drone and the structure's surface. This indicates the vertical distance from the drone to the sampling point. Indicates the pixel size of the camera sensor. Indicates the camera's focal length. Indicates the camera's field of view. Indicates the material correction factor. Partly based on the base resolution, Part of it is the field distortion factor. Part of it is a material correction factor.

[0076] S2. Based on the 3D spatial data of the BIM model, key inspection areas, and component inspection priorities, the initial inspection path is calculated using the improved A* algorithm.

[0077] In this embodiment, intelligent detection route planning is performed. Based on the three-dimensional spatial data of the BIM model, key detection areas, and component detection priorities, the improved A* algorithm is used to optimize the path and generate an initial detection path.

[0078] The core of the improved A* algorithm lies in the cost function, the specific expression of which is as follows:

[0079] ;

[0080] ;

[0081] ;

[0082] ;

[0083] in, Represents a node n Overall priority Indicates the distance from the starting point to the node. n The actual cost, Represents a node n Heuristic cost estimation to the destination Indicates the priority item for BIM components. Represents a node n Priority of BIM component inspection within radius R This indicates the shooting quality constraints. This represents the drone camera orientation vector. Represents the surface normal vector of the component. Indicates the angle of incidence during shooting. Indicates the minimum angle of incidence for shooting. Indicates the angle penalty coefficient. All represent weighting coefficients, achieved by adding a flight status item. and This improves both detection efficiency and the number of effective detection points.

[0084] When selecting the next node to traverse, choose the node with the highest overall priority (i.e., the smallest value).

[0085] S3. The initial detection path is optimized using a multi-objective optimization algorithm. Combined with the UAV's safe flight parameters, a 3D optimized path is generated. The 3D optimized flight path is then dynamically adjusted using BIM data to obtain the optimal detection path for the UAV. The specific steps are as follows:

[0086] S301. The initial detection path is optimized by a multi-objective optimization algorithm. The distance and path of each waypoint in the initial detection path to the structure surface are optimized to generate a three-dimensional optimized path.

[0087] In this embodiment, a multi-objective optimization model is established by using a multi-objective optimization algorithm, which includes minimizing flight time and maximizing detection coverage. The distance and path of each detection point on the structure surface in the initial detection path are optimized. Based on the shooting quality requirements and the flight safety constraints of the UAV's safe flight parameters, the optimal detection scheme is solved using a genetic algorithm, and a three-dimensional optimized path that takes into account both detection efficiency and safety is automatically generated.

[0088] The expression for the multi-objective optimization model is as follows:

[0089] ;

[0090] in, Indicates the optimization objective. The objective function represents minimizing the flight time. This represents the objective function that maximizes detection coverage. This indicates safe flight parameters for drones, including the distance between the drone and the structure surface, flight speed, and camera shooting angle.

[0091] In this embodiment, the genetic algorithm used for multi-objective optimization is the NSGA-II algorithm, and the specific steps are as follows:

[0092] D1: Initialize the population;

[0093] Based on the feasible space defined by the BIM model, N initial feasible routes are randomly generated as individuals (i.e. chromosomes). Each route consists of a series of path point codes. The path points are used as genes to obtain an initialized population.

[0094] D2: Quick Nondominated Sort;

[0095] Calculate the three objective function values ​​for each individual in the population, and sort all individuals in the population into a stratified population.

[0096] D3: Calculate congestion level;

[0097] Based on the stratified population, within each non-dominated layer, calculate the crowding degree of each individual in the target space;

[0098] The crowding level reflects the density of other individuals around an individual and is used to maintain population diversity;

[0099] D4: Selection, crossover, and mutation;

[0100] The binary tournament selection method is adopted to prioritize individuals with higher levels in the stratified population to generate the parent population. A simulated binary crossover operator is used to randomly swap some path points of two parent individuals with a certain probability to generate offspring. A polynomial mutation operator is used to randomly change the position of a certain path point in the offspring individuals with a small probability to introduce new genetic characteristics and avoid premature convergence.

[0101] D5: Elite Retention Strategy;

[0102] The parent and offspring populations are merged to form a mixed population of size 2N. The mixed population is then re-sorted using fast non-dominated sorting and crowding calculation to obtain the sorted mixed population and its crowding.

[0103] Based on the sorted mixed population and the crowding degree of the mixed population, select the top N best individuals to form a new parent population, and retain the excellent individuals;

[0104] D6: Terminate inspection;

[0105] Repeat steps D2 to D5. In response to reaching the preset maximum number of generations or the convergence of the solution set, output the Pareto optimal solution set of the final population, and select the final route according to actual preferences to obtain the three-dimensional optimized path.

[0106] S302. Based on 3D optimized path, use UAVs for flight detection;

[0107] The methods of drone flight detection mentioned include: single-drone flight detection and multi-drone collaborative detection;

[0108] S303 responds to the UAV's flight process by using the UAV's onboard sensors to monitor environmental changes in real time and dynamically adjusts the 3D optimized path in combination with BIM data to obtain the optimal detection path for the UAV.

[0109] In this embodiment, a UAV is used for flight detection based on a 3D optimized path. During the flight, the UAV’s onboard sensors monitor environmental changes in real time. When an unmodeled obstacle is detected, the flight path is dynamically adjusted to achieve dynamic adjustment of the 3D optimized path. The camera shooting parameters are also automatically adjusted according to the lighting conditions to ensure the quality of the collected data and obtain the optimal detection path of the UAV.

[0110] The methods for drone flight detection include: single-drone flight detection and multi-drone collaborative detection.

[0111] S4. Based on the optimal inspection path, acquire inspection data, associate the inspection data with the BIM model, mark the defect locations, obtain the inspected BIM model, and generate a structured inspection report. The specific steps are as follows:

[0112] S401. Based on the optimal detection route, use a drone to collect data and obtain detection data including image and point cloud data.

[0113] S402. Associate the inspection data with the components in the BIM model to obtain the associated BIM model;

[0114] S403. Based on the associated BIM model, use image recognition algorithms to mark the defect locations, obtain the inspected BIM model, and generate a structured inspection report.

[0115] In this embodiment, as Figure 2 As shown, in order to associate the inspection data with the BIM model, UAVs are used to collect data according to the optimal inspection route to obtain inspection data containing images and point cloud data. The inspection data is then associated with the components in the BIM model to obtain the associated BIM model. Based on the image recognition algorithm, the defect location is automatically marked. A data update mechanism is used to update the associated BIM model and generate a structured inspection report to support subsequent maintenance decisions.

[0116] This embodiment uses a pre-trained YOLOv5 model to identify defects, employs the SIFT algorithm for feature extraction to ensure a matching error of less than 0.5 pixels, and transforms image feature points to the BIM coordinate system through spatial transformation during the BIM association process, then matches the specific component ID through the nearest neighbor algorithm.

[0117] Practical applications show that linking the detection data with the BIM model improves the crack location accuracy from ±15cm to ±2cm.

[0118] The data update mechanism automatically generates BIM custom attributes, recording information such as defect type, size, and inspection date. It also supports outputting inspection results in JSON format, including fields such as defect number, type, size, and most recent inspection date. The above data is updated to the BIM model in real time, providing support for subsequent maintenance decisions.

[0119] Example 2

[0120] In this embodiment, as Figure 3 As shown, a structural defect intelligent detection system based on BIM and UAV collaboration is provided, including:

[0121] The BIM data processing module is used to parse and extract model information, including: geometric data extraction and semantic analysis, generating BIM data, generating a 3D mesh, and obtaining 3D spatial data;

[0122] The BIM data processing module obtains model data, including geometric dimensions and component attributes, from the BIM software via an API interface. This module uses the Octree algorithm to mesh the model in 3D space, with a default mesh precision of 0.05m × 0.05m × 0.05m. In practice, the intelligent structural defect detection system can automatically identify specific component types; for example, when extracting the coordinates of steel beam welds, it will identify the "Steel_Beam_Weld" family type in the BIM.

[0123] In this embodiment, the intelligent structural defect detection system based on BIM and UAV collaboration also includes: a flight path planning module, used to generate and optimize detection flight paths, including: a path generation algorithm, acquiring a defect knowledge base, generating an initial path, using a multi-objective optimizer for multi-objective optimization, and performing compliance checks to determine whether it meets flight safety constraints.

[0124] The algorithm generates the basic path, employing the NSGA-II algorithm for multi-objective optimization. The optimization objective function is set to minimize flight time and maximize detection coverage, while ensuring a safe distance of no less than 0.3m. The final output flight path file adopts the DJI Waypoint 2.0 protocol format. The real-time control system achieves dynamic obstacle avoidance through RTK positioning and LiDAR point cloud, with RTK positioning accuracy reaching ±1cm and LiDAR point cloud refresh rate of 10Hz. During flight, the system automatically adjusts the camera ISO (sensitivity) according to the light intensity, with an adjustment range of 100-1600.

[0125] The flight control module is used to execute and adjust flight missions, including: sensor data fusion, dynamic obstacle avoidance, and dynamic adjustment of the flight path when unmodeled obstacles are detected based on real-time environmental changes, thereby adjusting flight parameters and storing them in the flight log library.

[0126] In this embodiment, the intelligent structural defect detection system based on BIM and UAV collaboration also includes a data analysis module for processing detection data and updating the BIM model. This module employs a data update mechanism to automatically generate custom BIM attributes, recording information such as defect type, size, and detection date. The system supports outputting detection results in JSON format, including fields such as defect number, type, size, and most recent detection date. This data is updated to the BIM model in real time, providing support for subsequent maintenance decisions.

[0127] In this embodiment, as Figure 4 As shown, the specific implementation steps for route generation are as follows: First, import the model. The system supports input of BIM models in various formats and automatically performs model preprocessing, including repairing broken surfaces and removing decorative components.

[0128] During the inspection requirements configuration phase, users can set the key inspection components and shooting parameters through the configuration file. For example, they can set "steel beam welds" and "concrete cracks" as priority inspection objects and specify a resolution of 20mm / pixel and a shooting angle of 45°.

[0129] The route optimization process involves establishing and solving an optimization model;

[0130] Through actual experiments, the intelligent structural defect detection system in this embodiment can significantly reduce the detection time of a 200m² wall from 120 minutes to 45 minutes. During the flight execution phase, the drone is controlled via the MAVLink protocol. When encountering sudden obstacles, the system will perform online replanning to ensure flight safety.

[0131] In this embodiment, as Figure 5 As shown, a typical arch bridge route planning is performed, and the optimal detection path is automatically generated based on the curve characteristics of the arch bridge. The main arch rib uses a B-spline curve to accurately fit the centerline route, achieving full coverage of the entire surface of the structure without dead angles.

[0132] Compared to traditional grid-based flight paths, this invention has three significant advantages: First, it effectively avoids collision risks in complex structural environments through BIM-driven intelligent path planning; second, it automatically optimizes shooting angles and distances based on component geometric features to ensure the quality of inspection data; and third, it achieves precise spatial matching between inspection data and the BIM model, providing a reliable basis for subsequent structural health assessments.

Claims

1. A method for intelligent detection of structural defects based on BIM and UAV collaboration, characterized in that, Includes the following steps: S1. Obtain the BIM model of the structure to be inspected and parse it to obtain BIM data. Divide the BIM model into three-dimensional meshes to obtain three-dimensional spatial data. Based on the component type, determine the component inspection priority, mark the key inspection areas, and set the safe flight parameters of the UAV. The calculation expression for the UAV safe distance parameter in the UAV safe flight parameters is as follows: in, This indicates the minimum distance between the drone and the structure's surface. This indicates the vertical distance from the drone to the sampling point. Indicates the pixel size of the camera sensor. Indicates the camera's focal length. Indicates the camera's field of view. Indicates the material correction factor; S2. Based on the 3D spatial data of the BIM model, key inspection areas, and component inspection priorities, the initial inspection path is calculated using the improved A* algorithm. The cost function of the improved A* algorithm is as follows: in, Represents a node n Overall priority Indicates the distance from the starting point to the node. n The actual cost, Represents a node n Heuristic cost estimation to the destination Indicates the priority item for BIM components. Represents a node n Priority of BIM component inspection within radius R This indicates the shooting quality constraints. This represents the drone camera orientation vector. Represents the surface normal vector of the component. Indicates the angle of incidence during shooting. Indicates the minimum angle of incidence for shooting. Indicates the angle penalty coefficient. All represent weighting coefficients; S3. The initial detection path is optimized using a multi-objective optimization algorithm. Combined with UAV safe flight parameters, a 3D optimized path is generated. This 3D optimized path is then dynamically adjusted using BIM data to obtain the optimal UAV detection path. Specifically: A multi-objective optimization model is established by using a multi-objective optimization algorithm to minimize flight time and maximize detection coverage. The distance and path of each detection point on the structure surface in the initial detection path are optimized. Based on the shooting quality requirements and combined with the safe flight parameters of the UAV, the optimal detection scheme is solved by using a genetic algorithm to generate a three-dimensional optimized path. S4. Based on the optimal inspection path, obtain inspection data, associate the inspection data with the BIM model, mark the defect locations, obtain the inspected BIM model, and generate a structured inspection report.

2. The intelligent structural defect detection method based on BIM and UAV collaboration according to claim 1, characterized in that, S1 includes the following steps: S101. Obtain the BIM model of the structure to be inspected, parse the BIM model, and obtain BIM data containing geometric data, component attribute information, and historical inspection records. S102. Divide the BIM model into three-dimensional meshes, establish a detection space coordinate system, and obtain three-dimensional space data by generating a three-dimensional space map; S103. Based on component type, set the priority of load-bearing components and connection nodes to the highest level to determine component inspection priority; S104. Based on historical defect data in historical inspection records, mark key inspection areas in the BIM model and set safe flight parameters for UAVs.

3. The intelligent structural defect detection method based on BIM and UAV collaboration according to claim 1, characterized in that, S3 includes the following steps: S301. The initial detection path is optimized by a multi-objective optimization algorithm. The distance and path of each waypoint in the initial detection path to the structure surface are optimized to generate a three-dimensional optimized path. S302. Based on 3D optimized path, use UAVs for flight detection; The methods for drone flight detection include: single-drone flight detection and multi-drone collaborative detection; S303 responds to the UAV's flight process by using the UAV's onboard sensors to monitor environmental changes in real time and dynamically adjusts the 3D optimized path in combination with BIM data to obtain the optimal detection path for the UAV.

4. The intelligent structural defect detection method based on BIM and UAV collaboration according to claim 1, characterized in that, S4 includes the following steps: S401. Based on the optimal detection path, use a drone to collect data and obtain detection data containing image and point cloud data. S402. Associate the inspection data with the components in the BIM model to obtain the associated BIM model; S403. Based on the associated BIM model, use image recognition algorithms to mark the defect locations, obtain the inspected BIM model, and generate a structured inspection report.

Images