Unmanned aerial vehicle inspection bridge intelligent monitoring and evaluation system and method

Abstract

The invention discloses an unmanned aerial vehicle inspection bridge intelligent monitoring and evaluation system and method. Comprising a multi-source sensor data acquisition module, a data preprocessing and synchronous calibration module, a bridge structure feature extraction and defect identification module, a defect parameter quantification and risk assessment module, an inspection path adaptive planning module, a real-time communication and data transmission module and a historical data fusion and trend prediction module. By integrating vision, infrared, laser and inertia measurement multi-source sensors, structure surface defects are captured, internal hidden defects can be recognized, data deviation is effectively reduced by combining data preprocessing and synchronous calibration technologies, and the problems that a traditional single sensor is insufficient in detection dimension and low in precision are solved; structural feature extraction, defect identification and type classification are realized by adopting a deep learning algorithm, and the defect severity is quantified through a standardized formula, so that the subjectivity, missing detection and false detection of manual interpretation are avoided, and the objectivity and reliability of an evaluation result are improved.

Description

Technical Field

[0001] This invention relates to the field of bridge structure monitoring technology, specifically to an intelligent monitoring and evaluation system and method for bridge inspection by unmanned aerial vehicles (UAVs). Background Technology

[0002] As a core component of transportation infrastructure, the structural health of bridges directly affects traffic safety and transportation efficiency. With the increase in service life, the accumulation of loads, and the impact of environmental erosion, bridges are prone to structural defects such as cracks, corrosion, spalling, and deformation. If these defects are not detected and assessed in a timely manner, they may lead to structural failure.

[0003] Current bridge inspection and monitoring technologies face numerous bottlenecks: Traditional manual inspections rely on workers climbing to heights or erecting scaffolding, which is not only labor-intensive and inefficient, but also poses extremely high safety risks when operating on long-span bridges or high-altitude structural parts; existing drone inspections mostly use only a single visual sensor, resulting in limited data dimensions and difficulty in comprehensively capturing internal structural defects. Furthermore, defect identification largely depends on manual interpretation, leading to strong subjectivity, high rates of missed and false detections, and a lack of standardized quantitative analysis methods; inspection paths are mostly pre-set fixed routes, unable to be dynamically adjusted according to the actual structural form of the bridge, defect distribution, and environmental changes, resulting in blind spots; during data acquisition, multi-source sensors exhibit temporal and spatial synchronization deviations, affecting the accuracy of defect location and parameter measurement; existing systems can only achieve single defect detection, lacking the ability to fuse and analyze historical data and predict structural performance degradation trends, thus failing to provide forward-looking decision support for preventive bridge maintenance. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent monitoring and evaluation system and method for bridge inspection by unmanned aerial vehicles (UAVs) to solve the problems existing in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent monitoring and evaluation system for bridge inspection by unmanned aerial vehicles (UAVs), comprising a multi-source sensor data acquisition module, a data preprocessing and synchronous calibration module, a bridge structural feature extraction and defect identification module, a defect parameter quantification and risk assessment module, an inspection path adaptive planning module, a real-time communication and data transmission module, and a historical data fusion and trend prediction module;

[0006] The multi-source sensor data acquisition module is used to collect surface and internal feature data of the bridge structure, including a visual acquisition unit, an infrared thermal imaging unit, a laser ranging unit, and an inertial measurement unit.

[0007] The data preprocessing and synchronization calibration module is used to perform noise reduction, registration and synchronization processing on the raw data collected by multi-source sensors, including a data noise reduction unit, a spatiotemporal synchronization calibration unit and a data format standardization unit.

[0008] The bridge structure feature extraction and defect identification module is used to extract key features of the bridge structure and identify various structural defects, including a structural feature extraction unit, a defect candidate region detection unit, and a defect type confirmation unit.

[0009] The defect parameter quantification and risk assessment module is used to calculate key defect parameters and assess the degree of impact on structural safety, including a defect parameter measurement unit, a defect severity assessment unit, and a structural risk level determination unit.

[0010] The inspection path adaptive planning module is used to dynamically plan the optimal inspection path, including a bridge structure modeling unit, an initial path generation unit, a path dynamic adjustment unit, and an obstacle avoidance path planning unit.

[0011] The real-time communication and data transmission module is used to realize data interaction and command transmission between the UAV and the ground control center, including a data uplink transmission unit, a command downlink transmission unit and a data buffer unit;

[0012] The historical data fusion and trend prediction module is used to integrate data from previous inspections and analyze the structural performance degradation trend. It includes a historical data association unit, a structural performance degradation modeling unit, and a maintenance suggestion generation unit.

[0013] Furthermore, the visual acquisition unit of the multi-source sensor data acquisition module consists of a high-definition RGB camera and a telephoto macro lens, the infrared thermal imaging unit adopts an uncooled infrared focal plane detector, the laser ranging unit is based on the principle of pulsed laser ranging, and the inertial measurement unit integrates a three-axis accelerometer, a three-axis gyroscope and a three-axis magnetometer.

[0014] Furthermore, the spatiotemporal synchronization calibration unit of the data preprocessing and synchronization calibration module uses the GPS timestamp of the UAV flight control system as a reference, and realizes multi-source data time synchronization through a timestamp alignment algorithm. Based on the UAV attitude data and sensor installation parameters, a coordinate transformation model is established to realize multi-source data spatial synchronization.

[0015] Furthermore, the structural feature extraction unit of the bridge structural feature extraction and defect identification module uses an improved U-Net semantic segmentation model to segment key structural components of the bridge, the defect candidate region detection unit uses the YOLOv8 target detection algorithm and temperature threshold segmentation algorithm to locate defect candidate regions, and the defect type confirmation unit uses the MobileNetV3 classification network to achieve accurate classification of defect types.

[0016] Furthermore, the defect severity assessment unit of the defect parameter quantification and risk assessment module calculates the defect severity coefficient S using a weighted summation formula, which is:

[0017]

[0018] Where ω1, ω2, ω3, and ω4 are weights and their sum is 1, d is the actual key parameter of the defect, d0 is the allowable limit parameter of the defect, s is the actual area of ​​the defect, s0 is the reference area of ​​the structural component, l is the actual length of the defect, l0 is the reference length of the structural component, and e is the environmental impact coefficient.

[0019] Furthermore, the initial path generation unit of the inspection path adaptive planning module uses an improved A algorithm to plan the initial inspection path, the path dynamic adjustment unit corrects the inspection path according to newly identified high-risk defect areas or structural morphology deviations, and the obstacle avoidance path planning unit uses the RRT algorithm to generate local obstacle avoidance paths.

[0020] Furthermore, the structural performance degradation modeling unit of the historical data fusion and trend prediction module uses an exponential degradation model to establish a structural performance prediction formula, which is:

[0021] P(t) = P0·e -k·t +ε

[0022] Where P(t) is the structural performance index at predicted time t, P0 is the initial service state performance index, k is the performance degradation coefficient, t is the service time, and ε is the error correction term.

[0023] Furthermore, the real-time communication and data transmission module adopts a dual-mode communication method of 5G and WiFi 6. The data uplink transmission unit adopts a block compression transmission strategy for large-capacity image data, and the data caching unit temporarily stores data when communication is interrupted and retransmits it after communication is restored.

[0024] Furthermore, the structural risk level determination unit of the defect parameter quantification and risk assessment module divides the risk into three levels: low, medium, and high, based on the defect severity coefficient S, and determines the overall structural risk level of the bridge and its individual components by combining the importance weights of the structural components.

[0025] An evaluation method for an intelligent monitoring and evaluation system for bridge inspection using unmanned aerial vehicles (UAVs) includes the following steps:

[0026] Step 1: System initialization and task configuration. The ground control center inputs inspection parameters, and the UAV completes module self-test and calibration.

[0027] Step 2: Multi-source data acquisition and preprocessing. The UAV flies along the initial path, and data is acquired by multiple sensors and processed by the preprocessing module.

[0028] Step 3: Structural feature extraction and defect identification, extracting bridge structural features and accurately identifying defect types;

[0029] Step 4: Defect quantification and risk assessment, calculate defect parameters and severity coefficients, and determine the risk level;

[0030] Step 5: Dynamically adjust and avoid obstacles in the inspection path, and dynamically optimize the inspection path based on the defect distribution and obstacles;

[0031] Step 6: Data transmission and historical data fusion: Transmit the data from this operation to the ground control center and associate it with historical data;

[0032] Step 7: Maintenance suggestion generation and system reset, generate maintenance decision report, and return the drone to home and reset.

[0033] Compared with the prior art, the beneficial effects of the present invention are:

[0034] This invention integrates multiple sensors, including vision, infrared, laser, and inertial measurement, to capture both surface defects and identify hidden internal defects. Combined with data preprocessing and synchronous calibration techniques, it effectively reduces data bias and solves the problems of insufficient detection dimensions and low accuracy associated with traditional single-sensor detection. Deep learning algorithms are used to extract structural features, identify defects, and classify types. Standardized formulas quantify defect severity, avoiding subjectivity and false negatives in manual interpretation, thus improving the objectivity and reliability of the assessment results. Based on the bridge structure model and real-time detection results, the inspection path is dynamically adjusted to balance comprehensive coverage with precise data collection in key areas. An obstacle avoidance algorithm avoids flying obstacles, solving the problems of blind spots and poor adaptability associated with fixed paths, while also avoiding the risks of manual high-altitude operations. Attached Figure Description

[0035] Figure 1 This is a system module diagram of the present invention;

[0036] Figure 2 This is a schematic diagram of the multi-source sensor data acquisition module of the present invention;

[0037] Figure 3 This is a schematic diagram of the data preprocessing and synchronous calibration module of the present invention;

[0038] Figure 4 This is a schematic diagram of the bridge structure feature extraction and defect identification module of the present invention;

[0039] Figure 5 This is a flowchart of the method of the present invention. Detailed Implementation

[0040] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0041] Please see Figure 1-5 This invention provides an intelligent monitoring and evaluation system for bridge inspection by unmanned aerial vehicles (UAVs), including a multi-source sensor data acquisition module, a data preprocessing and synchronous calibration module, a bridge structural feature extraction and defect identification module, a defect parameter quantification and risk assessment module, an inspection path adaptive planning module, a real-time communication and data transmission module, and a historical data fusion and trend prediction module.

[0042] The multi-source sensor data acquisition module is used to comprehensively collect surface and internal feature data of the bridge structure, providing multi-dimensional data sources for subsequent analysis. It includes a visual acquisition unit, an infrared thermal imaging unit, a laser ranging unit, and an inertial measurement unit.

[0043] The visual acquisition unit consists of a high-definition RGB camera and a telephoto macro lens, which are used for imaging the overall structure of the bridge and capturing details of local defects, respectively. The lens focal length can be automatically switched according to the inspection distance.

[0044] Infrared thermal imaging unit: Employs an uncooled infrared focal plane detector to detect defects invisible to the naked eye, such as internal voids in concrete bridges and hidden corrosion in steel structure welds, reflecting the internal state of the structure through differences in temperature field distribution;

[0045] Laser ranging unit: Based on the principle of pulsed laser ranging, it measures the distance between the UAV and the surface of the bridge structure in real time, providing data support for defect size quantification and UAV altitude-keeping flight;

[0046] Inertial Measurement Unit: Integrates a three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer to collect attitude angle, angular velocity, and acceleration data during the UAV's flight, which is used to correct attitude deviations in the sensor-collected data.

[0047] The data preprocessing and synchronization calibration module is used to perform noise reduction, registration, and synchronization processing on the raw data acquired by multiple sensors to ensure data consistency and reliability. It includes a data noise reduction unit, a spatiotemporal synchronization calibration unit, and a data format standardization unit.

[0048] Data noise reduction unit: median filtering is used to remove salt-and-pepper noise from RGB images, Gaussian filtering is used to smooth temperature field fluctuations from infrared thermal imaging data, moving average filtering is used to remove outliers from laser ranging data, and Kalman filtering is used to suppress random interference from inertial measurement data.

[0049] Spatiotemporal synchronization calibration unit: Based on the GPS timestamp of the UAV flight control system, the time synchronization of visual data, infrared data, laser data and inertial data is achieved through a timestamp alignment algorithm; based on the UAV attitude data and sensor installation parameters, a coordinate transformation model is established to uniformly map the data of each sensor to the bridge global coordinate system to achieve spatial synchronization;

[0050] Data format standardization unit: Converts preprocessed image data to JPEG2000 format, and distance and attitude data to JSON format, providing a unified interface for data calls in subsequent modules.

[0051] The bridge structure feature extraction and defect identification module is used to extract key features of the bridge structure from preprocessed data and accurately identify various structural defects. It includes a structural feature extraction unit, a defect candidate region detection unit, and a defect type confirmation unit.

[0052] Structural feature extraction unit: Using a deep learning-based semantic segmentation model, key structural components such as the main beam, piers, supports, guardrails, cables / suspenders, etc., of the bridge are segmented from RGB images to obtain the position, outline, and size range of each component;

[0053] Defect candidate region detection unit: For the segmented structural component images, the YOLOv8 target detection algorithm is used to initially locate defect candidate regions, such as cracks, corrosion, peeling, deformation, weld defects, etc.; For infrared thermal imaging data, temperature threshold segmentation and region growing algorithms are used to locate temperature abnormal regions, including suspected internal cavities and corrosion regions.

[0054] Defect type confirmation unit: Extracts multi-dimensional features such as texture features, shape features, and temperature features from candidate regions and inputs them into a lightweight classification network (MobileNetV3) to achieve accurate classification of defect types, such as vertical cracks in concrete, horizontal cracks, pitting corrosion in steel structures, surface corrosion, and weld cracks. At the same time, it removes false candidate regions, such as stains, shadows, and surface texture interference.

[0055] The defect parameter quantification and risk assessment module is used to accurately calculate key defect parameters and assess the impact of defects on bridge structural safety. It includes a defect parameter measurement unit, a defect severity assessment unit, and a structural risk level determination unit.

[0056] Defect parameter measurement unit: Based on the relationship between laser ranging data and image pixel calibration, calculates the geometric parameters of defects: for crack defects, measure length, width, and depth; for spalling / corrosion defects, measure area and perimeter; for deformation defects, measure displacement and deformation angle; for weld defects, measure defect length and width.

[0057] Defect Severity Assessment Unit: A defect severity assessment model is constructed, incorporating defect geometric parameters, structural component importance weights, and environmental factors affecting the defect's location. The defect severity coefficient S is calculated using a weighted summation formula. Wherein, ω1, ω2, ω3, and ω4 are the weights of defect characteristic parameters, defect area, defect length, and environmental impact, respectively, satisfying ω1+ω2+ω3+ω4=1; d is the actual key parameter of the defect, such as crack width and corrosion depth, and d0 is the limit allowable parameter of the corresponding defect, based on bridge design specifications; s is the actual area of ​​the defect, and s0 is the reference area of ​​the structural component where the defect is located; l is the actual length of the defect, and l0 is the reference length of the structural component where the defect is located; e is the environmental impact coefficient, which is 0.1 for dry environments, 0.3 for humid / corrosive environments, and 0.5 for extreme environments;

[0058] Structural risk level determination unit: The risk level is divided according to the severity coefficient S of the defect: S < 0.3 is low risk, 0.3 ≤ S < 0.6 is medium risk, and S ≥ 0.6 is high risk. The importance of the structural component where the defect is located is considered. For example, the main beam and the support are core components and have a weight of 1.2. The guardrail is a secondary component and has a weight of 0.8. Finally, the structural risk level of the bridge as a whole and each component is determined.

[0059] The adaptive inspection path planning module is used to dynamically plan the optimal inspection path based on bridge structural characteristics, defect distribution, and flight environment. It includes a bridge structure modeling unit, an initial path generation unit, a path dynamic adjustment unit, and an obstacle avoidance path planning unit.

[0060] Bridge structure modeling unit: Based on the bridge basic data collected in the early stage, including design drawings and historical inspection data, a simplified three-dimensional model of the bridge is constructed, and the location, size and key inspection areas of key structural components are marked.

[0061] Initial path generation unit: Using the improved A* algorithm, the initial inspection path that meets the flight constraints is planned in the 3D model, with the UAV take-off point as the starting point, the inspection points of each key component of the bridge as the waypoints, and the landing point as the end point.

[0062] Path dynamic adjustment unit: During the inspection process, if a new high-risk defect area is identified, the inspection path point for that area is automatically added, and the path is optimized to get closer to the defect area, thereby improving the accuracy of defect data collection; if a deviation between the bridge structure and the preset model is detected, the 3D model is corrected based on real-time laser ranging data, and the inspection path is adjusted synchronously.

[0063] Obstacle avoidance path planning unit: The laser ranging unit monitors obstacles in the flight path in real time, such as trees, cables and buildings. When the distance between the drone and the obstacle is less than the safety threshold, the fast exploration random tree (RRT*) algorithm is used to generate a local obstacle avoidance path. After avoiding the obstacle, the drone reconnects to the original planned path.

[0064] The real-time communication and data transmission module is used to realize data interaction and command transmission between the UAV and the ground control center, including a data uplink transmission unit, a command downlink transmission unit, and a data buffer unit.

[0065] Data uplink transmission unit: Adopts 5G and WiFi 6 dual-mode communication to transmit pre-processed sensor data, defect identification results, risk assessment results, and UAV flight status data to the ground control center in real time; For large-capacity image data, a block compression transmission strategy is adopted to ensure transmission efficiency;

[0066] Downlink transmission unit: Receives inspection mission commands (such as adjusting inspection area, modifying flight parameters, and emergency return) and parameter configuration commands (such as sensor acquisition frequency and defect identification threshold) sent by the ground control center, and transmits them to the UAV flight control system and various functional modules.

[0067] Data caching unit: When the communication signal is weak or interrupted, the collected raw data and processing results are temporarily stored in the UAV's local high-speed storage module. After the communication is restored, the data will be automatically retransmitted to the ground control center to avoid data loss.

[0068] The historical data fusion and trend prediction module integrates data from previous inspections, analyzes the trend of bridge structural performance degradation, and provides decision support for preventative maintenance. It includes a historical data association unit, a structural performance degradation modeling unit, and a maintenance suggestion generation unit.

[0069] Historical data association unit: Based on the unique identifier of the bridge structure, such as the bridge number and component number, the defect data and risk level of this inspection are associated and matched with historical inspection data to construct a full life cycle defect evolution dataset for each structural component.

[0070] Structural performance degradation modeling unit: An exponential degradation model is used to describe the change in bridge structural performance over time. Combining the severity coefficients of past defects with service time, a structural performance prediction formula is established.

[0071] P(t) = P0·e -k·t +ε

[0072] Wherein, P(t) is the structural performance index at the predicted time t, with a value range of 0-1, where 1 indicates structural integrity and 0 indicates structural failure; P0 is the initial service state performance index of the bridge, determined according to design standards; k is the structural performance attenuation coefficient, obtained by fitting historical defect evolution data, material properties, and load conditions; t is the service time of the bridge; and ε is the error correction term, obtained by fitting influencing factors such as environmental factors and maintenance measures.

[0073] Maintenance Recommendation Generation Unit: Based on the structural performance prediction results and risk level, combined with bridge maintenance specifications, it generates targeted maintenance recommendations, including maintenance priority (prioritizing high-risk components), maintenance methods (such as crack grouting, rust removal and coating, and structural reinforcement), recommended maintenance time windows, and outputs a structured maintenance decision report.

[0074] An evaluation method for an intelligent monitoring and evaluation system for bridge inspection using unmanned aerial vehicles (UAVs) includes the following steps:

[0075] Step 1: System Initialization and Task Configuration

[0076] The ground control center inputs the basic information of the bridge to be inspected (including bridge type, structural dimensions, and service life) and the inspection mission parameters (including inspection area, key monitoring components, and risk assessment standards). The UAV starts up and activates its various functional modules, completes sensor self-test, communication link establishment, and initial attitude calibration, and the multi-source sensor data acquisition module enters standby mode.

[0077] Step 2: Multi-source data acquisition and preprocessing

[0078] The UAV takes off according to the initial path generated by the inspection path adaptive planning module. The multi-source sensor data acquisition module simultaneously acquires RGB images, infrared thermal imaging data, laser ranging data, and inertial measurement data. The data preprocessing and synchronous calibration module performs noise reduction, spatiotemporal synchronization calibration, and format standardization on the raw data.

[0079] Step 3: Structural Feature Extraction and Defect Identification

[0080] The bridge structure feature extraction and defect identification module performs semantic segmentation of structural components on the preprocessed data, locates defect candidate regions, achieves accurate identification of defect types through multi-dimensional feature classification, and transmits the identification results to the defect parameter quantification and risk assessment module.

[0081] Step 4: Defect Quantification and Risk Assessment

[0082] The defect parameter quantification and risk assessment module calculates the defect geometric parameters based on the relationship between laser ranging data and image calibration, substitutes them into the defect severity assessment formula to obtain the S value, and determines the risk level in combination with the importance of structural components. The results are transmitted to the ground control center in real time.

[0083] Step 5: Dynamic adjustment and obstacle avoidance of the inspection path

[0084] The inspection path adaptive planning module dynamically adjusts the inspection path and generates an obstacle avoidance path based on the defect identification results (such as the discovery of high-risk defects) and the obstacles detected by the laser ranging unit, ensuring full inspection coverage and flight safety.

[0085] Step 6: Data Transmission and Integration with Historical Data

[0086] The real-time communication and data transmission module transmits all data from this inspection to the ground control center, while the historical data fusion and trend prediction module associates this data with historical data, substitutes it into the structural performance degradation prediction formula, and obtains the structural performance evolution trend.

[0087] Step 7: Maintenance suggestion generation and system reset

[0088] The historical data fusion and trend prediction module generates a maintenance decision report based on the risk level and performance prediction results; after the UAV completes the inspection mission, it returns to land and all modules are reset to standby status, waiting for the next inspection mission.

[0089] Example:

[0090] Implementation process:

[0091] Step 1: The ground control center inputs the basic information of the concrete beam bridge, including a span of 30m, a beam height of 1.8m, a service life of 10 years, and sets the inspection focus to the beam body, supports, and piers. The risk assessment standard adopts the "Technical Condition Assessment Standard for Highway Bridges". After the UAV is started, each sensor completes self-check, the inertial measurement unit calibrates the initial attitude of the UAV, and the communication module establishes a 5G connection with the ground control center.

[0092] Step 2: The inspection path adaptive planning module constructs a simplified 3D model based on the bridge design drawings and uses an improved A* algorithm to generate the initial path, specifically: takeoff point → left side of the beam → right side of the beam → pier → support → landing point; the UAV takes off according to the path, and multiple sensors collect data synchronously: the RGB camera captures images of the beam surface, automatically switches to a telephoto lens to capture suspected crack areas, the infrared thermal imager collects temperature field data of the beam, the laser rangefinder measures the distance between the UAV and the beam in real time, and the inertial measurement unit records the flight attitude.

[0093] Step 3: The data preprocessing and synchronization calibration module performs median filtering on the RGB image, Gaussian filtering on the infrared data, aligns all data using GPS timestamps, maps sensor data to the bridge's global coordinate system based on the attitude data of the inertial measurement unit, and converts it into a standardized format.

[0094] Step 4: The bridge structure feature extraction and defect identification module uses an improved U-Net model to segment the beam, pier, and bearing components. The YOLOv8 algorithm locates three candidate crack regions in the beam image. The MobileNetV3 network, combined with texture features, identifies two transverse cracks and one vertical crack. Infrared thermal imaging data is used for threshold segmentation to locate one temperature anomaly region, which is suspected to be a void inside the concrete.

[0095] Step 5: The defect parameter quantification and risk assessment module, based on the laser ranging data and image calibration relationship, calculates the transverse crack length as 2.3m and width as 0.3mm, the vertical crack length as 1.8m and width as 0.2mm, and the area of ​​the temperature anomaly region as 0.5m². 2 Substituting into the defect severity assessment formula, with weights ω1, ω2, ω3, and ω4, and an environmental impact coefficient e = 0.2 (for a humid environment), the calculated S values ​​are 0.28, 0.25, and 0.32, respectively. Considering the beam as the core component with a weight of 1.2, the beam is determined to be of medium risk, while the piers and supports are of low risk.

[0096] Step 6: The inspection path adaptive planning module detected tree obstacles near the temperature anomaly area of ​​the beam, at a distance of 3m. The RRT* algorithm was used to generate an obstacle avoidance path with an offset of 0.8m. After the UAV adjusted its flight attitude to avoid the obstacles, it approached the temperature anomaly area to collect additional data.

[0097] Step 7: The real-time communication and data transmission module transmits defect images, parameters, and risk levels to the ground control center in real time. Ground staff can view the inspection results through the control software. The historical data fusion and trend prediction module correlates the current defect data with the data from the previous two inspections, substitutes them into the structural performance degradation prediction formula, and obtains a degradation coefficient k = 0.02. It predicts that the beam performance index will be 0.85 after 5 years and generates maintenance recommendations: conduct non-destructive testing on areas with abnormal temperatures, grout cracks, and prioritize maintenance as medium.

[0098] Step 8: After the inspection mission is completed, the UAV returns and lands along the planned path, all modules are reset to standby status, and the inspection data is automatically archived to the ground control center database.

[0099] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A smart monitoring and evaluation system for bridge inspection using unmanned aerial vehicles (UAVs), characterized in that: It includes a multi-source sensor data acquisition module, a data preprocessing and synchronous calibration module, a bridge structure feature extraction and defect identification module, a defect parameter quantification and risk assessment module, an inspection path adaptive planning module, a real-time communication and data transmission module, and a historical data fusion and trend prediction module. The multi-source sensor data acquisition module is used to collect surface and internal feature data of the bridge structure, including a visual acquisition unit, an infrared thermal imaging unit, a laser ranging unit, and an inertial measurement unit. The data preprocessing and synchronization calibration module is used to perform noise reduction, registration and synchronization processing on the raw data collected by multi-source sensors, including a data noise reduction unit, a spatiotemporal synchronization calibration unit and a data format standardization unit. The bridge structure feature extraction and defect identification module is used to extract key features of the bridge structure and identify various structural defects, including a structural feature extraction unit, a defect candidate region detection unit, and a defect type confirmation unit. The defect parameter quantification and risk assessment module is used to calculate key defect parameters and assess the degree of impact on structural safety, including a defect parameter measurement unit, a defect severity assessment unit, and a structural risk level determination unit. The inspection path adaptive planning module is used to dynamically plan the optimal inspection path, including a bridge structure modeling unit, an initial path generation unit, a path dynamic adjustment unit, and an obstacle avoidance path planning unit. The real-time communication and data transmission module is used to realize data interaction and command transmission between the UAV and the ground control center, including a data uplink transmission unit, a command downlink transmission unit and a data buffer unit; The historical data fusion and trend prediction module is used to integrate data from previous inspections and analyze the structural performance degradation trend. It includes a historical data association unit, a structural performance degradation modeling unit, and a maintenance suggestion generation unit.

2. The intelligent monitoring and evaluation system for bridge inspection by unmanned aerial vehicles (UAVs) according to claim 1, characterized in that: The visual acquisition unit of the multi-source sensor data acquisition module consists of a high-definition RGB camera and a telephoto macro lens. The infrared thermal imaging unit uses an uncooled infrared focal plane detector. The laser ranging unit is based on the principle of pulsed laser ranging. The inertial measurement unit integrates a three-axis accelerometer, a three-axis gyroscope and a three-axis magnetometer.

3. The intelligent monitoring and evaluation system for bridge inspection by unmanned aerial vehicles (UAVs) according to claim 1, characterized in that: The spatiotemporal synchronization calibration unit of the data preprocessing and synchronization calibration module uses the GPS timestamp of the UAV flight control system as a reference, and realizes multi-source data time synchronization through a timestamp alignment algorithm. Based on the UAV attitude data and sensor installation parameters, a coordinate transformation model is established to realize multi-source data spatial synchronization.

4. The intelligent monitoring and evaluation system for bridge inspection by unmanned aerial vehicles (UAVs) according to claim 1, characterized in that: The structural feature extraction unit of the bridge structural feature extraction and defect identification module uses an improved U-Net semantic segmentation model to segment key structural components of the bridge. The defect candidate region detection unit uses the YOLOv8 target detection algorithm and temperature threshold segmentation algorithm to locate defect candidate regions. The defect type confirmation unit uses the MobileNetV3 classification network to achieve accurate classification of defect types.

5. The intelligent monitoring and evaluation system for bridge inspection by unmanned aerial vehicles (UAVs) according to claim 1, characterized in that: The defect severity assessment unit of the defect parameter quantification and risk assessment module calculates the defect severity coefficient S using a weighted summation formula, which is: Where ω1, ω2, ω3, and ω4 are weights and their sum is 1, d is the actual key parameter of the defect, d0 is the allowable limit parameter of the defect, s is the actual area of ​​the defect, s0 is the reference area of ​​the structural component, l is the actual length of the defect, l0 is the reference length of the structural component, and e is the environmental impact coefficient.

6. The intelligent monitoring and evaluation system for bridge inspection by unmanned aerial vehicles (UAVs) according to claim 1, characterized in that: The initial path generation unit of the inspection path adaptive planning module uses an improved A algorithm to plan the initial inspection path, the path dynamic adjustment unit corrects the inspection path according to newly identified high-risk defect areas or structural morphology deviations, and the obstacle avoidance path planning unit uses the RRT algorithm to generate local obstacle avoidance paths.

7. The intelligent monitoring and evaluation system for bridge inspection by unmanned aerial vehicles according to claim 1, characterized in that: The structural performance degradation modeling unit of the historical data fusion and trend prediction module uses an exponential degradation model to establish a structural performance prediction formula, which is: P(t)=P0·e -k·t +ε Where P(t) is the structural performance index at predicted time t, P0 is the initial service state performance index, k is the performance degradation coefficient, t is the service time, and ε is the error correction term.

8. The intelligent monitoring and evaluation system for bridge inspection by unmanned aerial vehicles (UAVs) according to claim 1, characterized in that: The real-time communication and data transmission module adopts a dual-mode communication method of 5G and WiFi 6. The data uplink transmission unit uses a block compression transmission strategy for large-capacity image data. The data caching unit temporarily stores data when communication is interrupted and retransmits it after communication is restored.

9. The intelligent monitoring and evaluation system for bridge inspection by unmanned aerial vehicles (UAVs) according to claim 1, characterized in that: The structural risk level determination unit of the defect parameter quantification and risk assessment module divides the risk into three levels: low, medium, and high, based on the defect severity coefficient S, and determines the overall structural risk level of the bridge and its individual components by combining the importance weights of structural components.

10. An evaluation method for an intelligent monitoring and evaluation system for bridge inspection by unmanned aerial vehicles (UAVs) according to any one of claims 1-9, characterized in that: Includes the following steps: Step 1: System initialization and task configuration. The ground control center inputs inspection parameters, and the UAV completes module self-test and calibration. Step 2: Multi-source data acquisition and preprocessing. The UAV flies along the initial path, and data is acquired by multiple sensors and processed by the preprocessing module. Step 3: Structural feature extraction and defect identification, extracting bridge structural features and accurately identifying defect types; Step 4: Defect quantification and risk assessment, calculate defect parameters and severity coefficients, and determine the risk level; Step 5: Dynamically adjust and avoid obstacles in the inspection path, and dynamically optimize the inspection path based on the defect distribution and obstacles; Step 6: Data transmission and historical data fusion: Transmit the data from this operation to the ground control center and associate it with historical data; Step 7: Maintenance suggestion generation and system reset, generate maintenance decision report, and return the drone to home and reset.

Images