Bridge pier measuring method and system with laser gyroscope measuring rod and unmanned aerial vehicle vision fused

Abstract

The invention discloses a laser gyroscope measuring rod and unmanned aerial vehicle vision fused bridge pier measuring method and system, and the method comprises the steps that an unmanned aerial vehicle carries a laser gyroscope measuring rod integrated with a gyroscope, a tilt angle sensor and a wireless communication module, and flies to a bridge pier measuring region; acquiring three-dimensional attitude data and inertial reference data of the unmanned aerial vehicle in real time through a laser gyroscope measuring rod; the unmanned aerial vehicle synchronously starts a binocular vision module and a laser range finder, performs multi-view scanning on the pier along a preset surrounding path, and obtains pier surface feature point cloud and real-time distance data; and performing space-time alignment on the inertial reference data, the visual point cloud and the real-time distance data. According to the invention, through laser gyroscope-visual point cloud space-time fusion, GAN enhanced rain and fog resistance verification and dynamic wind resistance closed-loop control, the problem of serious distortion of measurement data caused by perception degradation under complex meteorological conditions is solved, so that the real geometric features of the bridge pier can be accurately extracted.

Description

Technical Field

[0001] The present application relates to the technical field of bridge pier measurement, and in particular to a bridge pier measurement method and system that integrates a laser gyro measuring rod with unmanned aerial vehicle vision. Background Art

[0002] In the field of bridge structure inspection, UAV pier measurement technology has long faced the problem of data reliability under dynamic environmental interference, while the existing technology has proposed a dynamic perception method that combines UAV vision with vision.

[0003] For example, the Chinese patent with publication number CN116182837A proposes a positioning and mapping method based on tight coupling of visual laser radar and inertia, so as to obtain high-precision positioning trajectory and geometric structure information of the surrounding environment, and maintain good robustness and accuracy in typical laser radar degradation scenarios. This invention, based on the positioning and mapping method of tight coupling of visual laser radar and inertia, first processes the data of three sensors: autonomous positioning and mapping, laser radar autonomous positioning and mapping, and visual and inertial measurement unit IMU; secondly, constructs visual reprojection, IMU pre-integration and radar geometric constraints, and combines the three constraints to establish a nonlinear optimization problem; finally, uses the optimized posture and the current frame radar point cloud to update the voxel map, providing geometric structure information for subsequent processes. This invention is mainly used in the design and manufacturing of drones.

[0004] However, when the above technology is applied to bridge pier measurement, strong wind disturbances and the UAV's own motion noise will be coupled and superimposed, resulting in serious distortion of the measurement data, making it difficult to accurately extract the true geometric features of the bridge piers. This core problem further leads to the problem of perception degradation under complex meteorological conditions - rain, fog, low light and other environments cause visual features to become blurred, and traditional detection methods are forced to rely on a single sensor, and the measurement robustness is significantly reduced. Existing technologies mostly alleviate interference by restricting flight conditions or increasing hardware redundancy, but they have defects such as low efficiency and limited adaptability. Summary of the invention

[0005] In order to solve the above problems, an embodiment of the present invention provides a bridge pier measurement method integrating a laser gyro measuring rod with a drone vision, the method comprising:

[0006] The UAV is equipped with a laser gyro measuring rod that integrates a gyroscope, an inclination sensor, and a wireless communication module. It flies to the bridge pier measurement area and collects the UAV's three-dimensional attitude data and inertial reference data in real time through the laser gyro measuring rod.

[0007] The drone simultaneously starts the binocular vision module and laser rangefinder, performs multi-view scanning of the bridge pier along the preset orbital path, and obtains the characteristic point cloud and real-time distance data of the bridge pier surface;

[0008] Temporally and spatially align inertial reference data, visual point cloud, and real-time distance data;

[0009] The anti-interference extended Kalman filter algorithm is used to fuse multi-source data and separate the actual size data of the bridge piers from the UAV motion noise, generating the three-dimensional actual size data of the bridge piers and the size deviation heat map;

[0010] According to the confidence level of the size deviation heat map, the drone's local refined scanning path is dynamically planned, and obstacles are avoided in real time through the laser rangefinder until the measurement accuracy meets the standard.

[0011] Furthermore, the binocular vision module calls the image enhancement model based on the generative adversarial network in a rainy and foggy environment to dehaze and restore the texture of low-visibility images.

[0012] Furthermore, the spatiotemporal alignment method includes:

[0013] Compensate for the UAV's own posture jitter based on the high-frequency attitude data of the gyroscope;

[0014] Dynamically bind the camera coordinate system of the binocular vision module, the laser ranging coordinate system and the inertial reference coordinate system of the laser gyro measuring rod;

[0015] During the hovering phase of the UAV, the static data of the laser gyro measuring rod is used to perform online zero bias calibration on the binocular vision module and laser rangefinder.

[0016] Furthermore, the dynamic binding method includes:

[0017] Based on the inertial data of the laser gyro measurement rod, the pose estimation results of visual SLAM are forced to be corrected;

[0018] The indicators of visual SLAM include the number of feature point tracking, reprojection error, and the trace of the pose estimation covariance matrix. If any of the indicators exceeds the limit, the visual SLAM is judged to be invalid and a forced correction is triggered;

[0019] When visual SLAM fails, the gyroscope's posture is used as a reference, and the output of visual SLAM is forcibly corrected through the extended Kalman filter. During forced correction, the observation value is switched to the inertial posture, and the observation noise covariance is increased to reflect the drift characteristics of the inertial data. At the same time, the filter is updated. The filter update equation includes: ;

[0020] In the formula, is the posterior state estimate; is the prior state estimate; is the Kalman gain; is the observation vector; is the observation matrix.

[0021] Furthermore, the generation of the three-dimensional actual size data of the bridge pier and the size deviation heat map includes:

[0022] The drone flies along the axial direction of the pier at a constant distance, scanning the surface at a fixed frequency;

[0023] IMU data is aligned with laser point cloud timestamps, and visual images are used to assist in point cloud coloring;

[0024] Initialize the state covariance matrix of the filter and set the prior distribution of motion noise; take the actual size parameters of the bridge pier and the UAV motion noise as the joint state quantity;

[0025] Preprocess the scanned data, introduce an adaptive noise covariance matrix, and dynamically reduce the weight of abnormal jump points;

[0026] Iteratively execute the prediction-update step, decouple the static size data of the bridge pier from the dynamic motion noise of the drone through the observation model, and output the denoised three-dimensional actual size data of the bridge pier;

[0027] Based on the filtered point cloud data, the non-uniform B-spline surface fitting algorithm is used to generate the 3D model of the bridge pier, the reconstructed model is aligned with the design drawing coordinate system, and the ICP algorithm is used to optimize the registration accuracy;

[0028] Calculate the projection deviation in the normal direction of each point on the surface and map it to the HSV color space;

[0029] Outputs an interactive heatmap of size deviation.

[0030] Furthermore, the dynamic planning path includes an initialization phase and a feedback phase;

[0031] Initialization phase:

[0032] Load the 3D point cloud model of the target structure and divide the detection grid;

[0033] Preset defect feature library and establish mapping relationship between multi-sensor data and confidence value;

[0034] Online planning phase:

[0035] Real-time fusion of surface texture analysis results from visual cameras and laser ranging data to update defect confidence maps;

[0036] Calculate the regional stability score based on the variance of the current distance measurement data and dynamically adjust the path weight;

[0037] Generate flight paths based on an improved heuristic search algorithm;

[0038] After the rescan is completed, the task weight of the corresponding node is reduced to guide the drone to the new target;

[0039] Generate Pareto frontier solution sets for conflicting objectives;

[0040] Feedback phase:

[0041] Continuously collect sensor data during flight, and trigger path replanning if sudden obstacles and confidence changes are detected;

[0042] Perform point cloud integrity check on the completed scanned area and automatically scan the missing areas.

[0043] Furthermore, the bridge pier measurement method integrating the laser gyro measuring rod and the UAV vision also includes a wind speed compensation mechanism, which includes:

[0044] Automatically calculate the initial wind speed threshold based on the aerodynamic performance of the drone to form a dynamically adjustable wind speed threshold range;

[0045] When the wind speed continuously exceeds the current dynamic wind speed threshold:

[0046] Start duration cumulative counter;

[0047] If the accumulated value reaches the "continuous risk" level predefined in the fuzzy rule base, a response action is triggered;

[0048] According to the fuzzy controller output, it smoothly switches between [low speed maintenance mode] and [emergency avoidance mode];

[0049] A feedforward compensation mechanism is introduced to offset the posture deviation caused by continuous wind pressure;

[0050] Enable the kinematic chain inverse compensation algorithm to eliminate the coupling effect of body shaking on the detection data.

[0051] The bridge pier measurement system is a fusion of laser gyro measuring rod and UAV vision. The system includes:

[0052] A measurement module, in which the UAV is equipped with a laser gyro measuring rod with an integrated gyroscope, an inclination sensor and a wireless communication module, and flies to the bridge pier measurement area, and collects the UAV's three-dimensional attitude data and inertial reference data in real time through the laser gyro measuring rod;

[0053] A scanning and measuring module, in which the drone simultaneously starts a binocular vision module and a laser rangefinder, performs multi-view scanning of the bridge pier along a preset circling path, and obtains a feature point cloud and real-time distance data on the surface of the bridge pier;

[0054] A spatiotemporal alignment module, which performs spatiotemporal alignment on the inertial reference data, the visual point cloud and the real-time distance data;

[0055] A thermal map generation module, wherein the thermal map generation module adopts an anti-interference extended Kalman filter algorithm to fuse multi-source data and separate the actual size data of the bridge piers from the UAV motion noise, and generates the three-dimensional actual size data of the bridge piers and a thermal map of size deviation;

[0056] The re-scanning module dynamically plans the local refined scanning path of the UAV according to the confidence of the size deviation heat map, and avoids obstacles in real time through the laser rangefinder until the measurement accuracy reaches the standard.

[0057] The technical effects and advantages of the bridge pier measurement method and system provided by the present invention by integrating the laser gyro measuring rod with the vision of an unmanned aerial vehicle are as follows:

[0058] The present invention achieves a synergistic breakthrough in submillimeter accuracy of bridge detection, full meteorological data integrity and all-weather continuous operation capability through laser gyro-visual point cloud spatiotemporal fusion, GAN-enhanced anti-rain and fog verification and dynamic anti-wind closed-loop control, solving the problem of severe distortion of measurement data caused by sensor degradation under complex meteorological conditions, so that the true geometric features of bridge piers can be accurately extracted. The present invention solves the core defects of traditional video watermarking technology, such as easy erasure due to insufficient reversibility and confusion of forgery attribution caused by static keys, through multi-dimensional synergistic optimization of anti-attack capability, dynamic security mechanism and verification efficiency, and provides reliable technical guarantee for digital content copyright protection. The present invention is based on the spatiotemporal alignment and fusion of the laser gyro inertial reference and the visual point cloud to achieve efficient separation of the true size of the bridge pier and motion noise, and maintain submillimeter measurement accuracy under strong wind disturbance, which is significantly better than the traditional optical measurement method; combined with adversarial image enhancement and multi-sensor redundancy verification, it effectively overcomes the problem of feature loss in rainy and foggy environments and ensures data integrity under complex meteorological conditions; constructs a closed-loop detection optimization link, and through the synergy of dynamic path planning and intelligent wind resistance strategies, greatly improves the detection efficiency and operation continuity under sudden environmental interference, forming an all-weather, all-terrain bridge pier detection capability. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a flow chart of the bridge pier measurement method using the laser gyro measuring rod and drone vision fusion in Example 1;

[0060] Figure 2 It is a flow chart of the method for generating the three-dimensional actual size data and size deviation thermal map of the bridge pier in the first embodiment;

[0061] Figure 3 This is a schematic diagram of the connection between the laser gyro measuring rod and the bridge pier measurement system with UAV vision fusion in Example 2. DETAILED DESCRIPTION

[0062] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] Embodiment 1: See also Figure 1 As shown, the bridge pier measurement method of the embodiment of the present invention by integrating the laser gyro measuring rod with the vision of the unmanned aerial vehicle comprises:

[0064] The UAV is equipped with a laser gyro measuring rod that integrates a gyroscope, an inclination sensor, and a wireless communication module. It flies to the bridge pier measurement area and collects the UAV's three-dimensional attitude data and inertial reference data in real time through the laser gyro measuring rod.

[0065] The drone simultaneously starts the binocular vision module and laser rangefinder, performs multi-view scanning of the bridge pier along the preset orbital path, and obtains the characteristic point cloud and real-time distance data of the bridge pier surface;

[0066] Temporally and spatially align inertial reference data, visual point cloud, and real-time distance data;

[0067] The anti-interference extended Kalman filter algorithm is used to fuse multi-source data and separate the actual size data of the bridge piers from the UAV motion noise, generating the three-dimensional actual size data of the bridge piers and the size deviation heat map;

[0068] According to the confidence level of the dimensional deviation heat map, the drone's local refined scanning path is dynamically planned, and obstacles are avoided in real time through the laser rangefinder until the measurement accuracy meets the standard (compared with the design drawing).

[0069] The binocular vision module calls the image enhancement model based on the generative adversarial network in a rainy and foggy environment to dehaze and restore the texture of low-visibility images.

[0070] When the binocular vision module detects that the ambient humidity or particle concentration exceeds the preset threshold, the image enhancement process is automatically triggered:

[0071] First, the original image is grayscale equalized to suppress the uneven brightness caused by rain and fog. Then, the preprocessed image is input into the pre-trained generative adversarial network model. The generative adversarial network model learns the mapping relationship between rain and fog degradation characteristics and real textures through adversarial training of a large number of synthetic rain and fog images and clear image pairs, thereby achieving image dehazing and detail restoration. The processed output image significantly improves edge clarity and texture continuity.

[0072] The repaired image and the laser rangefinder data are synchronously entered into the feature matching link. By dynamically adjusting the weight of the stereo matching algorithm of binocular vision, high-frequency feature areas such as the edges of the pier structure and bolt holes in the enhanced image are preferentially selected as matching reference points to reduce noise interference. At the same time, the feature point cloud of the enhanced image is aligned in time and space with the inertial reference data of the laser gyro measurement rod, and the extended Kalman filter algorithm is used to compensate for the delay error introduced by image processing.

[0073] Spatiotemporal alignment methods include:

[0074] Compensate for the UAV's own posture jitter based on the high-frequency attitude data of the gyroscope;

[0075] Dynamically bind the camera coordinate system of the binocular vision module, the laser ranging coordinate system and the inertial reference coordinate system of the laser gyro measuring rod;

[0076] During the hovering phase of the UAV, the static data of the laser gyro measuring rod is used to perform online zero bias calibration on the binocular vision module and laser rangefinder.

[0077] The UAV pose estimation algorithm uses the PnP algorithm to estimate the pose, and outputs the real-time attitude angle (pitch, roll, yaw) and angular velocity of the UAV at a fixed frequency through the gyroscope, thereby constructing a pose time series; the data frames of the binocular camera and laser rangefinder carry precise timestamps; for each sampling point in the visual point cloud and laser ranging data, the pose at the corresponding moment is interpolated from the gyroscope's pose time series according to its timestamp.

[0078] If the drone shakes during the acquisition of a certain visual point cloud frame, causing the visual point cloud to be distorted, the observation time of each pixel in the point cloud (according to the row scanning sequence) is associated with the high-frequency attitude data (the posture at a certain moment in the posture time series), and then the time alignment can be completed through the reverse compensation displacement based on the kinematic model.

[0079] Dynamic binding methods include:

[0080] The inertial data of the laser gyro measurement rod is used as a benchmark to forcibly correct the pose estimation results of visual SLAM.

[0081] The indicators of visual SLAM include the number of feature point tracking, reprojection error and the trace of the pose estimation covariance matrix. If any of the indicators exceeds the limit, the visual SLAM is judged to be invalid and a forced correction is triggered. Before the visual SLAM is judged to be invalid (only anomalies occur), when the pose estimation of the visual SLAM is abnormal due to motion blur or feature loss, it automatically switches to the inertial dominant mode based on the motion state detection of the inertial data (such as the sudden change amplitude of angular velocity or acceleration). At this time, the pose output of the visual SLAM is suspended, and the high-frequency attitude data of the gyroscope is directly injected into the pose solution module to ensure the continuity of the coordinate system.

[0082] The origin of the camera coordinate system is dynamically aligned with the inertial reference coordinate system through Lie group transformation, and the distance observation value of the laser rangefinder is introduced as a spatial constraint. This process suppresses the cumulative error of pure inertial data and avoids positioning drift caused by short-term switching.

[0083] In the inertial-dominant mode, the stability of the visual point cloud is continuously monitored. When the motion state returns to a smooth state, the visual SLAM gradually re-participates in the pose solution, but accepts the real-time offset compensation of the inertial reference coordinate system to achieve a smooth transition. This mechanism ensures that the spatiotemporal alignment error of multi-source data during high-speed maneuvers remains within the acceptable range for engineering.

[0084] Exemplary:

[0085] During a bridge pier top scanning mission, the drone performed an emergency obstacle avoidance maneuver due to a sudden crosswind.

[0086] at this time:

[0087] When the vision module detected that five consecutive frames of point cloud matching failed (angular velocity exceeded the sample value of 28 / s, and acceleration exceeded 1.8g), the system immediately switched to the inertial-dominated mode, and used the 400Hz gyroscope data to force correction of the posture. The laser ranging observation value constrained the inertial coordinate system mapping and suppressed horizontal drift. After about four control cycles (sample value of 60ms), the visual point cloud matching success rate returned to 85%, and the system restored the visual-inertial fusion mode. Finally, the spatial offset error between the point cloud and the inertial data was controlled within the sample threshold of ±7cm.

[0088] When visual SLAM fails, the gyroscope's posture (the posture that has completed time alignment) is used as a reference, and the output of visual SLAM is forcibly corrected through the extended Kalman filter. During forced correction, the observation value is switched to the inertial posture, and the observation noise covariance is increased to reflect the drift characteristics of the inertial data. At the same time, the filter is updated. The filter update equation includes: ;

[0089] In the formula, is the posterior state estimate, that is, at time Optimal state estimation after fusion of observations; is the prior state estimate, that is, at time State estimates obtained only from predictions (without taking into account current observations); is the Kalman gain; is the observation vector; is the observation matrix.

[0090] After the forced correction, when the visual SLAM is restored (the number of feature tracking increases), the inertial pose during the forced correction is used as the initial guess for visual relocalization; the PnP algorithm is used to quickly match the feature points of the current frame with the historical map to complete the pose alignment.

[0091] Dynamic binding can maintain centimeter-level positioning accuracy when vision fails, preventing the UAV from losing control or interrupting the measurement. It also allows the UAV to perform high-speed maneuvering tasks (such as fast surround scanning), shortens the measurement time, ensures that the point cloud model is strictly aligned with the design drawing coordinate system, and supports automated defect detection.

[0092] like Figure 2 As shown in the figure, the generation of the three-dimensional actual size data of the pier and the thermal map of size deviation includes:

[0093] The drone flies along the axial direction of the pier at a constant distance, scanning the surface at a fixed frequency;

[0094] IMU data is aligned with laser point cloud timestamps, and visual images are used to assist in point cloud coloring;

[0095] Initialize the state covariance matrix of the filter and set the prior distribution of motion noise; take the actual size parameters of the bridge pier (length, width, curvature) and the UAV motion noise (yaw angle jitter and acceleration offset) as joint state quantities;

[0096] Preprocess the scanned data by introducing an adaptive noise covariance matrix to dynamically downgrade abnormal jump points (caused by surface reflection or temporary occlusion);

[0097] The prediction-update steps are iteratively performed, in which the observation update stage gives a higher weight to the visual-assisted registration result, decouples the static size data of the bridge pier from the dynamic motion noise of the drone through the observation model, and outputs the denoised three-dimensional actual size data of the bridge pier;

[0098] Based on the filtered point cloud data, the non-uniform B-spline surface fitting algorithm is used to generate the 3D model of the bridge pier, the reconstructed model is aligned with the design drawing coordinate system, and the ICP algorithm is used to optimize the registration accuracy;

[0099] Calculate the projection deviation in the normal direction of each point on the surface and map it to the HSV color space;

[0100] Output interactive 3D heat map (size deviation heat map).

[0101] At the same time, when the size deviation heat map is generated, the grid resolution of the heat map is dynamically adjusted according to the UAV flight altitude and the laser ranging accuracy, and the design drawing contour lines are superimposed on the deviation area for visual comparison.

[0102] Exemplary:

[0103] When a UAV encounters crosswind during flight, its attitude fluctuates violently. The algorithm uses IMU data for real-time correction to eliminate the impact of motion noise on the point cloud.

[0104] The reconstructed model shows that there is an inward depression of up to 8mm at the bottom of the pier, and the red area of ​​the thermal diagram is consistent with the concrete spalling location confirmed by manual inspection;

[0105] The output report automatically counts the area ratio of the area with excessive deviation, providing a quantitative basis for maintenance decision-making.

[0106] The dynamic planning path includes the initialization phase, the online planning phase, and the feedback phase;

[0107] Initialization phase:

[0108] Load the 3D point cloud model of the target structure and divide the detection grid;

[0109] Preset defect feature library and establish mapping relationship between multi-sensor data and confidence value;

[0110] Online planning phase:

[0111] Real-time fusion of surface texture analysis results from visual cameras and laser ranging data to update defect confidence maps;

[0112] Calculate the regional stability score based on the variance of the current distance measurement data and dynamically adjust the path weight;

[0113] Generate flight paths based on an improved heuristic search algorithm;

[0114] The improved heuristic search algorithm is based on the traditional A* algorithm principle, integrates multiple indicators such as defect confidence, ranging stability, and flight attitude constraints into the cost function, and designs a dynamic adjustment strategy, which includes:

[0115] When a high-confidence defect is detected, the task weight is increased to trigger a focused scan;

[0116] In narrow areas or when the distance measurement is unstable, the safety weight is increased and obstacle avoidance is given priority;

[0117] After the rescan is completed, the task weight of the corresponding node is reduced to guide the drone to the new target;

[0118] Generate a Pareto front solution set for conflicting goals (such as shortest path vs. highest confidence scan) for the decision system to choose;

[0119] Only the routes within a limited time in the future are planned, and periodic re-planning is performed based on real-time data to adapt to dynamic environments.

[0120] Feedback phase:

[0121] Continuously collect sensor data during flight, and trigger path replanning if sudden obstacles and confidence changes are detected;

[0122] Perform point cloud integrity check on the completed scanned area and automatically scan the missing areas.

[0123] Exemplary:

[0124] Take the detection of bridge piers as an example:

[0125] When the UAV was flying at a set distance from the pier, the laser ranging suddenly showed abnormal fluctuations. The system determined that it was interference from debris and automatically generated a detour path. After identifying high-confidence irregular features on the surface of the pier (compared with the design drawing), the UAV switched to spiral descent mode to obtain multi-angle detailed images while maintaining a safe range of pitch angles. After completing a single scan, it automatically compared the point cloud coverage and initiated a supplementary scan of the shadowed area.

[0126] Through the multi-constraint fusion cost function and dynamic weight mechanism, the traditional path search is upgraded from "geometric space optimization" to "task-safety-efficiency collaborative optimization", which solves the planning problems of target conflicts and time-varying environments in complex inspection scenarios.

[0127] The bridge pier measurement method integrating the laser gyro measuring rod and the UAV vision also includes a wind speed compensation mechanism, which includes:

[0128] The drone is equipped with a wind speed sensor to monitor the ambient wind speed in real time. When the wind speed exceeds the preset wind speed threshold:

[0129] Reduce the flight speed of the drone and limit its maneuverability;

[0130] The weight of laser gyro data in the extended Kalman filter algorithm is enhanced to suppress the visual point cloud jitter error caused by wind speed.

[0131] Automatically calculate the initial wind speed threshold based on the aerodynamic performance of the drone (such as maximum wind resistance level and gimbal stability) to form a dynamically adjustable wind speed threshold range;

[0132] When the wind speed continuously exceeds the current dynamic wind speed threshold:

[0133] Start the duration accumulation counter, the counter growth rate is positively correlated with the wind speed exceeding the standard (greater than the current dynamic wind speed threshold value);

[0134] If the accumulated value reaches the "continuous risk" level predefined in the fuzzy rule base, a response action is triggered.

[0135] According to the output of the fuzzy controller, it can smoothly switch between [low speed maintenance mode] and [emergency avoidance mode];

[0136] A feedforward compensation mechanism is introduced to offset the posture deviation caused by continuous wind pressure.

[0137] The feedforward compensation mechanism includes wind pressure modeling and feedforward command generation;

[0138] Wind Pressure Modeling:

[0139] Obtain three-dimensional wind speed vector in real time through airborne anemometer ;

[0140] Combined with the aerodynamic parameters of the drone (such as windward area and drag coefficient), calculate the torque generated by wind pressure on the aircraft body and calculate the torque The methods include: ;

[0141] In the formula, is the air density, is the drag coefficient, is the windward area, is the position vector of the wind pressure center relative to the center of gravity.

[0142] Feedforward instruction generation:

[0143] According to the predicted wind pressure moment , inversely solve the required motor thrust compensation, add the compensation to the output of the conventional PID controller to form the final motor control command.

[0144] Automatically increase the fusion weight of inertial navigation data to suppress periodic jitter noise in visual / laser data;

[0145] Enable the kinematic chain inverse compensation algorithm to eliminate the coupling effect of body shaking on the detection data.

[0146] The kinematic chain inverse compensation algorithm operation method includes:

[0147] Model the motion transfer from the body to the sensor (forward kinematics);

[0148] Based on the body motion data measured by the IMU, reverse calculate the theoretical motion that the sensor should produce;

[0149] The theoretical motion component is subtracted from the raw sensor data to obtain the true geometric information of the static target (bridge pier).

[0150] Exemplary:

[0151] In a canyon bridge detection task:

[0152] The drone encountered intermittent strong winds caused by the canyon effect (wind speed fluctuations of 6-15m / s), and the system dynamically calculated the optimal response threshold range to be 7.2-9.1m / s;

[0153] When continuous excessive wind conditions are detected, it automatically switches to canyon crossing mode:

[0154] The flight speed is reduced to the minimum value within the safety envelope, and the detection sensor is kept aimed at the key area of ​​the bridge pier;

[0155] After motion compensation of the point cloud data, the measurement error of the bridge deck joint is still controlled within ±1.5mm;

[0156] After the task is completed, a wind field-response correlation report is automatically generated to provide data support for subsequent task parameter optimization.

[0157] Through the technical path of fuzzy parameter processing-dynamic event judgment-multimodal response, we get rid of the dependence on fixed values ​​and realize autonomous robust control of detection operations in complex wind field environments.

[0158] Embodiment 2: like Figure 3 As shown, based on the same inventive concept as the bridge pier measurement method of the laser gyro measurement rod and the UAV visual fusion in the above-mentioned embodiment, the present application provides a bridge pier measurement system of the laser gyro measurement rod and the UAV visual fusion, and the system and method embodiments in the embodiments of the present application are based on the same inventive concept. The system includes:

[0159] Measurement module: In the measurement module, the UAV is equipped with a laser gyro measuring rod with integrated gyroscope, tilt sensor and wireless communication module, and flies to the bridge pier measurement area. The laser gyro measuring rod collects the UAV's three-dimensional attitude data and inertial reference data in real time;

[0160] Scanning and measurement module: In the scanning and measurement module, the drone simultaneously starts the binocular vision module and the laser rangefinder, performs multi-view scanning of the bridge pier along the preset orbital path, and obtains the surface feature point cloud and real-time distance data of the bridge pier;

[0161] The spatiotemporal alignment module performs spatiotemporal alignment of inertial reference data, visual point cloud and real-time distance data;

[0162] Thermal map generation module: The thermal map generation module adopts the anti-interference extended Kalman filter algorithm to fuse multi-source data and separate the actual size data of the bridge piers from the UAV motion noise, and generate the three-dimensional actual size data of the bridge piers and the size deviation thermal map;

[0163] The re-scanning module dynamically plans the local refined scanning path of the UAV according to the confidence of the size deviation heat map, and avoids obstacles in real time through the laser rangefinder until the measurement accuracy meets the standard.

[0164] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

[0165] What has been described above is only a preferred specific implementation manner of the embodiments of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can make equivalent substitutions or changes according to the technical scheme and concept of the present application within the technical scope disclosed in the present application, which should be covered by the protection scope of the present application.

Claims

1. A bridge pier measurement method integrating laser gyro measuring rod and UAV vision, characterized in that: include: The UAV is equipped with a laser gyro measuring rod that integrates a gyroscope, an inclination sensor, and a wireless communication module. It flies to the bridge pier measurement area and collects the UAV's three-dimensional attitude data and inertial reference data in real time through the laser gyro measuring rod. The drone simultaneously starts the binocular vision module and laser rangefinder, performs multi-view scanning of the bridge pier along the preset orbital path, and obtains the characteristic point cloud and real-time distance data of the bridge pier surface; Temporally and spatially align inertial reference data, visual point cloud, and real-time distance data; The anti-interference extended Kalman filter algorithm is used to fuse multi-source data and separate the actual size data of the bridge piers from the UAV motion noise, generating the three-dimensional actual size data of the bridge piers and the size deviation heat map; According to the confidence level of the size deviation heat map, the drone's local refined scanning path is dynamically planned, and obstacles are avoided in real time through the laser rangefinder until the measurement accuracy meets the standard.

2. The bridge pier measurement method of laser gyro measuring rod and drone vision fusion according to claim 1 is characterized in that: The binocular vision module calls the image enhancement model based on the generative adversarial network in a rainy and foggy environment to dehaze and restore the texture of low-visibility images.

3. The bridge pier measurement method of laser gyro measuring rod and drone vision fusion according to claim 1 is characterized in that: Spatiotemporal alignment methods include: Compensate for the UAV's own posture jitter based on the high-frequency attitude data of the gyroscope; Dynamically bind the camera coordinate system of the binocular vision module, the laser ranging coordinate system and the inertial reference coordinate system of the laser gyro measuring rod; During the hovering phase of the UAV, the static data of the laser gyro measuring rod is used to perform online zero bias calibration on the binocular vision module and laser rangefinder.

4. The bridge pier measurement method of laser gyro measuring rod and drone vision fusion according to claim 3 is characterized in that: Dynamic binding methods include: Based on the inertial data of the laser gyro measurement rod, the pose estimation results of visual SLAM are forced to be corrected; The indicators of visual SLAM include the number of feature point tracking, reprojection error, and the trace of the pose estimation covariance matrix. If any of the indicators exceeds the limit, the visual SLAM is judged to be invalid and a forced correction is triggered; When visual SLAM fails, the gyroscope's posture is used as a reference, and the output of visual SLAM is forcibly corrected through the extended Kalman filter. During forced correction, the observation value is switched to the inertial posture, and the observation noise covariance is increased to reflect the drift characteristics of the inertial data. At the same time, the filter is updated. The filter update equation includes: ； In the formula, is the posterior state estimate; is the prior state estimate; is the Kalman gain; is the observation vector; is the observation matrix.

5. The bridge pier measurement method of laser gyro measuring rod and drone vision fusion according to claim 1 is characterized in that: The generation of 3D actual size data and size deviation thermal map of bridge piers includes: The drone flies along the axial direction of the pier at a constant distance, scanning the surface at a fixed frequency; IMU data is aligned with laser point cloud timestamps, and visual images are used to assist in point cloud coloring; Initialize the state covariance matrix of the filter and set the prior distribution of motion noise; take the actual size parameters of the bridge pier and the UAV motion noise as the joint state quantity; Preprocess the scanned data, introduce an adaptive noise covariance matrix, and dynamically reduce the weight of abnormal jump points; Iteratively execute the prediction-update step, decouple the static size data of the bridge pier from the dynamic motion noise of the drone through the observation model, and output the denoised three-dimensional actual size data of the bridge pier; Based on the filtered point cloud data, the non-uniform B-spline surface fitting algorithm is used to generate the 3D model of the bridge pier, the reconstructed model is aligned with the design drawing coordinate system, and the ICP algorithm is used to optimize the registration accuracy; Calculate the projection deviation in the normal direction of each point on the surface and map it to the HSV color space; Outputs an interactive heatmap of size deviation.

6. The bridge pier measurement method of laser gyro measuring rod and drone vision fusion according to claim 1 is characterized in that: The dynamic planning path includes the initialization phase and the feedback phase; Initialization phase: Load the 3D point cloud model of the target structure and divide the detection grid; Preset defect feature library and establish mapping relationship between multi-sensor data and confidence value; Online planning phase: Real-time fusion of surface texture analysis results from visual cameras and laser ranging data to update defect confidence maps; Calculate the regional stability score based on the variance of the current distance measurement data and dynamically adjust the path weight; Generate flight paths based on an improved heuristic search algorithm; After the rescan is completed, the task weight of the corresponding node is reduced to guide the drone to the new target; Generate Pareto frontier solution sets for conflicting objectives; Feedback phase: Continuously collect sensor data during flight, and trigger path replanning if sudden obstacles and confidence changes are detected; Perform point cloud integrity check on the completed scanned area and automatically scan the missing areas.

7. The bridge pier measurement method of laser gyro measuring rod and drone vision fusion according to claim 1 is characterized in that: It also includes a wind speed compensation mechanism, which includes: Automatically calculate the initial wind speed threshold based on the aerodynamic performance of the drone to form a dynamically adjustable wind speed threshold range; When the wind speed continuously exceeds the current dynamic wind speed threshold: Start duration cumulative counter; If the accumulated value reaches the "continuous risk" level predefined in the fuzzy rule base, a response action is triggered; According to the fuzzy controller output, it smoothly switches between [low speed maintenance mode] and [emergency avoidance mode]; A feedforward compensation mechanism is introduced to offset the posture deviation caused by continuous wind pressure; Enable the kinematic chain inverse compensation algorithm to eliminate the coupling effect of body shaking on the detection data.

8. The bridge pier measurement system integrating laser gyro measuring rod and UAV vision is characterized by: The system includes: A measurement module, in which the UAV is equipped with a laser gyro measuring rod with an integrated gyroscope, an inclination sensor and a wireless communication module, and flies to the bridge pier measurement area, and collects the UAV's three-dimensional attitude data and inertial reference data in real time through the laser gyro measuring rod; A scanning and measuring module, in which the drone simultaneously starts a binocular vision module and a laser rangefinder, performs multi-view scanning of the bridge pier along a preset circumferential path, and obtains a feature point cloud and real-time distance data on the surface of the bridge pier; A spatiotemporal alignment module, which performs spatiotemporal alignment on the inertial reference data, the visual point cloud and the real-time distance data; A thermal map generation module, wherein the thermal map generation module adopts an anti-interference extended Kalman filter algorithm to fuse multi-source data and separate the actual size data of the bridge piers from the UAV motion noise, and generates the three-dimensional actual size data of the bridge piers and a thermal map of size deviation; The re-scanning module dynamically plans the local refined scanning path of the UAV according to the confidence of the size deviation heat map, and avoids obstacles in real time through the laser rangefinder until the measurement accuracy reaches the standard.

Images