An unmanned aerial vehicle inspection method and system based on road construction

Abstract

The application discloses a kind of unmanned aerial vehicle inspection method and system based on road construction, belong to road construction inspection technical field.This method obtains the partition construction plan information of road construction area and real-time spatial data collected by unmanned aerial vehicle, and the partition construction plan information includes the planned construction time window and construction machinery moving range of each construction section;Real-time spatial data is mapped with partition construction plan information to generate construction progress correlation data;According to construction progress correlation data, determine priority inspection sub-area and generate boundary closed line;According to the expected track point sequence of boundary closed line planning, the area proportion of adjacent track points is related to priority inspection sub-area;According to the expected speed vector and expected yaw angle of expected track point sequence.Calculation scheme can dynamically determine key inspection area according to road construction progress, and compensate yaw angle under wind disturbance.

Description

Technical Field

[0001] This invention relates to the field of road construction inspection technology, specifically to a method and system for unmanned aerial vehicle (UAV) inspection of road construction. Background Technology

[0002] Road construction areas are typically divided into multiple construction sections, each with different planned construction time windows and activity ranges for construction machinery. As a means of monitoring road construction, drone inspections need to be conducted on key work areas according to the construction progress. Current inspection methods mostly involve manual setup of fixed flight paths at ground stations, or drones performing bow-shaped scans according to preset rectangular areas and fixed intervals. Fixed flight paths do not change with the planned construction time window and the distribution of construction machinery during the inspection mission, easily leading to deviations between the inspection coverage and current construction hotspots. This results in some areas under construction not being fully scanned, or completed but unconstructed areas being repeatedly covered. Regarding flight path density, conventional bow-shaped scans usually use a uniform interval, without adjusting the scan line spacing according to the ratio of the priority inspection sub-area area to the area of ​​the circumscribed rectangle. When the priority inspection sub-area occupies a small proportion of the boundary rectangle, the fixed interval causes a large number of trackpoints to fall in invalid areas; when the proportion is large, the fixed interval is insufficient to form a sufficiently dense number of trackpoints in the key areas. Regarding the association of construction progress data, existing methods, after obtaining zonal construction plan information, fail to map the planned construction time window and the movement range of construction machinery with the real-time spatial data of UAVs along a time axis. This prevents the marking of spatial points corresponding to the current construction progress at the data level, resulting in the inability to automatically generate priority inspection sub-areas based on the current construction progress. In terms of yaw control, crosswinds in the road construction environment and airflow disturbances near construction machinery cause changes in the roll and pitch angles of the UAV. Conventional yaw control primarily relies on feedback correction based on yaw angle errors, without utilizing measured roll and pitch angles for feedforward compensation of yaw moments caused by wind disturbances. Under gusts or localized airflow, heading correction lags, easily leading to lateral drift. Under dynamic changes in construction progress and airflow disturbances, maintaining consistency between fixed flight paths and the current key construction area, as well as flight path tracking accuracy, is difficult. Summary of the Invention

[0003] To address the issues of fixed flight paths in existing road construction drone inspections, which cannot dynamically adjust key areas according to construction progress, and the lag in yaw correction under wind disturbance, this paper proposes a drone inspection method and system based on road construction. The method generates boundary closures for priority inspection sub-regions by mapping the construction progress timeline and aggregating hotspot grids. The spacing between track points is determined based on the area proportion of the priority inspection sub-regions. Furthermore, the equivalent yaw moment due to wind disturbance is calculated using measured attitude data, and feedforward compensation is applied to the desired yaw angle to generate corrective yaw angle control commands.

[0004] To achieve the above objectives, the present invention provides the following technical solution: The present invention provides a method and system for unmanned aerial vehicle (UAV) inspection based on road construction. The method includes: acquiring zoning construction plan information of the road construction area and real-time spatial data collected by the UAV, wherein the zoning construction plan information includes the planned construction time window and the movement range of construction machinery for each construction section; mapping the real-time spatial data to the zoning construction plan information using a construction progress time axis to generate construction progress-related data; determining priority inspection sub-areas based on the construction progress-related data and generating boundary closure lines for the priority inspection sub-areas; planning a desired trackpoint sequence based on the boundary closure lines, wherein the distance between adjacent trackpoints in the desired trackpoint sequence is related to the area proportion of the priority inspection sub-area; calculating the desired velocity vector and desired yaw angle between adjacent trackpoints based on the desired trackpoint sequence; acquiring measured attitude data during flight, correcting the desired yaw angle based on the measured attitude data, and generating a corrected yaw angle control command. This method unifies the association of construction plan time information, construction machinery spatial information, and UAV real-time spatial data, enabling dynamic matching of the inspection path with the construction progress, improving the inspection coverage efficiency and flight control accuracy of construction hotspot areas.

[0005] Preferably, obtaining the zoned construction plan information for the road construction area includes: parsing the overall schedule of the road construction project to obtain the planned construction time field and the construction machinery activity area field for each construction section; pairing the planned construction time field and the construction machinery activity area field to generate a construction section spatiotemporal mapping table; sorting the construction section spatiotemporal mapping table according to the start time of the planned construction time window, and using the sorted construction section spatiotemporal mapping table as the zoned construction plan information. Through the above processing, the construction plan information has indexable characteristics of temporal order and spatial range, which facilitates subsequent association with real-time spatial data.

[0006] Preferably, the real-time spatial data collected by the UAV includes: acquiring raw spatial data through positioning sensors and a laser scanner on the UAV; and performing coordinate transformation on the raw spatial data to generate real-time spatial data with absolute coordinates. A unified coordinate reference helps reduce spatial deviations between multi-source data and improves the reliability of hotspot identification and boundary extraction.

[0007] Preferably, mapping the real-time spatial data to the partitioned construction plan information along a construction progress timeline to generate construction progress-related data includes: extracting the planned construction time window from the partitioned construction plan information and determining the target planned construction time window to which the current moment belongs; using the construction segment number corresponding to the target planned construction time window as a construction progress status marker; matching the spatial coordinates of each data point in the real-time spatial data with the movement range of the construction machinery to generate a construction machinery impact marker; and writing the construction progress status marker and the construction machinery impact marker into the real-time spatial data to generate the construction progress-related data. Through timeline mapping and spatial matching, each data point in the real-time spatial data can carry construction progress status and machinery impact attributes, providing a basis for identifying priority inspection sub-areas.

[0008] Preferably, determining the priority inspection sub-region based on the construction progress correlation data and generating the boundary closure line of the priority inspection sub-region includes: filtering data points from the construction progress correlation data that cover the current moment within the planned construction time window, projecting the filtered data points onto a horizontal plane to generate a construction hotspot distribution map; performing rasterization processing on the construction hotspot distribution map, counting the number of data points in each grid, and marking grids with more than a preset number of data points as hotspot grids; merging adjacent hotspot grids to obtain the priority inspection sub-region, and performing polygon fitting on the outline of the priority inspection sub-region to generate the boundary closure line. This method can automatically identify the current densely constructed area and form a closed boundary for trajectory planning, reducing the subjectivity and time cost of manually dividing sub-regions.

[0009] Preferably, planning the desired waypoint sequence based on the boundary closure line includes: generating a bow-shaped reciprocating scan line based on the boundary closure line, wherein the initial spacing of the bow-shaped reciprocating scan line is a fixed proportion of the effective coverage width of the UAV; calculating the area ratio of the area of ​​the priority inspection sub-region to the area of ​​the rectangle circumscribed by the boundary closure line, adjusting the spacing of the bow-shaped reciprocating scan line according to the area ratio, wherein the larger the area ratio, the smaller the adjusted spacing; and connecting the intersections of the adjusted bow-shaped reciprocating scan line with the boundary closure line in the scanning order to generate the desired waypoint sequence. By adaptively adjusting the scanning spacing through the area ratio, the waypoint density within the priority inspection sub-region matches the region's shape complexity, improving inspection resolution and reducing redundant flight outside the boundary.

[0010] Preferably, calculating the expected velocity vector and expected yaw angle between adjacent waypoints based on the expected waypoint sequence includes: calculating the expected flight duration for each flight segment based on the spatial distance between adjacent waypoints in the expected waypoint sequence and the UAV's preset cruise speed; calculating the expected velocity vector based on the expected flight duration and the direction vector between adjacent waypoints; decomposing the expected velocity vector to the longitudinal and lateral axes of the UAV's body coordinate system; determining the turning direction of the expected yaw angle based on the direction of the lateral axis component; and determining the magnitude of the expected yaw angle based on the amplitude of the lateral axis component. This converts the waypoint sequence into executable flight control reference quantities for the UAV, ensuring consistency between the flight trajectory and the planned flight path.

[0011] Preferably, acquiring measured attitude data during flight includes: integrating the three-axis angular velocities output by the inertial measurement unit in the UAV attitude sensor over time to obtain a first attitude angle; correcting the three-axis accelerations output by the accelerometer in the UAV attitude sensor for gravity direction to obtain a second attitude angle; and performing complementary correction on the first and second attitude angles to generate the measured attitude data. Complementary correction can suppress angular velocity integral drift and instantaneous accelerometer noise, improving the stability of the measured attitude data.

[0012] Preferably, the method of correcting the desired yaw angle based on the measured attitude data and generating a corrected yaw angle control command includes: extracting the current roll angle and current pitch angle from the measured attitude data; calculating the wind disturbance equivalent yaw moment based on the current roll angle and current pitch angle; converting the wind disturbance equivalent yaw moment into a yaw angle feedforward compensation amount; and superimposing the yaw angle feedforward compensation amount with the desired yaw angle to generate the corrected yaw angle control command. This method uses the measured roll angle and pitch angle to estimate the impact of wind disturbance on yaw and performs feedforward compensation, which can reduce yaw deviation under wind disturbance and improve the stability of inspection flights.

[0013] As a technical solution of the present invention, a UAV inspection system based on road construction is provided, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of the aforementioned UAV inspection method based on road construction. This system can perform processes such as construction progress correlation, priority inspection sub-area identification, trajectory planning, and yaw angle correction, achieving efficient inspection of road construction areas.

[0014] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0015] The planned construction time windows and machinery movement ranges for each construction segment in the zonal construction plan information are mapped onto the construction progress timeline using real-time UAV spatial data. During the mapping process, a construction progress status marker is determined based on the planned construction time window to which the current moment belongs. The spatial coordinates of real-time spatial data points are matched with the machinery movement range to form a machinery influence marker, which is then written into the real-time spatial data to obtain construction progress association data. Based on the construction progress association data, data points covering the current moment within the planned construction time window are selected and projected onto a horizontal plane for raster aggregation. Adjacent hotspot rasteres are merged to obtain a priority inspection sub-region. The outline of this sub-region is then fitted with polygons to generate a boundary closure line. This process aligns the inspection range with the actual working area and machinery activity area of ​​the current road construction, reducing invalid scans of unconstructed and completed areas. When planning trackpoints, the spacing of the bow-shaped reciprocating scan lines is adjusted based on the ratio of the area of ​​the priority inspection sub-region to the area of ​​its circumscribed rectangle. A larger area ratio results in a smaller scan line spacing, and the spacing between adjacent trackpoints is adjusted accordingly. This allows the trackpoint density to vary with the spatial distribution of the priority inspection sub-region, avoiding redundant trackpoints in invalid areas with a fixed scan spacing. When the priority inspection sub-area accounts for a large proportion, the scan line spacing decreases, resulting in a denser distribution of waypoints within the key area; conversely, when the area is relatively small, the scan line spacing increases, reducing waypoints in invalid areas. During flight, measured attitude data is acquired, and the current roll and pitch angles are extracted. The equivalent yaw moment due to wind disturbance is calculated from these angles and converted into a yaw angle feedforward compensation. This compensation is then superimposed on the desired yaw angle to generate a corrected yaw angle control command. This feedforward compensation acts on yaw control before wind disturbances cause an increase in yaw angle error, compensating for the lag in response to crosswinds and local airflow disturbances in conventional yaw angle feedback control. This allows the UAV to maintain tracking of the desired velocity vector and desired waypoint sequence in road construction environments, reducing heading and lateral deviations. The feedforward compensation is calculated in real-time based on the roll and pitch angles and directly participates in yaw angle control when the aircraft's attitude changes due to gusts, without waiting for yaw angle error accumulation. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0017] Figure 1 This is a flowchart of a drone inspection method based on road construction.

[0018] Figure 2This is a flowchart of the process for acquiring road construction area zone construction plan information and real-time spatial data from drones;

[0019] Figure 3 This is a flowchart for generating priority inspection sub-areas related to construction progress. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] See Figure 1 This invention provides a method and system for unmanned aerial vehicle (UAV) inspection of road construction. The method includes: acquiring zoning construction plan information of the road construction area and real-time spatial data collected by the UAV, wherein the zoning construction plan information includes the planned construction time window and the movement range of construction machinery for each construction section; mapping the real-time spatial data and the zoning construction plan information onto a construction progress time axis to generate construction progress correlation data; determining priority inspection sub-areas based on the construction progress correlation data and generating boundary closure lines for the priority inspection sub-areas; planning a desired track point sequence based on the boundary closure lines, wherein the distance between adjacent track points in the desired track point sequence is related to the area ratio of the priority inspection sub-area; calculating the desired velocity vector and desired yaw angle between adjacent track points based on the desired track point sequence; acquiring measured attitude data during flight, correcting the desired yaw angle based on the measured attitude data, and generating a corrected yaw angle control command.

[0022] See Figure 2When obtaining the zoning construction plan information for road construction areas, the overall road construction project schedule is read from the road construction project management database. This schedule is a structured data table containing the construction section number, construction section name, planned start time, planned end time, and the set of boundary points for the construction machinery activity area. The overall schedule is parsed to extract the construction section number, planned start time, planned end time, and the set of boundary points for the construction machinery activity area from each row. These boundary points are sequentially arranged planar coordinate sequences, representing closed polygons representing the activity range of the construction machinery. The planned start and end times are merged into a planned construction time window, with the start and end times of the window being the planned start and end times, respectively. The construction machinery activity area field represents the closed polygon region corresponding to the boundary point set. The planned construction time window for each construction section is paired with the construction machinery activity area polygon, and this is written into a construction section spatiotemporal mapping table. This table is an associative array, where each element contains the construction section number, the planned construction time window start time, the planned construction time window end time, and a list of vertex coordinates for the construction machinery activity area polygon. The elements in the construction section spatiotemporal mapping table are sorted in ascending order according to the start time of the planned construction time window. When the start times of the planned construction time windows are the same, they are sorted in ascending order according to the construction section number. The sorted construction section spatiotemporal mapping table is output as the partition construction plan information.

[0023] When acquiring real-time spatial data collected by the UAV, raw spatial data is collected through the positioning sensor and laser scanner on the UAV. The positioning sensor is a Global Navigation Satellite System receiver, outputting the UAV's position coordinates in the absolute coordinate system and the corresponding timestamp. The laser scanner is a multi-line LiDAR, outputting raw point cloud data. The raw point cloud data includes the distance, scanning angle, and corresponding timestamp of each laser point in the laser scanner's coordinate system. The timestamps of the positioning sensor and laser scanner are synchronized by the UAV's onboard clock. Coordinate transformation is performed on the raw spatial data, converting each laser point in the raw point cloud data from the laser scanner's coordinate system to the UAV's body coordinate system, obtaining the point coordinates in the UAV's body coordinate system. Then, the point coordinates in the UAV's body coordinate system are converted back to the absolute coordinate system, obtaining the point coordinates in the absolute coordinate system. The coordinate transformation process uses rotation matrices and translation vectors. The specific transformation relationships are as follows:

[0024]

[0025] in, The coordinates of the laser point in the absolute coordinate system. The coordinates of the laser point in the UAV's body coordinate system. Let be the rotation matrix from the UAV's body coordinate system to its absolute coordinate system. Let be the translation vector of the origin of the UAV's body coordinate system in the absolute coordinate system. The attitude angle of the UAV at the moment of laser point acquisition is determined. The location is determined by the position coordinates of the laser point at the time of acquisition by the positioning sensor. When aligning the location data with the output of the positioning sensor, the nearest neighbor timestamp matching method is used. The timestamp of each laser point is matched with the timestamp of the location data output by the positioning sensor, and the positioning sensor location data with the smallest absolute time difference is selected as the aligned location data. The absolute coordinates of each laser point are combined with the aligned location data to generate real-time spatial data with absolute coordinates. Each data point in the real-time spatial data contains absolute coordinates, timestamp, and the original distance of the corresponding laser point.

[0026] In specific implementation, please refer to Figure 3 The planned construction time windows are extracted from the zonal construction plan information, which is a sorted spatiotemporal mapping table of construction sections. Each element in the spatiotemporal mapping table contains a construction section number, the start time of the planned construction time window, the end time of the planned construction time window, and a list of vertex coordinates of the polygon in the construction machinery activity area. The current time is read by the UAV's onboard clock and is the execution time of the construction progress timeline mapping. The current time is matched against all planned construction time windows in the zonal construction plan information. The matching method is to determine whether the current time is greater than or equal to the start time of the planned construction time window and less than the end time of the planned construction time window. The planned construction time window that meets the matching conditions is determined as the target planned construction time window. When multiple planned construction time windows cover the current time, the target planned construction time window with the latest start time is selected.

[0027] The construction segment number corresponding to the target planned construction time window is used as the construction progress status marker. The construction progress status marker is an integer value representing the sequence number of the construction segment at the current moment. Each data point in the real-time spatial data contains absolute coordinates and a timestamp. The spatial coordinates of each data point in the real-time spatial data are matched with the polygon of the construction machinery activity area corresponding to the target planned construction time window in the partitioned construction plan information. The matching method is to determine the inclusion relationship between the point and the polygon, and the ray casting method is used to determine whether the horizontal plane projection coordinates of each data point in the real-time spatial data are located inside the polygon of the construction machinery activity area. When the horizontal plane projection coordinates of a data point in the real-time spatial data are located inside the polygon of the construction machinery activity area, the construction machinery influence marker corresponding to that data point is set to 1; when the horizontal plane projection coordinates of a data point in the real-time spatial data are located outside the polygon of the construction machinery activity area, the construction machinery influence marker corresponding to that data point is set to 0.

[0028] The construction progress status marker and construction machinery influence marker are written into the real-time spatial data. Specifically, a construction progress status marker field and a construction machinery influence marker field are added to each data point in the real-time spatial data. The determined construction progress status marker and construction machinery influence marker are written into the corresponding fields respectively to generate construction progress association data. The construction progress association data retains the original absolute coordinates, timestamp, and original distance fields of the laser point in the real-time spatial data, and also includes the construction progress status marker and construction machinery influence marker fields.

[0029] In practice, data points covering the current moment within the planned construction time window are selected from the construction progress correlation data. The selection criteria are that the construction progress status marker in the construction progress correlation data is equal to the construction segment number corresponding to the target planned construction time window, and the construction machinery influence marker is equal to 1. The selected data points are then projected onto a horizontal plane. The projection method is to take the two-dimensional horizontal coordinates from the absolute coordinates of each data point, ignoring the height coordinates, to generate a construction hotspot distribution map.

[0030] The construction hotspot distribution map is rasterized. Orthogonal grids are divided within the horizontal plane of the map, with fixed grid side lengths determined by the area of ​​the polygon representing the construction machinery's activity zone and the point cloud density from the laser scanner. The number of data points within each grid is counted. For each grid, the number of construction hotspot distribution map data points falling within its range is counted. Grids with more than a preset number of data points are marked as hotspot grids. The preset number is a fixed threshold, set based on the density of non-construction background data points within the construction area; a preset number of 3 is used. When the number of data points within a grid is greater than 3, that grid is marked as a hotspot grid.

[0031] Adjacent hotspot grids are merged. Adjacent hotspot grids are those sharing the same edge or the same vertex. Hotspot grids belonging to the same connected region are merged to obtain a priority inspection sub-region. The priority inspection sub-region is a continuous region formed after the hotspot grids are merged. Polygon fitting is performed on the contour of the priority inspection sub-region. Polygon fitting uses a contour extraction algorithm to first extract the outer boundary contour of the priority inspection sub-region, and then approximate the outer boundary contour with straight line segments to generate a boundary closure line. The boundary closure line is a sequentially arranged sequence of planar coordinates, representing the closed polygon boundary of the priority inspection sub-region.

[0032] In practical implementation, a bow-shaped reciprocating scan line is generated based on the boundary closure line. The boundary closure line is the closed polygon boundary of the priority inspection sub-region, represented by ordered vertex coordinates. The circumscribed rectangle of the boundary closure line on the horizontal plane is obtained, and the circumscribed rectangle is determined by the minimum x-coordinate, maximum x-coordinate, minimum y-coordinate, and maximum y-coordinate of all vertices of the boundary closure line. The effective coverage width of the UAV is determined, which is determined by the field of view angle and flight altitude of the inspection payload. Let the effective coverage width of the UAV be... A fixed ratio of 0.8 is chosen to maintain a 20% overlap between adjacent scan lines, reducing the probability of missed scans due to flight attitude deviations. Initial spacing... Multiply the effective coverage width of the drone by 0.8, that is Within the circumscribed rectangle, a set of parallel scan lines is generated along the long side of the circumscribed rectangle. The spacing between adjacent scan lines is the initial spacing, and the scan line direction is perpendicular to the long side of the circumscribed rectangle. The portion where each scan line intersects with the boundary closure line forms the initial bow-shaped reciprocating scan line.

[0033] Optionally, calculate the area ratio of the priority inspection sub-region to the area of ​​the circumscribed rectangle of its boundary closure line. (Priority inspection sub-region area) To calculate the area of ​​the polygon enclosed by the boundary closure lines, use the polygon area formula. With the coordinates of the vertices of the boundary closure lines arranged in order, use the shoelace formula to calculate the area. The area of ​​the circumscribed rectangle of the boundary closure lines. The value is calculated as the length of the circumscribed rectangle multiplied by its width. The length of the circumscribed rectangle is equal to the maximum x-coordinate of the vertices of the boundary closure line minus the minimum x-coordinate, and the width of the circumscribed rectangle is equal to the maximum y-coordinate of the vertices of the boundary closure line minus the minimum y-coordinate. The spacing of the bow-shaped reciprocating scan lines is adjusted according to the area proportion. Calculate using the following formula:

[0034]

[0035] in, The adjusted spacing between the bow-shaped reciprocating scan lines. This is the initial spacing. Prioritize inspecting the area of ​​the sub-region. This represents the area of ​​the rectangle circumscribed by the boundary closure line. As the proportion of the area of ​​the priority inspection sub-region to the area of ​​the circumscribed rectangle increases... The spacing is reduced; the adjusted spacing is reduced.

[0036] It's understandable to use the adjusted spacing. Regenerate the bow-shaped reciprocating scan lines, with the spacing between adjacent scan lines within the bounding rectangle being [missing information]. The intersections of the adjusted bow-shaped reciprocating scan lines with the boundary closure lines are connected according to the scan sequence. Specifically, for each scan line, all intersections with the boundary closure line are calculated. The intersections of each scan line are sorted along the scan line direction, which alternates. The intersections of the first scan line with the boundary closure line are sorted from one side of the circumscribed rectangle to the other, the second scan line is sorted from the other side of the circumscribed rectangle to one side, and so on. The sorted intersections of all scan lines are connected sequentially according to the scan line order to generate the desired track point sequence. The spacing between adjacent track points in the desired track point sequence is related to the area proportion of the priority inspection sub-region; the larger the area proportion, the smaller the spacing, and the denser the track points.

[0037] In practical implementation, when calculating the expected velocity vector and expected yaw angle between adjacent waypoints based on the expected waypoint sequence, two adjacent waypoints are sequentially extracted from the expected waypoint sequence, and the spatial distance between adjacent waypoints is calculated. The spatial distance is obtained by taking the square root of the sum of the squares of the differences in the absolute coordinates of adjacent waypoints on the three coordinate axes. The UAV's preset cruise speed is read; the UAV's preset cruise speed is set by the UAV inspection task and is a fixed scalar. The expected flight time for each flight segment is obtained by dividing the spatial distance between adjacent waypoints by the UAV's preset cruise speed.

[0038] The desired velocity vector is calculated based on the expected flight duration and the direction vectors between adjacent waypoints. The direction vector between adjacent waypoints is obtained by subtracting the absolute coordinates of the current waypoint from the absolute coordinates of the next waypoint, and the magnitude of the direction vector is the spatial distance. The desired velocity vector is calculated using the following formula:

[0039]

[0040] in, For the desired velocity vector, This represents the direction vector between adjacent waypoints. The expected flight time for each route segment, The three components are the differences between the absolute coordinates of the next track point and the absolute coordinates of the current track point on the three coordinate axes.

[0041] The desired velocity vector is decomposed into the longitudinal and lateral axes of the UAV's body coordinate system. The longitudinal axis of the UAV's body coordinate system is the forward direction axis, and the lateral axis is the rightward direction axis. During decomposition, the current attitude angle of the UAV is obtained, and the desired velocity vector in the absolute coordinate system is transformed to the UAV's body coordinate system. The component of the transformed desired velocity vector on the longitudinal axis is the longitudinal axis component, and the component on the lateral axis is the lateral axis component.

[0042] The direction of the desired yaw angle is determined by the direction of the lateral axis component. When the lateral axis component is positive, the desired yaw angle is a right turn; when the lateral axis component is negative, the desired yaw angle is a left turn. The magnitude of the desired yaw angle is determined by the magnitude of the lateral axis component, which is determined by the arctangent of the ratio of the lateral axis component magnitude to the longitudinal axis component magnitude. The magnitude of the angle reflects the angle between the desired velocity vector and the longitudinal axis of the UAV. When the lateral axis component magnitude is zero, the magnitude of the desired yaw angle is zero.

[0043] In practical implementation, when acquiring measured attitude data during flight, the UAV attitude sensor includes an inertial measurement unit (IMU) and an accelerometer, both fixed to the UAV's body coordinate system. The IMU outputs three-axis angular velocities: roll angular velocity around the roll axis, pitch angular velocity around the pitch axis, and yaw angular velocity around the yaw axis. The three-axis angular velocities output by the IMU are integrated over time using a discrete-time integration method. The integration step size is determined by the IMU's sampling period, which is determined by the IMU's hardware parameters. The first attitude angle at each moment is obtained by adding the first attitude angle from the previous moment to the product of the current three-axis angular velocity and the sampling period. The first attitude angle includes the first roll angle, the first pitch angle, and the first yaw angle.

[0044] Gravity direction correction is applied to the three-axis acceleration output from the accelerometer in the UAV's attitude sensor. The accelerometer outputs the specific force along the three axes of the UAV's body coordinate system. When the UAV is in stable flight, the accelerometer measurement mainly consists of the component of the gravity vector in the UAV's body coordinate system. Based on the component of the gravity vector in the UAV's body coordinate system, and utilizing the geometric relationship between the gravity vector direction and the axes of the UAV's body coordinate system, the second pitch angle and the second roll angle in the second attitude angle are calculated. The second yaw angle in the second attitude angle is the first yaw angle from the first attitude angle. The second attitude angle is the attitude angle obtained by gravity direction correction from the accelerometer.

[0045] The first and second attitude angles are complementaryly corrected. The complementary correction uses a complementary filtering method. The first attitude angle is obtained by integrating the angular velocity, which is insensitive to high-frequency vibrations but suffers from integral drift. The second attitude angle is obtained by gravity direction correction, providing accurate long-term attitude measurements but is susceptible to interference from UAV body vibrations. The first and second attitude angles are then weighted and fused according to the complementary filtering transfer function to generate measured attitude data. The measured attitude data includes measured roll angle, measured pitch angle, and measured yaw angle.

[0046] When correcting the desired yaw angle based on measured attitude data, the current roll angle and current pitch angle are extracted from the measured attitude data. The current roll angle is the measured roll angle from the measured attitude data, and the current pitch angle is the measured pitch angle from the measured attitude data. The wind-induced equivalent yaw moment is calculated based on the current roll angle and current pitch angle. The wind-induced equivalent yaw moment represents the yaw moment generated by the aerodynamic asymmetry caused by the current roll angle and current pitch angle. The wind-induced equivalent yaw moment is calculated using the following formula:

[0047]

[0048] in, The equivalent yaw moment due to wind disturbance is expressed in Newton-meters. This is the current roll angle, in radians; The current pitch angle, in radians; is the first aerodynamic coupling coefficient from roll angle to yaw moment, in Newton-meter per radian; The second aerodynamic coupling coefficient, representing the pitch angle to yaw moment, is expressed in Newton-meters per radian (N·m / radian). The first and second aerodynamic coupling coefficients are calibrated in a ground-based wind tunnel before the UAV leaves the factory. The first aerodynamic coupling coefficient ranges from 0.05 N·m / radian to 0.35 N·m / radian, and the second aerodynamic coupling coefficient ranges from 0.03 N·m / radian to 0.28 N·m / radian. The corresponding calibration values ​​are selected based on the UAV model.

[0049] The equivalent yaw moment due to wind disturbance is converted into a yaw angle feedforward compensation. The conversion method involves dividing the equivalent yaw moment by the product of the UAV's yaw channel moment of inertia and the preset yaw angle acceleration response gain, yielding the yaw angle feedforward compensation value. This yaw angle feedforward compensation is then superimposed on the desired yaw angle to generate a corrected yaw angle control command. This command contains the corrected target yaw angle value. The UAV flight control system controls the yaw channel actuators according to the corrected yaw angle control command to achieve yaw angle correction.

[0050] The drone inspection system for road construction includes a memory, a processor, and a computer program stored in the memory and running on the processor. The memory is a non-volatile storage medium, and the processor is a central processing unit (CPU). When executing the computer program, the processor implements the steps of the drone inspection method for road construction described above.

[0051] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for unmanned aerial vehicle (UAV) inspection based on road construction, characterized in that, The method includes: The system acquires zoned construction plan information for road construction areas and real-time spatial data collected by drones. The zoned construction plan information includes the planned construction time window and the movement range of construction machinery for each construction section. The real-time spatial data is mapped to the partitioned construction plan information along a construction progress timeline to generate construction progress-related data. Based on the construction progress correlation data, priority inspection sub-areas are determined, and boundary closure lines of the priority inspection sub-areas are generated. The desired track point sequence is planned based on the boundary closure line, and the distance between adjacent track points in the desired track point sequence is related to the area ratio of the priority inspection sub-region. Calculate the desired velocity vector and desired yaw angle between adjacent waypoints based on the desired waypoint sequence; During flight, measured attitude data is acquired, and the desired yaw angle is corrected based on the measured attitude data to generate a corrected yaw angle control command.

2. The method for unmanned aerial vehicle (UAV) inspection based on road construction according to claim 1, characterized in that, The acquisition of the zoned construction plan information for the road construction area includes: Analyze the overall schedule of the road construction project to obtain the planned construction time field and the construction machinery activity area field for each construction section; The planned construction time field and the construction machinery activity area field are paired to generate a construction segment spatiotemporal mapping table; The construction segment spatiotemporal mapping table is sorted according to the start time of the planned construction time window, and the sorted construction segment spatiotemporal mapping table is used as the partition construction plan information.

3. The method for unmanned aerial vehicle (UAV) inspection based on road construction according to claim 1, characterized in that, The real-time spatial data collected by the drone includes: Raw spatial data is collected using positioning sensors and laser scanners on drones; The original spatial data is transformed to generate real-time spatial data with absolute coordinates.

4. The method for unmanned aerial vehicle (UAV) inspection based on road construction according to claim 1, characterized in that, The real-time spatial data is mapped to the partitioned construction plan information along a construction progress timeline to generate construction progress-related data, including: Extract the planned construction time window from the partitioned construction plan information and determine the target planned construction time window to which the current moment belongs; The construction segment number corresponding to the target planned construction time window is used as the construction progress status marker. The spatial coordinates of each data point in the real-time spatial data are matched with the movement range of the construction machinery to generate the construction machinery influence marker. The construction progress status marker and the construction machinery impact marker are written into the real-time spatial data to generate the construction progress association data.

5. The method for unmanned aerial vehicle (UAV) inspection based on road construction according to claim 1, characterized in that, The step of determining priority inspection sub-regions based on the construction progress correlation data and generating the boundary closure line of the priority inspection sub-regions includes: The data points that cover the current moment within the planned construction time window are selected from the construction progress correlation data, and the selected data points are projected onto the horizontal plane to generate a construction hotspot distribution map; The construction hotspot distribution map is rasterized, the number of data points in each grid is counted, and grids with more than a preset number of data points are marked as hotspot grids. The adjacent hotspot grids are merged to obtain the priority inspection sub-region, and the contour of the priority inspection sub-region is polygonally fitted to generate the boundary closure line.

6. The method for unmanned aerial vehicle (UAV) inspection based on road construction according to claim 1, characterized in that, The step of planning the desired track point sequence based on the boundary closure line includes: Based on the boundary closure line, a bow-shaped reciprocating scan line is generated, and the initial spacing of the bow-shaped reciprocating scan line is a fixed proportion of the effective coverage width of the UAV; Calculate the area ratio of the area of ​​the priority inspection sub-region to the area of ​​the rectangle circumscribed by the boundary closure line, and adjust the spacing of the bow-shaped reciprocating scan lines according to the area ratio. The larger the area ratio, the smaller the adjusted spacing. The intersections of the adjusted bow-shaped reciprocating scan line and the boundary closure line are connected in the scanning order to generate the desired track point sequence.

7. The method for unmanned aerial vehicle (UAV) inspection based on road construction according to claim 1, characterized in that, The step of calculating the desired velocity vector and desired yaw angle between adjacent waypoints based on the desired waypoint sequence includes: The expected flight time for each route segment is calculated based on the spatial distance between adjacent waypoints in the expected waypoint sequence and the preset cruise speed of the UAV. The desired velocity vector is calculated based on the desired flight duration and the direction vector between adjacent waypoints; The desired velocity vector is decomposed into the longitudinal and lateral axes of the UAV body coordinate system. The direction of the desired yaw angle is determined based on the direction of the lateral axis component, and the magnitude of the desired yaw angle is determined based on the magnitude of the lateral axis component.

8. The method for unmanned aerial vehicle (UAV) inspection based on road construction according to claim 1, characterized in that, The step of correcting the desired yaw angle based on the measured attitude data and generating a corrected yaw angle control command includes: Extract the current roll angle and current pitch angle from the measured attitude data; Calculate the wind disturbance equivalent yaw moment based on the current roll angle and the current pitch angle; The wind disturbance equivalent yaw moment is converted into a yaw angle feedforward compensation amount, and the yaw angle feedforward compensation amount is superimposed with the desired yaw angle to generate the corrected yaw angle control command.

9. A method for unmanned aerial vehicle (UAV) inspection based on road construction according to claim 1, characterized in that, The acquisition of measured attitude data during flight includes: The first attitude angle is obtained by integrating the three-axis angular velocities output by the inertial measurement unit in the UAV attitude sensor over time. The second attitude angle is obtained by correcting the three-axis acceleration output from the accelerometer in the UAV's attitude sensor for gravity direction. The first attitude angle and the second attitude angle are complementaryly corrected to generate the measured attitude data.

10. A drone inspection system for road construction, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the unmanned aerial vehicle (UAV) inspection method based on road construction as described in any one of claims 1 to 9.

Images