Complex terrain-oriented unmanned aerial vehicle cluster Beidou differential cooperative image control layout method

Abstract

The invention provides a complex terrain-oriented unmanned aerial vehicle cluster Beidou differential cooperative image control layout method, which comprises the following steps of: constructing a system consisting of an image control unmanned aerial vehicle cluster, a Beidou differential reference station and a ground control station, and uniformly calibrating space and time references; planning a photo control point and a route based on the DEM and a task demand; after the unmanned aerial vehicle flies to the target airspace, sensing the terrain by using a multi-source sensor, dynamically selecting a hovering or landing mode according to the gradient and the vegetation coverage, and executing positioning operation; beidou differential data are collected, and images are synchronously obtained; and processing data to calculate coordinates and elevation of image control points, and outputting surveying and mapping results. According to the invention, high-efficiency image control point layout under a complex terrain can be realized, the operation efficiency is improved, the labor cost and the safety risk are reduced, the high-precision measurement of the image control points is ensured, and the large-scale surveying and mapping requirements are met.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) surveying and mapping technology, and more specifically, to a method for deploying UAV swarms using BeiDou differential collaborative image control for complex terrain. Background Technology

[0002] In the field of UAV surveying, traditional ground control point (GCP) deployment methods mainly rely on manual operation. Operators must carry surveying equipment into the survey area, set markers at target locations, and complete coordinate measurements. This method is inefficient, especially in complex terrains such as mountainous areas, forests, and around water bodies, where manual passage is difficult, allowing only a small number of GCPs to be deployed per day, and carries high safety risks. With technological advancements, single-UAV-assisted GCP deployment has become a new option, but it still has many limitations, such as a limited number of targets that can be carried per trip, severe interference with positioning accuracy in areas with weak signals, and a disconnect between data acquisition and processing. Satellite positioning technology, especially BeiDou differential positioning technology, can improve positioning accuracy, but in complex terrain, the signal coverage of a single base station is limited, and the lack of differential data sharing among UAVs leads to inconsistent positioning accuracy when multiple UAVs are operating.

[0003] In implementing the embodiments of the present invention, the prior art has at least the following problems or defects: traditional methods have high labor costs, low efficiency and high safety risks, while existing UAV technology has made some improvements, but single-unit operation efficiency is low, endurance is short, positioning accuracy is greatly affected by environmental interference, and there is a lack of cluster collaborative control mechanism, making it difficult to meet the high-precision and high-efficiency surveying and mapping needs in complex terrain. Summary of the Invention

[0004] This invention provides a method for deploying BeiDou differential cooperative image control systems for UAV swarms in complex terrain, comprising: Construct a collaborative operation system consisting of an image-controlled UAV cluster, at least one BeiDou differential reference station, and a ground control station, and perform unified calibration of the spatial geometric relationship and time reference of the system; Based on the digital elevation model of the survey area and the requirements of the surveying task, the location of the image control points is planned and a collaborative flight route is planned for the image control UAV cluster. The system controls the image-controlled UAV cluster to fly to the preset image control point airspace and activates the onboard multi-source sensors to perceive the local terrain in real time in order to extract terrain feature parameters. Based on the comparison results between the terrain feature parameters and the preset threshold, the hovering operation mode or the landing operation mode is dynamically selected, and the UAV is controlled to perform the corresponding positioning operation. After the positioning operation is executed, BeiDou differential positioning data is collected, and the image acquisition unit is coordinated to perform synchronous image acquisition. The collected positioning data is processed to calculate the coordinates and elevations of the control points, and the surveying results are output.

[0005] Furthermore, the unified calibration of the spatial geometric relationships and time reference of the system includes: The image-controlled drone cluster is configured such that each drone is equipped with a high-precision GNSS receiver, an inertial measurement unit, a lidar and a multispectral camera, and an image control mark with a specific optical pattern is fixed on the lower part of the fuselage; The aforementioned BeiDou differential reference station is established to generate and broadcast real-time differential correction information; The ground control station is deployed and integrates a cluster control module, a mission planning module, and a data processing module. The spatial offset between the center of the image control marker and the phase center of the GNSS antenna in the UAV body coordinate system was determined. Perform high-precision time synchronization for all drones and base stations within the cluster; The interior orientation elements and optical distortion parameters of the multispectral camera are calibrated.

[0006] Furthermore, calibrating the spatial offset between the image control marker center and the GNSS antenna phase center in the UAV body coordinate system includes: in the calibration field, using a laser tracker to accurately measure the three-dimensional coordinates of the image control marker center and the GNSS antenna phase center in a unified measurement coordinate system when the UAV is in various preset attitudes, and calculating the fixed spatial offset vector between the two through coordinate transformation and data adjustment methods; the high-precision time synchronization of all UAVs and base stations in the cluster includes: using the high-stability crystal oscillator clock of the ground control station as the time reference, periodically sending precise time synchronization frames through a wireless communication link, and after each UAV and base station in the cluster receives the synchronization frame, calibrating its own local clock and compensating for transmission delay, so that the entire system maintains a unified time reference.

[0007] Furthermore, the step of planning the location of ground control points and the collaborative flight path for the ground control UAV swarm based on the digital elevation model of the survey area and the requirements of the surveying task includes: Based on the target scale and accuracy indicators of the surveying results, determine the layout density and distribution rules of the image control points; Import the digital elevation model of the survey area, automatically calculate the slope map based on the model, and combine it with the preset land cover type data to identify and exclude areas that do not meet the conditions for safe landing of UAVs, and generate an initial set of image control point positions. A swarm intelligence optimization algorithm is adopted, with the objective function of minimizing the total operation time of the cluster and balancing the load of each UAV. Under the premise of avoiding areas that do not meet the safe landing conditions, the initial set of control points is allocated to each UAV in the cluster, and the optimal flight path is generated for each UAV. Set a combination of terrain parameter thresholds to trigger hovering and landing operation modes.

[0008] Furthermore, the adoption of a swarm intelligence optimization algorithm, with the objective function of minimizing the total operation time of the cluster and balancing the load of each UAV, involves allocating the initial set of ground control points to each UAV within the cluster while avoiding areas that do not meet safe landing conditions, and generating the optimal flight path for each UAV. This includes: constructing a mathematical optimization model with minimizing the maximum task completion time as the primary objective and minimizing the total flight distance as the secondary objective, and using an improved ant colony algorithm for iterative solution. This algorithm takes the maximum endurance of the UAV, the spatial distribution structure of the ground control points, and the flight speed changes caused by terrain undulations as hard constraints, and finally outputs the task sequence and three-dimensional flight path for each UAV. The setting of terrain parameter threshold combinations for triggering hovering and landing operation modes includes: setting slope thresholds and vegetation cover thresholds for different geomorphic classification units within the survey area to determine the terrain complexity. These geomorphic classification units are obtained through cluster analysis based on digital elevation models and historical remote sensing images.

[0009] Furthermore, the step of activating the airborne multi-source sensor to perceive the local terrain in real time and extract terrain feature parameters includes: Control the airborne lidar to scan a predetermined area directly below the image control point to acquire high-density three-dimensional laser point cloud data; The three-dimensional laser point cloud data is denoised and interpolated to generate a local high-precision digital surface model. Based on the digital surface model, the center point slope, slope variance and surface roughness index of the landing area are calculated. The multispectral camera is controlled to image the same area simultaneously. The normalized vegetation index map of the image is calculated, and the vegetation pixels are extracted by threshold segmentation. Then, the percentage of vegetation coverage in the area is calculated.

[0010] Furthermore, the step of dynamically selecting a hovering operation mode or a landing operation mode based on the comparison result of the terrain feature parameters and a preset threshold, and controlling the UAV to perform the corresponding positioning operation includes: The real-time sensed slope value and vegetation coverage percentage are compared with the preset slope threshold and vegetation coverage threshold, respectively. If the real-time slope value is greater than the preset slope threshold, or the real-time vegetation coverage percentage is greater than the preset vegetation coverage threshold, then select the hovering operation mode, control the drone to accurately position itself at a constant safe height directly above the image control point, and activate the anti-shake positioning data acquisition program. If the real-time slope value is not greater than the preset slope threshold and the real-time vegetation coverage percentage is not greater than the preset vegetation coverage threshold, then the landing operation mode is selected, the drone is controlled to execute a graded slow descent procedure, and finally lands on the ground, and the automatic leveling system is activated to make the drone level.

[0011] Furthermore, the activation of the anti-shake positioning data acquisition program includes: in the hovering state, simultaneously recording multi-epoch BeiDou carrier phase observation data and high-frequency attitude data output by the inertial measurement unit, using a Kalman filter algorithm to perform tight combination processing on the two types of data, dynamically estimating and compensating for changes in the center position of the image control marker caused by the swaying of the UAV body, thereby obtaining high-precision static coordinates in motion; the execution of the graded slow descent program includes: the UAV descends from the cruising altitude to the first hovering altitude for preliminary terrain confirmation, then descends to the second hovering altitude for final landing confirmation, and after confirming safety, performs the final touchdown maneuver, and after touchdown, adjusts the landing gear height through the servo mechanism to make the fuselage level.

[0012] Furthermore, the synchronous image acquisition by the coordinated image acquisition unit includes: after the image-controlled UAV completes its own positioning data acquisition and enters a stable state, it sends a task ready signal containing its current precise position and unique identifier to the photogrammetric UAV performing aerial surveying tasks in the survey area through the cluster internal communication link; the photogrammetric UAV continuously listens to this signal during flight, and when its flight path is about to cover the position of the image control point, it adjusts its heading and attitude to ensure that its onboard aerial camera can clearly capture the image control mark from an orthogonal or specific tilt angle.

[0013] Furthermore, the data processing of the collected positioning data to calculate the coordinates and elevation of the control points includes: Gross errors are detected and eliminated for multiple sets of BeiDou positioning observations collected at a single control point location. Least squares estimation or robust estimation methods are used to calculate the most probable values ​​of the plane coordinates and geodetic height of the control point and their accuracy information. Using a high-precision geoid model covering the survey area, the calculated geodetic height data is converted into normal heights suitable for surveying results; The finalized control point numbers, planar coordinates, normal heights, coordinate accuracy indicators, operating modes, and acquisition timestamp information are structured, stored, and output as standard format control point result files.

[0014] The embodiments of the present invention have at least the following beneficial effects: 1. By constructing a system architecture that coordinates the operation of UAV swarms and BeiDou differential reference stations, efficient deployment of image control points in complex terrain was achieved. This method utilizes the cooperative flight capability of UAV swarms and the high-precision characteristics of BeiDou differential positioning technology to solve the problems of low efficiency and high safety risks of traditional manual deployment methods in complex terrain. At the same time, it overcomes the deficiency of insufficient positioning accuracy of single UAV operations in areas with weak signals, thereby improving operational efficiency and measurement accuracy.

[0015] 2. A location planning method for ground control points (GCPs) based on digital elevation models and surveying task requirements is adopted, combined with a swarm intelligence optimization algorithm to generate cooperative flight paths. This allows for dynamic selection of hovering or landing operation modes based on terrain features. This not only improves the flexibility and adaptability of GCP deployment but also ensures high-precision positioning operations under various terrain conditions. It effectively solves the problem that a single operation mode cannot meet surveying needs in complex terrain, thus enhancing the overall performance and reliability of the system.

[0016] 3. The system underwent unified calibration of its spatial geometry and time reference, including spatial offset calibration of image control markers and GNSS antennas, high-precision time synchronization between UAVs and base stations within the cluster, and calibration of the interior orientation elements and optical distortion parameters of the multispectral camera. These calibration measures ensured the accuracy and consistency of the measurement data, eliminated systematic errors, and thus achieved centimeter-level accuracy in calculating the coordinates and elevation of image control points. This met the high-precision requirements of large-scale surveying for image control points and overcame the problem of insufficient measurement accuracy caused by inaccurate calibration in existing technologies. Attached Figure Description

[0017] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings. Several embodiments of the invention are illustrated in the drawings by way of example and not limitation, wherein: Figure 1 This is a flowchart illustrating a method for deploying UAV swarms with BeiDou differential collaborative image control for complex terrain, provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of a working system provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of an image-controlled drone provided according to an embodiment of the present invention. Detailed Implementation

[0018] The principles and spirit of the invention will now be described with reference to several exemplary embodiments. It should be understood that these embodiments are provided merely to enable those skilled in the art to better understand and implement the invention, and are not intended to limit the scope of the invention in any way. Rather, these embodiments are provided to make the invention more thorough and complete, and to fully convey the scope of the invention to those skilled in the art.

[0019] Those skilled in the art will recognize that embodiments of the present invention can be implemented as a system, apparatus, device, method, or computer program product. Therefore, the present invention can be specifically implemented in the following forms: entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0020] It should be noted that the number of any elements in the accompanying drawings is for illustrative purposes only and not as a limitation, and any naming is for distinction only and has no limiting meaning.

[0021] The following is for reference. Figure 1 , Figure 1 This is a flowchart illustrating a method for deploying UAV swarms using BeiDou differential collaborative image control in complex terrain, as provided in an embodiment of the present invention. Figure 1 As shown, a method for deploying UAV swarms using BeiDou differential collaborative image control for complex terrain includes: S1. Construct a collaborative operation system consisting of an image-controlled UAV cluster, at least one BeiDou differential reference station, and a ground control station, and perform unified calibration of the spatial geometric relationship and time reference of the system; S2. Based on the digital elevation model of the survey area and the requirements of the surveying task, plan the location of the image control points and plan the collaborative flight route for the image control UAV cluster. S3. Control the image-controlled UAV cluster to fly to the preset image control point airspace, and activate the onboard multi-source sensor to perceive the local terrain in real time to extract terrain feature parameters; S4. Based on the comparison results between the terrain feature parameters and the preset threshold, dynamically select the hovering operation mode or the landing operation mode, and control the UAV to perform the corresponding positioning operation. S5. After the positioning operation is executed, collect BeiDou differential positioning data and coordinate the image acquisition unit to perform synchronous image acquisition. S6. Process the collected positioning data, calculate the coordinates and elevations of the control points, and output the surveying results.

[0022] like Figure 2As shown, this invention constructs a collaborative operation system to achieve efficient deployment of image control points (ARPCs) in complex terrain using a swarm of unmanned aerial vehicles (UAVs). An image control UAV swarm refers to a group of multiple UAVs, each equipped with high-precision positioning and sensing equipment, used to perform image control point deployment tasks at designated locations. The BeiDou differential reference station is a ground facility that provides high-precision differential positioning information, significantly improving the accuracy of UAV positioning. The ground control station serves as the command center of the entire system, responsible for mission planning, monitoring flight status, and processing collected data.

[0023] like Figure 3 As shown, each drone in the image-controlled drone swarm is equipped with a high-precision GNSS receiver, inertial measurement unit, lidar, and multispectral camera. These devices are used to acquire precise position information, measure the drone's attitude, perceive the three-dimensional structure of the terrain, and acquire the spectral information of the ground surface, respectively. Image control markers are fixed to the underside of the drones and have specific optical patterns, serving as reference points during image acquisition. The BeiDou differential reference station receives BeiDou satellite signals and performs differential processing to generate and broadcast real-time differential correction information to improve the accuracy of drone positioning. The ground control station integrates a swarm control module, a mission planning module, and a data processing module for unified management and scheduling of the drone swarm's operations. When planning the placement of image control points, the density and distribution rules of the image control points are determined based on the target scale and accuracy indicators of the survey results. Then, the digital elevation model (DEM) of the survey area is imported, and the slope map is automatically calculated based on this model. Combined with preset land cover type data, areas that do not meet the safe landing conditions for drones are identified and excluded, generating an initial set of image control point locations.

[0024] During system calibration, a laser tracker is used to accurately measure the three-dimensional coordinates of the image control point center and the GNSS antenna phase center in a unified measurement coordinate system when the UAV is in various preset attitudes. Through coordinate transformation and data adjustment methods, the fixed spatial offset vector between the two is calculated. For time synchronization, a high-stability crystal oscillator clock from the ground control station is used as the time reference. Precise time synchronization frames are periodically sent via a wireless communication link. Each UAV in the cluster and the base station receive this synchronization frame, calibrate their local clock, and compensate for transmission delays to maintain a unified time reference for the entire system. When planning the deployment locations of image control points, a swarm intelligence optimization algorithm is used to construct a mathematical optimization model with minimizing the maximum mission completion time as the primary objective and minimizing the total flight distance as a secondary objective. The maximum flight range of the UAV, the spatial distribution structure of the image control points, and the flight speed changes caused by terrain undulations are used as hard constraints. Finally, the mission sequence and three-dimensional flight path for each UAV are output. In the data processing stage, gross errors are detected and eliminated from multiple sets of BeiDou positioning observations collected at a single control point location. Least squares estimation or robust estimation methods are used to calculate the most probable values ​​of the plane coordinates and geodetic height of the control point and their accuracy information. Then, using a high-precision geoid model covering the survey area, the calculated geodetic height data is converted into normal heights suitable for surveying results.

[0025] In some embodiments, the unified calibration of the system's spatial geometric relationships and time reference includes: The image-controlled drone cluster is configured such that each drone is equipped with a high-precision GNSS receiver, an inertial measurement unit, a lidar and a multispectral camera, and an image control mark with a specific optical pattern is fixed on the lower part of the fuselage; The aforementioned BeiDou differential reference station is established to generate and broadcast real-time differential correction information; The ground control station is deployed and integrates a cluster control module, a mission planning module, and a data processing module. The spatial offset between the center of the image control marker and the phase center of the GNSS antenna in the UAV body coordinate system was determined. Perform high-precision time synchronization for all drones and base stations within the cluster; The interior orientation elements and optical distortion parameters of the multispectral camera are calibrated.

[0026] The spatial offset between the center of the geocontrol point and the phase center of the GNSS antenna refers to the relative positional relationship between the geometric center of the geocontrol point and the phase center of the GNSS antenna on the UAV in the UAV's body coordinate system. This parameter is crucial for accurately calculating the coordinates of the geocontrol point. High-precision time synchronization ensures that all UAVs and base stations within the cluster operate under a unified time reference, which is critical for the consistency of collaborative operations and data processing. The calibration of the interior orientation elements and optical distortion parameters of the multispectral camera is to correct camera imaging errors and improve the quality and usability of image data.

[0027] In a field-controlled UAV swarm configuration, each UAV carries a high-precision GNSS receiver to acquire accurate geographic location information, an inertial measurement unit (IMU) to measure the UAV's attitude and motion, a lidar to acquire 3D point cloud data of the terrain, and a multispectral camera to acquire multispectral images of the ground surface. Field control markers, with specific optical patterns, are fixed to the underside of the UAVs and serve as reference points during image acquisition. The BeiDou differential reference station receives BeiDou satellite signals, performs differential processing, generates and broadcasts real-time differential correction information to improve UAV positioning accuracy. The ground control station integrates a swarm control module, a mission planning module, and a data processing module for unified management and scheduling of the UAV swarm's operations.

[0028] During calibration, a laser tracker is used to precisely measure the three-dimensional coordinates of the image control marker center and the GNSS antenna phase center in a unified measurement coordinate system when the UAV is in various preset attitudes within the calibration field. Through coordinate transformation and data adjustment methods, the fixed spatial offset vector between the two is calculated. For time synchronization, a high-stability crystal oscillator clock from the ground control station is used as the time reference. Precise time synchronization frames are periodically sent via a wireless communication link. Each UAV in the cluster and the base station receive this synchronization frame, calibrate their local clock, and compensate for transmission delays to maintain a unified time reference for the entire system. For the calibration of the multispectral camera, a checkerboard calibration method is used to determine the camera's internal orientation elements, such as focal length, principal point coordinates, and optical distortion parameters, to ensure the accuracy and consistency of the image data.

[0029] When calibrating the spatial offset between the center of the image control marker and the phase center of the GNSS antenna, multiple measurement points with known coordinates can be set up in the calibration field. Using a laser tracker, the coordinates of the image control marker and the phase center of the GNSS antenna relative to these measurement points are measured under different UAV attitudes. Then, through coordinate transformation and least squares methods, the fixed spatial offset vector between the two can be accurately calculated. During time synchronization, a time synchronization accuracy detection mechanism can be set up. By comparing the timestamp differences between each UAV and the base station, the time synchronization strategy can be adjusted in real time to ensure that the system's time synchronization accuracy is within 10 milliseconds. For the calibration of the multispectral camera, multiple calibration experiments can be conducted under different ambient lighting conditions to obtain more accurate interior orientation elements and optical distortion parameters. Simultaneously, a camera calibration error model can be established for real-time correction of image data during actual operations, improving the quality and reliability of the image data.

[0030] In some embodiments, calibrating the spatial offset between the image control marker center and the GNSS antenna phase center in the UAV body coordinate system includes: in the calibration field, using a laser tracker to accurately measure the three-dimensional coordinates of the image control marker center and the GNSS antenna phase center in a unified measurement coordinate system when the UAV is in various preset attitudes, and calculating the fixed spatial offset vector between the two through coordinate transformation and data adjustment methods; the high-precision time synchronization of all UAVs and base stations in the cluster includes: using the high-stability crystal oscillator clock of the ground control station as the time reference, periodically sending precise time synchronization frames through a wireless communication link, and after each UAV and base station in the cluster receives the synchronization frame, calibrating its own local clock and compensating for transmission delay, so that the entire system maintains a unified time reference.

[0031] The spatial offset between the image control marker center and the GNSS antenna phase center in the UAV's body coordinate system was calibrated. This was achieved by measuring relevant data of the UAV under various preset attitudes using a laser tracker within the calibration field, and then calculating the fixed spatial offset vector between the two using coordinate transformation and data adjustment methods, thus ensuring measurement accuracy. Simultaneously, high-precision time synchronization was performed on all UAVs and base stations within the cluster to ensure the entire system operates under a unified time reference, which is crucial for collaborative operations and data processing consistency. Furthermore, the interior orientation elements and optical distortion parameters of the multispectral camera were calibrated to correct camera imaging errors and improve the quality and usability of image data.

[0032] During calibration, a laser tracker is used to precisely measure the three-dimensional coordinates of the image control marker center and the GNSS antenna phase center in a unified measurement coordinate system when the UAV is in various preset attitudes within the calibration field. A fixed spatial offset vector between the two is calculated using coordinate transformation and data adjustment methods. For time synchronization, high-precision time synchronization refers to using the high-stability crystal oscillator clock of the ground control station as the time reference. Precise time synchronization frames are periodically sent via a wireless communication link. Each UAV in the cluster and the base station receive this synchronization frame, calibrate their local clock, and compensate for transmission delays, ensuring the entire system maintains a unified time reference. For the calibration of the interior orientation elements and optical distortion parameters of the multispectral camera, a checkerboard calibration method is used to determine the camera's interior orientation elements, such as focal length, principal point coordinates, and optical distortion parameters, to ensure the accuracy and consistency of the image data.

[0033] When calibrating the spatial offset between the center of the image control marker and the phase center of the GNSS antenna, multiple measurement points with known coordinates can be set up in the calibration field. Using a laser tracker, the coordinates of the image control marker and the phase center of the GNSS antenna relative to these measurement points are measured under different UAV attitudes. Then, through coordinate transformation and least squares methods, the fixed spatial offset vector between the two can be accurately calculated. During time synchronization, a time synchronization accuracy detection mechanism can be set up. By comparing the timestamp differences between each UAV and the base station, the time synchronization strategy can be adjusted in real time to ensure that the system's time synchronization accuracy is within 10 milliseconds. For the calibration of the multispectral camera, multiple calibration experiments can be conducted under different ambient lighting conditions to obtain more accurate interior orientation elements and optical distortion parameters. Simultaneously, a camera calibration error model can be established for real-time correction of image data during actual operations, improving the quality and reliability of the image data.

[0034] In some embodiments, planning the location of ground control points and planning cooperative flight routes for the ground control UAV cluster based on the digital elevation model of the survey area and the requirements of the surveying task includes: Based on the target scale and accuracy indicators of the surveying results, determine the layout density and distribution rules of the image control points; Import the digital elevation model of the survey area, automatically calculate the slope map based on the model, and combine it with the preset land cover type data to identify and exclude areas that do not meet the conditions for safe landing of UAVs, and generate an initial set of image control point positions. A swarm intelligence optimization algorithm is adopted, with the objective function of minimizing the total operation time of the cluster and balancing the load of each UAV. Under the premise of avoiding areas that do not meet the safe landing conditions, the initial set of control points is allocated to each UAV in the cluster, and the optimal flight path is generated for each UAV. Set a combination of terrain parameter thresholds to trigger hovering and landing operation modes.

[0035] It should be noted that, in this invention, based on the Digital Elevation Model (DEM) of the survey area and the requirements of the surveying task, the planning of control point (COP) placement locations and the planning of collaborative flight routes for the COP swarm are aimed at ensuring that the UAVs can efficiently and safely complete the COP placement task in complex terrain. This process involves multiple steps, including determining the COP placement density and distribution rules, generating an initial set of COP locations, using a swarm intelligence optimization algorithm to allocate COPs and generate the optimal flight path, and setting terrain parameter threshold combinations. These steps together ensure that the UAV swarm can operate efficiently in complex terrain, while guaranteeing the accuracy and reliability of COP placement.

[0036] Specifically, the density and distribution rules of control points (COPs) refer to determining the distribution of COPs within the survey area based on the target scale and accuracy indicators of the survey results. For example, for a 1:2000 scale survey task, the spacing between COPs should not exceed 1000 meters. After importing the Digital Elevation Model (DEM) of the survey area, the slope map is automatically calculated based on the model, and combined with preset land cover type data, areas that do not meet the safe landing conditions for UAVs are identified and excluded, generating an initial set of COP locations. In this process, the slope map is obtained through DEM calculation and is used to represent the slope of each point in the survey area; land cover type data refers to the land cover conditions of different areas within the survey area, such as vegetation cover and water bodies. The swarm intelligence optimization algorithm is an optimization algorithm that simulates the behavior of biological swarms. It is used to allocate the initial set of COPs to each UAV in the swarm while avoiding areas that do not meet the safe landing conditions, and to generate the optimal flight path for each UAV. The objective function of this algorithm is to minimize the total operation time of the swarm and balance the load of each UAV. The terrain parameter threshold combination refers to the terrain parameter thresholds used to trigger hovering and landing operation modes, such as slope threshold and vegetation coverage threshold.

[0037] When planning the location of ground control points (GCPs), the resolution of the Digital Elevation Model (DEM) should not exceed 5 meters to ensure accurate representation of terrain features. When generating slope maps, slope analysis tools in GIS software can be used; input DEM data and output slope maps. For land cover type data, historical remote sensing imagery and field survey data can be combined to obtain vegetation cover and water distribution for different landform classification units through cluster analysis. When employing swarm intelligence optimization algorithms, a mathematical optimization model can be constructed with minimizing the maximum task completion time as the primary objective and minimizing the total flight distance as a secondary objective. The model's input parameters include the UAV's maximum endurance, the spatial distribution structure of GCPs, and flight speed variations caused by terrain undulations. An improved ant colony algorithm is used for iterative solving, ultimately outputting the task sequence and 3D flight path for each UAV. When setting terrain parameter threshold combinations, slope and vegetation cover thresholds can be set separately for different landform classification units within the survey area to determine terrain complexity. For example, for mountainous areas, the slope threshold can be set to 25°, and the vegetation cover threshold to 60%. These thresholds can be obtained through field testing and historical data analysis to ensure the safety and efficiency of drone operations under different terrain conditions.

[0038] In some embodiments, the adoption of a swarm intelligence optimization algorithm, with the objective function of minimizing the total operation time of the cluster and balancing the load of each UAV, involves allocating an initial set of ground control points to each UAV within the cluster and generating an optimal flight path for each UAV, while avoiding areas that do not meet safe landing conditions. This includes: constructing a mathematical optimization model with minimizing the maximum task completion time as the primary objective and minimizing the total flight distance as the secondary objective; using an improved ant colony algorithm for iterative solution; this algorithm takes the maximum endurance of the UAV, the spatial distribution structure of the ground control points, and the flight speed changes caused by terrain undulations as hard constraints; and finally outputting the task sequence and three-dimensional flight path for each UAV. The setting of terrain parameter threshold combinations for triggering hovering and landing operation modes includes: setting slope thresholds and vegetation cover thresholds for different geomorphic classification units within the survey area to determine the terrain complexity. These geomorphic classification units are obtained through cluster analysis based on digital elevation models and historical remote sensing images.

[0039] By constructing a mathematical optimization model to allocate control points and generate optimal flight paths for UAVs, the objective function is to minimize the total cluster operation time and balance the load on each UAV. This ensures that UAVs complete the control point deployment task while avoiding areas that do not meet safe landing conditions. Furthermore, by setting a combination of terrain parameter thresholds to trigger hovering and landing operation modes, the adaptability and operational efficiency of UAVs under different terrain conditions are further improved.

[0040] Specifically, swarm intelligence optimization algorithms are optimization algorithms that simulate the behavior of biological groups, used to find optimal solutions in complex environments. In this invention, the objective function of the algorithm is to minimize the total operation time of the swarm and balance the load of each UAV. The constructed mathematical optimization model prioritizes minimizing the maximum task completion time, with the shortest total flight range as a secondary objective. Input parameters include the maximum endurance of the UAVs, the spatial distribution structure of the control points, and the flight speed changes caused by terrain undulations. These parameters ensure that the model can comprehensively consider the operational capabilities of the UAVs and terrain conditions. The terrain parameter threshold combination refers to the slope critical value and vegetation cover critical value used to judge the terrain complexity. These thresholds are obtained through cluster analysis to distinguish different landform classification units, ensuring that the UAVs select appropriate operation modes under safe conditions.

[0041] When constructing the mathematical optimization model, an improved ant colony algorithm can be used for iterative solution. Input parameters include the maximum endurance of the UAV, the spatial distribution structure of the control points, and the flight speed changes caused by terrain undulations. In each iteration, the algorithm updates the mission sequence and 3D flight path of each UAV based on the current pheromone distribution and heuristic information. To improve the convergence speed and solution quality of the algorithm, a mechanism for dynamically adjusting the pheromone evaporation rate and heuristic factor can be introduced. When setting the combination of terrain parameter thresholds, slope and vegetation cover thresholds can be set separately for different geomorphic classification units within the survey area to determine terrain complexity. For example, for mountainous areas, the slope threshold can be set to 25°, and the vegetation cover threshold can be set to 60%. These thresholds can be obtained through field testing and historical data statistics to ensure the safety and efficiency of UAV operations under different terrain conditions. In practical applications, these parameters can be adjusted and optimized according to specific task requirements and terrain conditions to achieve the best operational results.

[0042] In some embodiments, activating the airborne multi-source sensor to perceive the local terrain in real time to extract terrain feature parameters includes: Control the airborne lidar to scan a predetermined area directly below the image control point to acquire high-density three-dimensional laser point cloud data; The three-dimensional laser point cloud data is denoised and interpolated to generate a local high-precision digital surface model. Based on the digital surface model, the center point slope, slope variance and surface roughness index of the landing area are calculated. The multispectral camera is controlled to image the same area simultaneously. The normalized vegetation index map of the image is calculated, and the vegetation pixels are extracted by threshold segmentation. Then, the percentage of vegetation coverage in the area is calculated.

[0043] It should be noted that the real-time perception of local terrain by airborne multi-source sensors in this invention to extract terrain feature parameters is for the purpose of dynamically selecting appropriate operating modes, such as hovering or landing, to ensure the safe and efficient operation of the UAV in complex terrain. This process uses airborne LiDAR and multispectral cameras to acquire the three-dimensional structure and vegetation cover information of the terrain, and then calculates terrain feature parameters, such as slope and vegetation cover, to provide a basis for subsequent operating mode selection. In this way, the UAV can flexibly adjust its operating strategy based on real-time terrain information, improving the adaptability and reliability of its operations.

[0044] Specifically, the airborne multi-source sensors include lidar and a multispectral camera. LiDAR scans a predetermined area directly below the ground control point, acquiring high-density 3D laser point cloud data. After denoising and interpolation, this data generates a high-precision local digital surface model, which is then used to calculate the center-point slope, slope variance, and surface roughness index of the landing area. The multispectral camera simultaneously images the same area, calculating the normalized vegetation index (NDI) map of the image and extracting vegetation pixels using threshold segmentation to calculate the percentage of vegetation cover in the area. These terrain feature parameters provide crucial information for the UAV to select appropriate operational modes. Real-time perception means that sensors continuously acquire data during the UAV's flight, ensuring that the UAV can make decisions based on the latest terrain information. Terrain feature parameters include slope and vegetation cover, which reflect the complexity of the terrain and the feasibility of the operation.

[0045] When initiating an airborne lidar scan, the scanning range can be set to a 10m x 10m area directly below the ground control point to ensure sufficient terrain information is acquired. The 3D lidar point cloud data acquired by the lidar needs to undergo denoising processing to remove outliers and noise points, and then a high-precision digital surface model is generated using an interpolation algorithm. When calculating slope, a difference method can be used to calculate the slope value at each point, and the average slope and slope variance for the entire area can be calculated. For multispectral cameras, the imaging range can be set to match the lidar scanning range, and vegetation cover can be assessed by calculating the Normalized Difference Vegetation Index (NDVI). NDVI is calculated based on the reflectivity of red and near-infrared light bands, effectively reflecting vegetation growth and coverage. By setting a vegetation cover threshold, the area can be divided into vegetated and non-vegetated areas. In practical applications, these parameters can be adjusted and optimized according to specific task requirements and terrain conditions to achieve the best operational results.

[0046] In some embodiments, dynamically selecting a hovering operation mode or a landing operation mode based on the comparison result of the terrain feature parameters and a preset threshold, and controlling the UAV to perform the corresponding positioning operation includes: The real-time sensed slope value and vegetation coverage percentage are compared with the preset slope threshold and vegetation coverage threshold, respectively. If the real-time slope value is greater than the preset slope threshold, or the real-time vegetation coverage percentage is greater than the preset vegetation coverage threshold, then select the hovering operation mode, control the drone to accurately position itself at a constant safe height directly above the image control point, and activate the anti-shake positioning data acquisition program. If the real-time slope value is not greater than the preset slope threshold and the real-time vegetation coverage percentage is not greater than the preset vegetation coverage threshold, then the landing operation mode is selected, the drone is controlled to execute a graded slow descent procedure, and finally lands on the ground, and the automatic leveling system is activated to make the drone level.

[0047] Specifically, terrain feature parameters refer to terrain information acquired and calculated through airborne multi-source sensors, such as slope values ​​and vegetation coverage percentage. These parameters reflect the complexity of the terrain and the feasibility of operations. Preset thresholds are parameter values ​​pre-set based on terrain conditions to determine whether the terrain is suitable for landing operations. For example, the slope threshold can be set to 25°, and the vegetation coverage threshold can be set to 60%. If the real-time monitored slope value is greater than the preset slope threshold, or the vegetation coverage percentage is greater than the preset vegetation coverage threshold, the hovering operation mode is selected; otherwise, the landing operation mode is selected. The hovering operation mode refers to the UAV maintaining a constant safe altitude directly above the control point to collect positioning data; while the landing operation mode refers to the UAV executing a graded, slow descent procedure, eventually landing on the ground and activating the automatic leveling system to level the aircraft.

[0048] In hovering mode, the UAV hovers at a height of 8-10 meters directly above the ground control point (GCP), maintaining attitude stability via the inertial measurement unit (IMU) to ensure attitude fluctuations do not exceed 0.5°. At this time, the anti-shake positioning data acquisition program is activated, simultaneously recording multi-epoch BeiDou carrier phase observation data and high-frequency attitude data output by the IMU. A Kalman filter algorithm is used to tightly combine these two types of data, dynamically estimating and compensating for changes in the GCP center position caused by UAV body sway, thereby obtaining high-precision static coordinates while in motion. In landing mode, the UAV descends from its cruising altitude to a first hovering height (e.g., 5 meters) for preliminary terrain confirmation, then descends to a second hovering height (e.g., 2 meters) for final landing confirmation. After safety is confirmed, the final touchdown maneuver is performed. Upon touchdown, the landing gear height is adjusted via a servo mechanism to achieve a level fuselage with a horizontal error not exceeding 0.1°. These specific operational steps and parameter settings ensure the safety and accuracy of the UAV's operations under different terrain conditions, improving the adaptability and reliability of the entire system.

[0049] In some embodiments, the activation of the anti-shake positioning data acquisition program includes: in the hovering state, simultaneously recording multi-epoch BeiDou carrier phase observation data and high-frequency attitude data output by the inertial measurement unit, using a Kalman filter algorithm to perform tight combination processing on the two types of data, dynamically estimating and compensating for changes in the center position of the image control marker caused by the shaking of the UAV body, thereby obtaining high-precision static coordinates in motion; the execution of the graded slow descent program includes: the UAV descends from the cruising altitude to the first hovering altitude for preliminary terrain confirmation, then descends to the second hovering altitude for final landing confirmation, and after confirming safety, performs the final touchdown maneuver, and after touchdown, adjusts the landing gear height through the servo mechanism to make the fuselage level.

[0050] Specifically, the anti-shake positioning data acquisition program refers to the simultaneous recording of multi-epoch BeiDou carrier phase observation data and high-frequency attitude data output by the IMU while the UAV is hovering. This data is processed using a Kalman filter algorithm to dynamically estimate and compensate for changes in the center position of the image control marker caused by UAV body sway. The Kalman filter algorithm is a recursive filter capable of processing observation and prediction data in real time, providing optimal estimates. The staged slow descent program refers to the process of the UAV descending to the ground in stages from its cruising altitude. First, the UAV descends to a first hovering altitude for preliminary terrain confirmation, then descends to a second hovering altitude for final landing confirmation, and performs the final touchdown maneuver after safety is confirmed. After touchdown, the landing gear height is adjusted via a servo mechanism to level the fuselage. The servo mechanism is a device that automatically adjusts the position of mechanical components based on input signals to ensure the fuselage is level after landing.

[0051] When implementing the anti-shake positioning data acquisition procedure, initial parameters for the Kalman filter algorithm can be set, including initial values ​​for the state vector and covariance matrix. The state vector includes the position coordinates of the image control marker center and the attitude angles of the IMU, while the covariance matrix reflects the uncertainty of these state variables. In each sampling period, the Kalman filter algorithm updates the state vector and covariance matrix based on observed and predicted data, thus providing high-precision position estimation. When implementing the graded slow descent procedure, the first hovering height can be set to 5 meters, and the second hovering height to 2 meters. During each hovering phase, the UAV acquires terrain information using lidar and a multispectral camera to confirm terrain flatness and safety. After touchdown, the servo mechanism adjusts the landing gear height based on feedback signals from the fuselage level sensor, ensuring that the fuselage level error does not exceed 0.1°. These specific operational steps and parameter settings ensure high-precision positioning and safe landing of the UAV in complex terrain, improving the adaptability and reliability of the entire system.

[0052] In some embodiments, the coordinated image acquisition unit performs synchronous image acquisition by: after the image-controlled UAV completes its own positioning data acquisition and enters a stable state, it sends a task ready signal containing its current precise position and unique identifier to the photogrammetric UAV performing aerial surveying tasks in the survey area through the cluster internal communication link; the photogrammetric UAV continuously listens to this signal during flight, and when its flight path is about to cover the position of the image control point, it adjusts its heading and attitude to ensure that its onboard aerial camera can clearly capture the image control mark from an orthogonal or specific tilt angle.

[0053] Specifically, coordinated image acquisition by the image acquisition unit refers to the process where, after the image-controlled UAV completes its own positioning data acquisition and enters a stable state, it sends a mission-ready signal containing its current precise location and unique identifier to the photogrammetric UAV performing aerial surveying tasks over the survey area via the cluster's internal communication link. The photogrammetric UAV continuously monitors this signal during flight, and when its flight path is about to cover the image control point's location, it adjusts its heading and attitude to ensure that its onboard aerial camera can clearly capture the image control point from an orthogonal or specific tilt angle. The cluster's internal communication link refers to the wireless link used for data transmission and communication between UAVs, ensuring real-time information transmission. The mission-ready signal contains the image-controlled UAV's current precise location and unique identifier, used to notify the photogrammetric UAV that the image control point is ready for image acquisition. An orthogonal or specific tilt angle refers to the camera angle at which the photogrammetric UAV photographs the image control point. An orthogonal angle means the camera is perpendicular to the ground, while a specific tilt angle means the camera is tilted at a certain angle to obtain more terrain information.

[0054] When implementing synchronized image acquisition using a coordinated image acquisition unit, the image-controlled UAV can be configured to send a task-ready signal via its wireless communication module after completing positioning data acquisition. This signal contains the UAV's precise location coordinates, such as longitude, latitude, and altitude, and a unique identifier, such as a serial number. During flight, the photogrammetric UAV receives this signal through its communication module and adjusts its flight path and camera attitude accordingly. For example, if the image control point is located in a mountainous area, the photogrammetric UAV can adjust to a specific tilt angle to capture more terrain details. During this adjustment process, the photogrammetric UAV can use its onboard Inertial Measurement Unit (IMU) and Global Positioning System (GPS) module to precisely control the camera's attitude and position. After acquisition, the image data is transmitted back to the ground control station in real time via a wireless communication link for subsequent data processing and analysis. This synchronized acquisition mechanism not only improves the accuracy of image data but also reduces errors caused by time differences, thereby enhancing the overall quality of the surveying and mapping results.

[0055] In some embodiments, the data processing of the collected positioning data to calculate the coordinates and elevation of the control points includes: Gross errors are detected and eliminated for multiple sets of BeiDou positioning observations collected at a single control point location. Least squares estimation or robust estimation methods are used to calculate the most probable values ​​of the plane coordinates and geodetic height of the control point and their accuracy information. Using a high-precision geoid model covering the survey area, the calculated geodetic height data is converted into normal heights suitable for surveying results; The finalized control point numbers, planar coordinates, normal heights, coordinate accuracy indicators, operating modes, and acquisition timestamp information are structured, stored, and output as standard format control point result files.

[0056] Gross error detection and removal refers to the process of first performing a quality check on multiple consecutive sets of BeiDou positioning observations during positioning data processing, removing outliers or gross errors to improve data accuracy and reliability. Least squares estimation or robust estimation methods are used to calculate the most probable values ​​and accuracy information of the plane coordinates and geodetic height of control points. Least squares estimation is a commonly used mathematical optimization method that solves for the optimal solution by minimizing the sum of squared errors; robust estimation is an estimation method with stronger robustness to outliers. A high-precision quasi-geoid model is a mathematical model used to convert geodetic height to orthographic height, reflecting the elevation difference between the geoid and the reference ellipsoid. This model allows the calculated geodetic height data to be converted into orthographic height suitable for surveying results. Structured storage and output of control point result files in a standard format refers to storing and outputting information such as the control point number, plane coordinates, orthographic height, coordinate accuracy indicators, operating mode used, and acquisition timestamp in a standardized format for easy use in subsequent surveying work.

[0057] When performing data processing, a gross error detection threshold can be set. For example, if the deviation of a set of BeiDou positioning observations from the rest exceeds a preset threshold, it is identified as a gross error and removed. During least squares estimation, an error equation system encompassing all observations can be constructed. Solving this system yields the optimal coordinates and elevation of the control points. For robust estimation, an iterative reweighted least squares method can be used to gradually reduce the impact of outliers, thereby obtaining more reliable estimation results. When using a high-precision quasi-geoid model for elevation transformation, a suitable model can be selected based on the specific location of the survey area, and the geodetic coordinates and geodetic height of the control points can be input to calculate the corresponding normal height.

[0058] Finally, detailed information on all geocontrol points is stored and output according to a preset standard format, such as a CSV or XML file. This file includes the geocontrol point's number, plane coordinates (X, Y), normal elevation (H), coordinate accuracy indicators such as horizontal mean square error and vertical mean square error, the operating mode used (hovering or landing mode), and the data acquisition timestamp. This detailed and standardized output method not only facilitates data management and use but also helps improve the efficiency and quality of surveying work.

[0059] The above description is merely an explanation of some preferred embodiments of the present invention and the technical principles employed. Those skilled in the art should understand that the scope of the invention as described in the embodiments of the present invention is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. A method for deploying UAV swarms using BeiDou differential collaborative image control for complex terrain, characterized in that, include: Construct a collaborative operation system consisting of an image-controlled UAV cluster, at least one BeiDou differential reference station, and a ground control station, and perform unified calibration of the spatial geometry and time reference of the system; Based on the digital elevation model of the survey area and the requirements of the surveying task, the location of the image control points is planned and a collaborative flight route is planned for the image control UAV cluster. The system controls the image-controlled UAV cluster to fly to the preset image control point airspace and activates the onboard multi-source sensors to perceive the local terrain in real time in order to extract terrain feature parameters. Based on the comparison results between the terrain feature parameters and the preset threshold, the hovering operation mode or the landing operation mode is dynamically selected, and the UAV is controlled to perform the corresponding positioning operation. After the positioning operation is executed, BeiDou differential positioning data is collected, and the image acquisition unit is coordinated to perform synchronous image acquisition. The collected positioning data is processed to calculate the coordinates and elevations of the control points, and the surveying results are output.

2. The method according to claim 1, characterized in that, The unified calibration of the spatial geometric relationships and time reference of the system includes: The image-controlled drone cluster is configured such that each drone is equipped with a high-precision GNSS receiver, an inertial measurement unit, a lidar and a multispectral camera, and an image control mark with a specific optical pattern is fixed on the lower part of the fuselage; The aforementioned BeiDou differential reference station is established to generate and broadcast real-time differential correction information; The ground control station is deployed and integrates a cluster control module, a mission planning module, and a data processing module. The spatial offset between the center of the image control marker and the phase center of the GNSS antenna in the UAV body coordinate system was determined. Perform high-precision time synchronization for all drones and base stations within the cluster; The interior orientation elements and optical distortion parameters of the multispectral camera are calibrated.

3. The method according to claim 2, characterized in that, The calibration of the spatial offset between the image control marker center and the GNSS antenna phase center in the UAV body coordinate system includes: in the calibration field, using a laser tracker to accurately measure the three-dimensional coordinates of the image control marker center and the GNSS antenna phase center in a unified measurement coordinate system when the UAV is in various preset attitudes, and calculating the fixed spatial offset vector between the two through coordinate transformation and data adjustment methods; the high-precision time synchronization of all UAVs and base stations in the cluster includes: using the high-stability crystal oscillator clock of the ground control station as the time reference, periodically sending precise time synchronization frames through a wireless communication link, and after each UAV and base station in the cluster receives the synchronization frame, calibrating its own local clock and compensating for transmission delay, so that the entire system maintains a unified time reference.

4. The method according to claim 1, characterized in that, The process of planning the location of ground control points and planning collaborative flight routes for the ground control UAV swarm based on the digital elevation model of the survey area and the requirements of the surveying task includes: Based on the target scale and accuracy indicators of the surveying and mapping results, determine the layout density and distribution rules of the image control points; Import the digital elevation model of the survey area, automatically calculate the slope map based on the model, and combine it with the preset land cover type data to identify and exclude areas that do not meet the conditions for safe landing of UAVs, and generate an initial set of image control point positions. A swarm intelligence optimization algorithm is adopted, with the objective function of minimizing the total operation time of the cluster and balancing the load of each UAV. Under the premise of avoiding areas that do not meet the safe landing conditions, the initial set of control points is allocated to each UAV in the cluster, and the optimal flight path is generated for each UAV. Set a combination of terrain parameter thresholds to trigger hovering and landing operation modes.

5. The method according to claim 4, characterized in that, The proposed swarm intelligence optimization algorithm aims to minimize the total operation time of the cluster and balance the load of each UAV. While avoiding areas that do not meet safe landing conditions, it allocates the initial set of control points to each UAV within the cluster and generates the optimal flight path for each UAV. This includes constructing a mathematical optimization model with minimizing the maximum task completion time as the primary objective and minimizing the total flight distance as the secondary objective. An improved ant colony algorithm is used for iterative solution. This algorithm takes the maximum flight range of the UAV, the spatial distribution structure of the control points, and the flight speed changes caused by terrain undulations as hard constraints, ultimately outputting the task sequence and three-dimensional flight path for each UAV. The setting of terrain parameter threshold combinations to trigger hovering and landing operation modes includes setting slope and vegetation coverage thresholds for different geomorphic classification units within the survey area, respectively, to determine the terrain complexity. These geomorphic classification units are obtained through cluster analysis based on digital elevation models and historical remote sensing imagery.

6. The method according to claim 1, characterized in that, The activation of the airborne multi-source sensor to perceive the local terrain in real time and extract terrain feature parameters includes: Control the airborne lidar to scan a predetermined area directly below the image control point to acquire high-density three-dimensional laser point cloud data; The three-dimensional laser point cloud data is denoised and interpolated to generate a local high-precision digital surface model. Based on the digital surface model, the center point slope, slope variance and surface roughness index of the landing area are calculated. The multispectral camera is controlled synchronously to image the same area. The normalized vegetation index map of the image is calculated, and the vegetation pixels are extracted by threshold segmentation. Then, the percentage of vegetation coverage in the area is calculated.

7. The method according to claim 1, characterized in that, The step of dynamically selecting a hovering operation mode or a landing operation mode based on the comparison result of the terrain feature parameters and a preset threshold, and controlling the UAV to perform the corresponding positioning operation includes: The real-time sensed slope value and vegetation coverage percentage are compared with the preset slope threshold and vegetation coverage threshold, respectively. If the real-time slope value is greater than the preset slope threshold, or the real-time vegetation coverage percentage is greater than the preset vegetation coverage threshold, then select the hovering operation mode, control the drone to accurately position itself at a constant safe height directly above the image control point, and activate the anti-shake positioning data acquisition program. If the real-time slope value is not greater than the preset slope threshold and the real-time vegetation coverage percentage is not greater than the preset vegetation coverage threshold, then select the landing operation mode, control the drone to execute a graded slow descent procedure, finally land on the ground, and activate the automatic leveling system to make the drone level.

8. The method according to claim 7, characterized in that, The activation of the anti-shake positioning data acquisition program includes: in the hovering state, simultaneously recording multi-epoch BeiDou carrier phase observation data and high-frequency attitude data output by the inertial measurement unit, using a Kalman filter algorithm to perform tight combination processing on the two types of data, dynamically estimating and compensating for changes in the center position of the image control marker caused by the swaying of the UAV body, thereby obtaining high-precision static coordinates in motion; the execution of the graded slow descent program includes: the UAV descends from the cruising altitude to the first hovering altitude for preliminary terrain confirmation, then descends to the second hovering altitude for final landing confirmation, and after confirming safety, performs the final touchdown maneuver, and after touchdown, adjusts the landing gear height through the servo mechanism to make the fuselage level.

9. The method according to claim 1, characterized in that, The coordinated image acquisition unit performs synchronous image acquisition by: after the image-controlled UAV completes its own positioning data acquisition and enters a stable state, it sends a task ready signal containing its current precise position and unique identifier to the photogrammetric UAV performing aerial surveying tasks in the survey area through the cluster internal communication link; the photogrammetric UAV continuously listens to this signal during flight, and when its flight path is about to cover the position of the image control point, it adjusts its heading and attitude to ensure that its onboard aerial camera can clearly capture the image control mark from an orthogonal or specific tilt angle.

10. The method according to claim 1, characterized in that, The data processing of the collected positioning data to calculate the coordinates and elevations of the control points includes: Gross errors are detected and eliminated for multiple sets of BeiDou positioning observations collected at a single control point location. Least squares estimation or robust estimation methods are used to calculate the most probable values ​​of the plane coordinates and geodetic height of the control point and their accuracy information. Using a high-precision geoid model covering the survey area, the calculated geodetic height data is converted into normal heights suitable for surveying results; The finalized control point numbers, planar coordinates, normal heights, coordinate accuracy indicators, operating modes, and acquisition timestamp information are structured, stored, and output as standard format control point result files.

Images