Unmanned aerial vehicle-mounted laser radar geographic surveying and mapping data acquisition system and method

Abstract

The invention relates to the technical field of geographic surveying and mapping, and discloses an unmanned aerial vehicle-mounted laser radar geographic surveying and mapping data acquisition system and method, and the system comprises an unmanned aerial vehicle platform, a laser radar sensor, a position attitude measurement unit and an airborne calculation unit. The airborne computing unit shunts the original point cloud into a look-ahead point set and a surveying and mapping point set through a view field partitioning module; the geometric prediction module solves a predicted incident angle of the surface micro-element by using the look-ahead point set in combination with the pose; the instruction generation module reversely calculates acquisition parameters including the pulse frequency and the flight ground speed according to the parameters, and generates space anchor point coordinates; the space synchronization execution module issues parameters when the unmanned aerial vehicle arrives at the anchor point; and the real-time feedback verification module triggers override interruption according to the actual density of the surveying and mapping point set. According to the method, adaptive parameter adjustment is realized through prospective perception and closed-loop feedback, the problem of non-uniform distribution of point clouds of complex terrains is solved, and the integrity and efficiency of geographic surveying and mapping data acquisition of the unmanned aerial vehicle-mounted laser radar are improved.

Description

Technical Field

[0001] This invention relates to the field of geographic surveying and mapping technology, specifically to an unmanned aerial vehicle (UAV)-borne lidar geographic surveying and mapping data acquisition system and method. Background Technology

[0002] With the rapid development of remote sensing technology, UAV platforms equipped with lidar sensors have become an important means of acquiring three-dimensional spatial information of the Earth's surface. In geographic mapping, this combination method, with its advantages of mobility, flexibility, and low cost, is widely used in terrain modeling and environmental monitoring, enabling the rapid completion of large-scale geographic mapping data collection tasks.

[0003] Existing workflows typically employ a preset, fixed parameter mode, where a constant laser pulse emission frequency and flight cruising speed are set before mission execution. In flat or gently undulating areas, this experience-based open-loop control method can maintain a relatively stable point cloud distribution, meeting basic data quality requirements.

[0004] However, existing technologies have significant limitations when facing complex scenarios with drastically changing terrain, such as mountains and cliffs. Fixed acquisition parameters are difficult to adapt to dynamic changes in surface slope, leading to an increase in the laser beam incident angle. This results in a significant reduction in the effective point cloud density per unit area, or even the creation of data holes. Furthermore, traditional systems lack the ability to perceive the terrain ahead and a location-based real-time feedback mechanism. They cannot adaptively compensate for fluctuations in flight speed or abrupt changes in terrain, easily causing uneven spatial distribution of mapping data. This often requires increasing the overlap rate or performing secondary flights to ensure data integrity, severely impacting operational efficiency.

[0005] Therefore, this invention proposes an unmanned aerial vehicle (UAV)-borne lidar geographic mapping data acquisition system and method to address the shortcomings of existing technologies. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides an unmanned aerial vehicle (UAV)-borne lidar geographic mapping data acquisition system and method, which solves the problems of uneven spatial density and poor acquisition integrity of lidar point cloud data caused by the difficulty in adaptively adjusting acquisition parameters according to surface geometric features under complex terrain conditions.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an unmanned aerial vehicle (UAV)-borne lidar geographic mapping data acquisition system, comprising an UAV platform, a lidar sensor, a position and attitude measurement unit, and an airborne computing unit; the airborne computing unit includes: The field-of-view partitioning module is used to receive the original laser point cloud data stream and split the original laser point cloud data stream into a look-ahead point set and a mapping point set according to the vertical scanning angle threshold. The geometric prediction module is used to combine real-time pose data to transform the look-ahead point set to the world coordinate system and calculate the predicted incident angle of the surface micro-element. The instruction generation module is used to calculate the acquisition parameters required to meet the target density based on the predicted incident angle, and to calculate the spatial anchor point coordinates used to trigger the acquisition parameters. The acquisition parameters include the pulse transmission frequency and the flight ground speed. The spatial synchronization execution module is used to associate the collected parameters with the spatial anchor point coordinates and store them in the spatial command queue, and to issue the collected parameters when it is determined that the real-time position of the UAV platform has reached the spatial anchor point coordinates. The real-time feedback verification module is used to calculate the actual spatial density based on the mapping point set, and send an overdrive interrupt signal when the actual spatial density is lower than the safety baseline threshold.

[0008] Preferably, the field of view partitioning module has a preset set of parameters that define the boundaries of the functional area. The logic of the field of view partitioning module to filter the look-ahead point set is that the vertical scanning angle of the sampling point is within the forward detection range. The look-ahead point set includes point cloud data projected by the laser beam in front of the flight direction of the UAV platform. The logic for the field-of-view partitioning module to filter the mapping point set is that the vertical scanning angle of the sampling point is within the mapping boundary angle range centered on the Nadir direction, and the mapping point set includes point cloud data projected by the laser beam directly below the UAV platform.

[0009] Preferably, the geometric prediction module is specifically used to: calculate the spatial normal vector of the foreground surface micro-element using the transformed look-ahead point set, define a virtual vertical down-looking laser beam vector, and calculate the angle between the spatial normal vector and the vertical down-looking laser beam vector as the predicted incident angle; The geometric prediction module is also used to traverse the adjacent scan points in the look-ahead point set, calculate the Euclidean distance difference between the adjacent scan points in the radar line-of-sight direction, and generate a true shadow area marker when the Euclidean distance difference exceeds the breakage threshold.

[0010] Preferably, the instruction generation module performs inverse calculation based on the equivalent spatial density model, which represents that the effective point cloud density per unit surface area is proportional to the pulse emission frequency and inversely proportional to the product of the flight ground speed, the scanning structure constant, and the cosine of the predicted incident angle. When the instruction generation module receives a true shadow area marker, it ignores the result of the reverse calculation and forces the output of a preset security mode parameter.

[0011] Preferably, the instruction generation module employs a tiered adjustment strategy when calculating the acquired parameters: Prioritize maintaining the flight ground speed at the preset cruise speed, and calculate the target pulse frequency that can compensate for the influence of terrain slope; When the calculated target pulse frequency exceeds the maximum pulse frequency supported by the lidar sensor, the pulse emission frequency is locked to the maximum pulse frequency, and the flight speed is reduced to meet the target density.

[0012] Preferably, the instruction generation module calculates the coordinates of the spatial anchor point in the following way: Extract the horizontal position of the corresponding surface micro-element that generates the predicted incident angle in the world coordinate system, and set the horizontal position as the coordinate of the spatial anchor point.

[0013] Preferably, the condition for the spatial synchronization execution module to determine that the real-time position of the UAV platform has reached the coordinates of the spatial anchor point is: Calculate the Euclidean distance between the real-time position of the UAV platform and the coordinates of the spatial anchor point; when the Euclidean distance is less than a preset trigger threshold, it is determined that the UAV has arrived.

[0014] The preset trigger threshold is determined by the sum of the basic positioning tolerance radius and the velocity feedforward term, where the velocity feedforward term is the product of the current flight speed modulus of the UAV platform and the comprehensive response time constant.

[0015] Preferably, the real-time feedback verification module uses a sliding window statistical method to calculate the actual spatial density, counts the number of survey points falling within a statistical window of a fixed size, and divides this number by the physical area of ​​the statistical window to obtain the actual spatial density.

[0016] Preferably, after receiving the overdrive interrupt signal, the space synchronization execution module clears the pending content in the space instruction queue and forcibly outputs preset safety mode parameters to adjust the hardware working state. The preset safety mode parameters include setting the pulse transmission frequency to the maximum pulse frequency and setting the flight ground speed to the minimum safe level flight speed.

[0017] This invention also provides a method for acquiring geographic mapping data using UAV-borne lidar, comprising the following steps: Receive the raw laser point cloud data stream and split the raw laser point cloud data stream into a look-ahead point set and a mapping point set according to the vertical scanning angle threshold; By combining real-time pose data, the look-ahead point set is transformed to the world coordinate system, and the predicted incident angle of the surface micro-element is calculated. Based on the predicted incident angle, the acquisition parameters that meet the target density requirements are calculated in reverse, and the coordinates of the spatial anchor point used to trigger the acquisition parameters are calculated. The acquisition parameters include the pulse transmission frequency and the flight ground speed. The collected parameters are associated with and packaged with the spatial anchor point coordinates, and then stored in the spatial instruction queue in sequence. The real-time position of the drone platform is monitored in real time. When it is determined that the distance between the real-time position of the drone platform and the coordinates of a certain spatial anchor point in the spatial command queue is less than a preset trigger threshold, the corresponding acquisition parameters are sent. The actual spatial density is calculated using the mapping point set. When the actual spatial density is lower than the safety threshold, the override interrupt mechanism is triggered to clear the pending content in the spatial instruction queue and force a switch to the safety mode parameters.

[0018] This invention provides an unmanned aerial vehicle (UAV)-borne lidar geographic mapping data acquisition system and method. It has the following beneficial effects: 1. This invention uses a field-of-view partitioning module to split the raw lidar data, and uses the look-ahead point set to pre-calculate the predicted incident angle of surface micro-elements, and then uses this to deduce the pulse frequency and ground speed that satisfy the target density. This feedforward control mechanism based on geometric prediction overcomes the problem of sparse point cloud distribution caused by drastic changes in terrain slope in traditional UAV geographic mapping operations, ensuring the uniformity and integrity of data acquisition in complex mountainous or cliff scenes, and effectively improving the accuracy of subsequent 3D reconstruction models; 2. This invention constructs an instruction queue containing acquisition parameters and spatial anchor point coordinates, and utilizes a spatial synchronization execution module to compare the real-time position of the UAV platform with the distance to the anchor point in real time, realizing parameter triggering logic based on position awareness. This mechanism avoids spatiotemporal misalignment caused by flight speed fluctuations, ensuring that the variable frequency and speed commands of the lidar can accurately act on the predetermined geographical area. This solves the technical problem of control lag caused by wind speed disturbances or maneuvering speed changes in existing surveying and mapping data acquisition systems, achieving accurate scanning coverage of specific terrain features; 3. This invention uses a real-time feedback verification module to calculate the current actual spatial density using the survey point set. When data quality is detected to be below the safety threshold, an override interrupt is triggered, forcing the system to switch to a high-frequency, low-speed safety mode. This dual-protection architecture, combining forward prediction and a posteriori correction, effectively addresses the risk of decreased data acquisition quality caused by sudden gusts of wind or model deviations. It not only ensures the operational safety of UAV-borne lidar under extreme conditions but also avoids rework due to data gaps, significantly improving the efficiency and reliability of geographic surveying tasks. Attached Figure Description

[0019] Figure 1 This is an architecture diagram of an UAV-borne lidar geographic mapping data acquisition system according to the present invention; Figure 2 This is a flowchart of a UAV-borne lidar geographic mapping data acquisition method according to the present invention; Figure 3The figure shows the experimental results comparing the actual spatial density under different control strategies of the present invention.

[0020] Among them, 10 is the UAV platform; 20 is the lidar sensor; 30 is the position and attitude measurement unit; 40 is the airborne computing unit; 41 is the field of view partitioning module; 42 is the geometric prediction module; 43 is the command generation module; 44 is the spatial synchronous execution module; and 45 is the real-time feedback verification module. Detailed Implementation

[0021] The technical solutions in 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, and 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.

[0022] Please see the appendix Figure 1 The present invention provides an unmanned aerial vehicle (UAV)-borne lidar geographic mapping data acquisition system, the main hardware components of which include an UAV platform 10, a lidar sensor 20, a position and attitude measurement unit 30, and an airborne computing unit 40.

[0023] The UAV platform 10 utilizes existing multi-rotor or vertical takeoff and landing fixed-wing aircraft, possessing a low-level flight control interface capable of responding to external speed commands to change its flight state. The lidar sensor 20 is an existing mechanical rotating or solid-state scanning ranging device, mounted on the UAV platform 10. Its scanning range covers the front and bottom of the flight direction and has an open parameter control interface, supporting adjustment of pulse emission frequency or scanning mode via external commands. The position and attitude measurement unit 30 is an existing high-precision integrated navigation device, integrating an inertial measurement unit (IMU) and a global navigation satellite system (GNSS), used for real-time output of high-frequency position, velocity, and attitude data. The onboard computing unit 40, as the core processing device, establishes bidirectional communication connections via a data bus with the lidar sensor 20, the position and attitude measurement unit 30, and the flight controller of the UAV platform 10, used to receive sensor data and send control commands.

[0024] The airborne computing unit 40 is internally configured with multiple modules that perform specific logical functions, and these modules are tightly coupled through data flow processes. The functional modules include a field-of-view partitioning module 41, a geometric prediction module 42, an instruction generation module 43, a spatial synchronous execution module 44, and a real-time feedback verification module 45.

[0025] The field-of-view partitioning module 41 is connected to the data output terminal of the lidar sensor 20 and is used to receive the raw lidar point cloud data stream. Based on a preset vertical scanning angle threshold, the field-of-view partitioning module 41 logically splits the raw data stream into two independent datasets: a look-ahead point set and a mapping point set. The look-ahead point set contains point cloud data projected by the laser beam in front of the UAV platform 10's flight direction, used for environmental perception of unreached areas; the mapping point set contains point cloud data projected by the laser beam directly below the UAV platform 10, used for subsequent map generation and real-time quality monitoring.

[0026] The geometric prediction module 42 is connected to the field-of-view partitioning module 41 and the position and attitude measurement unit 30. The geometric prediction module 42 receives the look-ahead point set from the field-of-view partitioning module 41 and, in conjunction with the real-time pose data output by the position and attitude measurement unit 30, transforms the look-ahead point set to the world coordinate system. The geometric prediction module 42 uses the transformed look-ahead point set to calculate the spatial normal vector of a foreground surface element and to calculate the predicted incident angle of the laser beam on that element. Simultaneously, the geometric prediction module 42 analyzes the spatial continuity of the look-ahead point set, identifies terrain-occupied areas, and outputs corresponding shadow area markers.

[0027] The instruction generation module 43 is connected to the geometric prediction module 42 and is used to receive the predicted incident angle and shadow area markers. Based on a preset equivalent spatial density model, the instruction generation module 43 calculates the acquisition parameters required to satisfy the target density according to the input predicted incident angle. The acquisition parameters include the pulse emission frequency of the lidar sensor 20 and the flight ground speed of the UAV platform 10. The instruction generation module 43 also calculates the spatial anchor point coordinates used to trigger the acquisition parameters based on the ground position corresponding to the predicted incident angle, and packages and associates the acquisition parameters with the spatial anchor point coordinates.

[0028] The spatial synchronization execution module 44 is connected to the instruction generation module 43 and the position and attitude measurement unit 30. The spatial synchronization execution module 44 receives the packaged and associated acquisition parameters and spatial anchor point coordinates, and stores them sequentially into an internally maintained spatial instruction queue. The spatial synchronization execution module 44 continuously reads the real-time position of the UAV platform 10 output by the position and attitude measurement unit 30. When it determines that the distance between the real-time position of the UAV platform 10 and the coordinates of a certain spatial anchor point in the queue is less than a preset trigger threshold, it immediately sends the corresponding acquisition parameters to the lidar sensor 20 or the UAV platform 10 to adjust the hardware operating status.

[0029] The real-time feedback verification module 45 is connected to both the field-of-view partitioning module 41 and the spatial synchronization execution module 44. The real-time feedback verification module 45 receives the mapping point set from the field-of-view partitioning module 41 and calculates the actual spatial density of the currently acquired data in real time. When the actual spatial density is lower than a preset safety threshold, the real-time feedback verification module 45 sends an override interrupt signal to the spatial synchronization execution module 44. In response to this override interrupt signal, the spatial synchronization execution module 44 clears the pending execution content in the spatial command queue and forcibly outputs preset safety mode parameters.

[0030] See attached document Figure 2 This invention provides a method for collecting geographic mapping data using UAV-borne lidar, comprising the following steps: S10, Field of view partitioning and data splitting: The airborne computing unit 40 receives the raw scanning data collected by the lidar sensor 20 in real time, and logically splits the raw scanning data into a look-ahead point set and a mapping point set according to the preset vertical scanning angle threshold. S20, Look-ahead geometric feature calculation: Extract point cloud data from the look-ahead point set, combine it with the real-time pose data output by the position and attitude measurement unit 30 to convert it to the world coordinate system, use the converted look-ahead point set to calculate the spatial normal vector of the ground micro-element in front, and calculate the predicted incident angle of the laser beam on the micro-element. At the same time, detect whether there is terrain occlusion and generate shadow area markers. S30, Adaptive control command generation: Based on the preset equivalent spatial density model, according to the predicted incident angle and shadow area marking, the acquisition parameters that meet the target density requirements are calculated in reverse. The acquisition parameters include the laser pulse emission frequency and flight ground speed. Also, according to the ground position corresponding to the predicted incident angle, the coordinates of the spatial anchor point used to trigger the acquisition parameters are calculated. S40, Spatial command queue construction: The collected parameters are associated with and packaged with the spatial anchor point coordinates, and stored in the spatial command queue in sequence. This spatial command queue is used to manage the control operations to be executed. S50, Spatial Synchronization Trigger Execution: Real-time monitoring of the current position of the UAV platform 10 output by the position and attitude measurement unit 30. When it is determined that the distance between the real-time position of the UAV platform 10 and the coordinates of a certain spatial anchor point in the queue is less than the preset trigger threshold, the corresponding acquisition parameters are immediately sent to the lidar sensor 20 or the UAV platform 10. S60, Closed-loop feedback verification: Calculates the actual spatial density of the currently collected data in real time using the data from the survey point set. When the actual spatial density is lower than the preset safety threshold, an override interrupt mechanism is triggered, clearing the pending content in the spatial command queue and forcibly switching to the safe mode parameters.

[0031] The UAV-borne LiDAR geographic mapping data acquisition method of the present invention is based on the aforementioned UAV-borne LiDAR geographic mapping data acquisition system. Specifically, the method is the working logic of the system during operation, wherein steps S10 to S60 correspond to the data processing process executed sequentially or in parallel by the field of view partitioning module 41, geometric prediction module 42, instruction generation module 43, spatial synchronization execution module 44, and real-time feedback verification module 45 inside the airborne computing unit 40.

[0032] See attached document Figure 1 In this embodiment, the field-of-view partitioning module 41 establishes a physical connection with the data output terminal of the lidar sensor 20 via a high-bandwidth data bus (e.g., Gigabit Ethernet or a dedicated LVDS interface) to receive the raw laser point cloud data stream output by the lidar sensor 20 in real time. In this embodiment, the lidar sensor 20 can be a mechanical rotating ranging device (e.g., a multi-beam mechanical rotating radar, which uses a motor to drive the laser transceiver components to rotate at high speed to achieve 360-degree scanning) or a solid-state scanning ranging device (e.g., a MEMS galvanometer radar or a prism scanning radar, which uses electronic components to control beam deflection). The lidar sensor 20 of the above types encapsulates the detected environmental information into a raw laser point cloud data stream. This data stream typically consists of a series of data packets containing spatial location information, reflection intensity, and timestamps, transmitted via UDP or TCP / IP protocols. This data acquisition and transmission process is a well-known technology familiar to those skilled in the art.

[0033] The field-of-view partitioning module 41 internally stores a preset vertical scanning angle threshold. This threshold serves as a benchmark parameter for logical judgment, defining the boundaries of different functional areas within the lidar's field of view. Specifically, this vertical scanning angle threshold is not a single value, but rather a set of boundary parameters defining an angle range. Its values ​​are set based on the typical flight altitude of the UAV platform 10 and the radar installation tilt angle to ensure that the field-of-view partitioning can cover effective forward warning and downward-looking operation areas. The field-of-view partitioning module 41 samples each point contained in the original lidar point cloud data stream. Perform point-by-point logical judgments and, based on the preset vertical scanning angle threshold, logically split the original data stream into two independent datasets: the look-ahead point set and the mapping point set.

[0034] In this embodiment, the vertical scanning angle in the lidar coordinate system is defined. The criteria for judgment (with the horizontal plane as 0 degrees, downward as positive, upward as negative, or according to the coordinate system defined by the specific radar protocol). Look-through point set. This includes point cloud data projected by a laser beam in front of the UAV platform 10's flight path. Its significance lies in utilizing the radar's far-field of view for environmental perception of unreached areas. The logical criterion for the field-of-view partitioning module 41 to select the look-ahead point set is the sampling point... Vertical scanning angle In the forward detection range The mathematical logic for this determination is as follows: ; In the formula, sampling points The laser point object currently being processed; For this sampling point The distance measurement value; For this sampling point Horizontal azimuth; and This refers to the vertical scanning angle threshold component that defines the look-ahead region. As a preferred method, Set to 30 degrees. The angle was set at 60 degrees. This angle range was chosen because: if the angle is less than 30 degrees, the detection point is too close to the horizontal position of the UAV, leaving insufficient response time for the onboard computing unit to adjust flight parameters; if the angle is greater than 60 degrees, the laser beam tends to be horizontal, the incident angle is too large, resulting in severe attenuation of echo energy, and it is easily interfered with by distant background noise, making it difficult to accurately reflect geometric changes in the surface terrain. Therefore, this range achieves an engineering balance between ensuring detection lead and data signal-to-noise ratio.

[0035] Meanwhile, the survey point set This includes point cloud data projected by a laser beam directly below the UAV platform 10. This data is used for subsequent map generation and real-time quality monitoring. The logical criterion for the field-of-view partitioning module 41 to filter the mapping point set is the sampling point. Vertical scanning angle The mapping boundary angle lies centered on the direction of Nadir. Within the specified range, the determination logic is as follows:

[0036] ; In the formula, The angle is typically set between 15 and 20 degrees. Within this range, the laser beam is incident on the ground almost perpendicularly, resulting in minimal beam distortion and the highest echo intensity, thus providing the most accurate ranging data.

[0037] In addition, to ensure data validity, the field-of-view partitioning module 41 also performs data cleaning operations in parallel while executing the aforementioned splitting logic. An effective range interval is set to address any outliers that may exist in the lidar data. .in, The preset minimum blind zone distance (e.g., set to 1.0 meter, this value is determined based on the drone's body size and installation location, used to filter out obstructions from the fuselage landing gear or near-field dust and noise). This is the preset maximum effective measurement distance (e.g., set to 90% of the nominal range of the lidar sensor 20, i.e., the physical limit distance at which the signal-to-noise ratio meets the mapping requirements, such as 200 meters). When the sampling point... Distance value or At that time, the field of view partitioning module 41 will sample the point. The data is marked as invalid and removed, and is not written into any of the above point sets. Through the above processing, the field of view partitioning module 41 decouples the original data stream of a single physical channel into two logical data streams with clear functional directions, laying the data foundation for the subsequent parallel processing architecture.

[0038] See attached document Figure 1 In this embodiment, the geometric prediction module 42 acts as a data aggregation and geometric solution node, establishing communication connections with both the field-of-view partitioning module 41 and the position and attitude measurement unit 30. Upon receiving the look-ahead point set output by the field-of-view partitioning module 41... Subsequently, to assign geospatial attributes to these discrete point clouds, the geometric prediction module 42 synchronously acquires the real-time pose data output by the position and attitude measurement unit 30 via a high-frequency interface. Considering the difference between the scanning frequency of the lidar (typically 10Hz-20Hz) and the output frequency of the inertial navigation system (typically 100Hz-200Hz), to ensure strict alignment of multi-source data in the time dimension, this embodiment employs a timestamp-based linear interpolation algorithm to align the high-frequency pose data to the precise time of each laser point's generation. The real-time pose data includes the three-dimensional position coordinates of the UAV platform 10 in a world coordinate system (e.g., the North-East-Ground NED coordinate system). and the attitude rotation matrix describing the body's attitude. In this embodiment, the aircraft attitude specifically refers to the angular position relationship between the UAV's body coordinate system and the navigation coordinate system (world coordinate system). It is usually determined by a set of Euler angle parameters, namely roll angle, pitch angle, and yaw angle. Its significance lies in describing the tilt and rotation state of the UAV relative to the horizontal plane and due north direction during flight, which directly determines the instantaneous pointing of the airborne sensors.

[0039] To eliminate the impact of dynamic changes in UAV flight attitude on terrain geometric feature extraction, the geometric prediction module 42 utilizes pre-calibrated extrinsic parameters of the lidar sensor 20 relative to the center of the UAV platform 10, combined with real-time pose data, to transform the look-ahead point set to the world coordinate system. This process is based on the principle of rigid body transformation, aiming to establish an absolute geographic reference system that does not change with the movement of the UAV. In this step, to accurately distinguish the physical meaning of different rotational transformations, the following definitions are made: Define the attitude rotation matrix that varies with time. This is the static extrinsic parameter rotation matrix describing the lidar installation angle. For any sampling point in the look-ahead point set defined in the lidar's local coordinate system... Its coordinates are transformed to the world coordinate system. The calculation formula is as follows: ; In the formula, This represents the rotation matrix representing the calibration extrinsic parameters of the lidar sensor 20 relative to the machine body; This represents the original coordinate vector of the sampling point in the local coordinate system of the lidar. This represents the extrinsic parameter vector representing the translation of the lidar sensor 20 relative to the center of the aircraft. This represents the real-time position vector of the UAV platform 10 in the world coordinate system. The meaning of this formula is that, through two-stage transformation, the laser point is first transformed from the radar coordinate system to the body coordinate system (operation within parentheses), and then the point in the body coordinate system is transformed to the absolute geographic coordinate system in combination with the current attitude and position, thereby obtaining a set of point cloud data with absolute geographic coordinates.

[0040] After obtaining the transformed point cloud data, it is not treated as isolated points, but rather as discrete samples of a continuous surface. Based on this concept of infinitesimalization, the geometric prediction module 42 uses the transformed look-ahead point set to calculate the spatial normal vector of the foreground surface infinitesimal elements. Specifically, it uses the current sampling point in the transformed point cloud... Centered on, the search radius is The set of all points within a spherical neighborhood (e.g., 0.5 meters to 1.0 meters). For neighborhood point sets Construct the covariance matrix The covariance matrix is ​​used to characterize the local spatial distribution features of the infinitesimal element. The mathematical expression is: ; In the formula, For neighborhood point set The number of points in the middle, Let be the geometric centroid of the neighborhood point set. In practical engineering implementations, to avoid computational divergence or insufficient statistical significance due to an insufficient number of points, when... If the number of points is less than a preset minimum threshold (e.g., 3 points), the system marks the point as invalid and skips the calculation. This is achieved by analyzing the covariance matrix. Perform eigenvalue decomposition to obtain three eigenvalues. And its corresponding eigenvectors. Based on the principles of Principal Component Analysis (PCA), the minimum eigenvalue... The corresponding eigenvector represents the direction of least data change, which is the spatial normal vector of the surface element. To ensure the physical consistency of the spatial normal vector direction (uniformly pointing upwards towards the Earth's surface), it is necessary to determine... The sign of the dot product with the world coordinate system's perpendicular axis (z-axis); if negative, it indicates a negative sign. Perform the reverse operation.

[0041] After acquiring the surface geometric properties, in order to quantitatively assess the impact of terrain on surveying operations, the geometric prediction module 42 further calculates the predicted incident angle of the laser beam on the micro-element. The predicted incident angle here... This is not the observation angle at the current moment, but specifically refers to the angle between the virtual vertical downward-looking laser beam and the surface normal when the UAV platform 10 flies directly above this point to perform vertical mapping operations. The geometric prediction module 42 defines a virtual beam vector that is always vertically downward (i.e., parallel to the gravity vector) in the world coordinate system. (For example, the value is) Predicting the angle of incidence The calculation formula is as follows: ; In the formula, To prevent extremely small positive numbers with a denominator of zero (e.g., taking 10) -6 ), Represents the magnitude of a vector. This indicates the dot product operation. The predicted incident angle directly reflects the relative slope of the ground ahead: the larger the incident angle, the steeper the ground, resulting in a smaller effective echo area and lower echo intensity during future vertical observations. Therefore, more intensive scanning is required to ensure data quality.

[0042] Meanwhile, considering the potential data blind spots caused by complex terrain, the geometric prediction module 42 analyzes the spatial continuity of the look-ahead point set to identify terrain-occluded areas and output corresponding shadow area markers. In actual flight scenarios, the presence of cliffs, steep slopes, or tall buildings on the ground surface can cause significant abrupt changes in the radial distance between adjacent LiDAR scan lines. The geometric prediction module 42 traverses adjacent scan points in the look-ahead point set. and (Arranged in scan line order), calculate the Euclidean distance difference in the radar line-of-sight direction. The Euclidean distance difference The calculation formula is: In the formula, and Corresponding to adjacent sampling points and The original laser ranging value. With the preset fracture threshold A comparison is made. In this embodiment, the fracture threshold... It is based on lidar angular resolution and current measured distance The dynamically set judgment parameters are typically set to 3 to 5 times the theoretical point spacing (e.g., between 0.5 meters and 5.0 meters) to effectively distinguish between normal terrain undulations and discontinuous terrain abrupt changes. If the calculated... Then it is determined that in and There are terrain obstructions (i.e., blind spots). In this case, the geometric prediction module 42 marks the corresponding surface location as an untrusted area and generates a Boolean-type shadow area marker. (For example, set to True). This flag indicates that the system cannot accurately determine the true terrain conditions of the area (there may be deep pits or steep slopes). To ensure flight and data acquisition safety, the subsequent control module will adopt the most conservative control strategy for this marked area. Conversely, when the calculated Euclidean distance difference... Less than or equal to the fracture threshold When the terrain changes between adjacent scan points are within the normal range of undulation and there is no line-of-sight obstruction, a false shadow area marker is generated. (For example, if set to False), the system will recognize the predicted geometric information of the area as valid and continue to execute the regular acquisition parameter calculation process.

[0043] See attached document Figure 1 In this embodiment, the instruction generation module 43, as the core decision-making unit of the system, establishes a data communication connection with the geometric prediction module 42. The main task of this module is to receive the predicted incident angle calculated and output by the geometric prediction module 42. and the corresponding shaded area markers This geometric environmental perception information is then transformed into specific control commands at the hardware level. To achieve this transformation, the command generation module 43 internally stores a preset equivalent spatial density model. In this embodiment, the preset equivalent spatial density model is not a statistical black-box model, but a deterministic physical mapping relationship constructed based on the geometric principle of lidar triangulation imaging. This equivalent spatial density model quantifies the functional relationship between the actual landing density of the point cloud and the lidar acquisition parameters, flight dynamics parameters, and surface geometric properties, and is used to describe how to compensate for the density attenuation caused by terrain by adjusting hardware parameters at a specific incident angle. Its mathematical expression is as follows: ; In the formula, This represents the effective point cloud density per unit area of ​​the land surface (unit: points / square meter). The pulse emission frequency variable (unit: Hz) represents the pulse excitation rate of the lidar at any given moment. The ground speed variable (unit: m / s) represents the horizontal movement rate of the UAV platform at any given time. The scanning structure constant (unit: meters) is a parameter that characterizes the effective coverage width of the lidar perpendicular to the flight direction. It is determined by the lidar's beam distribution and field of view. Let be the predicted angle of incidence. This formula explicitly reveals the physical causal relationship: as the terrain becomes steeper (i.e., ... Increase), cosine term If the flight speed and emission frequency remain unchanged, the spacing between the laser beams projected onto the ground will increase significantly, resulting in a decrease in the actual effective point cloud density.

[0044] Based on the above physical model, the instruction generation module 43 executes the reverse solution logic, calculating the acquisition parameters required to satisfy the target density according to the input predicted incident angle. The system pre-sets the target density that the surveying operation must meet. (For example, it is set to 100 points / square meter, and this value is determined according to the accuracy requirements of the surveying task). During the calculation process, it is considered that the terrain may approach vertical almost infinitely (i.e.,...). Direct calculation can cause the denominator to approach zero, leading to numerical overflow. Therefore, the instruction generation module 43 introduces a protection threshold. In this embodiment, the protection threshold... It is a dimensionless numerical lower bound constant (preferably ranging from 0.05 to 0.1, corresponding to a maximum effective calculation angle of approximately 84 to 87 degrees) used to ensure that the divisor is not zero. The instruction generation module 43 first calculates the frequency-to-velocity ratio factor required to compensate for the terrain slope. : ; get Then, the instruction generation module 43 executes parameter optimization logic to determine the specific acquisition parameters, which specifically include the pulse emission frequency of the lidar sensor 20. and the flight ground speed of the unmanned aerial vehicle platform 10 .at this time, and They are respectively used as the model variables mentioned above. and The specific target solution within the current control cycle. To maximize operational efficiency while ensuring data quality, the system adopts a tiered adjustment strategy with frequency priority and speed as a safety net: ; In the formula, The preset cruise speed refers to the ideal flight speed (e.g., 10 m / s) for the drone to perform standard operations on flat terrain. The maximum pulse frequency refers to the highest pulse emission rate (e.g., 300kHz) that the lidar sensor 20 hardware can physically support. The significance of this logic is that when the predicted terrain slope is gentle and the required parameters are within the radar's performance range, maintaining the standard cruise speed is prioritized. The point cloud density is maintained by dynamically increasing the pulse frequency, thus ensuring operational efficiency. However, if the predicted terrain slope is too steep, even using the maximum pulse frequency may not be sufficient. If the density requirement cannot be met, the system will lock the maximum frequency and extend the scanning dwell time in the area by forcibly reducing the flight ground speed, thereby ensuring that the final point cloud density meets the standard.

[0045] In addition, the input is marked with a shaded area. The instruction generation module 43 is configured with the highest priority security overlay logic. When the received instruction... When the value is True, it indicates the existence of unknown terrain blind spots or fault zones ahead. In this case, the instruction generation module 43 ignores the above reverse calculation results and forcibly outputs the preset safety mode parameters (i.e., and ,in (The minimum safe level flight speed allowed for drones) to maximize data redundancy in response to potential terrain abrupt changes.

[0046] After completing the parameter calculation, the command generation module 43 also calculates the coordinates of the spatial anchor point used to trigger the acquisition parameters based on the surface location corresponding to the predicted incident angle. Since the geometric prediction module 42 analyzes the look-ahead area located in front of the UAV, in order to ensure that the control command takes effect accurately when the UAV actually flies over this area, the command generation module 43 extracts the horizontal position of the corresponding surface micro-element that generated the predicted incident angle in the world coordinate system. And set it as the spatial anchor point coordinates. : ; The final instruction generation module 43 packages and associates the collected parameters with the spatial anchor point coordinates to generate a command containing... The space instruction package, in which This is the command sequence number. This packaging and association operation ensures a strict binding between control parameters and geospatial location, enabling the UAV to achieve precise location-based frequency and speed control, rather than traditional time-based control. Through the above processing, the system completes a logical closed loop from terrain perception to control decision-making.

[0047] See attached document Figure 1 In this embodiment, the spatial synchronization execution module 44 serves as a crucial bridge connecting the decision-making algorithm and the physical execution mechanism, and is connected to both the command generation module 43 and the position and attitude measurement unit 30. This spatial synchronization execution module 44 aims to solve the spatial misalignment problem caused by traditional time-axis control in non-uniform flight conditions. By introducing a geofence triggering mechanism, it ensures that control commands are accurately executed at predetermined geographical locations. The spatial synchronization execution module 44 receives the packaged and associated acquisition parameters and spatial anchor point coordinates. This data packet typically contains... Quadruple information. To effectively manage these discrete instructions generated in the order of the flight path, the space synchronization execution module 44 stores them sequentially into an internally maintained space instruction queue. In this embodiment, the spatial command queue adopts a first-in, first-out (FIFO) data structure design to ensure that the execution of commands follows the preset flight path planning order of the UAV. Each node in the queue stores the control parameters to be executed and their corresponding geographical triggering conditions.

[0048] To enable position-based command triggering, the spatial synchronization execution module 44 continuously reads the real-time position of the UAV platform 10 output by the position and attitude measurement unit 30. As those skilled in the art will know, the position and attitude measurement unit 30 integrates a Global Navigation Satellite System (GNSS) and an Inertial Measurement Unit (IMU). It utilizes the Extended Kalman Filter (EKF) algorithm to fuse the absolute coordinate data provided by the GNSS with the high-frequency acceleration and angular velocity data provided by the IMU, thereby calculating in real time high-frequency (e.g., above 100Hz), low-latency, and smooth three-dimensional position coordinates. (Real-time location) can maintain the continuity of positioning output through inertial recursion even when satellite signals are momentarily blocked, which is a well-known technology in this field.

[0049] After obtaining the real-time location, the spatial synchronization execution module 44 executes the proximity determination logic to calculate the current position of the UAV. Spatial anchor point coordinates corresponding to the queue head instruction Euclidean distance between When the distance between the real-time position of the drone platform 10 and the coordinates of a spatial anchor point in the queue is less than a preset trigger threshold, the condition is met. When a certain condition is met, the trigger signal is activated. In this embodiment, a preset trigger threshold is used. It is not a single, fixed distance scalar, but a dynamic judgment parameter that includes communication latency compensation. Considering the physical delay between command issuance and hardware response, a trigger threshold is set to compensate for the spatial lag (overshoot) caused by this delay. The calculation formula introduces a velocity feedforward term: ; In the formula, Use the basic positioning tolerance radius (e.g., 0.5 meters). The current flight speed modulus of the drone (by...) (obtained by first-order differential) This is the overall response time constant (including communication link delay and actuator response time, typically in the range of 100ms to 200ms). The meaning of this formula is that the faster the flight speed, the earlier the system needs to trigger the command (i.e., when farther from the target anchor point), to allow sufficient time for the hardware to complete the state transition, thereby ensuring that the hardware parameters are adjusted and in place when the UAV arrives at the anchor point.

[0050] Once the triggering conditions are met, the spatial synchronization execution module 44 immediately sends the corresponding acquisition parameters to the lidar sensor 20 or the UAV platform 10 to adjust the hardware operating status. Specifically, the spatial synchronization execution module 44 performs a command distribution operation: on the one hand, it adjusts the pulse transmission frequency... The command is sent to the lidar sensor 20 via a sensor control interface (such as a UDP command packet), instructing it to reset its internal timer to change the repetition rate; on the other hand, the flight ground speed is... The command is sent to the flight control computer of the UAV platform 10 via a flight control protocol (such as MAVLink), instructing it to adjust the thruster power to change the cruise speed. After the command is sent, the space synchronization execution module 44 removes the command node from the space command queue and immediately begins monitoring the trigger status of the next command node, thereby achieving fully autonomous continuous space parameter-variable scanning.

[0051] See attached document Figure 1 In this embodiment, the real-time feedback verification module 45 serves as the closed-loop quality monitoring unit of the system, connected to both the field-of-view partitioning module 41 and the spatial synchronization execution module 44. This real-time feedback verification module 45 aims to construct a reverse verification link from the execution end to the decision-making end to address fluctuations in acquisition quality caused by gust disturbances or prediction model deviations during actual flight. During operation, the real-time feedback verification module 45 receives the mapping point set from the field-of-view partitioning module 41. Based on the filtering logic of the aforementioned field-of-view partitioning module 41, this set of survey points specifically refers to the laser beam projected directly below the UAV platform 10 (i.e., the sampling points). Vertical scanning angle The mapping boundary angle lies centered on the direction of Nadir. The dataset contains point cloud data within a certain range, thus it can objectively reflect the actual landing points of the current core surveying area.

[0052] To quantitatively evaluate the current operational effectiveness, the real-time feedback verification module 45 calculates the actual spatial density of the currently collected data in real time. In this embodiment, the system uses a sliding window statistical method for calculation. Specifically, a fixed-size statistical window (e.g., a 1m x 1m rectangular area) is defined on the projection plane centered on the current horizontal position of the UAV, and the number of survey points falling within this window is counted. Actual spatial density The calculation formula is as follows: ; In the formula, This indicates the number of valid laser points contained in the current statistics window; This indicates the preset physical area of ​​the statistics window (unit: square meters, and...). This formula calculates the number of valid measurement points actually acquired per unit area of ​​the land surface at the current moment. This calculation is performed continuously at a high frequency (e.g., 10Hz) to ensure instantaneous capture of data quality fluctuations.

[0053] After obtaining the actual spatial density, the real-time feedback verification module 45 executes the quality judgment logic, comparing the value with the preset safety baseline threshold set by the system. A comparison is made. In this embodiment, a preset safety baseline threshold is used. This is the minimum density requirement to ensure that subsequent terrain reconstruction does not result in data gaps. As a preferred approach, this safety baseline threshold is typically set at the target density. 80% to 90% (for example, if the target density is 100 points / square meter, then the safety baseline threshold is set at 80 points / square meter) is used to effectively intercept substantial quality declines caused by sudden environmental changes while tolerating small statistical fluctuations.

[0054] If detected This indicates that the current control parameters are valid and the system is maintaining normal operation. However, when the actual spatial density is lower than the preset safety threshold (i.e., This indicates that despite prior forecasting, the actual data acquisition results still fell short of expectations (e.g., a sudden strong crosswind caused a rapid increase in the drone's ground speed, unexpectedly widening the scan line spacing). In this situation, to prevent further deterioration of data quality, the real-time feedback verification module 45 sends an override interrupt signal to the spatial synchronization execution module 44. .

[0055] The overdrive interrupt signal It has the highest system logic priority. In response to the overdrive interrupt signal, the spatial synchronization execution module 44 immediately suspends processing the current normal instruction stream and executes emergency intervention logic: on the one hand, it clears the pending content in the spatial instruction queue, because the parameters generated based on previous predictions in the queue may no longer be applicable to the current adverse operating conditions, and continued execution may lead to error accumulation; on the other hand, it forcibly outputs preset safety mode parameters. In this embodiment, the preset safety mode parameters are a set of extreme parameter combinations that sacrifice operational efficiency for absolute data quality, specifically including the maximum pulse frequency of the lidar. and the minimum safe level flight speed allowed for drones Through this mechanism, the system can quickly switch its operational status to the most conservative mode, using a combination of extremely low-speed flight and high-frequency scanning to instantly compensate for point cloud density until it reaches the actual spatial density. Once the system returns to normal, the overdrive state is released, thus achieving fully autonomous fail-safe control.

[0056] To verify the effectiveness of the UAV-borne lidar geographic mapping data acquisition system and method of the present invention in complex terrain, a mine restoration operation face in a mountainous area was selected as the test scenario. The scenario is 500 meters long and the terrain features include gentle road sections (slope <10°), steep road sections (slope about 60°), and local vertical cliffs (with terrain obstruction blind spots).

[0057] 1. System configuration and parameter initialization: Unmanned Aerial Vehicle Platform 10: A multi-rotor aircraft with a low-level control interface is selected, with a maximum ground speed of [missing information]. The minimum safe level flight speed is set at 10 m / s (as the standard cruise speed). Set to 2 m / s.

[0058] LiDAR sensor 20: Utilizes a mechanical rotating ranging device; maximum supported pulse emission frequency by hardware. The frequency is 300 kHz. Its scanning structure constant has been calibrated. It is 50m (a parameter representing the effective coverage width perpendicular to the flight direction).

[0059] Airborne computing unit 40: Preset target density for surveying operations 100 points / m 2 Preset safety baseline threshold Set to 80 points / m 2 .

[0060] 2. Work process execution: When the UAV platform 10 flies along the predetermined route, the collaborative working logic of the various modules of the onboard computing unit 40 is as follows: Phase 1 (Smooth Area): The field-of-view partitioning module 41 receives the raw data stream and splits it into a look-ahead point set. The geometric prediction module 42 solves the look-ahead point set, determines that the terrain ahead is flat, and predicts the incident angle. .

[0061] Instruction generation module 43, based on formula Perform reverse calculation. Because... To meet The calculated pulse emission frequency .at this time And the calculated ground speed of flight Maintain standard cruising speed (10m / s). The spatial synchronization execution module 44 sends this parameter to the hardware, and the system is in standard cruise mode.

[0062] Phase Two (Encountering a Steep Slope): The drone approaches an area with a slope of 60°. The geometric prediction module 42 calculates the predicted angle of incidence using the look-ahead point set. .

[0063] Instruction generation module 43 performs the calculation again: at this time, the cosine term... To maintain 100 points / m 2 Target density, required frequency and velocity ratio factor Double.

[0064] According to the frequency-first strategy, the geometric prediction module 42 prioritizes increasing the frequency: the calculation requires... Boosted to 100kHz. Due to (300kHz), the system does not require speed adjustment, i.e., the flight ground speed. Still remain as (10m / s). Instruction generation module 43 calculates the spatial anchor point coordinates of the steep slope starting point. and the instruction package The signal is sent to the spatial synchronization execution module 44. When the UAV platform 10 reaches the anchor point, the lidar sensor 20 automatically up-frequencys to compensate for the density attenuation caused by the terrain slope.

[0065] Phase Three (Encountering the Cliff Blind Spot): The geometric prediction module 42 detected the Euclidean distance difference between adjacent scan points in the look-ahead point set. Exceeding the fracture threshold If terrain occlusion is detected, a true shadow area marker is generated. The instruction generation module 43 detected... If true, the highest priority security override logic is triggered, forcibly outputting the preset security mode parameters: that is... Set as (300kHz), and flight ground speed Set as (2 m / s). The spatial synchronization execution module 44 executes this command. The UAV platform 10 flies at extremely low speed above the cliff and scans at the highest frequency to ensure that a large amount of redundant data is acquired to cover blind spots.

[0066] Closed-loop verification: Throughout the process, the real-time feedback verification module 45 calculates the actual spatial density in real time using the survey point set. Data shows that even in areas with drastic terrain changes, It remained stable at 95 to 110 points / m 2 No overdrive interrupt signal was triggered during this period.

[0067] To verify the adaptive control effect of the UAV-borne lidar geographic mapping data acquisition system of the present invention in complex terrain, typical test scenarios including flat land, steep slopes and cliffs were constructed in a simulation environment, and a comparative experiment was conducted with a traditional fixed parameter acquisition system.

[0068] 1. Experimental instructions and environment setup: Simulation scenario construction: The test flight route is 300 meters long, and the terrain geometry features are as follows: 0m-100m (flat area): The terrain slope is approximately 0°, simulating plain surveying; 100m-200m (steep slope area): The terrain slope is 60°, simulating mountain climbing operations; 200m-250m (Cliff Blind Zone): A terrain fault zone exists (simulated cliff), triggering the system's shadow zone marker. ; 250m-300m (gradual recovery zone): The terrain has been restored to a flat state.

[0069] Initialization parameters of the UAV-borne LiDAR geographic mapping data acquisition system: The following parameters are all loaded into the airborne computing unit 40 as the basis for algorithm operation: Target density 100 points / m 2 Safety baseline threshold 80 points / m 2 ; Scan structure constants 50m; Hardware limitations: Maximum pulse transmission frequency Standard cruising speed Minimum safe level flight speed .

[0070] 2. Experimental steps: This experiment was divided into a control group (traditional system) and an experimental group (the system of this invention). Both groups of systems flew along the same route. The specific steps are as follows: Step S1: Control group test (fixed parameter mode)

[0071] The drone is programmed to operate according to traditional surveying logic. Throughout the process, its ground speed is locked. Constantly set at 10 m / s, pulse transmission frequency locked. The frequency is kept constant at 50kHz (this parameter can only meet the requirements of 100 points / m on flat ground). 2 (Requirements). Record the actual spatial density of the entire flight segment. .

[0072] Step S2: Experimental group testing (adaptive parameter mode)

[0073] The airborne computing unit 40 of the present invention is activated.

[0074] In the 0-100m range: Instruction generation module 43 calculates... Output , .

[0075] At 100-200m: the geometric prediction module 42 predicts a slope of 60°, and the instruction generation module 43 compensates for this. The attenuation will Dynamically boosted to 100kHz, maintained .

[0076] At 200-250m: Geometric prediction module 42 identifies terrain occlusion ( This triggers the security overwrite logic. The system forces output. (Maximum frequency) (Minimum speed).

[0077] Record the actual spatial density of the entire flight segment .

[0078] Step S3: Data Comparison and Analysis The actual spatial density recorded in the two sets of experiments Compared with the preset safety baseline threshold Compare the data to assess its stability and quality.

[0079] 3. Analysis of experimental results (refer to the appendix) Figure 3 ): See attached document Figure 3 The figure shows the relationship between flight distance and actual space density: Flat area (0-100m): The curves of the experimental group (solid line) and the control group (dashed line) are highly coincident, and the actual spatial density is consistently 100 points / m². 2 Nearby. This demonstrates that the invention maintains standard high-efficiency operation under normal terrain conditions.

[0080] Steep slope areas (100-200m): Control group (dashed line): Due to the lack of parameter adjustment, the steeper terrain led to an increase in the projected area, and the actual spatial density decreased significantly to approximately 50 points / m². 2 Far below the safety threshold (As shown by the dotted line), the data collected at this time cannot be used for high-precision modeling.

[0081] Experimental group (solid line): Thanks to the advance sensing of the geometric prediction module 42 and the precise triggering of the spatial synchronization execution module 44, the system automatically doubled the pulse emission frequency. The curve shows that the actual spatial density remained at 100 points / m². 2 The baseline completely eliminates the negative impact of terrain slope.

[0082] Cliff blind spot (200-250m): The experimental group curve showed a significant jump, with the actual spatial density exceeding 300 points / m². 2This is because the system detected terrain obstruction features and automatically switched to a high-frequency, low-speed safety mode. Although this strategy increases the operation time for the current flight segment, the high-density oversampling mechanism ensures complete capture of blind spot features, effectively avoiding the data gaps or missing data acquisition problems common in traditional operation methods.

[0083] 4. Conclusion: Experiments have shown that, compared with traditional fixed parameter schemes, this invention, by constructing a physical mapping model and a spatial synchronization mechanism, can ensure that the actual spatial density compliance rate is increased from 66% to 100% even under conditions of drastic changes in terrain slope (0° to 60°) and the presence of blind spots, thus achieving high-quality mapping under all terrain conditions.

[0084] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A UAV-borne lidar geographic mapping data acquisition system, characterized in that, It includes an unmanned aerial vehicle (UAV) platform, a lidar sensor, a position and attitude measurement unit, and an onboard computing unit; the onboard computing unit includes: The field-of-view partitioning module is used to receive the original laser point cloud data stream and split the original laser point cloud data stream into a look-ahead point set and a mapping point set according to the vertical scanning angle threshold. The geometric prediction module is used to combine real-time pose data to transform the look-ahead point set to the world coordinate system and calculate the predicted incident angle of the surface micro-element. The instruction generation module is used to calculate the acquisition parameters required to meet the target density based on the predicted incident angle, and to calculate the spatial anchor point coordinates used to trigger the acquisition parameters. The acquisition parameters include the pulse transmission frequency and the flight ground speed. The spatial synchronization execution module is used to associate the collected parameters with the spatial anchor point coordinates and store them in the spatial command queue, and to issue the collected parameters when it is determined that the real-time position of the UAV platform has reached the spatial anchor point coordinates. The real-time feedback verification module is used to calculate the actual spatial density based on the mapping point set, and send an overdrive interrupt signal when the actual spatial density is lower than the safety baseline threshold.

2. The UAV-borne lidar geographic mapping data acquisition system according to claim 1, characterized in that, The field of view partitioning module has a preset set of parameters that define the boundaries of the functional area. The logic of the field of view partitioning module to filter the look-ahead point set is that the vertical scanning angle of the sampling point is within the forward detection range. The look-ahead point set includes point cloud data projected by the laser beam in front of the flight direction of the UAV platform. The logic for the field-of-view partitioning module to filter the mapping point set is that the vertical scanning angle of the sampling point is within the mapping boundary angle range centered on the Nadir direction, and the mapping point set includes point cloud data projected by the laser beam directly below the UAV platform.

3. The UAV-borne lidar geographic mapping data acquisition system according to claim 1, characterized in that, The geometric prediction module is specifically used to: calculate the spatial normal vector of the foreground surface micro-element using the transformed look-ahead point set, define a virtual vertical down-looking laser beam vector, and calculate the angle between the spatial normal vector and the vertical down-looking laser beam vector as the predicted incident angle. The geometric prediction module is also used to traverse the adjacent scan points in the look-ahead point set, calculate the Euclidean distance difference between the adjacent scan points in the radar line-of-sight direction, and generate a true shadow area marker when the Euclidean distance difference exceeds the breakage threshold.

4. The UAV-borne lidar geographic mapping data acquisition system according to claim 3, characterized in that, The instruction generation module performs reverse calculations based on the equivalent spatial density model. The equivalent spatial density model represents that the effective point cloud density per unit surface area is directly proportional to the pulse emission frequency and inversely proportional to the product of the flight ground speed, the scanning structure constant, and the cosine of the predicted incident angle. When the instruction generation module receives a true shadow area marker, it ignores the result of the reverse calculation and forces the output of a preset security mode parameter.

5. The UAV-borne lidar geographic mapping data acquisition system according to claim 4, characterized in that, The instruction generation module employs a tiered adjustment strategy when calculating the acquired parameters: Prioritize maintaining the flight ground speed at the preset cruise speed, and calculate the target pulse frequency that can compensate for the influence of terrain slope; When the calculated target pulse frequency exceeds the maximum pulse frequency supported by the lidar sensor, the pulse emission frequency is locked to the maximum pulse frequency, and the flight speed is reduced to meet the target density.

6. The UAV-borne lidar geographic mapping data acquisition system according to claim 1, characterized in that, The instruction generation module calculates the coordinates of the spatial anchor point in the following way: Extract the horizontal position of the corresponding surface micro-element that generates the predicted incident angle in the world coordinate system, and set the horizontal position as the coordinate of the spatial anchor point.

7. The UAV-borne lidar geographic mapping data acquisition system according to claim 1, characterized in that, The condition for the spatial synchronization execution module to determine that the real-time position of the UAV platform has reached the coordinates of the spatial anchor point is: Calculate the Euclidean distance between the real-time position of the UAV platform and the coordinates of the spatial anchor point; when the Euclidean distance is less than a preset trigger threshold, it is determined that the UAV has arrived. The preset trigger threshold is determined by the sum of the basic positioning tolerance radius and the velocity feedforward term, where the velocity feedforward term is the product of the current flight speed modulus of the UAV platform and the comprehensive response time constant.

8. The UAV-borne lidar geographic mapping data acquisition system according to claim 1, characterized in that, The real-time feedback verification module uses a sliding window statistical method to calculate the actual spatial density, counts the number of survey points falling within a statistical window of a fixed size, and divides this number by the physical area of ​​the statistical window to obtain the actual spatial density.

9. The UAV-borne lidar geographic mapping data acquisition system according to claim 1, characterized in that, After receiving the overdrive interrupt signal, the space synchronization execution module clears the pending content in the space instruction queue and forcibly outputs preset safety mode parameters to adjust the hardware working state. The preset safety mode parameters include setting the pulse transmission frequency to the maximum pulse frequency and setting the flight ground speed to the minimum safe level flight speed.

10. A method for acquiring geographic mapping data using an unmanned aerial vehicle (UAV)-borne lidar, applied to the system described in any one of claims 1 to 9, characterized in that, Includes the following steps: Receive the raw laser point cloud data stream and split the raw laser point cloud data stream into a look-ahead point set and a mapping point set according to the vertical scanning angle threshold; By combining real-time pose data, the look-ahead point set is transformed to the world coordinate system, and the predicted incident angle of the surface micro-element is calculated. Based on the predicted incident angle, the acquisition parameters that meet the target density requirements are calculated in reverse, and the coordinates of the spatial anchor point used to trigger the acquisition parameters are calculated. The acquisition parameters include the pulse transmission frequency and the flight ground speed. The collected parameters are associated with and packaged with the spatial anchor point coordinates, and then stored in the spatial instruction queue in sequence. The real-time position of the drone platform is monitored in real time. When it is determined that the distance between the real-time position of the drone platform and the coordinates of a certain spatial anchor point in the spatial command queue is less than a preset trigger threshold, the corresponding acquisition parameters are sent. The actual spatial density is calculated using the mapping point set. When the actual spatial density is lower than the safety threshold, the override interrupt mechanism is triggered to clear the pending content in the spatial instruction queue and force a switch to the safety mode parameters.

Images