3D Modeling and UAV Landing Area Selection Method Based on Laser Point Cloud Data

Abstract

This invention relates to the field of unmanned aerial vehicle (UAV) automation, and particularly to a method for 3D modeling and UAV landing area selection based on laser point cloud data. The invention includes: acquisition and dynamic compensation of laser radar point cloud data, multi-scale adaptive registration of 3D point clouds, terrain feature extraction and evaluation, selection of the optimal landing point, path planning based on an improved A algorithm, and a real-time environmental data update and dynamic adjustment mechanism. Kalman filtering is used to dynamically compensate the point cloud data; the 3D-NDT algorithm is used for high-precision 3D environmental modeling; and the RANSAC algorithm is used to extract key terrain features to comprehensively evaluate terrain complexity and safe distance, selecting the optimal landing point. The improved A algorithm is then used to plan the optimal path for the UAV, ensuring path safety and flight efficiency. By updating environmental data in real time and dynamically adjusting the flight path and landing point, the invention adapts to sudden environmental changes, ensuring safe and stable autonomous landing of the UAV in complex and ever-changing environments.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) automation, and in particular to a method for 3D modeling and UAV landing area selection based on laser point cloud data, for autonomous landing missions of UAVs in complex terrain and low-visibility environments such as nighttime low light. Background Technology

[0002] With the rapid development of drone technology, drones are being widely used in various tasks, including logistics and transportation, and environmental monitoring. Autonomous flight and autonomous landing are key technologies for drones, especially at night or in complex terrain conditions, where safe landing becomes particularly important.

[0003] In traditional drone landing systems, visual sensors are typically the primary data source, assisting the drone in environmental perception. However, visual systems heavily rely on external lighting conditions, leading to a significant reduction in their effectiveness at night or in low-light environments. Furthermore, complex terrain (such as rugged mountainous areas or densely vegetated regions) poses challenges to visual sensors, making it difficult to guarantee the safety and accuracy of autonomous landing.

[0004] The introduction of LiDAR (Light Detection and Ranging) technology has provided a new solution for environmental perception of unmanned aerial vehicles (UAVs). LiDAR generates high-precision 3D point cloud data by emitting laser pulses and receiving reflected light, enabling it to operate effectively under any lighting conditions. However, how to efficiently select UAV landing areas using LiDAR point cloud data remains a complex technical problem. Existing LiDAR data processing methods suffer from high computational costs and poor real-time performance, and traditional terrain feature extraction methods have limited capabilities in handling complex terrain, failing to meet the needs of rapid and safe UAV landing.

[0005] Furthermore, during autonomous flight of a drone, dynamic environmental factors (such as moving obstacles or changes in wind speed) can affect the drone's path planning. Therefore, real-time updates of environmental data and adjustments to the flight path remain unresolved challenges in current technologies.

[0006] Based on the above background, this invention aims to provide an efficient and reliable method for UAV 3D environment modeling and optimal landing point selection by utilizing lidar point cloud data, improved algorithms, and dynamic environment update mechanisms. This method is particularly suitable for autonomous UAV landing missions at night or in complex terrain conditions. Summary of the Invention

[0007] This invention aims to address the challenges of autonomous landing of unmanned aerial vehicles (UAVs) at night and in complex terrain conditions. It proposes a method for 3D environment modeling and optimal landing point selection based on LiDAR point cloud data. By integrating multi-sensor data (such as IMU and LiDAR) and combining improved point cloud registration algorithms, terrain feature extraction methods, and path planning algorithms, this invention enables real-time perception of complex terrain environments during UAV flight, selection of the optimal landing area, and planning of a safe autonomous landing path.

[0008] This invention provides a systematic solution, mainly including the following aspects:

[0009] (1) Data acquisition: Point cloud data is acquired using lidar, and pose information of the human-machine interface is acquired through IMU sensor.

[0010] (2) Data preprocessing: The point cloud data is dynamically compensated by combining Kalman filtering with IMU data to correct data errors caused by UAV motion. The displacement and attitude data provided by the IMU can effectively reduce motion distortion in the point cloud and ensure the spatiotemporal consistency and accuracy of the data.

[0011] (3) 3D Point Cloud Registration: Based on the improved 3D Normal Distribution Transform (3D-NDT) algorithm, multi-scale adaptive registration of point clouds is performed. This algorithm provides higher accuracy in complex terrain while reducing the computational burden in simpler terrain, ensuring the real-time performance of the system. Through high-precision point cloud registration, an accurate 3D model of the environment in which the UAV is located is established.

[0012] (4) Terrain Feature Extraction and Evaluation: The RANSAC plane fitting algorithm was used to extract terrain features from the point cloud, including key terrain parameters such as slope, roughness, undulation, and safety distance. The robustness of the RANSAC algorithm can effectively handle noise and outliers in the point cloud data, ensuring the extraction of stable terrain features under complex terrain conditions. Through the quantitative evaluation of these features, the possible landing area of ​​the UAV was determined.

[0013] (5) Optimal Landing Point Selection: This invention proposes a multi-parameter-based landing point selection mechanism by comprehensively evaluating terrain complexity and safety distance. A comprehensive evaluation index is derived by weighted calculation of terrain complexity (including slope, roughness, and undulation) and safety distance. The area with the smallest comprehensive evaluation index is selected as the optimal landing point, thereby improving the safety and accuracy of UAV landing.

[0014] (6) Path Planning and Autonomous Landing: An improved A* path planning algorithm is used to generate the optimal path for the UAV from its current position to the best landing point. This algorithm combines obstacle risk assessment and flight energy consumption optimization mechanisms to ensure that the UAV can safely and efficiently plan the optimal path in complex environments. During flight, the algorithm can dynamically adjust the path based on the latest environmental data to ensure that the UAV avoids obstacles and lands safely.

[0015] (7) Dynamic update and adjustment mechanism: During the flight of the UAV, the lidar continuously scans the terrain, and the system continuously adjusts the optimal landing point and path planning by updating the point cloud data in real time. This mechanism can effectively cope with environmental changes, such as moving obstacles or sudden changes in wind speed, to ensure that the UAV always flies on the optimal path and lands safely.

[0016] The technical advantages of this invention are as follows: By deeply processing lidar point cloud data and comprehensively evaluating various terrain features (slope, roughness, undulation, and safe distance), this invention proposes a dynamic and adaptive autonomous landing method for unmanned aerial vehicles (UAVs). This method is applicable not only to low-visibility environments at night but also to various complex terrain conditions, such as mountainous areas and forests with dense obstacles. This invention effectively solves the limitations of traditional UAV autonomous landing systems in nighttime and complex terrain conditions. By combining IMU and LiDAR data, it provides a high-precision 3D environment modeling and terrain feature extraction method, which can accurately identify and select safe landing points. The improved A* path planning algorithm ensures path safety and energy efficiency optimization for UAVs in complex environments. The technical solution of this invention is applicable to various autonomous flight missions of UAVs, especially in complex flight environments such as nighttime, mountainous areas, and forests, and has broad application prospects. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating the hardware and software system workflow of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0020] This invention proposes an autonomous landing system for unmanned aerial vehicles (UAVs) that combines LiDAR point cloud data, inertial measurement unit (IMU) data, improved algorithms, and real-time dynamic adjustments. In complex and dynamic environments, this system can accurately perceive environmental information, select a safe landing point, and plan the optimal path. The specific implementation steps of this invention are as follows.

[0021] Step 1: Data Acquisition. During flight, drones need precise environmental perception, especially in complex terrain, areas with dense obstacles, or low visibility. Environmental perception systems are crucial in these situations. Traditional visual sensors rely on lighting conditions and perform poorly at night or in low-light environments. LiDAR, on the other hand, is a sensor perfectly suited for these scenarios, providing high-resolution 3D point cloud data under any lighting conditions.

[0022] Point cloud data contains the three-dimensional coordinates of objects around the drone, which form the terrain model and the distribution of obstacles. However, due to the movement of the drone during flight, these movements affect the accuracy of the point cloud, especially changes in the drone's speed, acceleration, and attitude, which can lead to distortion or inaccuracy in the point cloud.

[0023] Therefore, this invention introduces an IMU (Inertial Measurement Unit) to assist in correcting the motion errors of the UAV. The IMU can provide the UAV's real-time position information (such as acceleration, angular velocity, etc.).

[0024] Step 2, data preprocessing, involves combining the Kalman filter algorithm with LiDAR point cloud data for dynamic compensation, thereby eliminating motion errors during flight. The specific data processing and compensation flow is as follows:

[0025] IMU data integration: The IMU provides the drone's real-time motion status, including acceleration, angular velocity, attitude angle, and other information. This data is synchronized with LiDAR data to obtain accurate motion information at every moment.

[0026] Kalman filtering for state prediction: Kalman filtering is a recursive estimation method that predicts the UAV's motion state and then corrects it using actual observation data. The formulas for state prediction and covariance matrix prediction are as follows:

[0027]

[0028] P k|k-1 =F·P k-1|k-1 ·F T +Q+Q imu (2)

[0029] State Update and Correction: The second step of Kalman filtering is to correct the point cloud data using Kalman gain. Compared to traditional Kalman gain, we dynamically adjust the Kalman gain by combining displacement information provided by the IMU and point cloud data observed by LiDAR, as shown in the following formula:

[0030] K k =P k|k-1 ·H T ·(H·P k|k-1 ·H T +R+R imu ) -1 (3)

[0031] Then, state and covariance matrix updates are performed using Kalman gain. Traditional Kalman filtering relies solely on sensor observations for state updates, which cannot handle rapid attitude changes in dynamic systems such as UAVs. We introduce a rotation correction Δθ provided by the IMU. imu This is used to make additional corrections to the state, especially for dynamic compensation of rotation angles. This allows for a more accurate accounting of the actual motion of systems such as UAVs, particularly when there are attitude changes. The update formula is as follows:

[0032]

[0033] P k|k =(IK k ·H)·P k|k-1 (5)

[0034] This dynamic compensation mechanism is particularly suitable for UAVs operating at high speeds or with significant attitude changes, such as obstacle avoidance or rapid navigation. The compensated point cloud data provides an accurate 3D environmental description, eliminating errors caused by motion. This provides a solid data foundation for subsequent terrain modeling, feature extraction, and path planning.

[0035] Step 3, point cloud registration, is the process of aligning point cloud data collected at different times. During drone flight, the LiDAR continuously collects new point cloud data over time. To generate a complete 3D map, the system needs to accurately register these point cloud data collected at different times.

[0036] Traditional point cloud registration methods struggle to handle both complex and simple terrains simultaneously. Complex terrain requires high-precision processing, while using high precision in simple terrain leads to unnecessary computational overhead. Therefore, this invention proposes a multi-scale adaptive registration method based on 3D-NDT (Normal Distribution Transform).

[0037] The 3D-NDT algorithm divides point cloud data into multiple regions and models each region using a Gaussian distribution, expressed as:

[0038]

[0039] Furthermore, this invention incorporates an adaptive multi-scale mechanism into the system, dynamically adjusting the registration scale by calculating the complexity parameter λi for each region. For complex terrain, the system increases the resolution to ensure the accuracy of the point cloud data; for relatively flat areas, the system decreases the resolution to reduce computation. The adaptively adjusted covariance matrix is:

[0040]

[0041] This mechanism ensures that the system can operate efficiently under different terrain conditions. The registration process is completed iteratively. In each iteration, the system calculates the error between the source point cloud and the target point cloud, and adjusts the pose of the point cloud according to the error until the error meets the preset convergence condition.

[0042] Finally, the point cloud registration is completed iteratively. In each iteration, the system first calculates the error between the target point cloud and the source point cloud, and then adjusts the pose of the source point cloud according to the error value, so that the error between the two gradually decreases. Through continuous iterative adjustments, the error eventually meets the preset convergence condition, and the registration of the source point cloud and the target point cloud is completed.

[0043] The error is calculated based on the following formula:

[0044]

[0045] In this formula, E(T) represents the registration error. The system gradually adjusts the position and orientation of the source point cloud by minimizing this error in order to achieve optimal alignment with the target point cloud.

[0046] In each iteration, the source point cloud is adjusted using the transformation matrix T. This adjustment is based on the current registration error; the system corrects the transformation matrix T according to the magnitude of the error until the error reaches a preset convergence condition. Ultimately, the registration error between the source and target point clouds approaches its minimum, achieving accurate registration. Iteration effectively handles noise and outliers in the point cloud data, while also addressing feature extraction in complex terrain, ensuring high-precision point cloud registration.

[0047] After registration is complete, the map is updated using the strategy in the activity window, adding the newly added point clouds to the point cloud map. The update formula is as follows:

[0048] M k+1 =αM k +(1-α)N k (9)

[0049] This multi-scale adaptive registration method is particularly effective in complex 3D environments (such as mountains and forests). It not only ensures accuracy but also saves computational resources when dealing with simple shapes. During UAV missions, real-time and efficient registration ensures the accuracy of environmental maps, providing a basis for subsequent terrain feature extraction and landing point selection.

[0050] Step 4, Terrain Feature Extraction and Evaluation: When selecting a landing area, the UAV must understand the detailed characteristics of the terrain, including slope, roughness, and altitude variations. Traditional terrain feature extraction methods may be affected by noise and outliers, leading to inaccurate results. Therefore, this invention uses the RANSAC (Random Sample Consensus) algorithm for terrain feature extraction, ensuring accurate extraction of useful information even from noisy data.

[0051] Before feature extraction, the LiDAR point cloud data is first processed into a virtual grid. Each grid cell contains several point cloud data points, representing the terrain features within that cell. The entire terrain area is divided into N×M grid cells, each cell having a size of Δx×Δy. Then:

[0052] Before feature extraction, the LiDAR point cloud data is first processed into a virtual grid. Each grid cell contains several point cloud data points, representing the terrain features within that cell. The entire terrain region is divided into N×M grid cells, each cell having a size of Δx×Δy. The specific definitions are as follows:

[0053] G(i,j)={p i,j |x min +i·Δx≤x <x min +(i+1)·Δx,y min +j·Δy≤y <y min +(j+1)·Δy} (10)

[0055] Here, the boundary of each grid cell G(i,j) is defined by x. min and y min The term "determination" refers to the terrain region within a given cell, containing point cloud data used to describe the terrain features of that region.

[0056] The RANSAC algorithm selects the optimal planar model from a large amount of point cloud data and removes outliers through random sampling and consistency checks. The specific steps are as follows:

[0057] RANSAC first randomly selects three points from the point cloud to perform plane fitting, and the fitted plane equation is:

[0058] z = a·x + b·y + c

[0059] Where a, b, and c are the parameters of the fitted plane.

[0060] The system determines whether points belong to the fitted model by calculating the distances between other points and the fitted plane. If the distance from a point to the plane is less than a preset threshold, the point is considered an interior point; otherwise, it is considered noise or an outlier. RANSAC continuously samples randomly until it finds a planar model containing the most interior points.

[0061] After obtaining the planar model, the system further calculates the key parameters of the terrain:

[0062] Slope represents the inclination of terrain, and its formula is:

[0063]

[0064] Roughness represents the smoothness of the terrain, and the calculation formula is:

[0065]

[0066] Undulation represents the elevation variation of terrain, and the calculation formula is:

[0067] U = z max -z min (13)

[0068] The safe distance represents the distance between a grid cell G(i,j) and the nearest obstacle, and is calculated using the following formula:

[0069] D s =min(||G(i,j)-O k ||)(14)

[0070] Using the RANSAC algorithm, the system can accurately extract flat areas suitable for UAV landing, ensuring that useful terrain features can still be identified even in noisy environments. This feature extraction result will directly affect the subsequent landing point selection.

[0071] Step 5: In complex terrain, selecting the optimal landing point is crucial for the safe landing of the UAV. The optimal landing point not only requires flat terrain but also needs to consider the distribution of surrounding obstacles and safe distances. Therefore, this invention employs a multi-parameter comprehensive evaluation method to assess each potential landing point.

[0072] The system combines two main parameters, terrain complexity and safety, and calculates a comprehensive evaluation index Li for each region through weighted calculation.

[0073] Terrain complexity assessment: Terrain complexity is determined by the weighted sum of slope, roughness, and undulation, as shown in the formula:

[0074] T i =λ θ ·θ+λ R ·R+λ U ·U(15)

[0075] Where λθ, λR, and λU are weighting coefficients, representing the influence of slope, roughness, and undulation on terrain complexity, respectively.

[0076] Safe distance assessment: The safe distance is determined by calculating the distance between the landing point and surrounding obstacles, using the following formula:

[0077]

[0078] Where Ds represents the distance between the landing point and the nearest obstacle. The greater the safe distance, the safer the area.

[0079] Comprehensive Assessment: The system takes into account both terrain complexity and safety distance. The final assessment index Li is determined by the following formula:

[0080] L i =T i +τ·S i (17)

[0081] Where τ is the security weighting coefficient, representing the importance of security relative to terrain complexity.

[0082] Through this multi-parameter comprehensive evaluation method, the system can accurately select suitable landing areas and avoid landing in areas with dense obstacles or uneven terrain. This mechanism ensures the safety of the drone in complex environments.

[0083] Step 6: After determining the optimal landing point, the UAV needs a safe and efficient path to reach it. The A* algorithm is a heuristic search algorithm that finds the path with the minimum cost between a given starting point and a destination. Its core is to find the optimal path from the starting point to the target point by estimating the total cost f(n) of the path. The path cost calculation formula for the A* algorithm is as follows:

[0084] f(n) = g(n) + h(n) (18)

[0085] Where: f(n): represents the estimated total cost from the starting point to the target point via node n, g(n): represents the actual path cost from the starting point to node n, and h(n): represents the heuristic estimated cost from node n to the target point.

[0086] In the traditional A* algorithm, h(n) is typically estimated using Euclidean or Manhattan distance to determine the distance from node n to the target point. This method usually only considers path length, neglecting obstacle distribution and flight energy consumption. Therefore, this invention improves the A* algorithm by considering not only path length but also obstacle risk assessment and energy consumption optimization.

[0087] The improved A* algorithm incorporates three factors during path planning: path length, obstacle risk, and flight energy consumption, as detailed below:

[0088] Path cost function: The cost function of the A* algorithm is:

[0089] f′(n)=g(n)+h(n)+λ r ·R(n)(19)

[0090] Where g(n) is the actual cost from the starting point to the current node n, h(n) is the heuristic estimate from the current node to the target point, R(n) is the obstacle risk assessment value, representing the density of obstacles along the path, and λr is the weight coefficient of the risk assessment.

[0091] Energy Consumption Optimization: Addressing the limited energy resources of drones, energy consumption optimization is introduced during path planning. The energy consumption of each path node is calculated using an improved cost function based on factors such as flight distance, altitude changes, and ambient wind speed.

[0092] f″(n)=f′(n)+λ e ·E(n)(20)

[0093] Where E(n) is the estimated flight energy consumption, representing the energy consumption of the UAV from its current position to the target point, and λe is the energy consumption optimization weight.

[0094] By introducing a flight energy consumption optimization mechanism, the flight energy consumption of UAVs can be effectively reduced, especially in long-distance flight missions, where it can significantly improve endurance.

[0095] This invention addresses path planning in three-dimensional environments, particularly in complex terrain or areas with dense obstacles. Path planning needs to consider not only obstacles on a plane but also the distribution of obstacles at different height levels. The specific implementation is as follows:

[0096] Hierarchical path planning: The 3D environment is divided into multiple height layers, each corresponding to a 2D path planning problem. By running the A* algorithm on each height layer, the optimal path is selected, and the transformation cost between different height layers is considered to finally generate the 3D path.

[0097] Altitude-level transition cost: If a drone needs to transition between different altitude levels, the energy consumption and safety associated with altitude changes must be considered. The formula for calculating the altitude-level transition cost is as follows:

[0098] C h (n)=λ h ·Δh(21)

[0099] Among them, C h (n): Represents the height-level transformation cost at node n, λ h : Weight parameters for height layer transformation, Δh: Height difference between the current node and the target node.

[0100] Through altitude layering and transformation mechanisms, drones can flexibly plan paths in a three-dimensional environment, avoid high-risk areas, and ensure flight safety.

[0101] Implementation process of drone path planning:

[0102] 1. Acquire the drone's starting point, target point, and environmental information (such as obstacle distribution, wind speed, etc.), and divide the three-dimensional environment into multiple height layers.

[0103] 2. Based on the A* algorithm, path planning is performed at each height level to calculate the optimal path from the starting point to the target point, taking into account factors such as path length, risk, and energy consumption.

[0104] 3. Based on the conversion costs between different altitude layers, generate a three-dimensional path and select the path with the lowest cost as the final flight path.

[0105] During the drone's flight, environmental information (such as the addition of new obstacles or changes in wind speed) is updated in real time, and the flight path is dynamically adjusted based on the latest data.

[0106] Step 6: In real-world applications, the environment can be dynamically changing. For example, obstacles may move, and wind speed may increase, all of which can affect the drone's flight mission. Therefore, this invention designs a dynamic update and adjustment mechanism to ensure that the drone can adjust its flight path and landing point based on real-time environmental information during flight.

[0107] LiDAR and IMU provide real-time updated environmental information, including terrain changes and obstacle movement. The system continuously updates the 3D environment model based on this real-time data and recalculates the optimal landing point and path planning.

[0108] When the system detects changes in the environment (such as the appearance of new obstacles or changes in wind speed), it dynamically adjusts the flight path based on the latest data. The purpose of path adjustment is to ensure that the drone always chooses a safe and optimal route to reach the target point, while ensuring that the drone's energy efficiency is not significantly reduced.

Claims

1. A method for 3D modeling and UAV landing area selection based on laser point cloud data, characterized in that, Includes the following steps: Step 1, Acquire point cloud data: Use a lidar system to collect 3D point cloud data of the environment surrounding the drone; Step 2, Data Preprocessing: Perform Kalman filtering and coordinate transformation on the acquired 3D point cloud data; Step 3, Construct a 3D point cloud map: Based on the preprocessed point cloud data, the 3D Normal Distribution Transform (3D-NDT) algorithm is used for point cloud registration to generate a high-precision 3D point cloud map. The method for point cloud registration using the 3D Normal Distribution Transform (3D-NDT) algorithm is as follows: 1) The 3D-NDT algorithm divides point cloud data into multiple regions and models each region using a Gaussian distribution, expressed as: （6） in, : represents the probability distribution of the i-th region in the point cloud, x: the point to be registered. : Represents the mean vector of the i-th region. : represents the covariance matrix of the i-th region, d: the dimension of the point cloud, : indicates transpose; 2) An adaptive multi-scale registration mechanism was added to the basic 3D-NDT algorithm. The registration scale was dynamically adjusted by calculating the complexity parameter λi of each region. The adaptively adjusted covariance matrix is ​​as follows: （7） in, : Represents the covariance matrix after multi-scale adjustment. : Represents a multi-scale function, : Represents the region complexity parameter; 3) The registration process is completed iteratively. In each iteration, the error between the target point cloud and the source point cloud is calculated, and the pose of the point cloud is adjusted according to the error until the error meets the preset convergence condition. The registration is completed by minimizing the error between the target point cloud and the source point cloud according to the following formula: （8） in, : Represents the error function, : Represents a point in the target point cloud. : represents a point in the source point cloud, T: transformation matrix, t: represents the coordinates of a point in the target point cloud, s: represents the coordinates of a point in the source point cloud; Step 4, extract terrain features: Perform virtual grid processing on the 3D point cloud map, and extract terrain parameters, including slope, roughness, undulation and safety distance, through sliding window plane fitting. Step 5, Select the best landing area: Calculate the best landing point selection index based on terrain complexity and safety distance, and select the area with the smallest index as the landing area; Step 6, Path Planning and Autonomous Landing: Based on the selected optimal landing area, the improved A* path planning algorithm is used to plan the UAV's landing path and perform autonomous landing.

2. The method for 3D modeling and UAV landing area selection based on laser point cloud data according to claim 1, characterized in that, The three-dimensional point cloud data in step 1 consists of the three-dimensional coordinates of objects around the UAV, which form the terrain model and the distribution of obstacles.

3. The method for 3D modeling and UAV landing area selection based on laser point cloud data according to claim 1, characterized in that, Step 2 shown is as follows: 1) State Prediction: The motion state of the UAV is predicted by combining Kalman filtering with IMU data, and the process noise covariance matrix of the IMU is introduced. First, based on the state and input from the previous moment, predict the current state of the drone: （1） in, : Represents the state prediction at the current time k; F: State transition matrix, describing the state changes of the system from time k-1 to time k. B: State estimate at time k-1; ∠B: Control input matrix. : Control input, Displacement correction amount provided by the IMU; Then, the covariance matrix is ​​predicted: （2） in: : Represents the state covariance matrix at prediction time k. : Represents the state covariance matrix at time k-1, Q: Transpose of the state transition matrix; Q: Process noise covariance matrix. IMU noise covariance matrix; 2) Dynamic compensation for point cloud errors: Combining IMU noise information and LiDAR point cloud data, the Kalman gain is dynamically adjusted, and the pose and rotation of the point cloud are corrected using IMU attitude data to dynamically compensate for point cloud errors; the process is as follows: first, calculate the Kalman gain: （3） in, : represents the Kalman gain, H: observation matrix, R: Transpose of the observation matrix; Measurement noise covariance matrix. IMU measurement noise covariance matrix; Then the state prediction is updated using the observed data: （4） in, : The state after the update at time k : Represents the actual observed value at time k. Rotational correction amount provided by the IMU; Finally, update the state covariance matrix at time k: （5） in : The updated state covariance matrix, I: The identity matrix.

4. The method for 3D modeling and UAV landing area selection based on laser point cloud data according to claim 1, characterized in that, The method for generating the 3D point cloud map in step 3 is as follows: Based on the registered point cloud, at each time step, a sliding window local map update is performed using the following formula: （9） in, : Represents a local map in the next frame. : Current local map, : New point cloud data, α: smoothing coefficient, k: time step, used to represent the state update of the algorithm at different times.

5. The method for 3D modeling and UAV landing area selection based on laser point cloud data according to claim 1, characterized in that, Step 4 shown is as follows: First, the lidar point cloud data is processed into a virtual grid, with each grid cell containing several point cloud data points, representing the terrain features within that cell. The entire terrain area is divided into N×M grid cells, each cell having a size of Δx×Δy. Then: （10） in, : The i-th and j-th grid cell, Horizontal grid cell size, Vertical grid cell size; Secondly, for the point cloud data in each grid cell, the RANSAC algorithm is used for plane fitting. The RANSAC random sampling algorithm is used to fit a plane model that fits the point cloud data even when there is noise and outliers in the data.

6. The method for 3D modeling and UAV landing area selection based on laser point cloud data according to claim 1, characterized in that, Step 5 shown is as follows: Based on the planar model obtained by RANSAC fitting, the terrain parameters of the grid cells are calculated. The terrain parameters include: slope, roughness, undulation, and safety distance. The calculation formulas are as follows: Slope: The inclination of the fitted plane, representing the steepness of the terrain, calculated using the following formula: （11） in, : Represents the slope of the area; a and b: represent the parameters of the plane equation; Roughness: The average distance from each point in the point cloud to the fitted plane, representing the smoothness of the terrain. The calculation formula is: （12） Where R represents terrain roughness, : represents the height of the i-th point in the point cloud, N: represents the number of points in the grid cell; Undulation: The difference in elevation between the highest and lowest points in a local area, representing the degree of terrain undulation. The calculation formula is: （13） Where U represents the undulation of the terrain. and : Represents the highest and lowest points within the grid area; Safety distance: grid cell Distance to the nearest obstacle is calculated by determining the distance between the center point of each grid cell and the surrounding hazardous grid cells to ensure a safe landing. （14） in: : Indicates a safe distance, : Represents the center of the grid of surrounding obstacles; Based on the above parameters, the terrain complexity index Ti and safety index Si of the grid cells are calculated to evaluate the optimal landing point; the terrain complexity index Ti is obtained by weighting the slope, roughness, and undulation. （15） in, , , : Indicates slope, roughness, and undulation; The safety index Si is determined by the safety distance Ds: （16） Combining the terrain complexity index and the safety index, the landing evaluation index Li for each grid cell is finally calculated: （17） in, : Indicates the landing assessment index for this grid cell. : Represents the security weighting coefficient.

7. The method for 3D modeling and UAV landing area selection based on laser point cloud data according to claim 1, characterized in that, The process for selecting a landing area is as follows: 1) For the current flight area of ​​the UAV, the terrain is modeled and features are extracted based on lidar point cloud data to obtain the terrain complexity index Ti and safety distance index Si for each grid cell; 2) Calculate the comprehensive landing assessment index Li for all grid cells and sort them from smallest to largest. Generally, the smaller the Li value, the flatter the terrain, the fewer the obvious obstacles, and the farther away from potential danger zones. 3) Select the grid cell with the lowest Li value as the optimal landing point; if the area cannot meet the actual landing requirements, select the next area with a smaller Li value; 4) Send the coordinates of the optimal landing point to the UAV flight control system and perform path planning based on the selected area to ensure that the UAV can land safely.

8. The method for 3D modeling and UAV landing area selection based on laser point cloud data according to claim 1, characterized in that, The improved A* path planning algorithm shown in step 6 is as follows: Using the A* algorithm, its formula is... （18） 1) During path planning, the path length is considered, and obstacle risk assessment is incorporated. The risk of each path node is determined by the density and distance of surrounding obstacles. The improved cost function is: （19） in, : Represents the risk assessment value of node n, : Represents the risk weight parameter; 2) Energy consumption optimization was introduced during the path planning process. The energy consumption of each path node was calculated based on factors such as its flight distance, flight altitude changes, and ambient wind speed. The improved cost function is as follows: （20） in, : Represents the flight energy consumption of node n, : Represents the energy consumption weighting parameter; 3) Hierarchical path planning: The 3D environment is divided into multiple height layers, each corresponding to a 2D path planning problem; by running the A* algorithm on each height layer, the optimal path is selected, and the transformation cost between different height layers is combined to finally generate the 3D path; 4) Height-layer conversion cost: The formula for calculating the height-layer conversion cost is as follows: （21） in, : Represents the height-level transformation cost at node n. Weight parameters for height layer transformation. The height difference between the current node and the target node.

9. The method for 3D modeling and UAV landing area selection based on laser point cloud data according to claim 1, characterized in that, The implementation process for planning the landing path of a drone: 1) Acquire the starting point, target point, and environmental information of the UAV, and divide the three-dimensional environment into multiple height layers; 2) Based on the A* algorithm, path planning is performed at each height level to calculate the optimal path from the starting point to the target point, taking into account factors such as path length, risk, and energy consumption; 3) Combine the conversion costs between different altitude layers to generate a three-dimensional path, and select the path with the lowest cost as the final flight path; 4) During the drone's flight, environmental information is updated in real time, and the flight path is dynamically adjusted based on the latest data.

Images