A vertical take-off and landing facility unmanned aerial vehicle image site selection method and system for low-altitude economy

Abstract

The application provides a vertical take-off and landing facility unmanned aerial vehicle image site selection method and system facing low-altitude economy, and belongs to the technical field of image data processing and modeling. The method obtains high-overlap-rate images and laser point cloud data by means of oblique photogrammetry and laser radar scanning of a building roof by an unmanned aerial vehicle, constructs a real scene three-dimensional grid model and an obstacle surface three-dimensional model after pretreatment, filters a preliminary layout scheme based on the model according to the available area net height of the roof, carries out airspace safety simulation and noise propagation simulation on each scheme, calculates a minimum interval distance and a noise intensity distribution map, and filters a final site selection scheme according to airspace safety and noise constraints. The application realizes automatic and accurate site selection of the vertical take-off and landing facility in a complex roof environment.

Description

Technical Field

[0001] This invention belongs to the field of image data processing and modeling technology, specifically relating to an image-based location method and system for vertical take-off and landing (VTOL) facilities for low-altitude economic purposes. Background Technology

[0002] With the rapid rise of the low-altitude economy, vertical take-off and landing (VTOL) facilities, as key infrastructure for scenarios such as drone logistics and urban air traffic, face technical challenges in site selection and planning due to complex spatial constraints and multi-dimensional safety regulations. Existing low-altitude facility site selection technologies generally suffer from the problem of limited data acquisition dimensions. Traditional measurement methods, such as total stations or satellite remote sensing, struggle to obtain high-precision three-dimensional spatial information in complex scenarios such as building rooftops and urban canyons. In particular, there is a significant lack of information on the representation of fine structures such as rooftop equipment and the outlines of surrounding buildings, resulting in a large deviation between the basic site selection data and the actual scenario.

[0003] At the modeling and analysis level, traditional methods often rely on two-dimensional CAD drawings or simplified three-dimensional models, failing to effectively construct accurate digital twins that include elements such as the surface morphology of obstacles and airspace boundaries. This results in calculations of key parameters such as clearance height and safe distances for flight paths deviating from actual requirements. Furthermore, existing technologies lack multi-dimensional collaborative evaluation capabilities, often analyzing only a single dimension such as structural load-bearing capacity or airspace safety. They fail to couple environmental constraints such as noise propagation and electromagnetic environment with aviation safety regulations, leading to the frequent failure of site selection schemes due to conflicts in multiple constraints.

[0004] Especially in urban built environments, the site selection of vertical take-off and landing facilities must simultaneously meet multiple constraints such as building structural feasibility, civil aviation safety regulations, and urban planning requirements. However, in traditional technical processes, data collection, model building, and safety assessment are fragmented and have a low degree of automation, making it difficult to achieve an efficient closed loop from spatial feature extraction to site selection decision-making. Summary of the Invention

[0005] This invention overcomes the shortcomings of the prior art and provides a method and system for image-based site selection of vertical take-off and landing facilities for low-altitude economies.

[0006] To achieve the above objectives, the technical solution adopted by this invention is: a method for image-based site selection of vertical take-off and landing facilities for low-altitude economies, comprising the following steps:

[0007] Step S1: Obtain remote sensing data by using a drone to perform oblique photogrammetry and lidar scanning on the building roof; the remote sensing data includes high overlap images and lidar point cloud data;

[0008] Step S2: Preprocess the remote sensing data, and establish a real-scene 3D mesh model and an obstacle surface 3D model based on the preprocessed remote sensing data; the real-scene 3D mesh model is used to characterize the surface morphology of building roofs; the obstacle surface 3D model is used to define the airspace boundary for safe flight.

[0009] Step S3: Based on the real-world 3D mesh model and the 3D model of the obstruction surface, select a preliminary layout scheme according to the available clearance height of the roof area;

[0010] Step S4: For each preliminary layout scheme, perform airspace safety simulation and noise propagation simulation to obtain the minimum interval distance and noise intensity distribution map; airspace safety simulation is based on the distance detection between the flight path and the real-world 3D mesh model and the 3D model of the obstacle surface; noise propagation simulation is based on the calculation of the noise intensity distribution on the building roof;

[0011] Step S5: Based on the minimum spacing distance and noise intensity distribution map, select a site selection scheme according to airspace safety constraints and noise constraints.

[0012] Furthermore, the preprocessing in step S2 includes:

[0013] Image distortion correction and aerial triangulation refinement are performed on high overlap images;

[0014] Noise removal, data downsampling, and coordinate registration are performed on the laser point cloud data.

[0015] Furthermore, the methods for establishing a realistic 3D mesh model include:

[0016] Dense matching is performed on the preprocessed high overlap image to generate a dense point cloud;

[0017] Surface reconstruction is performed on dense point clouds to generate triangular meshes;

[0018] The texture information of high overlap images is projected onto a triangular mesh to generate a realistic 3D mesh model.

[0019] Furthermore, based on the preprocessed high overlap rate image and laser point cloud data, three-dimensional features including planar features, contour features and elevation features are extracted;

[0020] Based on three-dimensional features, laser point cloud data is classified into roof plan laser point cloud, roof equipment laser point cloud, and surrounding building laser point cloud through clustering and threshold judgment.

[0021] Furthermore, methods for establishing a three-dimensional model of the obstacle surface include:

[0022] Based on the laser point cloud of the roof equipment and the laser point cloud of the surrounding buildings obtained after classification, the area is divided into several independent regions according to spatial distribution.

[0023] For each independent region, the least squares method is used to fit it to a plane;

[0024] All fitted planes are spliced ​​together according to their spatial positions to form a continuous three-dimensional model of the obstacle surface.

[0025] Furthermore, preliminary layout schemes are selected based on the available clearance height of the roof area, including:

[0026] Generate a vector surface of the usable area of ​​the roof based on a realistic 3D mesh model;

[0027] Based on the available area vector surface of the roof, a digital elevation model is constructed using the inverse distance weighted interpolation method, and the clearance height of each grid cell is calculated to generate a clearance height raster map.

[0028] Based on the landing pad dimensions, the available area of ​​the roof is divided into a regular grid.

[0029] Grids with a clearance height greater than a preset threshold are selected as preliminary layout schemes.

[0030] Furthermore, airspace safety simulation includes:

[0031] Based on the approach and climb phases of the flight path, approach flight path models and climb flight path models are established respectively.

[0032] Calculate the minimum distance between the flight path and the vertices on the surface of the real-world 3D mesh model;

[0033] Calculate the minimum distance between the flight path and the 3D model of the obstruction surface;

[0034] Take the smaller of the two minimum distances as the minimum interval distance.

[0035] Furthermore, the noise propagation simulation includes:

[0036] Define the sound power level, spectral characteristics, and spatial location of the noise source;

[0037] Establish a noise propagation attenuation model that includes geometric divergence attenuation, air absorption attenuation, and ground reflection attenuation;

[0038] A noise intensity distribution map is generated based on the attenuation model.

[0039] Furthermore, the methods for selecting site selection schemes based on airspace safety constraints and noise constraints include:

[0040] Set the minimum interval distance threshold and the peak noise threshold for the outer area;

[0041] Preliminary layout schemes that simultaneously meet the minimum interval distance threshold and the noise peak threshold of the peripheral area are selected.

[0042] Another technical solution provided by the present invention: an image-based location system for vertical take-off and landing facilities for low-altitude economy, based on the above method, including a data acquisition and processing module, a three-dimensional modeling module, a preliminary screening module, a simulation module and a decision module connected in sequence;

[0043] The system comprises the following modules: a data acquisition and processing module for acquiring and preprocessing remote sensing data; a 3D modeling module for constructing realistic 3D mesh models and 3D models of obstacle surfaces; a preliminary screening module for screening preliminary layout schemes; a simulation module for conducting airspace safety simulations and noise propagation simulations; and a decision-making module for screening and outputting site selection schemes.

[0044] This invention addresses the shortcomings of the prior art and has the following beneficial effects:

[0045] This invention utilizes unmanned aerial vehicles (UAVs) to perform oblique photogrammetry and lidar scanning on building roofs, acquiring remote sensing data encompassing high-overlap images and laser point cloud data. After preprocessing, it constructs a realistic 3D mesh model and a 3D model of obstruction surfaces. This comprehensively captures the morphology of building roof surfaces, equipment distribution, and the 3D spatial characteristics of the surrounding environment, providing a digital spatial benchmark based on real-world scenarios for site selection analysis. This changes the traditional reliance on 2D drawings or manual measurements that leads to information gaps, achieving a shift from fuzzy qualitative analysis to precise quantitative modeling. It effectively solves the problem of incomplete spatial information acquisition in complex roof environments, freeing site selection assessment from subjective judgment biases.

[0046] Based on a three-dimensional model, this invention performs airspace safety simulation and noise propagation simulation on the preliminary layout scheme. By calculating the minimum distance between the flight path and the three-dimensional model and constructing a noise propagation model that includes geometric divergence attenuation, air absorption attenuation, and ground reflection attenuation, the invention quantitatively evaluates the site selection scheme from two dimensions: aviation safety regulations and environmental noise impact. It transforms abstract safety requirements into calculable specific parameters, breaking through the limitations of traditional methods that rely solely on experience or single-dimensional evaluation. This enables the site selection process to be more scientific and standardized, ensuring that the site selection scheme has verifiable technical basis in terms of structural safety and environmental adaptability, and solving the problem that traditional methods cannot accurately quantify safety regulations.

[0047] This invention combines UAV-based 3D modeling technology with airspace safety and noise propagation quantification simulation methods. The 3D model provides realistic spatial constraints for simulation, and simulation analysis is used to infer key safety parameters of the model, forming a complete technical chain from data acquisition and model building to safety assessment. In the complex and confined spaces of urban building rooftops, it can quickly and automatically complete the entire process from spatial feature extraction to safety feasibility judgment. Compared with the shortcomings of existing technologies, such as the disconnect between modeling and assessment or the single dimension, it achieves a dual improvement in site selection efficiency and reliability. It effectively solves the technical problem of traditional methods being unable to translate abstract regulations into specific engineering judgments in non-standard sites, enabling precise matching between aviation safety requirements and complex rooftop environments. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a flowchart of an image-based site selection method for vertical take-off and landing facilities (VTOL) geared towards the low-altitude economy.

[0050] Figure 2 This is an architecture diagram of a UAV image-based location selection system for vertical take-off and landing facilities geared towards the low-altitude economy. Detailed Implementation

[0051] 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. 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.

[0052] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein. Therefore, the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0053] Exemplary method:

[0054] like Figure 1 As shown, a method for UAV image-based site selection for vertical take-off and landing facilities aimed at the low-altitude economy includes the following steps:

[0055] Step S1: Obtain remote sensing data by using a drone to perform oblique photogrammetry and lidar scanning on the building roof; the remote sensing data includes high overlap images and lidar point cloud data;

[0056] Step S2: Preprocess the remote sensing data, and establish a real-scene 3D mesh model and an obstacle surface 3D model based on the preprocessed remote sensing data; the real-scene 3D mesh model is used to characterize the surface morphology of building roofs; the obstacle surface 3D model is used to define the airspace boundary for safe flight.

[0057] Step S3: Based on the real-world 3D mesh model and the 3D model of the obstruction surface, select a preliminary layout scheme according to the available clearance height of the roof area;

[0058] Step S4: For each preliminary layout scheme, perform airspace safety simulation and noise propagation simulation to obtain the minimum interval distance and noise intensity distribution map; airspace safety simulation is based on the distance detection between the flight path and the real-world 3D mesh model and the 3D model of the obstacle surface; noise propagation simulation is based on the calculation of the noise intensity distribution on the building roof;

[0059] Step S5: Based on the minimum spacing distance and noise intensity distribution map, select a site selection scheme according to airspace safety constraints and noise constraints.

[0060] As a core infrastructure in low-altitude economy scenarios such as drone logistics and urban air traffic, the site selection of vertical take-off and landing (VTOL) facilities must balance structural load-bearing capacity and aviation safety regulations in complex spaces such as urban building rooftops. The site selection of such facilities faces technical requirements including dense distribution of equipment on building rooftops, irregular spatial forms, and compliance with multi-dimensional constraints such as clearance height, safe distance for flight paths, and noise control. Efficient and accurate site selection schemes rely on the digital mapping of the three-dimensional rooftop environment and safety regulations. This embodiment proposes a site selection scheme based on drone imagery. It acquires rooftop remote sensing data through oblique photography and LiDAR scanning, constructs a realistic 3D model and an obstruction surface model, and combines airspace safety simulation and noise propagation analysis to achieve automated site selection decisions for VTOL facilities in complex scenarios.

[0061] Below, each step will be explained in detail.

[0062] For example, the drone is configured with a multi-rotor structure, equipped with an oblique photography camera and a LiDAR sensor. The oblique photography camera includes one vertical lens and four oblique lenses, used to simultaneously capture images of ground features from different angles; the LiDAR sensor has a scanning frequency range of 100-500kHz and a ranging range of 0.5-200m, used to acquire three-dimensional position information of the ground surface. The drone also carries a Global Navigation Satellite System (GNSS) receiver module, supporting dual-mode positioning of the BeiDou Navigation Satellite System and the Global Positioning System, used to simultaneously acquire the drone's position and time information during flight.

[0063] For example, high overlap images are Earth observation images where adjacent images overlap in certain areas, and are divided into orthophoto images and oblique images. Orthophoto images are taken at a vertical angle, with the shooting angle controlled at 0±5° (based on the horizontal plane downwards), to eliminate the influence of topographic undulations and projection differences. Oblique images are taken at an oblique angle, with the shooting angle controlled at 30-60° (based on the horizontal plane downwards), to supplement the side contour information of ground features.

[0064] Laser point cloud data is a dataset composed of a large number of discrete three-dimensional points. Each discrete point contains three-dimensional coordinate information (X, Y, Z). The X and Y coordinates represent the position of the ground feature on the horizontal plane, and the Z coordinate represents the elevation of the ground feature. It fully represents the surface geometry of the ground feature, including the undulation of building roofs, the spatial location of existing equipment, and the height profile of surrounding buildings.

[0065] For example, oblique photogrammetry ensures image overlap by planning and executing automated gridded flight paths and controlling drones to fly along preset paths.

[0066] The method for planning and executing automated gridded flight routes involves determining the flight route coverage area by considering the building rooftops and surrounding areas, including the entire building rooftop area and a surrounding area of ​​50-200m. Flight route spacing is calculated through... ,in Line spacing (m); The ground coverage width (m) of a single image; Lateral overlap rate (unitless, 0.6-0.8); ensures that images acquired along adjacent flight paths can form an effective overlap, avoiding areas where data is missed.

[0067] The planned flight path includes a flight altitude of 20-50m above the building roof and a speed of 3-8m / s. The vertical and oblique lenses of the tilt camera acquire images at 0.5-2s intervals, meeting the requirements of a 70-90% forward overlap and a 60-80% lateral overlap. Forward overlap refers to the overlap ratio between two adjacent images along the same flight path, while lateral overlap refers to the overlap ratio between corresponding images along adjacent flight paths.

[0068] After filtering out invalid images with motion blur, overexposure, or occlusion, high overlap images, including orthophotos and oblique images, are obtained.

[0069] For example, a lidar scan is performed by a drone flying at low altitude, controlling the lidar sensor to emit and receive laser beams, and calculating the three-dimensional coordinates of the building roof and surrounding ground surfaces.

[0070] The flight parameters for lidar scanning include a flight altitude of 5-30m above the building roof, a flight speed of 1-5m / s, and a scanning angle controlled between -30° and 30° (based on the UAV's flight direction).

[0071] The drone flies along a preset path, which is consistent with the flight path of the oblique photogrammetry, ensuring spatial matching between the laser point cloud data and the high-overlap image. During flight, the lidar sensor continuously emits laser beams, receives the beams reflected from the ground surface, and records the time difference between emission and reception.

[0072] Methods for calculating three-dimensional coordinates include using formulas Calculate the distance from the lidar sensor to the ground surface, where This indicates the distance from the lidar sensor to the ground surface, expressed in meters (m). This represents the speed of laser light in air, with a value of 3 × 10⁻⁶. 8 m / s; The time difference between laser beam emission and reception is expressed in seconds (s). Combined with the real-time position and attitude information of the UAV (obtained by the global navigation satellite system receiving module), the distance is converted into three-dimensional coordinates (X, Y, Z) of the ground surface. A large number of three-dimensional coordinate points are collected to form laser point cloud data.

[0073] High-overlap imagery and laser point cloud data employ a unified time and spatial reference. The time reference is synchronized with the Global Navigation Satellite System (GNSS) receiver module, recording timestamps with an accuracy controlled within 1 ms to ensure that the time difference between high-overlap imagery acquisition and laser point cloud data acquisition is less than 1 second. The 2000 National Geodetic Coordinate System (with both plane and elevation coordinates in meters) is used. The GNSS receiver module acquires the UAV's 3D coordinates in real time and correlates them with the high-overlap imagery and laser point cloud data, ensuring that the spatial location information of both types of remote sensing data is based on the same coordinate system and avoiding deviations in subsequent data processing due to coordinate system differences.

[0074] For example, preprocessing methods for high overlap images include image distortion correction and aerial triangulation to eliminate errors in the original image.

[0075] Image distortion correction is based on the intrinsic parameters of the tilt camera (focal length, principal point coordinates, radial distortion coefficient, and tangential distortion coefficient) in step S1. The high overlap rate image is geometrically corrected through the camera calibration model to eliminate image distortion caused by lens optical distortion. After correction, the geometric deviation of a single image is controlled within 0.01m.

[0076] Aerial triangulation densification uses distortion-corrected, high-overlap-rate images as input. The SIFT algorithm identifies corresponding feature points (such as stable ground features like roof edges and wall corners) between different images. Bundle adjustment is then used to calculate the extrinsic parameters (3D position and attitude angles) of each image, constructing a 3D control network covering the building roof and surrounding area. The densification results achieve a planar accuracy of no less than 0.05m and an elevation accuracy of no less than 0.1m, ensuring consistency between the spatial location of the images and the actual ground features.

[0077] For example, the real-scene 3D mesh model is constructed based on preprocessed high-overlap images, intuitively representing the surface morphology of building roofs and surrounding features. The steps are as follows:

[0078] Using the extrinsic parameters of the images obtained by aerial triangulation, dense matching is performed on high overlap images to generate a dense point cloud (point density of 50-200 points / m²) covering building rooftops and surrounding areas. 2 Each point contains three-dimensional coordinates (X, Y, Z) and color information.

[0079] Surface reconstruction is performed on dense point clouds. A Poisson surface reconstruction algorithm is used to generate continuous triangular meshes with a mesh resolution that matches the density of the dense point cloud (0.05-0.2m) to ensure that details such as roof texture and equipment corners can be clearly expressed.

[0080] The texture information of high overlap images is projected onto a triangular mesh, and color equalization processing is used to eliminate image exposure differences, so that the texture of the mesh model is consistent with the visual appearance of actual ground features.

[0081] The final generated realistic 3D mesh model covers the entire roof area of ​​the building and the surrounding 50-100m area.

[0082] For example, the preprocessing method for laser point cloud data includes sequential noise point removal, data downsampling, and coordinate registration.

[0083] Noise point removal employs a statistical filtering algorithm, calculating the average distance between each discrete point in the laser point cloud data and its surrounding 10-50 neighboring laser points. and standard deviation Set the filter threshold to to Outlier laser points (mainly due to measurement errors or bird interference) that are more than 5% away from the threshold are removed, with the removal rate controlled between 0.5% and 5%.

[0084] Data downsampling employs a voxel grid method, dividing the laser point cloud data into cubic voxels with a resolution of 0.05-0.2m. Each voxel retains one laser point (the closest point to the center of the voxel), reducing the amount of data (retaining 30-50% of the original points) while maintaining the integrity of geometric features.

[0085] For coordinate registration, select 3-10 feature points with the same name (which need to be clearly identified in both the real-world 3D mesh model and the laser point cloud data, such as the vertices of roof equipment and corner points). Calculate the transformation matrix using the iterative nearest point algorithm to align the laser point cloud data with the real-world 3D mesh model. After registration, the spatial position deviation should be ≤0.3m to ensure that the two types of data are based on the same coordinate system.

[0086] For example, methods for extracting three-dimensional features include extracting planar features, contour features, and elevation features based on preprocessed high overlap imagery and laser point cloud data.

[0087] The planar features are fitted from the laser point cloud data using the Random Sample Consensus Algorithm (RANSAC) to obtain parameters such as the slope and elevation of the roof plan. The planar fitting residual is ≤0.05m.

[0088] Contour features are extracted using edge detection (Canny operator) on high overlap images to extract the contour lines of roof equipment and surrounding buildings (such as the rectangular boundaries of equipment and the eaves of buildings), with a contour accuracy of ≤0.1m.

[0089] Elevation features are based on the Z-coordinate of laser point cloud data. The maximum, minimum and distribution range of elevation values ​​are statistically analyzed to distinguish the elevation differences between the roof plane, protruding equipment and surrounding buildings.

[0090] For example, the classification method for laser point cloud data includes classifying the laser point cloud data into roof plan laser point cloud, roof equipment laser point cloud, and surrounding building laser point cloud based on extracted three-dimensional features and through clustering and threshold judgment.

[0091] A roof planar laser point cloud refers to laser point cloud data whose elevation is within the main body of the building roof (within ±0.5m), whose reflection intensity is uniform (variance coefficient ≤0.1), and which conforms to planar characteristics, corresponding to the flat area of ​​the roof.

[0092] The laser point cloud of roof equipment refers to laser point cloud data with an elevation more than 0.5m above the roof plane, high reflection intensity (10%-30% higher than the roof plane), and a discrete and clustered distribution, corresponding to roof equipment such as air conditioning units and communication antennas.

[0093] The laser point cloud of surrounding buildings refers to the laser point cloud data that has an elevation difference of ≥3m from the laser point cloud of the building roof and is located within a range of 50m-100m around the building roof, corresponding to the roof or wall of the adjacent building.

[0094] For example, the method for constructing a 3D model of an obstruction surface includes constructing a 3D model of the obstruction surface based on the classified laser point cloud of the roof equipment and the laser point cloud of the surrounding buildings, and defining the airspace boundary for safe flight. The steps are as follows:

[0095] The laser point cloud of the rooftop equipment and the laser point cloud of the surrounding buildings are divided into several independent regions according to their spatial distribution (e.g., a single device, a section of wall). Each region is fitted to a plane using the least squares method, and the fitting formula (sloping plane equation) is as follows: ,in , The elevation (m) of the obstruction's surface at the plane coordinates (x, y); , Let x and y be the slope coefficients (unitless) in the x and y directions of the k-th segmented inclined plane, respectively, satisfying... ; For the constant term (m) of the k-th slice slope; The range of plane coordinates (m×m) corresponding to the k-th segmented inclined plane.

[0096] All fitted planes must satisfy the following conditions. ( The maximum allowable slope is set to 0.2-0.5 to ensure that the model can reflect the tilt state of the obstruction.

[0097] All the segmented planes are spliced ​​together according to their spatial positions to form a continuous three-dimensional model of the obstruction surface. The deviation between the model boundary and the actual obstruction edge is ≤0.2m.

[0098] Step S3 generates a vector surface of the available area of ​​the roof based on step S2, constructs a net height raster map, and selects preliminary layout schemes.

[0099] For example, the usable area vector surface of the roof is the usable area after excluding the space occupied by existing equipment on the roof. Specifically, a boundary point recognition algorithm based on K-nearest neighbors is used to select 10-30 neighboring points (within the range of K values) for each laser point in the laser point cloud on the roof plane, and calculate the mean angle between the normal vectors of the laser point and its neighboring points. And set a threshold for boundary point determination. It is 75°; when a certain laser point Greater than When the point is determined to be the boundary point of the roof plane, all boundary points are summarized to form a roof plane boundary point set (point density 2-8 points / m) to ensure coverage of the irregular edges of the roof plane.

[0100] Based on the set of boundary points of the roof plane, the Alpha Shapes algorithm is used for polygon fitting, and the parameters are controlled. (Value range 0.5-2m) Adjust the fitting effect: The smaller the value, the closer the fitted polygon is to the boundary point set, preserving more edge details; The larger the value, the smoother the polygon, which can eliminate local noise. During fitting, it is necessary to ensure that a simply connected closed region is generated, which completely contains all the roof plane laser point clouds in step S2, and finally obtain the initial roof plane region vector surface.

[0101] Extract the rooftop equipment laser point clouds obtained from the classification in step S2. For each rooftop equipment laser point cloud, calculate its projection range on the horizontal plane using the minimum bounding rectangle algorithm to obtain the equipment projection rectangle. The side length of the equipment projection rectangle must cover all horizontal coordinates (X, Y) of the corresponding rooftop equipment laser point cloud. Perform a Boolean difference operation between the initial rooftop planar region vector plane and all equipment projection rectangles: ,in Vector surface of the available area of ​​the roof (m) 2 ); For the initial roof plan area vector surface (m) 2 n represents the number of laser point clouds on the rooftop equipment (unitless); The horizontal projection rectangle (m) of the point cloud of the i-th device 2 ); The union operation (m) represents the union of projected rectangles from multiple devices. 2 ); Confirmation is required after calculation. Not empty, otherwise readjust. Alternatively, check the laser point cloud classification results of the rooftop equipment.

[0102] For example, a headroom grid is used to characterize the vertical clearance margin of each location within the available area of ​​the roof relative to the surface of the obstruction.

[0103] Specifically, after generating the available roof area vector surface, the roof area vector surface is converted into a digital elevation model (DEM) in the form of a regular grid. The DEM grid resolution (res) is set to 1-2m, and square grid cells are divided according to this resolution. The position is identified by the plane coordinates (X, Y) of the grid cell center. Inverse distance weighted interpolation (IDW) is used to calculate the elevation value of each grid cell based on the roof plane laser point cloud obtained in step S2. The interpolation formula is: ,in The elevation value (m) of the raster cell at coordinates (x, y); m is the number of roof plan point clouds involved in the interpolation (unitless), ranging from 10 to 20. Let be the elevation value (m) of the j-th rooftop plane point cloud; Let the horizontal distance (in meters) from the j-th rooftop plane point cloud to the center (x, y) of the raster cell be calculated using the formula: , , (The planar coordinates of the point cloud of the j-th roof plane). This is the distance weighting index (unitless, ranging from 1.5 to 3.0; the larger the value, the greater the influence of nearest neighbors on the interpolation result).

[0104] After completing the DEM construction, a 3D spatial interpolation calculation is performed to calculate the clearance height of each grid cell. The clearance height is defined as the difference between the elevation (DEM elevation) of a location within the usable roof area and the elevation of the corresponding obstruction surface. Based on the 3D obstruction surface model generated in step S2, the obstruction surface segmentation area belonging to the center of each grid cell is determined. Substitute the corresponding slope equation into the equation of the piecewise slope ( The elevation of the obstruction surface at that location was calculated. Then calculate the clearance height using the formula. ,in The net height (m) of the grid cell at coordinates (x, y); The DEM elevation (m) at this location; The elevation (m) of the obstruction surface at that location.

[0105] A net clearance height raster map is generated by arranging all raster cells according to planar coordinates and using gradient color coding. Areas with a net clearance height greater than 5m are represented in green, areas with a net clearance height between 0m and 5m are represented in yellow, and areas with a net clearance height less than 0m are represented in red. The raster map also includes annotations for coordinates, raster resolution, and net clearance height units.

[0106] For example, the initial layout options are selected based on the available roof area vector plane and the requirements for vertical take-off and landing facilities.

[0107] Specifically, this involves determining the dimensions of the landing pad for the vertical take-off and landing facility, including the length of the landing pad. With width The values ​​range from 10m to 20m, and a square landing pad is used. = (This is to adapt to the take-off and landing attitudes of the aircraft in different directions.)

[0108] Based on the set landing pad dimensions, the available area of ​​the roof is divided into a regularly laid-out grid. The side length of each grid is equal to the side length of the landing pad, ensuring that the grid area exactly accommodates one landing pad; the center coordinates of the grid... As the center coordinates of the landing pad corresponding to this grid, the boundary of the layout grid must be completely located inside the vector plane of the available area of ​​the roof.

[0109] Filter grids that meet the clearance height requirements and set a clearance height threshold. The clearance height is 2-4m (the specific value can be adjusted according to the aircraft model, such as 2-3m for small aircraft and 3-4m for medium aircraft). All greater than If the layout grid meets the clearance height requirements, it is determined to be a valid layout location.

[0110] Convert all valid layout locations into a preliminary layout scheme and mark the center coordinates of the landing pad, the side length of the landing pad, and the clearance height range of the corresponding area (i.e., the minimum and maximum clearance height range within the grid).

[0111] For example, a standard flight path for an aircraft includes an approach phase and a climb phase: the approach phase is the path the aircraft takes from high altitude to the landing pad, with the initial approach altitude set to 50-100m, the approach speed range to 8-15m / s, and the approach glide angle range to 3-5° (the glide angle is the angle between the flight path and the horizontal plane). The endpoint of the approach path is the center coordinate of the landing pad in the preliminary layout scheme. ,in The takeoff and landing pad elevation is obtained from the DEM data in step S3; the climb phase is the path the aircraft takes from the takeoff and landing pad to high altitude, and the starting climb altitude is set as... The climb termination altitude ranges from 100 to 200 meters, the climb speed ranges from 10 to 20 meters per second, and the climb elevation angle ranges from 8 to 12 degrees (the elevation angle is the angle between the flight path and the horizontal plane). The starting point of the climb path is the center coordinate of the takeoff and landing pad. Meanwhile, the horizontal deviation of the flight path is set to allowable range of ±2m to ensure coverage of minute attitude deviations during aircraft operation and improve the comprehensiveness of simulation verification.

[0112] Center of the take-off and landing pad A local coordinate system is established with the origin as the reference point. The X-axis of the local coordinate system is along the approach or climb direction, the Y-axis is perpendicular to the X-axis (horizontal direction), and the Z-axis is consistent with the Z-axis of the 2000 National Geodetic Coordinate System (vertical direction). Based on the approach and climb phases of the flight path, approach flight path models and climb flight path models are established respectively.

[0113] The approach trajectory model is Where t is the approach time (s), and its value ranges from 0 to 1. ; To approximate the total duration (s), by The approach start altitude and glide speed were calculated. Approach velocity (m / s); Approach glide angle (rad); Approach starting height (m); This represents the maximum horizontal offset (m) during the approach phase, with a value range of ±2m. , , These are the three-dimensional coordinates (m) of the spacecraft at any time t during the approach phase.

[0114] The climb trajectory model is Where t is the climb time (s), and its value ranges from 0 to 1. ; The total climb time (s) is denoted by . The climbing termination height and climbing speed were calculated. The climbing speed is (m / s). The elevation angle (rad) represents the climb angle. The final climb height (m); This represents the maximum horizontal offset (m) during the climbing phase, with a value range of ±2m. , , These are the three-dimensional coordinates (m) of the aircraft at any time t during the climb phase.

[0115] Extract the surface vertex set of the real-world 3D mesh model generated in step S2. (n is the number of vertices), and the equations of the piecewise inclined planes of the 3D model of the obstruction surface. ; on the approach and climb trajectories at time intervals Discrete points are taken from 0.1 to 0.5 seconds to form a trajectory point set. (m is the number of trajectory points).

[0116] Calculate the minimum distance between the trajectory points and the real-world 3D mesh model, including for each trajectory point. Calculate its relationship with all mesh vertices. Euclidean distance ,in Let be the distance (m) from the i-th trajectory point to the j-th grid vertex. Take all... The minimum value in the distance is taken as the distance between the trajectory point and the real-world 3D mesh model. Then take all trajectory points The minimum value in is denoted as (m).

[0117] Calculate the minimum distance between the trajectory point and the surface model of the obstacle, including for each trajectory point. Determine the surface segment k of the obstruction to which it belongs, and substitute it into the corresponding inclined plane equation to calculate the slope of the segment. elevation of The formula for the perpendicular distance from a trajectory point to the segment is: ,in Let be the distance (in meters) from the i-th trajectory point to the k-th obstruction surface segment. Take all... The minimum value in the equation is taken as the distance between the trajectory point and the surface model of the obstacle. Then take all trajectory points The minimum value in is denoted as (m).

[0118] Finally, and The smaller value in the equation is defined as the minimum spacing distance for this initial layout scheme. A minimum spacing distance report is generated, which includes the respective distances for the approach and climb phases. , The minimum interval distance is also specified, along with the flight path parameters (speed, angle, altitude).

[0119] An exemplary noise source type is defined as a composite source of aerodynamic and mechanical noise. The sound power level range of the noise source is set to 105-120 dB, the noise spectrum characteristics are a 1 / 3 octave band spectrum, and the center frequency covers 250-8000 Hz. Among them, the sound power level in the 250-2000 Hz band accounts for 60-70% (dominated by aerodynamic noise), and the sound power level in the 2000-8000 Hz band accounts for 30-40% (dominated by mechanical noise). The spatial location of the noise source is set to 1m-3m directly above the center of the take-off and landing pad in the preliminary layout scheme (corresponding to the noise radiation center when the aircraft is hovering, taking off, or landing), and is denoted as the noise source coordinates. ,in Elevation of the noise source (m). The elevation of the landing pad is shown in meters (m). The noise source radiation characteristics are set to omnidirectional radiation, that is, the noise propagates uniformly in the horizontal direction, and the vertical sound pressure level attenuation coefficient ranges from 0.5 to 1 dB / m (as the propagation height increases, the noise energy diffuses in the horizontal direction, resulting in attenuation in the vertical direction).

[0120] A mathematical model for noise propagation and attenuation is constructed. The total attenuation includes geometric divergence attenuation, air absorption attenuation, and ground reflection attenuation. The noise intensity (sound pressure level) at each point is obtained by calculating the total attenuation.

[0121] Geometric divergence decay ( The noise source is spherical wave radiation, and its energy dissipates with distance during propagation, manifesting as... ,in, The geometrical divergence attenuation (dB); The straight-line distance (m) from the point to the noise source. , The three-dimensional coordinates of the point are (m); This is a reference distance (m), with a value of 1m (industry standard reference distance).

[0122] Air absorption attenuation ( Noise is generated by the energy loss due to the vibration and friction of air molecules as it propagates through the air. ,in Air absorption attenuation (dB); The air absorption coefficient (dB / m) ranges from 0.01 to 0.05 dB / m (typical values ​​in the 250-8000 Hz frequency band at 25℃ and standard atmospheric pressure). The distance (m) is the straight-line distance from the point to the noise source.

[0123] Ground reflection attenuation ( For buildings with hard roofs (concrete or metal), noise is reflected off the ground and then superimposed on the direct sound, resulting in attenuation. The attenuation ranges from 3 to 6 dB (hard ground has a higher reflection coefficient and therefore a smaller attenuation; soft ground has a larger attenuation, and this is suitable for roof scenarios).

[0124] Total attenuation ( The sum of the three: .

[0125] Noise intensity (sound pressure level) The calculation formula is: ,in The noise sound pressure level (dB) at a point. 11 (dB) is the sound power level of the noise source; 11 (dB) is the conversion constant between the sound power level and the sound pressure level under spherical wave radiation (correction value at a reference distance of 1m).

[0126] The method for calculating noise intensity distribution includes determining the entire roof area (the roof area vector surface coverage range in step S3) and the surrounding horizontal area of ​​50m-100m, with a vertical height range of [missing information]. -2 (ground level, The elevation of the take-off and landing pad is (m) to +20 (above the roof); Divide the area into a regular three-dimensional grid with a grid resolution of 2-5m. For the center of each grid cell, substitute the noise propagation and attenuation model to calculate the noise sound pressure level. When the grid cell is located on the roof plane ( (This needs to be combined with the revised version) scope Calculate the distance from the point to the noise source: ,in Pick , , Three typical values ​​are calculated separately, and the average of the three calculation results is taken as the final noise sound pressure level of the grid cell to ensure coverage of all possible locations of the noise source in the vertical interval and improve the comprehensiveness of the simulation results.

[0127] The noise sound pressure levels of all grid units in the horizontal direction (i.e., the roof plane) are statistically analyzed and classified into levels according to the sound pressure level range: ≤60dB (low noise zone), 60dB-70dB (medium noise zone), 70dB-80dB (high noise zone), and >80dB (high noise zone). Different colors (such as blue, green, yellow, and red) are used to mark the areas of each level to generate a horizontal noise intensity distribution map. At the same time, the noise sound pressure level variation curve in the vertical direction (i.e., the vertical line from the center of the landing pad) is extracted, and the noise attenuation law at different heights is marked as supplementary data for the distribution map.

[0128] The output noise intensity distribution map should be labeled with coordinates, grid resolution (m), noise sound pressure level unit (dB), and noise source parameters (sound power level, location), and should also include statistics on the area proportion of each noise level region.

[0129] For example, step S5 extracts parameter indicators from the simulation results of step S4 for each preliminary layout scheme, including:

[0130] The minimum clearance distance, directly taken from the output of the airspace safety simulation verification in step S4, refers to the closest distance (m) between the aircraft's standard flight path (approach and climb phases) and the real-world 3D mesh model and obstacle surface model. This indicator is directly obtained in step S4 through the distance calculation between the trajectory points and the 3D model, reflecting airspace safety.

[0131] The noise peak value in the peripheral area was obtained from the noise intensity distribution map of the entire roof area and the surrounding 50-100m horizontal area, which was extracted during the noise propagation simulation verification in step S4. Step S4 had already divided this area into grid cells with a resolution of 2-5m. In this step, the peripheral area was further limited to grid cells more than 15m away from the center of the landing pad (avoiding the strong noise core area directly below the landing pad and focusing on the actual impact on the surrounding environment). The noise sound pressure level of all grid cells within this range was read, and the maximum value was taken as the noise peak value in the peripheral area.

[0132] For example, the preliminary layout scheme is screened for site selection based on the minimum interval distance and the noise peak in the surrounding area, including airspace safety constraints and noise constraints.

[0133] Among the airspace safety constraints, a minimum clearance distance of ≥7m is required, and this is directly compared with the minimum clearance distance; if it is not met, the site is excluded due to collision risk. For noise constraints, the peak noise level in the surrounding area must be ≤60dB, and this is directly compared with the extracted peak noise level in the surrounding area; if it is not met, the site is excluded due to excessive interference to the surrounding environment. Only preliminary layout schemes that simultaneously meet both airspace safety and noise constraints are included in the site selection scheme set.

[0134] After screening, the site selection schemes are sorted according to fixed rules, prioritizing the sorting by airspace safety redundancy (i.e., the difference between the minimum interval distance and 7m) from largest to smallest. The larger the difference, the more sufficient the airspace safety reserve. If the differences are the same, they are sorted by the noise peak value of the surrounding area from smallest to largest. The smaller the noise value, the less impact on the surrounding environment.

[0135] The final site selection report includes a list of site selection options (labeled with the option number, landing pad center coordinates, minimum clearance distance, and peak noise level in the surrounding area). In addition to the site selection option list, the following can also be attached simultaneously based on the label number:

[0136] The real-scene 3D mesh model (overlaid with landing pad layout) is taken from the real-scene 3D mesh model generated in step S2. The landing pad boundaries of each site selection scheme (with center coordinates and dimensions marked) are overlaid in the model to intuitively show the spatial location of the landing pad on the building roof and surrounding environment.

[0137] The roof clearance height grid map (marking the site selection area) is taken from the clearance height grid map generated in step S3. The take-off and landing pad range of each site selection scheme is marked with special color blocks in the map. Combined with the color gradient of the grid map, the airspace safety redundancy corresponding to the site selection scheme is presented intuitively.

[0138] The noise intensity distribution map (corresponding to the noise peak in the peripheral area) is taken from the noise intensity distribution map generated in step S4. The peripheral area beyond 15m from the center of the take-off and landing pad is marked with a dashed box in the map, and the noise peak position corresponding to each site selection scheme is marked. Combined with the color markings in the map, the noise impact range of different schemes is visualized.

[0139] Exemplary system:

[0140] like Figure 2 As shown, a UAV image location system for vertical take-off and landing facilities oriented towards the low-altitude economy includes a data acquisition and processing module, a 3D modeling module, a preliminary screening module, a simulation module, and a decision module connected in sequence.

[0141] The data acquisition and processing module is used to acquire remote sensing data of the building roof and surrounding area, and to preprocess the remote sensing data, including:

[0142] The data acquisition unit is configured to control a drone equipped with an oblique photography camera and a lidar sensor to perform oblique photogrammetry and lidar scanning on the building roof along a preset route, acquire high overlap images and lidar point cloud data, and simultaneously record time and space references.

[0143] The data preprocessing unit is configured to perform distortion correction and aerial triangulation densification on high overlap images, noise removal, downsampling, and coordinate registration on laser point cloud data, eliminate errors, and unify data benchmarks.

[0144] The 3D modeling module is used to construct realistic 3D models and obstacle surface models based on preprocessed remote sensing data, intuitively representing roof morphology and airspace boundaries, including:

[0145] The real-scene 3D modeling unit is configured to generate dense point clouds by dense matching of preprocessed high-overlap rate images, and generate real-scene 3D mesh models by surface reconstruction and texture projection, covering the roof and surrounding areas.

[0146] The 3D feature extraction unit is configured to extract planar features, contour features, and elevation features from preprocessed high overlap images and laser point cloud data.

[0147] The obstacle modeling unit is configured to classify laser point cloud data based on three-dimensional features and construct a three-dimensional model of the obstacle surface to define the safe flight airspace boundary.

[0148] The preliminary screening module is used to calculate the clear height of the available roof area based on the 3D model and to screen preliminary layout schemes that meet basic airspace requirements, including:

[0149] The available area analysis unit is configured to generate a vector surface of the available area of ​​the roof based on the roof plan point cloud and the equipment point cloud, through boundary recognition and Boolean operations, excluding the space occupied by equipment.

[0150] The clearance height calculation unit is configured to build a digital elevation model based on the available area vector surface and generate a clearance height raster map through inverse distance weighted interpolation and obstruction elevation calculation.

[0151] The layout filtering unit is configured to divide the grid into regular grids based on the landing pad size parameters, and filter the grids with a clearance height greater than a preset threshold as the initial layout scheme.

[0152] The simulation module is used to simulate airspace safety and noise propagation for the preliminary layout scheme, including:

[0153] The airspace safety simulation unit is configured to build approach and climb trajectory models, calculate the minimum clearance between the flight path and the real-world 3D model and obstacle models, and assess collision risk.

[0154] The noise propagation simulation unit is configured to define the characteristics of the noise source, build a propagation model that includes geometric divergence, air absorption and ground reflection attenuation, calculate the noise intensity and generate a distribution map.

[0155] The decision module is used to select the final site selection scheme based on simulation results, including:

[0156] The constraint screening unit is configured to set a minimum interval distance and a peak threshold for peripheral noise to screen schemes that simultaneously meet airspace safety and noise constraints.

[0157] The scheme sorting unit is configured to sort the selected schemes according to spatial safety redundancy and noise peak value to determine the priority order.

[0158] Based on the preferred embodiments of the present invention described above, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A method for image-based location selection of vertical take-off and landing (VTOL) facilities for low-altitude economies, characterized in that, Includes the following steps: Step S1: Obtain remote sensing data by using a drone to perform oblique photogrammetry and lidar scanning on the building roof; the remote sensing data includes high overlap images and lidar point cloud data; Step S2: Preprocess the remote sensing data, and establish a real-scene 3D mesh model and an obstacle surface 3D model based on the preprocessed remote sensing data; the real-scene 3D mesh model is used to characterize the surface morphology of building roofs; the obstacle surface 3D model is used to define the airspace boundary for safe flight. Step S3: Based on the real-world 3D mesh model and the obstacle surface 3D model, a preliminary layout scheme is selected according to the available clearance height of the roof area; Step S4: For each preliminary layout scheme, perform airspace safety simulation and noise propagation simulation to obtain the minimum spacing distance and noise intensity distribution map; The airspace safety simulation is based on distance detection between the flight path and the real-world 3D mesh model and the 3D model of the obstacle surface; The noise propagation simulation is based on the calculation of the noise intensity distribution on the building roof; Step S5: Based on the minimum spacing distance and noise intensity distribution map, select a site selection scheme according to airspace safety constraints and noise constraints; The airspace safety simulation includes: Based on the approach and climb phases of the flight path, approach flight path models and climb flight path models are established respectively. Calculate the minimum distance between the flight path and the vertices on the surface of the real-world 3D mesh model; Calculate the minimum distance between the flight path and the 3D model of the obstruction surface; Take the smaller of the two minimum distances as the minimum interval distance; The noise propagation simulation includes: Define the sound power level, spectral characteristics, and spatial location of the noise source; Establish a noise propagation attenuation model that includes geometric divergence attenuation, air absorption attenuation, and ground reflection attenuation; A noise intensity distribution map is generated based on the attenuation model.

2. The method according to claim 1, characterized in that, The preprocessing described in step S2 includes: Image distortion correction and aerial triangulation are performed on the high overlap rate images; The laser point cloud data is subjected to noise point removal, data downsampling, and coordinate registration.

3. The method according to claim 1, characterized in that, The method for establishing the real-scene 3D mesh model includes: Dense matching is performed on the preprocessed high overlap image to generate a dense point cloud; The surface of the dense point cloud is reconstructed to generate a triangular mesh. The texture information of the high overlap image is projected onto the triangular mesh to generate a realistic 3D mesh model.

4. The method according to claim 1, characterized in that, Based on preprocessed high overlap images and laser point cloud data, three-dimensional features including planar features, contour features and elevation features are extracted. Based on the aforementioned three-dimensional features, the laser point cloud data is classified into roof plan laser point cloud, roof equipment laser point cloud, and surrounding building laser point cloud through clustering and threshold judgment.

5. The method according to claim 4, characterized in that, The method for establishing the three-dimensional model of the obstruction surface includes: Based on the laser point cloud of the roof equipment and the laser point cloud of the surrounding buildings obtained after classification, the area is divided into several independent regions according to spatial distribution. For each independent region, the least squares method is used to fit it to a plane; All fitted planes are spliced ​​together according to their spatial positions to form a continuous three-dimensional model of the obstacle surface.

6. The method according to claim 1, characterized in that, The preliminary layout scheme selected based on the available clearance height of the roof area includes: Generate a vector surface of the usable area of ​​the roof based on a realistic 3D mesh model; Based on the available area vector surface of the roof, a digital elevation model is constructed using the inverse distance weighted interpolation method, and the clearance height of each grid cell is calculated to generate a clearance height grid map. Based on the landing pad dimensions, the available area of ​​the roof is divided into a regular grid. Grids with a clearance height greater than a preset threshold are selected as preliminary layout schemes.

7. The method according to claim 1, characterized in that, The method for selecting site selection schemes based on airspace safety constraints and noise constraints includes: Set the minimum interval distance threshold and the peak noise threshold for the outer area; Preliminary layout schemes that simultaneously meet the minimum interval distance threshold and the noise peak threshold of the peripheral area are selected.

8. A UAV image-based location system for vertical take-off and landing facilities oriented towards a low-altitude economy, based on the method of any one of claims 1-7, characterized in that, It includes a data acquisition and processing module, a 3D modeling module, a preliminary screening module, a simulation module, and a decision-making module that are connected sequentially; The data acquisition and processing module is used to acquire and preprocess remote sensing data; the 3D modeling module is used to construct a real-scene 3D mesh model and an obstacle surface 3D model; the preliminary screening module is used to screen preliminary layout schemes; the simulation module is used to perform airspace safety simulation and noise propagation simulation; and the decision module is used to screen and output site selection schemes.

Images