Terrain survey method and device based on unmanned aerial vehicle cooperation, and electronic equipment

Abstract

The invention provides a terrain surveying method and device based on unmanned aerial vehicle cooperation and electronic equipment, and relates to the technical field of engineering surveying and complex terrain surveying and mapping. The method comprises the following steps: determining a first surveying flight route of an unmanned aerial vehicle in combination with ground elevation data of a to-be-surveyed area, controlling the unmanned aerial vehicle to collect initial area topographic data, identifying and classifying the initial area topographic data, and determining a weak texture false topographic area; re-planning a second survey flight route of the unmanned aerial vehicle according to the weak texture false terrain area, starting the millimeter wave radar, and scanning and collecting supplementary terrain survey data at different angles; fusing the supplementary topographic survey data at different angles to obtain local topographic survey data; and splicing and fusing the initial regional topographic data and the local topographic survey data to obtain real topographic survey data. According to the technical scheme, the measuring precision and the surveying efficiency of the topographic surveying data of the complex topographic region can be improved, and the reliability and the stability of the real topographic surveying data are improved.

Description

Background Art

[0002] Complex terrain (such as cliffs, dense forests, snow on the plateau, etc.) poses a severe challenge to traditional terrain survey technology. Complex terrain mapping is an important research direction in current engineering survey and geographic information system construction. However, traditional terrain survey technology is often difficult to obtain efficient and accurate data in environments such as high altitude and high drop areas, vegetation or snow cover, large surface fluctuations, and inconvenient transportation. For example, manual ground measurement methods require manual deployment of points and point-by-point data collection, which is not only time-consuming and labor-intensive, but also has the risk of missing key terrain features.

[0003] At present, the relevant technology mainly relies on the LiDAR (Light Detection And Ranging, LiDAR) and optical cameras carried by drones to measure terrain data. However, in areas with high altitudes and high drop, large terrain undulations, dense vegetation and snow cover, or weak surface texture, there are often situations where the terrain data measurement accuracy is insufficient, the survey efficiency is low, key terrain features are missed, or some key terrain feature data are false terrain feature data, resulting in poor reliability and stability of the measured terrain data. Summary of the invention

[0004] The purpose of the embodiments of the present disclosure is to provide a terrain survey method based on drone collaboration, a terrain survey device based on drone collaboration, an electronic device and a computer-readable storage medium, so as to improve the measurement accuracy and survey efficiency of terrain survey data in complex terrain areas and improve the reliability and stability of real terrain survey data.

[0005] Other features and advantages of the present disclosure will become apparent from the following detailed description, or may be learned in part by the practice of the present disclosure.

[0006] According to a first aspect of an embodiment of the present disclosure, a terrain survey method based on UAV collaboration is provided, comprising: Acquire ground elevation data of the area to be surveyed, and determine first survey flight routes of at least two of the UAVs in combination with the ground elevation data; Based on the first survey flight route, the UAV is controlled to collect initial area terrain data, where the initial area terrain data is collected in real time by a laser radar and an optical camera carried by the UAV; Identify and classify the initial area terrain data to determine a weak texture false terrain area in the initial area terrain data; Replanning second survey flight routes respectively corresponding to at least two of the UAVs according to the weak-texture false terrain area, wherein the second survey flight routes are used to control at least two of the UAVs to turn on millimeter-wave radars and scan and collect supplementary terrain survey data of the weak-texture false terrain area at different angles; Fusing the supplementary terrain survey data at different angles to obtain local terrain survey data of the weak-texture false terrain area; The initial area terrain data and the local terrain survey data are spliced ​​and fused to obtain the real terrain survey data of the area to be surveyed.

[0007] According to a second aspect of an embodiment of the present disclosure, a terrain survey device based on drone collaboration is provided, comprising: An initial flight survey module, used to obtain ground elevation data of the area to be surveyed, and determine a first survey flight route of at least two of the UAVs in combination with the ground elevation data; A terrain data acquisition module, used for controlling the UAV to acquire terrain data of an initial area based on the first survey flight route, wherein the terrain data of the initial area is acquired in real time by a laser radar and an optical camera carried by the UAV; A false terrain region identification module, used to identify and classify the initial region terrain data, and determine a weak texture false terrain region in the initial region terrain data; A survey route updating module, used for re-planning second survey flight routes corresponding to at least two of the drones respectively according to the weak texture false terrain area, wherein the second survey flight routes are used for controlling at least two of the drones to turn on millimeter wave radars and scan and collect supplementary terrain survey data of the weak texture false terrain area at different angles; A supplementary survey data fusion module, used to fuse the supplementary terrain survey data at different angles to obtain local terrain survey data of the weak texture false terrain area; The real terrain data output module is used to splice and fuse the initial area terrain data and the local terrain survey data to obtain the real terrain survey data of the area to be surveyed.

[0008] According to a third aspect of an embodiment of the present disclosure, there is provided an electronic device, comprising: a processor; and a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, any one of the above-mentioned terrain survey methods based on drone collaboration is implemented.

[0009] According to a fourth aspect of an embodiment of the present disclosure, there is provided a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for terrain surveying based on UAV collaboration as described in any one of the above items is implemented.

[0010] The technical solution provided by the embodiments of the present disclosure may have the following beneficial effects: The terrain survey method based on UAV collaboration in the example embodiment of the present disclosure, on the one hand, can effectively reduce the problems of UAV route overlap and insufficient blind area coverage by dynamically planning the survey flight routes of multiple UAVs based on the ground elevation data of complex terrain areas, so that the collaborative operation of multiple UAVs can cover complex terrain areas more efficiently and evenly, and improve the efficiency of surveying terrain features in the survey area; in addition, the initial area terrain data is collected in real time by using the laser radar and optical camera carried by the UAV, so that richer and more multi-dimensional terrain information can be obtained under different terrain conditions, the spatial perception ability of the target area can be enhanced, and the precision and accuracy of the measured terrain feature data can be improved; on the other hand, by identifying and classifying the initial area terrain data, weak texture false terrain areas and real terrain areas can be effectively distinguished, thereby avoiding the inclusion of misidentified false terrain features into the terrain modeling data, which helps to improve the authenticity and integrity of the terrain data. Furthermore, based on the identified weak-texture false terrain area, the flight routes of multiple UAVs are replanned, and the UAVs are controlled to enable millimeter-wave radars to perform multi-angle scanning to collect supplementary data, so that the UAVs can break through the penetration limitations of traditional lidars under complex terrain conditions, thereby making up for the blind spots in data collection, achieving comprehensive perception of the surface terrain, avoiding omission or misidentification of key terrain feature data, and further improving the precision and accuracy of the measured terrain feature data, and improving the reliability and stability of real terrain survey data; on the other hand, by fusing the supplementary terrain data collected at different angles and efficiently splicing and fusing them with the initial area terrain data, it can effectively reduce data splicing errors and model gaps, and optimize the spatial continuity and accuracy of terrain data; combined with real-time data processing and multi-source data fusion strategies, the density and integrity of point cloud data can be significantly improved, thereby effectively reducing information loss and data redundancy problems during data collection.

[0011] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the accompanying drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0013] Figure 1The flowchart of the terrain survey method based on UAV collaboration according to some embodiments of the present disclosure is schematically shown.

[0014] Figure 2 A flowchart of generating a first survey flight route according to some embodiments of the present disclosure is schematically shown.

[0015] Figure 3 The flowchart of dividing the survey sub-areas according to some embodiments of the present disclosure is schematically shown.

[0016] Figure 4 The following schematically shows a flow chart of determining a weak-texture false terrain area according to some embodiments of the present disclosure.

[0017] Figure 5 A schematic diagram of a process for determining spatial features of a point cloud according to some embodiments of the present disclosure is schematically shown.

[0018] Figure 6 A flowchart of generating a second survey flight route according to some embodiments of the present disclosure is schematically shown.

[0019] Figure 7 The flowchart of generating local terrain survey data according to some embodiments of the present disclosure is schematically shown.

[0020] Figure 8 The schematic diagram shows the composition of a terrain survey device based on UAV cooperation according to some embodiments of the present disclosure.

[0021] Fig. 9 A schematic diagram of the structure of a computer system of an electronic device according to some embodiments of the present disclosure is schematically shown.

[0022] Fig.10 A schematic diagram of a computer-readable storage medium according to some embodiments of the present disclosure is schematically shown.

[0023] In the drawings, the same or corresponding reference numerals represent the same or corresponding parts. DETAILED DESCRIPTION

[0024] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with this specification. Instead, they are merely examples of devices and methods consistent with some aspects of this specification as detailed in the appended claims.

[0025] Furthermore, the drawings are only schematic illustrations and are not necessarily drawn to scale. The block diagrams shown in the drawings are only functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0026] In this example embodiment, a terrain survey method based on UAV collaboration is first provided. This terrain survey method based on UAV collaboration can be applied to UAV equipment or a host computer device of a UAV equipment, and can also be applied to a server of a terrain survey control center. This example embodiment does not make any special limitations on this. Figure 1 The following schematically shows a flow chart of a terrain survey method based on UAV collaboration according to some embodiments of the present disclosure. Figure 1 As shown, the terrain survey method based on UAV collaboration may include the following steps: Step S110, obtaining ground elevation data of the area to be surveyed, and determining a first survey flight route of at least two of the UAVs in combination with the ground elevation data; Step S120, controlling the UAV to collect initial area terrain data based on the first survey flight route, wherein the initial area terrain data is collected in real time by a laser radar and an optical camera carried by the UAV; Step S130, identifying and classifying the initial area terrain data, and determining a weak texture false terrain area in the initial area terrain data; Step S140, replanning second survey flight routes corresponding to at least two of the UAVs respectively according to the weak-texture false terrain area, wherein the second survey flight routes are used to control at least two of the UAVs to turn on millimeter-wave radars and scan and collect supplementary terrain survey data of the weak-texture false terrain area at different angles; Step S150, fusing the supplementary terrain survey data at different angles to obtain local terrain survey data of the weak-texture false terrain area; Step S160: splicing and fusing the initial area terrain data and the local terrain survey data to obtain real terrain survey data of the area to be surveyed.

[0027] According to the terrain survey method based on UAV collaboration in this example embodiment, on the one hand, by dynamically planning the survey flight routes of multiple UAVs based on the ground elevation data of complex terrain areas, the problems of overlapping UAV routes and insufficient blind area coverage can be effectively reduced, so that the collaborative operation of multiple UAVs can cover complex terrain areas more efficiently and evenly, and the survey efficiency of terrain features in the survey area can be improved; in addition, the initial area terrain data is collected in real time using the laser radar and optical camera carried by the UAV, which can obtain richer and multi-dimensional terrain information under different terrain conditions, enhance the spatial perception ability of the target area, and improve the precision and accuracy of the measured terrain feature data; on the other hand, by identifying and classifying the initial area terrain data, weak texture false terrain areas and real terrain areas can be effectively distinguished, thereby avoiding the inclusion of misidentified false terrain features into the terrain modeling data, which helps to improve the authenticity and integrity of the terrain data. Furthermore, based on the identified weak-texture false terrain areas, the flight routes of multiple UAVs are replanned, and the UAVs are controlled to enable millimeter-wave radars to perform multi-angle scanning and collect supplementary data, so that the UAVs can break through the penetration limitations of traditional lidars under complex terrain conditions, thereby making up for the blind spots in data collection, achieving comprehensive perception of the surface terrain, avoiding omission or misidentification of key terrain feature data, and further improving the precision and accuracy of the measured terrain feature data, and improving the reliability and stability of real terrain survey data; on the other hand, by fusing the supplementary terrain data collected at different angles and efficiently splicing and fusing them with the initial area terrain data, it can effectively reduce data splicing errors and model gaps, and optimize the spatial continuity and accuracy of terrain data; combined with real-time data processing and multi-source data fusion strategies, the density and integrity of point cloud data can be significantly improved, thereby effectively reducing information loss and data redundancy problems during data collection.

[0028] Next, the terrain survey method based on UAV collaboration in this example embodiment will be further described.

[0029] In step S110, ground elevation data of the area to be surveyed is acquired, and a first survey flight route of at least two of the UAVs is determined in combination with the ground elevation data.

[0030] In an example embodiment of the present disclosure, ground elevation data refers to data used to characterize the terrain undulation characteristics of the target area. For example, the ground elevation data can be obtained through remote sensing images, digital elevation models (DEM), and digital surface models (DSM). Specifically, the ground elevation data can be obtained by analyzing satellite remote sensing images to obtain DEM or DSM data, or through historical surveying and mapping data or existing geographic information system (GIS) data. Of course, the ground elevation data can also be obtained through preliminary aerial surveys. This embodiment does not specifically limit the method for obtaining the ground elevation data.

[0031] The drone may be a multi-rotor drone or a fixed-wing drone equipped with multiple sensors. The sensors carried by the drone may include one or more combinations of lidar, optical cameras, millimeter-wave radars, and infrared cameras. It is understandable that the drone may also be equipped with navigation equipment such as an inertial navigation system (INS) and a global positioning system (GPS) for precise positioning and attitude control. This embodiment does not specifically limit the type and configuration of the drone.

[0032] When determining the first survey flight route of the UAV, the terrain of the survey area can be analyzed based on the ground elevation data to extract terrain undulation parameters including elevation change rate, slope distribution, height gradient and surface roughness; through the analysis of terrain undulation parameters, a dynamic path planning algorithm (such as A* algorithm, Dijkstra algorithm, ant colony algorithm or genetic algorithm, etc.) is used to dynamically grid the survey area. The grid division can set different division densities based on the complexity of the terrain, so that a higher density grid division can be used in high-undulation or complex terrain areas, and a lower density grid division can be used in flat areas, thereby optimizing route planning and improving survey efficiency.

[0033] In step S120, the drone is controlled to collect initial area terrain data based on the first survey flight route. The initial area terrain data is collected in real time by a laser radar and an optical camera carried by the drone.

[0034] In an example embodiment of the present disclosure, a laser radar can be used to emit laser pulses and receive multiple echo signals to obtain multi-level terrain data including vegetation layers and surface layers. The laser radar can adopt multi-echo acquisition technology. Specifically, the laser pulses emitted by the laser radar will generate multiple echo signals when encountering different ground objects (such as tree crowns, tree trunks, and ground surfaces), wherein the first echo signal can be used to extract vegetation upper layer information, the middle echo signal can be used to extract middle layer structure information, and the last echo signal can be used to extract exposed surface information. In order to improve data acquisition accuracy, the laser radar can adopt high-frequency emission and multi-angle scanning technology. Of course, the emission frequency and scanning angle of the laser radar can also be dynamically adjusted according to the complexity of the terrain. This embodiment does not specifically limit the specific configuration of the laser radar.

[0035] The optical camera can be used to obtain high-resolution image data of the area to be surveyed. The image data may include but is not limited to red, green, and blue (RGB) images, multispectral images, or full-color images. The optical camera can use automatic exposure, automatic focus, and real-time image enhancement functions to stably obtain clear image data under different lighting conditions. In order to enhance the spatial synchronization of the data, the data acquisition process of the optical camera and the lidar is synchronized. The synchronization method can be based on timestamp alignment or an external trigger mechanism. This embodiment does not specifically limit the synchronization method.

[0036] In step S130, the initial area terrain data is identified and classified to determine a weak texture false terrain area in the initial area terrain data.

[0037] In an example embodiment of the present disclosure, the initial regional terrain data may include image data and LiDAR point cloud data, and the initial regional terrain data may be preprocessed. For example, the preprocessing may be operations such as noise removal, texture enhancement, data standardization, and point cloud density equalization to improve data quality. Specifically, the texture enhancement processing of image data may adopt a texture feature extraction algorithm based on a gray-level co-occurrence matrix (GLCM), an edge detection algorithm, or a local binary pattern (LBP) algorithm; the processing of LiDAR point cloud data may adopt a denoising and sparseness method based on voxel downsampling and normal vector filtering. For example, in one embodiment, an adaptive filtering algorithm based on local point cloud density may be used to identify and remove noise points in LiDAR point cloud data. This method assumes that noise points are usually isolated and do not have a sufficient number of neighboring points around them. By calculating the local density of each point, low-density noise points can be effectively identified and removed. Specifically, it can be expressed by the relationship: ; in, Can indicate point The local density of Can indicate point The domain collection, It can represent the adjustment parameter of density calculation, which is used to control the size of the neighborhood. Can indicate point and Point When it is detected that the local density is less than a certain set threshold, the point can be considered as a noise point and needs to be removed.

[0038] Weak texture false terrain areas refer to areas where, during terrain surveying, the texture features in terrain data are weak, terrain feature information is incomplete, or false terrain features exist because sensors (such as lidar, optical cameras, etc.) are limited by environmental factors or their own technical characteristics. This area usually shows missing texture details, fuzzy terrain contours, or insufficient data density, and is easily misidentified as real terrain, affecting the accuracy and integrity of terrain data. For example, weak texture false terrain areas can be areas with high vegetation coverage, areas covered by snow and glaciers, or areas blocked by complex terrain such as steep cliffs, cliffs, canyons, etc.

[0039] After preprocessing, the texture features of the image data and the spatial features of the point cloud data can be extracted respectively. Texture feature extraction methods can include grayscale difference analysis, gradient direction distribution analysis, and edge texture distribution analysis, and point cloud spatial features can include point cloud density, normal vector distribution, and curvature change rate. Based on the above features, the terrain data can be classified and identified through clustering analysis or classification algorithms (such as K-means algorithm, support vector machine SVM or convolutional neural network CNN), and then the areas with texture features below the preset threshold and abnormal spatial features can be identified as weak texture false terrain areas.

[0040] In step S140, second survey flight routes corresponding to at least two of the UAVs are replanned according to the weak-texture false terrain area, and the second survey flight routes are used to control at least two of the UAVs to turn on millimeter-wave radars and scan and collect supplementary terrain survey data of the weak-texture false terrain area at different angles.

[0041] In an exemplary embodiment of the present disclosure, the second survey flight route refers to a flight path replanned during the terrain survey based on UAV collaboration in order to provide more accurate and comprehensive data supplement to the weak texture false terrain area identified during the initial survey process. The flight route is obtained by dynamically adjusting the flight attitude of the UAV on the basis of the first survey flight route, in combination with the spatial distribution characteristics, terrain complexity and data collection requirements of the weak texture false terrain area, so as to achieve refined supplementary survey of the target area.

[0042] Supplementary terrain survey data refers to data obtained through re-surveying during the initial terrain survey process to compensate for and improve terrain information in the case of blind spots, incomplete data, or erroneous data caused by factors such as terrain complexity, environmental occlusion, and sensor performance limitations. This data is mainly used to enhance the integrity, accuracy, and reliability of the terrain model, and ensure a comprehensive and accurate representation of the terrain features of the weak textured false terrain area.

[0043] For weak-textured false terrain areas, the second survey flight routes corresponding to at least two UAVs can be re-planned. For example, the spatial distribution characteristics of the weak-textured false terrain areas can be analyzed to extract the regional boundary range, terrain undulation characteristics and spatial distribution density, and a dynamic path planning algorithm based on the cost function can be used to generate an optimized second survey flight route. The path planning can comprehensively consider the remaining power of the UAV, the flight environment and the working parameters of the millimeter-wave radar to ensure that the UAV fully covers the target area at different angles.

[0044] Under the second survey flight route, the drone is controlled to enable the millimeter-wave radar to scan at different angles to collect supplementary terrain survey data in the weak-textured false terrain area. The millimeter-wave radar has strong penetration and anti-interference capabilities, and can penetrate vegetation and snow to obtain surface feature information. According to the terrain complexity of the target area, the millimeter-wave radar's transmission frequency, power and scanning angle can be dynamically adjusted to supplement the lidar's measurement blind spots in complex terrain.

[0045] In step S150, the supplementary terrain survey data at different angles are fused to obtain local terrain survey data of the weak-texture false terrain area.

[0046] In step S160, the initial area terrain data and the local terrain survey data are spliced ​​and fused to obtain the real terrain survey data of the area to be surveyed.

[0047] In an exemplary embodiment of the present disclosure, the initial regional terrain data may be fused with the supplementary terrain survey data. For example, the splicing fusion of the initial regional terrain data and the local terrain survey data may be achieved by the following relationship: ; in, It can represent the global point cloud data corresponding to the real terrain survey data obtained by splicing and fusion. and can represent the rotation matrix, and can represent the translation vector, It can represent the point cloud data corresponding to the initial regional terrain data. It can represent the point cloud data corresponding to the local terrain survey data. It can represent the Euclidean distance between point clouds. The optimization process minimizes the error after stitching by continuously adjusting the rotation and translation parameters, thereby generating a seamless point cloud model, that is, obtaining the real terrain survey data corresponding to the area to be surveyed.

[0048] By fusing the initial regional terrain data with the supplementary terrain survey data, using a multi-view data registration algorithm to align data from different angles, and achieving high-precision data stitching based on feature point matching and point cloud density optimization algorithms, a continuous and complete three-dimensional terrain model is finally generated, realizing the collection of real terrain survey data of the surveyed area, thereby improving the measurement accuracy and survey efficiency of terrain survey data in complex terrain areas, and improving the reliability and stability of real terrain survey data.

[0049] Next, the contents of step S110 to step S160 are described in detail.

[0050] In an exemplary embodiment of the present disclosure, Figure 2 The steps in step S110 are used to determine the first survey flight routes of at least two UAVs in combination with the ground elevation data, referring to Figure 2 As shown, it may specifically include: Step S210, determining the terrain relief parameters of the area to be surveyed according to the ground elevation data, wherein the terrain relief parameters include elevation change rate, slope distribution, height gradient and surface roughness; Step S220, dynamically gridding the area to be surveyed based on the terrain relief parameters to determine a survey sub-area corresponding to the area to be surveyed; Step S230, allocating the drone to each of the survey sub-areas to obtain an initial survey flight route; Step S240, performing overlapping area and blind area analysis on the initial survey flight route, and adjusting the route spacing and scanning angle of the UAV based on the analysis results to generate a first survey flight route.

[0051] Among them, terrain relief parameters refer to terrain feature data extracted from ground elevation data. For example, terrain relief parameters may include elevation change rate, slope distribution, elevation gradient, and surface roughness. Specifically, elevation change rate can be used to indicate the severity of terrain height changes, slope distribution can be used to reflect the steepness of the terrain, elevation gradient can be used to describe the spatial variation trend of elevation differences, and surface roughness can be used to measure the undulation and irregularity of the surface. By analyzing terrain relief parameters, we can fully understand the terrain complexity and surface characteristics of the area to be surveyed.

[0052] Dynamic meshing refers to the process of flexibly adjusting the density and size of the mesh according to the complexity of the regional terrain to achieve refined coverage of complex terrain areas. Specifically, an adaptive meshing algorithm can be used to divide complex terrain areas into high-density small meshes and flat areas into low-density large meshes. For example, the three-dimensional space of the regional terrain can be refined layer by layer based on the quadtree or octree algorithm, or the slope threshold-based method can be used to encrypt the mesh density in areas with larger slopes. By dynamically meshing the area to be surveyed, the rationality of the drone route and the integrity of the survey data can be effectively improved.

[0053] After completing the dynamic grid division, a drone can be assigned to each survey sub-area. In one embodiment, the different sub-areas can be assigned to each drone using a task allocation algorithm (such as a partitioning method, a load balancing algorithm) based on the flight performance parameters of each drone (such as flight speed, maximum flight altitude, load capacity, and endurance) and the area and terrain complexity of each survey sub-area. During the task allocation process, the task load can be dynamically adjusted in combination with the real-time flight environment (such as meteorological conditions, wind speed, and air pressure) to ensure that multiple drones perform survey tasks efficiently and in a coordinated manner.

[0054] The initial survey flight route refers to the flight route of the drone formed by allocating the survey sub-areas to be measured to the drone. The initial survey flight route may include parameters such as flight altitude, route direction, scanning angle and flight speed. For example, the flight altitude can be adjusted according to the terrain undulations to avoid collisions with obstacles; the route direction can be optimized according to the distribution of the survey sub-areas to ensure that the routes are not overlapped; the scanning angle and flight speed can be dynamically matched according to the sensor performance (such as the scanning frequency of the laser radar and the resolution of the optical camera). In some optional embodiments, the flight route can be generated by a path planning algorithm. For example, the path planning algorithm can select the A* algorithm, or the Dijkstra algorithm, the ant colony algorithm, etc. This embodiment is not limited to this, so as to minimize flight energy consumption and avoid blind spots and overlapping areas.

[0055] After completing the initial flight route planning, the initial survey flight route is analyzed for overlapping areas and blind spots. The overlapping area and blind spot analysis can use an evaluation algorithm based on route overlap and point cloud coverage density to detect route overlap and area coverage between different drones. If overlapping routes or blind spots are found in the area, the route spacing and scanning angle can be adjusted dynamically. Specifically, route spacing adjustment can be based on terrain complexity and sensor coverage, and scanning angle adjustment can be based on surface slope and terrain feature change trends to ensure the integrity and consistency of data collection.

[0056] Through terrain undulation parameter analysis, dynamic grid division, route optimization and adjustment, and mission load balancing, UAVs can flexibly and efficiently perform survey tasks in complex terrain environments, thereby ensuring the comprehensiveness, accuracy, and stability of terrain data measurement. This not only improves the efficiency of terrain data survey and collection, but also enhances the ability to accurately perceive complex terrain features.

[0057] Optionally, you can Figure 3 The steps in step S220 implement the dynamic grid division of the area to be surveyed based on the terrain relief parameters, determine the survey sub-area corresponding to the area to be surveyed, and refer to Figure 3 As shown, it may specifically include: Step S310, determining the complex terrain area and the flat terrain area of ​​the area to be surveyed according to the terrain undulation parameter and a preset terrain undulation threshold; Step S320, gridding the complex terrain area with a first division density, and gridding the flat terrain area with a second division density, to obtain a survey sub-area corresponding to the area to be surveyed; wherein the first division density is greater than the second division density.

[0058] Among them, the preset terrain undulation threshold refers to a critical value pre-set based on the characteristics of terrain changes and survey requirements, which is used to distinguish the complexity of the terrain. Specifically, when the elevation change rate, slope distribution or elevation gradient of the terrain area is greater than or equal to the preset terrain undulation threshold, the current terrain area can be determined as a complex terrain area; otherwise, it can be determined as a flat terrain area. The terrain undulation threshold can be set through historical survey data analysis or standard experience of terrain characteristics. Of course, it can also be dynamically adjusted according to real-time survey data. This embodiment does not limit the threshold setting method.

[0059] After completing the division of complex terrain and flat terrain areas, different division densities can be used to grid the areas. Specifically, the complex terrain area is gridded with the first division density. Since the density value of the first division density is higher, the grid units divided are smaller, which can more finely cover the areas with large terrain undulations and complex landforms, ensuring the comprehensive acquisition of detailed terrain features; while for flat terrain areas, the second division density is used for grid division. Since the density value of the second division density is lower, the grid units divided are larger, which effectively reduces unnecessary data redundancy and improves the efficiency of drone surveys. The grid division method can use a quadtree or an adaptive division method based on a slope threshold to dynamically adjust the size and distribution of the grid units to better adapt to terrain changes.

[0060] Through differentiated grid division processing, multiple survey sub-areas covering the entire area to be surveyed are formed. The survey sub-areas have differentiated division density and size, which can realize refined and efficient data collection for complex terrain areas and simple terrain areas; UAV survey tasks can be reasonably allocated according to the division results of these survey sub-areas to ensure the balance and efficiency of the survey tasks, thereby improving the integrity and accuracy of the measured terrain data of the area to be surveyed.

[0061] In an exemplary embodiment of the present disclosure, Figure 4 The steps in step S130 are used to identify and classify the initial regional terrain data, determine the weak texture false terrain area in the initial regional terrain data, and refer to Figure 4 As shown, it may specifically include: Step S410, preprocessing the initial area terrain data, the initial area terrain data including image data and laser point cloud data, the preprocessing at least including noise removal, texture enhancement processing, data standardization and point cloud density equalization; Step S420, extracting image texture features from the preprocessed image data, and extracting point cloud spatial features from the preprocessed laser point cloud data; Step S430, determining a weak-texture false terrain region as a terrain region where the image texture feature is lower than a preset texture threshold and the point cloud spatial feature belongs to an abnormal spatial feature.

[0062] The preprocessing of the initial area terrain data may include noise removal, texture enhancement processing, data standardization and point cloud density equalization, and the initial area terrain data may include image data and laser point cloud data.

[0063] For image data and laser point cloud data, noise removal algorithms can be used to eliminate outliers. For example, isolated points and noise points can be removed by statistical filtering (SOR) or radius filtering (ROR). Statistical filtering can calculate the distance between each point in the image or point cloud data and the neighboring points, and remove points that deviate too much based on the standard deviation; radius filtering can remove points with insufficient number of neighboring points based on a set radius range.

[0064] After noise removal, the image data and laser point cloud data can be standardized. For example, the multi-source data collected by different devices can be unified into the same coordinate system through coordinate unification, scale normalization and spatial alignment. Coordinate unification can use geographic coordinate system (WGS-84) or UTM coordinate system. Scale normalization can normalize the data to a consistent scale through scaling. Spatial alignment can align the point cloud data through the iterative closest point (ICP) algorithm to ensure the consistency of the data in spatial position.

[0065] For laser point cloud data, point cloud density equalization can also be used to solve the problem of uneven distribution of point cloud data. Especially in areas with large terrain fluctuations or severe occlusion, point cloud data may be dense or sparse. Point cloud density equalization can be achieved through voxel grid downsampling (Voxel Grid Filter) and adaptive density compensation algorithm. For example, voxel grid downsampling can divide point cloud data into voxel grids of fixed size, retain a representative point in each voxel, reduce data redundancy and balance point cloud density. Adaptive density compensation algorithm dynamically adjusts the sampling rate according to the local point cloud density, downsampling in dense areas and interpolating points in sparse areas. Interpolation can use triangulated irregular network (TIN) or kriging interpolation to generate supplementary points by fitting neighborhood points.

[0066] For optical image data, the recognizability of terrain texture information can be improved through texture enhancement processing. For example, texture enhancement processing can enhance image contrast through histogram equalization, enhance edge details through Laplace sharpening, and smooth image noise through Gaussian filtering. In addition, edge features of optical image data can be extracted through edge detection algorithms (such as Canny operator and Sobel operator), which is helpful for the subsequent recognition of weak texture areas.

[0067] After preprocessing, features can be extracted from image data and laser point cloud data respectively. For image data, the main feature is image texture features, which refer to the spatial distribution of pixel grayscale or color changes in the image. They can be extracted using algorithms such as gray-level co-occurrence matrix (GLCM), local binary patterns (LBP) or histogram of oriented gradient (HOG). GLCM can quantify the contrast, uniformity and correlation of image texture, LBP can be used to detect local texture patterns, and HOG can be used to capture edge information and shape features. For laser point cloud data, point cloud spatial features can be extracted, which can include point cloud density, normal vector distribution and curvature change rate. Point cloud density can be obtained by calculating the number of points in a unit volume, normal vector distribution can be calculated by fitting the point cloud surface, and curvature change rate can be used to identify the changing trend of surface morphology. By extracting multi-dimensional features, the spatial and texture characteristics of terrain data can be characterized from different dimensions.

[0068] After feature extraction, the image texture features and point cloud spatial features can be comprehensively analyzed based on the multimodal data fusion algorithm; multimodal data fusion can adopt weighted fusion, feature-level fusion or decision-level fusion. Weighted fusion can linearly combine texture features and spatial features by assigning different weights to different data sources; feature-level fusion can map multimodal data to a unified feature space for joint analysis; decision-level fusion can classify each data source separately, and finally integrate the classification results. Through multimodal fusion analysis, abnormal areas in terrain data can be more accurately identified.

[0069] In the comprehensive analysis process, clustering algorithms or classification algorithms can be used to classify feature data. Clustering algorithms (such as K-means and DBSCAN) can automatically cluster data based on feature similarity and identify areas with obvious abnormal features; classification algorithms (such as support vector machines (SVM) and convolutional neural networks (CNN)) can be trained based on labeled samples to accurately distinguish between weak texture false terrain areas and real terrain areas. For areas where image texture features are lower than a preset threshold and spatial features are abnormal, they are marked as weak texture false terrain areas. The threshold can be set based on historical data experience or dynamic adjustment, and the classification model can also be continuously optimized through a real-time feedback mechanism.

[0070] By identifying areas of false terrain with weak textures and using the identification results as input data for subsequent flight path planning, as key target areas for subsequent UAV supplementary surveys, it is possible to ensure targeted supplementation of areas with missing terrain data and errors, further improving the integrity and accuracy of real terrain survey data.

[0071] Optionally, you can Figure 5 The steps in step S420 are used to extract the point cloud spatial features of the pre-processed laser point cloud data, referring to Figure 5 As shown, it may specifically include: Step S510, performing multi-echo classification on the laser signal in the laser point cloud data to obtain a multi-echo signal, wherein the multi-echo signal includes an initial echo signal, an intermediate echo signal, and a terminal echo signal; Step S520, determining echo layer point cloud features corresponding to each echo layer point cloud data in the multi-echo signal, wherein the echo layer point cloud features include point cloud density, normal vector distribution and curvature change rate; Step S530: constructing point cloud spatial features of the laser point cloud data according to the echo layer point cloud features.

[0072] Among them, multi-echo signals refer to the multiple echo responses generated by the laser pulses emitted by the lidar when encountering different object surfaces (such as the terrain of the vegetation-covered area, which can include tree crowns, trunks and the ground). Multi-echo signals can usually include initial echo signals (first echo), intermediate echo signals and terminal echo signals (last echo); the initial echo signal can reflect the upper structure of the vegetation (such as the crown), the intermediate echo signal can reflect the internal structure of the vegetation (such as tree trunks and shrubs), and the terminal echo signal can reflect the surface information after penetrating the vegetation. Through the layered processing of multi-echo signals, the terrain of the vegetation-covered area can be decomposed layer by layer to obtain more detailed ground object information.

[0073] For example, the intensity of the echo signal is related to factors such as the reflectivity and distance of the target object in the terrain area. The laser radar determines the relative positions of different objects by analyzing the time delay and intensity of the echo signal. For example, the separation of multiple echo signals can be achieved through the following relationship: ; in, It can indicate the signal strength of the echo signal; It can represent the received power; It can indicate the transmit power; It can represent the receiving area; It can represent the reflectivity of the target in the terrain area; It can represent the relative distance between the lidar and the target object in the terrain area.

[0074] The echo layer point cloud features may include point cloud density, normal vector distribution, and curvature change rate. Point cloud density refers to the number of points per unit volume, which is used to measure the density of data distribution; normal vector distribution can be used to describe the spatial orientation change of the point cloud surface and reflect the geometric characteristics of the terrain surface; curvature change rate can reflect the curvature and undulation characteristics of the point cloud surface, which helps to identify terrain boundaries and mutation features. Point cloud density can be calculated by setting the radius neighborhood to calculate the number of points in the neighborhood, normal vectors can be calculated by local plane fitting algorithms (such as principal component analysis), and curvature change rate can be calculated by local surface fitting or curvature analysis methods (such as Gaussian curvature or mean curvature).

[0075] When constructing the point cloud spatial features of laser point cloud data, the point cloud features of each echo layer can be combined for comprehensive analysis to form a multi-dimensional spatial feature expression. Specifically, a multi-scale feature extraction algorithm can be used to extract point cloud density, normal vector distribution, and curvature change rate at different scales to capture terrain change information at different levels. In addition, voxel grid downsampling can be used to perform structured processing on point cloud data to optimize the balance of data distribution and processing efficiency; of course, in order to enhance the expression ability of spatial features, deep learning models (such as PointNet++) can also be combined to perform feature learning on point cloud data to automatically extract complex spatial features, but this embodiment is not limited to this.

[0076] In the process of point cloud spatial feature extraction, in order to improve the integrity and accuracy of terrain data, a multi-echo signal weighted fusion strategy can be used to weight and integrate the data of different echo layers. Specifically, different weights can be assigned according to the intensity, echo sequence and echo interval of each echo signal, giving priority to the surface data of the terminal echo signal, and using the initial and intermediate echo signals to supplement the vegetation structure information to achieve multi-level and multi-scale data fusion; this weighted fusion strategy can effectively avoid the information loss or error accumulation caused by a single echo data, and improve the reliability of terrain data.

[0077] When identifying complex terrain and vegetation-covered areas, in order to further enhance the spatial feature analysis of point cloud data, the point cloud data can be combined with a terrain classification algorithm to classify objects. For example, the object classification algorithm can use a rule-based classification method (such as slope, curvature and density threshold judgment) or a machine learning-based classification method (such as random forest, support vector machine, convolutional neural network) to automatically classify and identify terrain data. Through the classification algorithm, the surface, vegetation and other objects can be effectively distinguished, thereby improving the accuracy and stability of spatial feature extraction.

[0078] Ultimately, through the refined classification processing and spatial feature extraction of multi-echo laser point cloud data, the terrain surface structure and object distribution can be fully and accurately described, providing high-quality data support for subsequent terrain data analysis and model construction.

[0079] In an optional implementation, the abnormal spatial features may be determined by the following steps: The terrain gradient distribution estimation data of the current terrain area can be determined by combining the echo layer point cloud features corresponding to the intermediate echo signal and the terminal echo signal; the echo layer point cloud features corresponding to the initial echo signal are compared with the terrain gradient distribution estimation data; if the terrain gradient corresponding to the echo layer point cloud features corresponding to the initial echo signal is inconsistent with the terrain gradient distribution estimation data, and the point cloud density of the initial echo signal is lower than the preset density value, the change rate of the normal vector distribution is greater than the preset change rate threshold, and the curvature change rate is in an abnormal state, then it is determined that the point cloud spatial features of the current terrain area are abnormal spatial features.

[0080] Among them, terrain gradient distribution estimation data refers to a set of spatial data calculated based on the analysis of the spatial characteristics of the terrain surface, which is used to describe the degree and trend of terrain undulation changes. This data mainly reflects the undulation characteristics and spatial distribution characteristics of the terrain surface at different scales through quantitative analysis of terrain elevation changes, namely terrain gradient, slope distribution, curvature characteristics and other parameters. In flat areas, the gradient of point cloud data is small and the terrain elevation changes slowly; while in areas where the terrain turns, boundaries or curvature changes greatly, the gradient value of point cloud data is large and the terrain elevation changes dramatically. For example, the gradient of point cloud data can be calculated by the following relationship: ; in, Can indicate point The gradient at , , Can represent points separately Partial derivatives in the x, y, and z directions, representing the point The rate of change in each direction.

[0081] The curvature of point cloud data can be calculated using the following relationship: ; in, It can represent the curvature of the point cloud surface. It can represent the determinant of the Hessian matrix, which characterizes the curvature of the quadratic surface. It can be expressed as the square of the gradient, which characterizes the rate at which the surface changes.

[0082] Based on the extracted spatial feature data, a terrain gradient analysis algorithm can be used to construct terrain gradient distribution estimation data. For example, terrain gradient analysis can reflect the undulating characteristics of the terrain by calculating the elevation differences and slope changes in local areas. Specifically, the slope calculation algorithm can be used to calculate the slope distribution of point cloud data, and the gradient direction histogram can be used to analyze the trend of surface direction changes. The curvature analysis can be combined to extract the surface mutation boundary, thereby comprehensively describing the undulating morphology and spatial distribution characteristics of the surface.

[0083] After obtaining the terrain gradient distribution estimation data, the echo layer point cloud features corresponding to the initial echo signal can be compared with the terrain gradient distribution estimation data to effectively avoid abnormal data caused by terrain gradient distribution. When it is determined that the terrain gradient corresponding to the echo layer point cloud features corresponding to the initial echo signal is inconsistent with the terrain gradient distribution estimation data, if the echo layer point cloud features corresponding to the initial echo signal have abnormal problems, they can be determined as abnormal spatial features; specifically, the difference between the point cloud density and the normal vector change rate of the initial echo signal and the terrain gradient data can be analyzed in detail. If the point cloud density corresponding to the initial echo signal is lower than the preset density threshold, the normal vector distribution change rate is higher than the preset change rate threshold, and the curvature change rate is in the abnormal range, it can be determined that there is data anomaly or terrain feature misjudgment in the area. It can be understood that the comparison process can use a difference detection algorithm based on threshold judgment, or a machine learning classification algorithm (such as support vector machine or random forest) can be used to automatically identify abnormal data.

[0084] Optionally, a multi-scale comparison strategy can be introduced, which can analyze local features and global features by calculating point cloud density, normal vector distribution and curvature change rate at different scales. Multi-scale comparison can effectively avoid the errors caused by single-scale analysis and enhance the accuracy of terrain data anomaly identification. For example, the pyramid analysis method can be used to extract features and make comprehensive judgments at coarse and fine scales respectively; during the comparison process, if the point cloud spatial features corresponding to the initial echo signal are detected to be significantly different from the terrain gradient distribution estimation data, the area can be marked as an abnormal spatial feature area. The abnormal area may be due to misjudgment of terrain information caused by vegetation occlusion, insufficient laser radar penetration or data noise. For the marked abnormal area, the system will automatically feedback to the path planning module to re-plan the drone's supplementary survey flight route to ensure the integrity and accuracy of the data.

[0085] Optionally, an adaptive threshold adjustment mechanism can be introduced during the comparison process; the adaptive threshold mechanism can dynamically adjust the point cloud density threshold, normal vector change rate threshold, and curvature change rate threshold by analyzing the data distribution in real time, to avoid misjudgment or missed judgment caused by overly strict or loose fixed threshold settings. This mechanism can be implemented through fuzzy logic algorithm or Bayesian dynamic updating algorithm to improve the flexibility and accuracy of anomaly detection.

[0086] Through the above steps, abnormal spatial features in terrain data can be accurately identified, the integrity and accuracy of terrain data can be further optimized, and a reliable data foundation can be provided for subsequent data fusion and three-dimensional terrain modeling.

[0087] In an exemplary embodiment of the present disclosure, Figure 6 To implement step S140 of replanning the second survey flight routes corresponding to at least two UAVs respectively according to the weak texture false terrain area, refer to Figure 6 As shown, it may specifically include: Step S610, performing spatial distribution analysis on the point cloud data corresponding to the weak texture false terrain area to determine the regional spatial features corresponding to the weak texture false terrain area, wherein the regional spatial features include the regional boundary range, the regional area and the terrain undulation change features; Step S620, determining the survey priority corresponding to each of the weak texture pseudo terrain regions based on the regional spatial features; Step S630, obtaining equipment parameters and flight attitude data corresponding to at least two of the UAVs, and performing a local flight attitude update on the first survey flight route corresponding to the UAV in combination with the survey priority, the regional spatial characteristics, the equipment parameters and the flight attitude data to obtain a second survey flight route.

[0088] Among them, spatial distribution analysis refers to a detailed analysis of the spatial range, morphology and density characteristics of the identified weak texture false terrain area, so as to provide a basis for the subsequent optimization of the UAV flight path. For example, the content of spatial distribution analysis can be the regional boundary range, regional area and terrain fluctuation characteristics. Specifically, the regional boundary range can be accurately determined by the boundary extraction algorithm of point cloud data (such as α-Shape algorithm or convex hull algorithm); the regional area can be obtained by spatial projection calculation; the terrain fluctuation characteristics can be calculated by slope analysis, curvature change rate and elevation change rate. These analysis results can fully reflect the distribution characteristics and terrain complexity of the weak texture false terrain area, and provide key parameters for path planning.

[0089] Survey priority refers to the priority of measurement based on a comprehensive assessment of the importance, complexity and urgency of the region. The survey priority assessment can be weighted by using a multi-factor decision-making model (such as the analytic hierarchy process or the fuzzy comprehensive evaluation method) to calculate factors such as the terrain complexity, area size, and degree of data loss in each region. For example, the greater the terrain undulation and the sparser the point cloud density, the higher the survey priority, and the supplementary survey tasks need to be prioritized. The priority ranking result will directly affect the task allocation and flight path planning of the drone, and the high-priority areas will be given priority for supplementary surveys.

[0090] Equipment parameters may include the flight speed, endurance, load capacity, flight altitude limit, sensor working status (lidar, millimeter wave radar, etc.), etc. of the drone, and flight attitude data may include the real-time position, heading angle, pitch angle and roll angle of the drone. These parameters can be obtained in real time through the flight control system of the drone, or updated in real time through the inertial navigation system and global positioning system carried by the drone. The purpose of obtaining this data is to fully understand the flight status and operating capabilities of each drone and provide real-time reference for path planning.

[0091] The local flight attitude update of the first survey flight route corresponding to the UAV can be performed in combination with the survey priority, regional spatial characteristics, equipment parameters and flight attitude data. The update process can be based on a real-time path optimization algorithm (such as A* algorithm, genetic algorithm or ant colony algorithm) to make local adjustments to the initial flight route to meet the needs of supplementary surveys. Specifically, the flight altitude, route direction and flight speed of the UAV can be dynamically adjusted according to the spatial distribution characteristics of the weak texture pseudo-terrain area, while optimizing the scanning angle and sensor working mode; the UAV will lower the flight altitude and slow down the flight speed in complex terrain areas to improve data collection accuracy; in areas with simpler terrain, the flight speed and scanning angle will be appropriately increased to improve survey efficiency.

[0092] Optionally, flight attitude update can also involve optimization of the obstacle avoidance strategy and dynamic obstacle avoidance function of the UAV. Based on real-time environmental information (such as wind speed, temperature, and obstacle position), the flight attitude is adjusted through obstacle avoidance algorithms such as the Dynamic Window Approach (DWA) to ensure that the UAV can perform tasks safely and stably in complex environments. The obstacle avoidance strategy can combine lidar obstacle avoidance and visual obstacle avoidance to detect obstacles in real time and dynamically adjust the route.

[0093] The second survey flight route generated through the above steps can more accurately cover the weak-textured false terrain area, and realize efficient supplementary scanning of the initial survey blind spots and data-missing areas; combined with the dynamic update of the real-time status and environmental information of the drone, the optimized flight route ensures the stability of multi-drone collaborative operations and the integrity of data collection, further improving the accuracy and efficiency of terrain surveys.

[0094] In an exemplary embodiment of the present disclosure, Figure 7 The steps in the above are used to fuse the supplementary terrain survey data at different angles to obtain the local terrain survey data of the weak texture false terrain area. Figure 7 As shown, it may specifically include: Step S710, preprocessing the supplementary topographic survey data at different angles, and determining key feature points in the preprocessed supplementary topographic survey data, wherein the key feature points include elevation mutation points, boundary points, and curvature feature points; Step S720, performing multi-view data registration on the pre-processed supplementary topographic survey data according to the key feature points to obtain spatially aligned supplementary topographic survey data; Step S730, weightedly fusing the spatially aligned supplementary terrain survey data with the initial area terrain data corresponding to the weak texture false terrain area based on a predetermined allocation weight to obtain fused terrain survey data; Step S740: Density compensation and data smoothing are performed on the fused terrain survey data to obtain local terrain survey data of the weak-texture pseudo terrain area.

[0095] Among them, the supplementary topographic survey data may include point cloud data obtained by millimeter wave radar and lidar, and image data obtained by optical cameras. The supplementary topographic survey data may be preprocessed, for example, the preprocessing may include format standardization, time synchronization and spatial coordinate unification, etc., to ensure the consistency and fusion of the supplementary topographic survey data. Format standardization can unify data of different formats (such as LAS format, PLY format or TIFF format) into a standard format through data conversion tools; time synchronization can be based on a global timestamp (GPS timestamp or internal timestamp) to ensure the time consistency of data collected by multiple drones; spatial coordinate unification uses inertial navigation system and global positioning system data to perform spatial registration of point cloud and image data to ensure that the data is aligned under a unified geographic spatial reference system.

[0096] After completing the preprocessing of the supplementary topographic survey data, the key feature points in the preprocessed supplementary topographic survey data can be extracted. For example, the key feature points may include elevation mutation points, boundary points, and curvature feature points. Elevation mutation points can be used to describe the sharp change in terrain height, usually located in areas such as cliffs, ridges, or gullies. Boundary points can be used to represent the junction of different landforms (such as vegetation and exposed surfaces). Curvature feature points can be used to detect the concave and convex changes of the surface, which can effectively reflect the terrain undulations. The method of extracting elevation mutation points can use a slope analysis algorithm (such as a gradient-based Sobel operator), boundary points can be extracted by an α-Shape algorithm or a convex hull algorithm (Concave Hull), and curvature feature points can be obtained by calculating Gaussian curvature and mean curvature.

[0097] Multi-view data registration refers to adjusting the spatial relationship between data sets by comparing key feature points so that they can be seamlessly connected in the same coordinate system. Feature point-based registration methods such as iterative closest point algorithm and global feature matching algorithm can be used. For example, iterative closest point algorithm can align data by minimizing the point-to-point distance, which is suitable for local alignment of dense point cloud data; global feature matching algorithm can be applied to sparse point clouds or data with large rotation deviations. In addition, deep learning registration models (such as PointNetLK model) can be used for automated multi-view data registration to further improve registration accuracy.

[0098] After completing the multi-view data registration, the supplementary terrain survey data after spatial alignment and the initial regional terrain data of the weak texture pseudo terrain area can be weighted fused based on the preset distribution weights. Weighted fusion refers to assigning different weight coefficients according to the accuracy, reliability and redundancy of different data sources, and integrating multi-source data. Millimeter wave radar data has strong penetration ability and can be assigned higher weights in vegetation-covered areas; LiDAR data has high accuracy and can be assigned higher weights in exposed surface areas; optical image data can be used for texture enhancement and assigned higher weights in texture-rich areas. Weighted fusion can be implemented using a Bayesian fusion algorithm, a Kalman filter algorithm, or a deep learning fusion algorithm (such as a multimodal neural network) to improve the accuracy and stability of the fused data. This example embodiment does not specifically limit the weighted fusion method.

[0099] Density compensation and data smoothing can be performed on the supplementary topographic survey data after weighted fusion. Density compensation is aimed at the problem of uneven density of point cloud data. The data distribution of the supplementary topographic survey data is balanced through voxel grid downsampling and adaptive density compensation algorithm. Sparse areas can be supplemented by interpolation, and dense areas can be optimized by downsampling. Data smoothing can eliminate data noise and boundary mutations through the Moving Least Square (MLS) surface reconstruction algorithm or Gaussian smoothing filter to ensure data continuity and smoothness.

[0100] Through multi-view data registration, weighted fusion, density compensation and data smoothing processing, continuous, complete and refined local terrain survey data of weak-textured pseudo-topography areas are finally generated, which can accurately and comprehensively reflect the real terrain information of weak-textured pseudo-topography areas and provide a reliable data foundation for subsequent global terrain data splicing and three-dimensional terrain model construction.

[0101] It should be noted that, although the steps of the method in the present disclosure are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0102] In addition, in this exemplary embodiment, a terrain survey device based on drone cooperation is also provided. Figure 8 As shown, the terrain survey device 800 based on UAV cooperation includes: an initial flight survey module 810, a terrain data acquisition module 820, a false terrain area identification module 830, a survey route update module 840, a supplementary survey data fusion module 850 and a real terrain data output module 860. Among them: The initial flight survey module 810 is used to obtain ground elevation data of the area to be surveyed, and determine a first survey flight route of at least two of the UAVs in combination with the ground elevation data; A terrain data acquisition module 820 is used to control the UAV to acquire initial area terrain data based on the first survey flight route, where the initial area terrain data is acquired in real time by a laser radar and an optical camera carried by the UAV; A false terrain region identification module 830 is used to identify and classify the initial region terrain data, and determine a weak texture false terrain region in the initial region terrain data; A survey route updating module 840 is used to re-plan second survey flight routes corresponding to at least two of the drones according to the weak-texture false terrain area, wherein the second survey flight routes are used to control at least two of the drones to turn on millimeter-wave radars and scan and collect supplementary terrain survey data of the weak-texture false terrain area at different angles; A supplementary survey data fusion module 850 is used to fuse the supplementary terrain survey data at different angles to obtain local terrain survey data of the weak texture false terrain area; The real terrain data output module 860 is used to splice and fuse the initial area terrain data and the local terrain survey data to obtain the real terrain survey data of the area to be surveyed.

[0103] In an exemplary embodiment of the present disclosure, based on the aforementioned scheme, the initial flight survey module 810 is configured to: determine the terrain undulation parameters of the area to be surveyed according to the ground elevation data, the terrain undulation parameters including the elevation change rate, slope distribution, height gradient and surface roughness; dynamically grid the area to be surveyed based on the terrain undulation parameters to determine the survey sub-areas corresponding to the area to be surveyed; allocate the UAV to each of the survey sub-areas to obtain an initial survey flight route; perform overlapping area and blind spot analysis on the initial survey flight route, and adjust the route spacing and scanning angle of the UAV based on the analysis results to generate a first survey flight route.

[0104] In an exemplary embodiment of the present disclosure, based on the aforementioned scheme, the initial flight survey module 810 is configured to: determine the complex terrain area and the flat terrain area of ​​the area to be surveyed according to the terrain undulation parameter and a preset terrain undulation threshold; grid-divide the complex terrain area with a first division density, and grid-divide the flat terrain area with a second division density to obtain a survey sub-area corresponding to the area to be surveyed; wherein the first division density is greater than the second division density.

[0105] In an exemplary embodiment of the present disclosure, based on the aforementioned scheme, the false terrain area identification module 830 is configured to: preprocess the initial area terrain data, the initial area terrain data includes image data and laser point cloud data, and the preprocessing includes at least noise removal, texture enhancement processing, data standardization and point cloud density equalization; extract image texture features in the preprocessed image data, and extract point cloud spatial features of the preprocessed laser point cloud data; determine the weak-texture false terrain area as a terrain area where the image texture features are lower than a preset texture threshold and the point cloud spatial features belong to abnormal spatial features.

[0106] In an exemplary embodiment of the present disclosure, based on the aforementioned scheme, the false terrain area identification module 830 is configured to: perform multi-echo classification on the laser signal in the laser point cloud data to obtain a multi-echo signal, wherein the multi-echo signal includes an initial echo signal, an intermediate echo signal, and a terminal echo signal; determine the echo layer point cloud features corresponding to each echo layer point cloud data in the multi-echo signal, wherein the echo layer point cloud features include point cloud density, normal vector distribution, and curvature change rate; and construct the point cloud spatial features of the laser point cloud data according to the echo layer point cloud features.

[0107] In an exemplary embodiment of the present disclosure, based on the aforementioned scheme, the false terrain area identification module 830 is configured to: determine the terrain gradient distribution estimation data of the current terrain area in combination with the echo layer point cloud features corresponding to the intermediate echo signal and the terminal echo signal; compare the echo layer point cloud features corresponding to the initial echo signal with the terrain gradient distribution estimation data; if the terrain gradient corresponding to the echo layer point cloud features corresponding to the initial echo signal is inconsistent with the terrain gradient distribution estimation data, and the point cloud density of the initial echo signal is lower than a preset density value, the rate of change of the normal vector distribution is greater than a preset change rate threshold, and the curvature change rate is in an abnormal state, then it is determined that the point cloud spatial features of the current terrain area are abnormal spatial features.

[0108] In an exemplary embodiment of the present disclosure, based on the aforementioned scheme, the survey route update module 840 is configured to: perform spatial distribution analysis on the point cloud data corresponding to the weak texture false terrain area, determine the regional spatial characteristics corresponding to the weak texture false terrain area, and the regional spatial characteristics include regional boundary range, regional area and terrain undulation change characteristics; determine the survey priority corresponding to each of the weak texture false terrain areas based on the regional spatial characteristics; obtain equipment parameters and flight attitude data corresponding to at least two of the drones, and perform a local flight attitude update on the first survey flight route corresponding to the drone in combination with the survey priority, the regional spatial characteristics, the equipment parameters and the flight attitude data to obtain a second survey flight route.

[0109] In an exemplary embodiment of the present disclosure, based on the aforementioned scheme, the supplementary survey data fusion module 850 is configured to: pre-process the supplementary topographic survey data at different angles, and determine the key feature points in the pre-processed supplementary topographic survey data, wherein the key feature points include elevation mutation points, boundary points and curvature feature points; perform multi-perspective data registration on the pre-processed supplementary topographic survey data according to the key feature points to obtain spatially aligned supplementary topographic survey data; perform weighted fusion of the spatially aligned supplementary topographic survey data with the initial regional topographic data corresponding to the weak texture false terrain area based on a predetermined allocation weight to obtain fused topographic survey data; perform density compensation and data smoothing on the fused topographic survey data to obtain local topographic survey data of the weak texture false terrain area.

[0110] The specific details of each module of the above-mentioned terrain survey device based on drone collaboration have been described in detail in the corresponding terrain survey method based on drone collaboration, so they will not be repeated here.

[0111] It should be noted that although several modules or units of the terrain survey device based on drone collaboration are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.

[0112] In addition, in an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above-mentioned terrain survey method based on drone collaboration is also provided.

[0113] Those skilled in the art will appreciate that various aspects of the present disclosure may be implemented as systems, methods or program products. Therefore, various aspects of the present disclosure may be specifically implemented in the following forms, namely: complete hardware embodiments, complete software embodiments (including firmware, microcode, etc.), or embodiments combining hardware and software aspects, which may be collectively referred to herein as "circuits", "modules" or "systems".

[0114] Refer to the following Fig. 9 hereinafter describes an electronic device 900 according to such an embodiment of the present disclosure. Fig. 9 The electronic device 900 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0115] like Fig. 9As shown, the electronic device 900 is in the form of a general computing device. The components of the electronic device 900 may include but are not limited to: at least one processing unit 910, at least one storage unit 920, a bus 930 connecting different system components (including the storage unit 920 and the processing unit 910), and a display unit 940.

[0116] The storage unit stores a program code, which can be executed by the processing unit 910, so that the processing unit 910 performs the steps according to various exemplary embodiments of the present disclosure described in the above “Exemplary Method” section of this specification. For example, the processing unit 910 can perform the following steps: Figure 1 In step S110, the ground elevation data of the area to be surveyed is obtained, and the first survey flight routes of at least two UAVs are determined in combination with the ground elevation data; step S120, based on the first survey flight route, the UAV is controlled to collect initial area terrain data, and the initial area terrain data is obtained by real-time collection of the laser radar and optical camera carried by the UAV; step S130, the initial area terrain data is identified and classified, and the weak texture pseudo terrain area in the initial area terrain data is determined; step S140, according to the weak texture pseudo terrain area, the second survey flight routes corresponding to at least two UAVs are re-planned, and the second survey flight routes are used to control at least two UAVs to turn on the millimeter wave radar, and scan and collect the supplementary terrain survey data of the weak texture pseudo terrain area at different angles; step S150, the supplementary terrain survey data at different angles are integrated to obtain the local terrain survey data of the weak texture pseudo terrain area; step S160, the initial area terrain data and the local terrain survey data are spliced ​​and integrated to obtain the real terrain survey data of the area to be surveyed.

[0117] The storage unit 920 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 921 and / or a cache memory unit 922 , and may further include a read-only memory unit (ROM) 923 .

[0118] The storage unit 920 may also include a program / utility 924 having a set (at least one) of program modules 925, such program modules 925 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0119] Bus 930 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0120] The electronic device 900 may also communicate with one or more external devices 970 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 900, and / or communicate with any device that enables the electronic device 900 to communicate with one or more other computing devices (e.g., routers, modems, etc.). This communication may be performed through an input / output (I / O) interface 950. In addition, the electronic device 900 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) through a network adapter 960. As shown, the network adapter 960 communicates with other modules of the electronic device 900 through a bus 930. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 900, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0121] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the embodiment of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the embodiment of the present disclosure.

[0122] In an exemplary embodiment of the present disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the above method of the present specification is stored. In some possible embodiments, various aspects of the present disclosure may also be implemented in the form of a program product, which includes a program code, and when the program product is run on a terminal device, the program code is used to enable the terminal device to perform the steps according to various exemplary embodiments of the present disclosure described in the above "Exemplary Method" section of the present specification.

[0123] refer to Fig.10As shown, a program product 1000 for implementing the terrain survey method based on UAV collaboration according to an embodiment of the present disclosure is described, which can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto, and in this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus, or a device.

[0124] The program product may adopt any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0125] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0126] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0127] Program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0128] In addition, the above-mentioned figures are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present disclosure, and are not intended to be limiting. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0129] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the embodiment of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the embodiment of the present disclosure.

[0130] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are to be considered exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

[0131] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A terrain survey method based on UAV collaboration, characterized in that: include: Acquire ground elevation data of the area to be surveyed, and determine first survey flight routes of at least two of the UAVs in combination with the ground elevation data; Based on the first survey flight route, the UAV is controlled to collect initial area terrain data, where the initial area terrain data is collected in real time by a laser radar and an optical camera carried by the UAV; Identify and classify the initial area terrain data to determine a weak texture false terrain area in the initial area terrain data; Replanning second survey flight routes respectively corresponding to at least two of the UAVs according to the weak-texture false terrain area, wherein the second survey flight routes are used to control at least two of the UAVs to turn on millimeter-wave radars and scan and collect supplementary terrain survey data of the weak-texture false terrain area at different angles; Fusing the supplementary terrain survey data at different angles to obtain local terrain survey data of the weak-texture false terrain area; The initial area terrain data and the local terrain survey data are spliced ​​and fused to obtain the real terrain survey data of the area to be surveyed.

2. The terrain survey method based on UAV collaboration according to claim 1 is characterized in that: The step of determining the first survey flight routes of at least two UAVs in combination with the ground elevation data comprises: Determining the terrain relief parameters of the area to be surveyed according to the ground elevation data, wherein the terrain relief parameters include elevation change rate, slope distribution, height gradient and surface roughness; Dynamically grid the area to be surveyed based on the terrain relief parameters to determine a survey sub-area corresponding to the area to be surveyed; Allocating the drone to each of the survey sub-areas to obtain an initial survey flight route; The initial survey flight route is analyzed for overlapping areas and blind spots, and the route spacing and scanning angle of the UAV are adjusted based on the analysis results to generate a first survey flight route.

3. The terrain survey method based on UAV collaboration according to claim 2 is characterized in that: The dynamically gridding the area to be surveyed based on the terrain relief parameters to determine the survey sub-area corresponding to the area to be surveyed includes: Determining the complex terrain area and the flat terrain area of ​​the area to be surveyed according to the terrain undulation parameter and a preset terrain undulation threshold; Performing grid division on the complex terrain region with a first division density, and performing grid division on the flat terrain region with a second division density, to obtain a survey sub-region corresponding to the area to be surveyed; The first partition density is greater than the second partition density.

4. The terrain survey method based on UAV collaboration according to claim 1 is characterized in that: The identifying and classifying the initial area terrain data to determine the weak texture false terrain area in the initial area terrain data includes: Preprocessing the initial area terrain data, the initial area terrain data including image data and laser point cloud data, the preprocessing at least including noise removal, texture enhancement processing, data standardization and point cloud density equalization; Extracting image texture features from the preprocessed image data, and extracting point cloud spatial features from the preprocessed laser point cloud data; A terrain area where the image texture feature is lower than a preset texture threshold and the point cloud spatial feature belongs to an abnormal spatial feature is determined as a weak-texture false terrain area.

5. The terrain survey method based on UAV collaboration according to claim 4 is characterized in that: The step of extracting the point cloud spatial features of the preprocessed laser point cloud data comprises: Performing multi-echo classification on the laser signal in the laser point cloud data to obtain a multi-echo signal, wherein the multi-echo signal includes an initial echo signal, an intermediate echo signal, and a terminal echo signal; Determine the echo layer point cloud features corresponding to each echo layer point cloud data in the multi-echo signal, wherein the echo layer point cloud features include point cloud density, normal vector distribution and curvature change rate; The point cloud spatial features of the laser point cloud data are constructed according to the echo layer point cloud features.

6. The terrain survey method based on UAV collaboration according to claim 5 is characterized in that: The method further comprises: Determine terrain gradient distribution estimation data of the current terrain area by combining the echo layer point cloud features corresponding to the intermediate echo signal and the terminal echo signal; Comparing the echo layer point cloud features corresponding to the initial echo signal with the terrain gradient distribution estimation data; If the terrain gradient corresponding to the echo layer point cloud feature corresponding to the initial echo signal is inconsistent with the terrain gradient distribution estimation data, and the point cloud density of the initial echo signal is lower than the preset density value, the change rate of the normal vector distribution is greater than the preset change rate threshold, and the curvature change rate is in an abnormal state, then it is determined that the point cloud spatial feature of the current terrain area is an abnormal spatial feature.

7. The terrain survey method based on UAV collaboration according to claim 1 is characterized in that: The replanning of the second survey flight routes respectively corresponding to at least two of the UAVs according to the weak texture false terrain area includes: Performing spatial distribution analysis on the point cloud data corresponding to the weak texture false terrain area to determine the regional spatial features corresponding to the weak texture false terrain area, wherein the regional spatial features include the regional boundary range, the regional area and the terrain undulation change features; Determine the survey priority corresponding to each of the weak-texture pseudo terrain areas based on the regional spatial characteristics; The device parameters and flight attitude data corresponding to at least two of the UAVs are obtained, and the first survey flight route corresponding to the UAV is locally updated in flight attitude in combination with the survey priority, the regional spatial characteristics, the device parameters and the flight attitude data to obtain a second survey flight route.

8. The terrain survey method based on UAV collaboration according to claim 1 is characterized in that: The fusing of the supplementary terrain survey data at different angles to obtain local terrain survey data of the weak-texture false terrain area includes: Preprocessing the supplementary topographic survey data at different angles, and determining key feature points in the preprocessed supplementary topographic survey data, wherein the key feature points include elevation mutation points, boundary points, and curvature feature points; Performing multi-view data registration on the pre-processed supplementary topographic survey data according to the key feature points to obtain spatially aligned supplementary topographic survey data; Based on a predetermined allocation weight, weighted fusion is performed on the spatially aligned supplementary terrain survey data and the initial area terrain data corresponding to the weak texture false terrain area to obtain fused terrain survey data; Density compensation and data smoothing are performed on the fused terrain survey data to obtain local terrain survey data of the weak-texture pseudo terrain area.

9. A terrain survey device based on UAV collaboration, characterized in that: include: An initial flight survey module, used to obtain ground elevation data of the area to be surveyed, and determine a first survey flight route of at least two of the UAVs in combination with the ground elevation data; A terrain data acquisition module, used for controlling the UAV to acquire terrain data of an initial area based on the first survey flight route, wherein the terrain data of the initial area is acquired in real time by a laser radar and an optical camera carried by the UAV; A false terrain region identification module, used to identify and classify the initial region terrain data, and determine a weak texture false terrain region in the initial region terrain data; A survey route updating module, used for re-planning second survey flight routes respectively corresponding to at least two of the drones according to the weak texture false terrain area, wherein the second survey flight routes are used for controlling at least two of the drones to turn on millimeter wave radars and scan and collect supplementary terrain survey data of the weak texture false terrain area at different angles; A supplementary survey data fusion module, used to fuse the supplementary terrain survey data at different angles to obtain local terrain survey data of the weak texture false terrain area; The real terrain data output module is used to splice and fuse the initial area terrain data and the local terrain survey data to obtain the real terrain survey data of the area to be surveyed.

10. An electronic device, characterized in that: include: processor; as well as A memory having computer-readable instructions stored thereon, wherein the computer-readable instructions, when executed by the processor, implement the terrain survey method based on UAV collaboration as described in any one of claims 1 to 8.

Images