Unmanned aerial vehicle three-dimensional aerial survey data processing method and system

Abstract

The invention relates to the technical field of aerial survey data processing, in particular to an unmanned aerial vehicle three-dimensional aerial survey data processing method and system. The method comprises the following steps: carrying an optical camera, a laser radar and a global navigation satellite system receiver on an unmanned aerial vehicle, and carrying out terrain multi-source data acquisition and preprocessing on terrain aerial survey corresponding to the unmanned aerial vehicle; an unmanned aerial vehicle aerial survey terrain standard image, unmanned aerial vehicle aerial survey terrain three-dimensional point cloud standard data and an unmanned aerial vehicle aerial survey flight positioning track are obtained; performing aerial survey three-dimensional panoramic matching splicing and terrain texture feature point detection on the corresponding unmanned aerial vehicle aerial survey terrain standard image at each aerial survey time point to obtain an unmanned aerial vehicle aerial survey three-dimensional terrain texture feature point set; and performing three-dimensional Poisson reconstruction and position positioning labeling on terrain aerial survey corresponding to the unmanned aerial vehicle, and generating an unmanned aerial vehicle aerial survey geographic target labeling three-dimensional model. According to the invention, efficient and accurate processing of aerial survey data can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of aerial survey data processing, and in particular to a method and system for processing three-dimensional aerial survey data of an unmanned aerial vehicle. Background Art

[0002] Drones are low-cost, efficient, and flexible, capable of completing large-scale, high-precision aerial surveys in a short period of time. They are widely used in fields such as geographic information systems (GIS), engineering construction, urban planning, agricultural monitoring, and environmental protection. The application of drones in three-dimensional aerial surveys primarily relies on high-precision sensors (such as high-resolution cameras and lidar) to acquire three-dimensional coordinate data of ground objects, enabling terrain modeling, mapping, and analysis. However, existing three-dimensional data processing methods mostly focus on optimizing a single link, such as image dedistortion, point cloud filtering, and image matching, but often overlook the coordination and efficiency of the entire data processing process. Furthermore, despite the obvious advantages of drone three-dimensional aerial surveying technology, the aerial survey data captured by drones typically contains a large amount of image or point cloud data. The data volume is large and high-dimensional, making it difficult to process, resulting in low accuracy or incomplete data. Summary of the Invention

[0003] Based on this, it is necessary for the present invention to provide a method and system for processing three-dimensional aerial survey data of a UAV to solve at least one of the above technical problems.

[0004] To achieve the above-mentioned purpose, a method for processing UAV 3D aerial survey data includes the following steps: Step S1: An optical camera, a laser radar, and a global navigation satellite system receiver are mounted on a UAV, and the optical camera, the laser radar, and the global navigation satellite system receiver are used to collect and preprocess terrain multi-source data for the terrain aerial survey corresponding to the UAV, thereby obtaining a standard image of the UAV aerial survey terrain, standard three-dimensional point cloud data of the UAV aerial survey terrain, and a flight positioning trajectory of the UAV aerial survey; Step S2: performing aerial survey 3D panoramic matching and stitching on the corresponding drone aerial survey terrain standard image at each aerial survey time point based on the drone aerial survey flight positioning trajectory to generate a drone aerial survey 3D panoramic terrain space; performing terrain texture feature point detection on the drone aerial survey 3D panoramic terrain space to obtain a drone aerial survey 3D terrain texture feature point set; Step S3: performing a three-dimensional Poisson reconstruction on the terrain surveyed by the UAV based on the three-dimensional point cloud standard data of the UAV aerial survey terrain and the three-dimensional terrain texture feature point set of the UAV aerial survey terrain to generate a three-dimensional surface model of the UAV aerial survey terrain; Step S4: obtaining the corresponding terrain aerial survey geographic target through the terrain aerial survey corresponding to the UAV, and positioning and labeling the UAV aerial survey terrain three-dimensional surface model based on the terrain aerial survey geographic target to generate the UAV aerial survey geographic target labeled three-dimensional model.

[0005] Furthermore, step S1 includes the following steps: Step S11: by equipping the drone with an optical camera, a laser radar, and a global navigation satellite system receiver; Step S12: using an optical camera, a laser radar, and a global navigation satellite system receiver to collect terrain multi-source data for the terrain aerial survey corresponding to the UAV, so as to extract a visible light image of the surface texture corresponding to the UAV aerial survey through the optical camera, obtain corresponding terrain three-dimensional point cloud data through the laser radar, and obtain corresponding real-time flight position point set of the UAV aerial survey through the global navigation satellite system receiver; Step S13: performing image preprocessing on the visible light image of the surface texture corresponding to the UAV aerial survey, using an adaptive filtering method to remove pixel noise corresponding to the visible light image of the surface texture, and using image enhancement technology to improve the contrast and clarity of the image, thereby obtaining a standard image of the terrain measured by the UAV aerial survey; Step S14: performing outlier filtering on the standard three-dimensional point cloud data of the drone aerial survey terrain to obtain the standard three-dimensional point cloud data of the drone aerial survey terrain; Step S15: Determine the flight trajectory of the terrain aerial survey corresponding to the UAV based on the real-time flight position point set of the UAV aerial survey to obtain the UAV aerial survey flight positioning trajectory.

[0006] Furthermore, step S15 includes the following steps: Step S151: determining the time interval for each flight position point in the real-time flight position point set of the UAV aerial survey, and obtaining the aerial survey recording time interval corresponding to each flight position point; Step S152: performing adjacent position space dimension analysis on each flight position point in the UAV aerial survey real-time flight position point set to obtain the aerial survey space distance and aerial survey space direction between adjacent flight position points; Step S153: performing a position point spatial distribution analysis on the aerial survey spatial distance and aerial survey spatial direction between adjacent flight position points based on the aerial survey recording time interval corresponding to each flight position point, so as to obtain the spatial distribution characteristics of the UAV aerial survey position points; Step S154: Obtaining the terrain elevation, slope, and landform type corresponding to the UAV aerial survey, and performing a terrain flight constraint analysis based on the terrain elevation, slope, and landform type corresponding to the UAV aerial survey to generate the terrain flight constraints for the UAV aerial survey, including the restrictions imposed by terrain undulations and obstacle distribution on the UAV aerial survey range; Step S155: Determine the flight trajectory of the terrain aerial survey corresponding to the UAV based on the UAV aerial survey terrain flight constraints and in combination with the spatial distribution characteristics of the UAV aerial survey position points to obtain the UAV aerial survey flight positioning trajectory.

[0007] Furthermore, the spatial distribution characteristics of the drone aerial survey position points in step S153 specifically include the aerial survey flight curvature and the aerial survey flight angle corresponding to each flight position point.

[0008] Furthermore, step S2 includes the following steps: Step S21: performing image edge depth mining on the standard UAV aerial survey terrain image corresponding to each aerial survey time point to obtain edge feature descriptors corresponding to each UAV aerial survey terrain image, including edge contour, edge color first-order moment, edge color second-order moment, and edge color third-order moment; Step S22: performing three-dimensional space edge matching positioning on the edge feature descriptors corresponding to each UAV aerial survey terrain image based on the UAV aerial survey flight positioning trajectory, and obtaining the UAV matching splicing positioning corresponding to each UAV aerial survey terrain image in three-dimensional space; Step S23: Obtain the corresponding UAV flight latitude, longitude, altitude, and attitude angle by matching and stitching the UAVs corresponding to each UAV aerial survey terrain image in three-dimensional space, and perform three-dimensional space matrix transformation on the UAV flight latitude, longitude, altitude, and attitude angle to generate the corresponding aerial survey coordinate matrix and aerial survey rotation matrix for each UAV aerial survey terrain image in three-dimensional space; Step S24: performing aerial survey 3D panoramic matching and stitching on the corresponding UAV aerial survey terrain standard image at each aerial survey time point based on the aerial survey coordinate matrix and the aerial survey rotation matrix, and performing gap optimization processing on the stitched 3D panoramic terrain space to generate the UAV aerial survey 3D panoramic terrain space; Step S25: performing terrain texture feature point detection on the UAV aerial survey 3D panoramic terrain space to obtain a UAV aerial survey 3D terrain texture feature point set.

[0009] Furthermore, step S3 includes the following steps: Step S31: performing point cloud and texture depth analysis on the standard 3D point cloud data of the UAV aerial survey terrain and the texture feature point set of the UAV aerial survey terrain, using a voxelization method to divide the corresponding standard 3D point cloud data of the UAV aerial survey terrain into uniform small cubes, and calculating the corresponding point density and distribution direction within each small cube. At the same time, using the Laplace operator to analyze the energy distribution of the texture feature points at different frequencies, the UAV aerial survey 3D point cloud and texture depth feature set are obtained; Step S32: performing geographic coordinate conversion on the standard three-dimensional point cloud data of the UAV aerial survey terrain and the three-dimensional terrain texture feature point set of the UAV aerial survey based on the UAV aerial survey three-dimensional point cloud and texture depth feature set and in combination with the geophysical model and the UAV flight attitude model, so as to convert the point cloud data and the terrain texture feature point set from the UAV coordinate system to the same geographic coordinate system, thereby obtaining the UAV aerial survey point cloud data and the UAV aerial survey terrain texture feature point set in the same three-dimensional geographic coordinate system; Step S33: Perform three-dimensional Poisson reconstruction on the terrain surveyed by the UAV based on the UAV aerial survey point cloud data and the UAV aerial survey terrain texture feature point set in the same three-dimensional geographic coordinate system to generate a three-dimensional surface model of the UAV aerial survey terrain.

[0010] Furthermore, step S33 includes the following steps: Perform geometric connection analysis on the UAV aerial survey point cloud data and the UAV aerial survey terrain texture feature point set in the same three-dimensional geographic coordinate system to obtain the topological connection relationship between each UAV aerial survey point cloud and the geometric connection relationship between each UAV aerial survey texture feature point; Based on the topological connection relationship between each UAV aerial survey point cloud and the geometric connection relationship between each UAV aerial survey texture feature point, a three-dimensional Poisson reconstruction is performed on the terrain aerial survey corresponding to the UAV, so as to quickly construct the approximate outline corresponding to the UAV aerial survey terrain according to the density distribution of each UAV aerial survey point cloud, and use each UAV aerial survey texture feature point for texture mapping and detail supplementation to generate a three-dimensional surface model of the UAV aerial survey terrain.

[0011] Furthermore, step S4 includes the following steps: Step S41: performing ground object type identification on the terrain aerial survey corresponding to the UAV to obtain the UAV aerial survey ground object type, including building, road and vegetation ground object types; Step S42: performing geographic target detection on the terrain aerial survey corresponding to the UAV based on the type of the UAV aerial survey ground object to obtain the corresponding terrain aerial survey geographic target; Step S43: Positioning and labeling the three-dimensional surface model of the terrain surveyed by the UAV based on the terrain aerial survey geographic target to generate a three-dimensional model of the terrain surveyed by the UAV geographic target labeling.

[0012] Furthermore, the terrain aerial survey geographical targets in step S42 include building distribution, road vertical and horizontal distribution, bridges, electric poles and flower garden geographical targets.

[0013] Furthermore, the present invention also provides a UAV 3D aerial survey data processing system for executing the UAV 3D aerial survey data processing method described above. The UAV 3D aerial survey data processing system includes: The aerial survey terrain multi-source data acquisition module is used to equip the UAV with an optical camera, a lidar, and a global navigation satellite system receiver, and use the optical camera, lidar, and global navigation satellite system receiver to collect and pre-process the terrain multi-source data corresponding to the UAV aerial survey, thereby obtaining the UAV aerial survey terrain standard image, the UAV aerial survey terrain 3D point cloud standard data, and the UAV aerial survey flight positioning trajectory; The aerial survey terrain texture feature detection module is used to perform aerial survey 3D panoramic matching and stitching on the corresponding drone aerial survey terrain standard image at each aerial survey time point based on the drone aerial survey flight positioning trajectory to generate the drone aerial survey 3D panoramic terrain space; the module also performs terrain texture feature point detection on the drone aerial survey 3D panoramic terrain space to obtain the drone aerial survey 3D terrain texture feature point set; The terrain 3D Poisson reconstruction module is used to perform 3D Poisson reconstruction on the terrain corresponding to the drone based on the standard 3D point cloud data of the drone aerial survey terrain and the 3D terrain texture feature point set of the drone aerial survey terrain, thereby generating a 3D surface model of the drone aerial survey terrain; The aerial survey geographic target positioning and labeling module is used to obtain the corresponding terrain aerial survey geographic target through the terrain aerial survey corresponding to the UAV, and to position and label the UAV aerial survey terrain three-dimensional surface model based on the terrain aerial survey geographic target, thereby generating a UAV aerial survey geographic target labeled three-dimensional model.

[0014] Beneficial effects of the present invention: 1. Compared with the existing technology, the method for processing three-dimensional aerial survey data of unmanned aerial vehicles proposed in the present invention has the beneficial effect of carrying out multi-source data collection for terrain aerial survey by equipping the unmanned aerial vehicle with an optical camera, a laser radar (LiDAR) and a global navigation satellite system (GNSS) receiver. These devices work together to ensure accurate measurement and data capture of the terrain. The optical camera can provide high-resolution standard images, the laser radar obtains accurate three-dimensional point cloud data of ground objects through laser beam emission and reception, and the GNSS receiver ensures the precise positioning of the unmanned aerial vehicle during flight. By combining these different sensor data, not only can the unmanned aerial vehicle's track be accurately located in space, but also high-quality terrain images and three-dimensional point cloud data can be generated. By integrating and synchronously processing data from different sources, the consistency and accuracy of the data in the subsequent processing links are ensured, the accuracy of the overall terrain surveying and mapping is improved, and the occurrence of problems such as data dislocation and distortion is avoided. Secondly, the terrain standard image at each aerial survey time point is used for three-dimensional panoramic matching and stitching based on the UAV's flight positioning trajectory. The precise data of the aerial survey flight positioning trajectory can be used to obtain the relative position of the UAV at each aerial survey time point, thereby ensuring the correct stitching of the image to form a seamless three-dimensional panoramic terrain space. This process uses image matching technology to ensure smooth transitions between images and seamless connections of seams, minimizing stitching errors caused by changes in camera perspective, imaging angle and illumination. Once the image stitching is completed, the three-dimensional panoramic terrain space will be subjected to terrain texture feature point detection. Through the identification of texture feature points, the surface features of ground objects can be more accurately extracted, thereby providing key data support for subsequent three-dimensional reconstruction. Texture feature points represent specific geographic information in three-dimensional space, usually reflecting the detailed features of the terrain, such as the outline of buildings, the boundaries of roads, etc. This process not only improves the accuracy and completeness of data processing, but also lays the foundation for subsequent modeling and target recognition. Then, three-dimensional Poisson reconstruction is performed based on the three-dimensional point cloud data of the drone and the terrain texture feature point set. Poisson reconstruction is a technology widely used in three-dimensional model generation. By performing algorithmic processing on discrete point cloud data, a smooth and continuous surface model is restored. This process effectively converts the three-dimensional point cloud data collected by the drone into a complete three-dimensional surface model, which can fill the gaps and irregularities in the point cloud data, ensuring that the generated three-dimensional surface model is not only highly accurate, but also has strong authenticity and expressiveness.Finally, the terrain aerial survey data obtained by UAV aerial survey is used to identify the corresponding geographic targets and annotate the generated three-dimensional surface model. The key to this stage is how to accurately identify geographic targets (such as buildings, roads, bridges and other ground features) and accurately annotate their positions in three-dimensional space. By identifying and analyzing feature points, each geographic target can be accurately located in the three-dimensional model, generating the final geographic target annotated three-dimensional model. This can help researchers, engineers or decision makers intuitively understand the spatial distribution and characteristics of the terrain, thereby improving the accuracy of aerial survey data processing.

[0015] 2. The UAV three-dimensional aerial survey data processing system proposed in the present invention is generally composed of an aerial survey terrain multi-source data acquisition module, an aerial survey terrain texture feature detection module, a terrain three-dimensional Poisson reconstruction module, and an aerial survey geographic target positioning and annotation module. It can implement any UAV three-dimensional aerial survey data processing method described in the present invention, and is used to combine the operations between the computer programs running on each module to implement the UAV three-dimensional aerial survey data processing method. The internal structures of the system cooperate with each other, which can greatly reduce duplication of work and manpower investment, and can quickly and effectively provide a more accurate and efficient UAV three-dimensional aerial survey data processing process, thereby simplifying the operating procedures of the UAV three-dimensional aerial survey data processing system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Other features, objects and advantages of the present invention will become more apparent upon reading the detailed description of non-limiting embodiments thereof made with reference to the following drawings: Figure 1 This is a schematic flow chart of the steps of the method for processing three-dimensional aerial survey data of a UAV according to the present invention; Figure 2 for Figure 1 Detailed step flow diagram of step S1; Figure 3 for Figure 2 Detailed step flow chart of step S15. DETAILED DESCRIPTION

[0017] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0018] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0019] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0020] To achieve this, please refer to Figures 1 to 2 The present invention provides a method for processing three-dimensional aerial survey data of a UAV, the method comprising the following steps: Step S1: An optical camera, a laser radar, and a global navigation satellite system receiver are mounted on a UAV, and the optical camera, the laser radar, and the global navigation satellite system receiver are used to collect and preprocess terrain multi-source data for the terrain aerial survey corresponding to the UAV, thereby obtaining a standard image of the UAV aerial survey terrain, standard three-dimensional point cloud data of the UAV aerial survey terrain, and a flight positioning trajectory of the UAV aerial survey; Step S2: performing aerial survey 3D panoramic matching and stitching on the corresponding drone aerial survey terrain standard image at each aerial survey time point based on the drone aerial survey flight positioning trajectory to generate a drone aerial survey 3D panoramic terrain space; performing terrain texture feature point detection on the drone aerial survey 3D panoramic terrain space to obtain a drone aerial survey 3D terrain texture feature point set; Step S3: performing a three-dimensional Poisson reconstruction on the terrain surveyed by the UAV based on the three-dimensional point cloud standard data of the UAV aerial survey terrain and the three-dimensional terrain texture feature point set of the UAV aerial survey terrain to generate a three-dimensional surface model of the UAV aerial survey terrain; Step S4: obtaining the corresponding terrain aerial survey geographic target through the terrain aerial survey corresponding to the UAV, and positioning and labeling the UAV aerial survey terrain three-dimensional surface model based on the terrain aerial survey geographic target to generate the UAV aerial survey geographic target labeled three-dimensional model.

[0021] In the embodiment of the present invention, please refer to Figure 1FIG. 1 is a flow chart showing the steps of the method for processing three-dimensional aerial survey data of a UAV according to the present invention. In this example, the method for processing three-dimensional aerial survey data of a UAV includes the following steps: Step S1: An optical camera, a laser radar, and a global navigation satellite system receiver are mounted on a UAV, and the optical camera, the laser radar, and the global navigation satellite system receiver are used to collect and preprocess terrain multi-source data for the terrain aerial survey corresponding to the UAV, thereby obtaining a standard image of the UAV aerial survey terrain, standard three-dimensional point cloud data of the UAV aerial survey terrain, and a flight positioning trajectory of the UAV aerial survey; In an embodiment of the present invention, aerial terrain survey data is collected using a drone equipped with an optical camera, a laser radar (LiDAR), and a global navigation satellite system (GNSS) receiver. The optical camera is responsible for acquiring large-scale images of the terrain surface and providing two-dimensional image data. The LiDAR generates accurate three-dimensional point cloud data of the terrain surface by emitting laser beams and receiving reflected signals, thereby capturing three-dimensional information about the terrain. The GNSS receiver is responsible for recording the drone's positioning information in real time and calculating the drone's precise position and attitude by receiving satellite signals. The synchronized operation of all devices ensures high accuracy in the data collection process. After data collection is completed, the obtained images, point cloud data, and positioning trajectory are pre-processed. The images undergo geometric correction to eliminate distortion and are orthorectified using known camera parameters. The point cloud data undergoes filtering, denoising, and coordinate conversion to ensure accuracy and consistency. The positioning trajectory is then resolved using GNSS signals to remove errors and calibrate the precise position of each collection point. Ultimately, standard images of the drone's aerial terrain survey, standard three-dimensional point cloud data of the drone's aerial terrain survey, and the drone's aerial survey flight positioning trajectory are obtained.

[0022] Step S2: performing aerial survey 3D panoramic matching and stitching on the corresponding drone aerial survey terrain standard image at each aerial survey time point based on the drone aerial survey flight positioning trajectory to generate a drone aerial survey 3D panoramic terrain space; performing terrain texture feature point detection on the drone aerial survey 3D panoramic terrain space to obtain a drone aerial survey 3D terrain texture feature point set; In an embodiment of the present invention, three-dimensional panoramic stitching is performed on the terrain standard images corresponding to each aerial survey time point according to the aerial survey flight positioning trajectory of the drone. During the stitching process, feature points of each image are first detected, and a feature matching algorithm (such as SIFT or SURF algorithm) is used to accurately align the images. Then, by combining the positioning trajectory information and adopting a timestamp-based interpolation method, the images of each time point are stitched at the correct spatial position to form a continuous drone aerial survey three-dimensional panoramic terrain space. Next, image processing technology is used to detect terrain texture feature points on the stitched panoramic image. Through edge detection, corner detection and other methods, representative texture feature points of the terrain surface are identified to form a three-dimensional terrain texture feature point set, and finally a three-dimensional terrain texture feature point set of the drone aerial survey is obtained.

[0023] Step S3: performing a three-dimensional Poisson reconstruction on the terrain surveyed by the UAV based on the three-dimensional point cloud standard data of the UAV aerial survey terrain and the three-dimensional terrain texture feature point set of the UAV aerial survey terrain to generate a three-dimensional surface model of the UAV aerial survey terrain; In an embodiment of the present invention, three-dimensional Poisson reconstruction is performed based on three-dimensional point cloud data and a three-dimensional terrain texture feature point set obtained from drone aerial survey. First, an accurate terrain point cloud model is obtained by filtering and removing noise from the three-dimensional point cloud data. Then, a Poisson surface reconstruction algorithm is used. This algorithm can generate a smooth three-dimensional surface by solving the Poisson equation based on the normal vector information of the point cloud data. In order to improve the accuracy and stability of the surface reconstruction, the terrain texture features extracted from the texture feature point set are weighted so that the reconstructed surface not only accurately reflects the geometric shape of the terrain, but also better reflects the texture characteristics of the terrain, ultimately generating a three-dimensional surface model of the drone aerial survey terrain.

[0024] Step S4: obtaining the corresponding terrain aerial survey geographic target through the terrain aerial survey corresponding to the UAV, and positioning and labeling the UAV aerial survey terrain three-dimensional surface model based on the terrain aerial survey geographic target to generate the UAV aerial survey geographic target labeled three-dimensional model.

[0025] In an embodiment of the present invention, geographic target recognition is performed on terrain data obtained by drone aerial survey. First, existing geographic information system (GIS) data and terrain models are used for comparative analysis. Specific targets (such as buildings, roads, bridges, etc.) in the terrain surface model are automatically identified and classified through image recognition technology and deep learning methods. The location information of these targets is matched with the aerial survey flight trajectory and positioning data to obtain a precise geographic location. Subsequently, based on the identified geographic target information, the three-dimensional surface model of the drone aerial survey is positionally labeled. Accurate three-dimensional coordinates are assigned to each target using a three-dimensional coordinate system, and these coordinates are labeled with corresponding target features. The resulting three-dimensional model of geographic target annotation for drone aerial survey not only has the accurate spatial position of the geographic target, but also includes the specific morphological characteristics of the target.

[0026] Furthermore, step S1 includes the following steps: Step S11: by equipping the drone with an optical camera, a laser radar, and a global navigation satellite system receiver; Step S12: using an optical camera, a laser radar, and a global navigation satellite system receiver to collect terrain multi-source data for the terrain aerial survey corresponding to the UAV, so as to extract a visible light image of the surface texture corresponding to the UAV aerial survey through the optical camera, obtain corresponding terrain three-dimensional point cloud data through the laser radar, and obtain corresponding real-time flight position point set of the UAV aerial survey through the global navigation satellite system receiver; Step S13: performing image preprocessing on the visible light image of the surface texture corresponding to the UAV aerial survey, using an adaptive filtering method to remove pixel noise corresponding to the visible light image of the surface texture, and using image enhancement technology to improve the contrast and clarity of the image, thereby obtaining a standard image of the terrain measured by the UAV aerial survey; Step S14: performing outlier filtering on the standard three-dimensional point cloud data of the drone aerial survey terrain to obtain the standard three-dimensional point cloud data of the drone aerial survey terrain; Step S15: determining the flight trajectory of the terrain aerial survey corresponding to the UAV based on the real-time flight position point set of the UAV aerial survey, and obtaining the UAV aerial survey flight positioning trajectory.

[0027] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed step flow diagram of step S1 in FIG. 1 , in this embodiment, step S1 includes the following steps: Step S11: by equipping the drone with an optical camera, a laser radar, and a global navigation satellite system receiver; In an embodiment of the present invention, an optical camera, a laser radar (LiDAR), and a global navigation satellite system (GNSS) receiver are mounted on a drone to achieve subsequent terrain aerial survey data collection. First, the optical camera is installed on the drone's fuselage to capture ground images at a specific angle and resolution, ensuring that the images have sufficient resolution and clarity for subsequent texture analysis. The LiDAR utilizes the principle of laser pulse emission and reception reflection to measure the time required for the light pulse to be emitted from the laser and returned to the receiver, calculates the corresponding distance information, and generates high-precision three-dimensional point cloud data. The GNSS receiver continuously receives signals from satellites, accurately determines the real-time position of the drone, and provides higher positioning accuracy through differential technology. These three devices work simultaneously to ensure the comprehensiveness and high precision of terrain data, forming a multi-source data acquisition system.

[0028] Step S12: using an optical camera, a laser radar, and a global navigation satellite system receiver to collect terrain multi-source data for the terrain aerial survey corresponding to the UAV, so as to extract a visible light image of the surface texture corresponding to the UAV aerial survey through the optical camera, obtain corresponding terrain three-dimensional point cloud data through the laser radar, and obtain corresponding real-time flight position point set of the UAV aerial survey through the global navigation satellite system receiver; In an embodiment of the present invention, the collection of multi-source terrain data is completed through the collaboration of an optical camera, a lidar, and a global navigation satellite system receiver. The optical camera captures visible light images of the ground texture. These images can provide accurate texture information for subsequent surface texture analysis through the high-resolution characteristics of the camera. The lidar generates detailed three-dimensional ground point cloud data by emitting laser beams and receiving reflected signals. The spatial position and height information of each point in the point cloud can be used to create a digital elevation model of the terrain. The global navigation satellite system receiver records the flight position point set of the drone in real time, and provides corresponding geographic coordinates for each image and point cloud data to ensure that all data have a consistent spatial reference.

[0029] Step S13: performing image preprocessing on the visible light image of the surface texture corresponding to the UAV aerial survey, using an adaptive filtering method to remove pixel noise corresponding to the visible light image of the surface texture, and using image enhancement technology to improve the contrast and clarity of the image, thereby obtaining a standard image of the terrain measured by the UAV aerial survey; In an embodiment of the present invention, image preprocessing is performed on visible light images of surface textures collected by a drone. First, an adaptive filtering algorithm is used to remove noise from the original image. The adaptive filter dynamically adjusts the filter intensity based on the characteristics of the local area of the image, so that the image details are preserved while the noise is effectively eliminated, thereby improving image quality. Next, image enhancement techniques, such as contrast stretching and histogram equalization, are applied to enhance the brightness and contrast of the image, making the surface features in the image clearer and facilitating subsequent texture analysis. The enhanced image becomes a standard terrain image, significantly improving image quality and visualization, ultimately resulting in a standard drone aerial survey terrain image.

[0030] Step S14: performing outlier filtering on the standard three-dimensional point cloud data of the drone aerial survey terrain to obtain the standard three-dimensional point cloud data of the drone aerial survey terrain; In an embodiment of the present invention, outlier filtering is performed on the three-dimensional point cloud data of the terrain collected by the UAV aerial survey. First, outliers that are significantly different from the majority of data points in the three-dimensional point cloud data are identified through statistical analysis. These outliers are invalid data caused by environmental interference, equipment errors, etc. Then, a density-based filtering method (such as the DBSCAN algorithm) or a distance-based filtering method (such as the K-nearest neighbor algorithm) is applied to remove the outliers. The key to the filtering process is to determine which points belong to noise and delete them by setting a reasonable threshold, while retaining most of the real ground points, thereby ensuring the accuracy and reliability of the three-dimensional point cloud data. The processed three-dimensional point cloud data is called standard data, which represents a more accurate terrain surface, and finally the standard three-dimensional point cloud data of the UAV aerial survey terrain is obtained.

[0031] Step S15: determining the flight trajectory of the terrain aerial survey corresponding to the UAV based on the real-time flight position point set of the UAV aerial survey, and obtaining the UAV aerial survey flight positioning trajectory.

[0032] In an embodiment of the present invention, the flight trajectory of the UAV is determined based on the real-time flight position point set of the UAV aerial survey. First, the flight position at each moment is determined based on the real-time positioning information provided by the global navigation satellite system receiver and the flight data of the UAV, such as speed and heading. Then, the complete flight trajectory of the UAV is constructed by interpolation of timestamps and position points. During the trajectory determination process, a smoothing algorithm (such as Kalman filtering or B-spline curve fitting) is used to eliminate trajectory jitter caused by positioning errors and flight state fluctuations to ensure the smoothness and accuracy of the trajectory. The generated flight trajectory is the motion path of the UAV, which can accurately represent the position changes of the UAV during the terrain aerial survey process, and finally the UAV aerial survey flight positioning trajectory is obtained.

[0033] Furthermore, step S15 includes the following steps: Step S151: determining the time interval for each flight position point in the real-time flight position point set of the UAV aerial survey, and obtaining the aerial survey recording time interval corresponding to each flight position point; Step S152: performing adjacent position space dimension analysis on each flight position point in the UAV aerial survey real-time flight position point set to obtain the aerial survey space distance and aerial survey space direction between adjacent flight position points; Step S153: performing a position point spatial distribution analysis on the aerial survey spatial distance and aerial survey spatial direction between adjacent flight position points based on the aerial survey recording time interval corresponding to each flight position point, so as to obtain the spatial distribution characteristics of the UAV aerial survey position points; Step S154: Obtaining the terrain elevation, slope, and landform type corresponding to the UAV aerial survey, and performing a terrain flight constraint analysis based on the terrain elevation, slope, and landform type corresponding to the UAV aerial survey to generate the terrain flight constraints for the UAV aerial survey, including the restrictions imposed by terrain undulations and obstacle distribution on the UAV aerial survey range; Step S155: Determine the flight trajectory of the terrain aerial survey corresponding to the UAV based on the UAV aerial survey terrain flight constraints and in combination with the spatial distribution characteristics of the UAV aerial survey position points to obtain the UAV aerial survey flight positioning trajectory.

[0034] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 2 Detailed step flow diagram of step S15 in the embodiment, step S15 includes the following steps: Step S151: determining the time interval for each flight position point in the real-time flight position point set of the UAV aerial survey, and obtaining the aerial survey recording time interval corresponding to each flight position point; In an embodiment of the present invention, the time intervals for a set of real-time flight location points of a UAV aerial survey are determined. First, the timestamps of all flight location points need to be obtained from the flight control system of the UAV. Each flight location point consists of its three-dimensional coordinates (X, Y, Z) and corresponding time information (T). The time interval is determined based on the time difference between each pair of adjacent flight location points, which can be obtained by calculating the difference between the timestamp of the current flight location point and the timestamp of the previous location point. Through this method, the time intervals of aerial survey records corresponding to all flight location points can be determined. For example, if the timestamp of the current flight location point is T2 and the timestamp of the previous flight location point is T1, then the time interval of aerial survey records for this location point is T2-T1. On this basis, the obtained time interval provides a key time reference for subsequent analysis, and ultimately the time interval of aerial survey records corresponding to each flight location point is obtained.

[0035] Step S152: performing adjacent position space dimension analysis on each flight position point in the UAV aerial survey real-time flight position point set to obtain the aerial survey space distance and aerial survey space direction between adjacent flight position points; In the embodiment of the present invention, for each flight position point in the real-time flight position point set of the UAV aerial survey, an adjacent position spatial dimension analysis is performed. First, the spatial distance between adjacent position points is calculated. The spatial distance can be calculated using the Euclidean distance formula between two points in a three-dimensional coordinate system, that is: ,in,( , , ) and( , , ) are the three-dimensional coordinates of adjacent flight position points. Then, based on the spatial coordinates of these two position points, the flight direction is calculated, that is, the spatial direction angle between the position points is calculated. The direction angle can be obtained by the coordinate difference between the two points combined with vector operations, and then the spatial direction of the flight path is obtained. Through these analyses, the aerial survey spatial distance and spatial direction between each pair of adjacent position points can be obtained in detail, and finally the aerial survey spatial distance and spatial direction between adjacent flight position points can be obtained.

[0036] Step S153: performing a position point spatial distribution analysis on the aerial survey spatial distance and aerial survey spatial direction between adjacent flight position points based on the aerial survey recording time interval corresponding to each flight position point, so as to obtain the spatial distribution characteristics of the UAV aerial survey position points; In an embodiment of the present invention, based on the aerial survey recording time interval corresponding to each flight position point, the aerial survey spatial distance and aerial survey spatial direction between adjacent flight position points are further analyzed to perform spatial distribution analysis of the position points. At this time, the spatial distribution analysis is based on the spatial position changes between the flight position points, and the flight curvature and flight angle are calculated in combination with the time interval obtained in the previous step. The specific operation is: first, the relationship between the time interval and the spatial distance is used to calculate the flight speed per unit time. Then, the curvature of the flight path, that is, the degree of curvature of the path, can be further calculated through the relationship between the speed and the spatial path. At the same time, the heading angle change is calculated to obtain the aerial survey flight angle. The aerial survey flight curvature and aerial survey flight angle corresponding to each flight position point can be obtained, and finally the spatial distribution characteristics of the drone aerial survey position points are obtained.

[0037] Step S154: Obtaining the terrain elevation, slope, and landform type corresponding to the UAV aerial survey, and performing a terrain flight constraint analysis based on the terrain elevation, slope, and landform type corresponding to the UAV aerial survey to generate the terrain flight constraints for the UAV aerial survey, including the restrictions imposed by terrain undulations and obstacle distribution on the UAV aerial survey range; In an embodiment of the present invention, terrain data involved in the drone aerial survey process is obtained, including terrain elevation, slope, and landform type. Terrain elevation is typically acquired using LiDAR (Light Detection and Ranging) or remote sensing imagery. Specifically, in ground surveying or remote sensing technology, real-time data collected by drone sensors is used to generate an elevation map. Slope information can also be inferred using a digital elevation model (DEM), which provides the degree of inclination at each surface location. Landform type classification is performed based on high-resolution remote sensing imagery and topographic features, combined with artificial intelligence classification technology. Then, based on this terrain information, a terrain flight constraint analysis is performed, primarily analyzing the impact of terrain undulation on drone flight. Areas with greater undulations may pose a higher risk to the drone's flight attitude and range, especially in areas with steep ascents or descents, where flight altitude must be more precisely controlled. Furthermore, the distribution of obstacles (such as buildings and mountains) must also be taken into consideration to develop flight altitude and trajectory constraints to ensure that the drone can successfully avoid these obstacles and complete the aerial survey mission. Ultimately, the drone's terrain flight constraints for aerial survey are generated.

[0038] Step S155: Determine the flight trajectory of the terrain aerial survey corresponding to the UAV based on the UAV aerial survey terrain flight constraints and in combination with the spatial distribution characteristics of the UAV aerial survey position points to obtain the UAV aerial survey flight positioning trajectory.

[0039] In the embodiment of the present invention, the flight trajectory is determined by combining the terrain flight constraints based on the UAV aerial survey with the previously obtained spatial distribution characteristics of the flight position points. First, all the terrain flight constraints are combined with the spatial distribution characteristics of the position points, and a path planning algorithm (such as The flight trajectory is optimized using the Dijkstra algorithm (or Dijkstra algorithm). This process needs to consider the influence of terrain undulations and the limitations of obstacles. The optimal flight trajectory is calculated under these constraints. The flight path of the UAV is adjusted according to the spatial distribution characteristics of each flight position point (including curvature, angle, etc.) to ensure that the UAV can complete the aerial survey mission within the limitations of the terrain while avoiding obstacles on the ground. During this process, the adjustment of the flight trajectory should also be fed back to the UAV control system in real time to ensure the accuracy and safety of the flight, and finally the UAV aerial survey flight positioning trajectory is obtained.

[0040] Furthermore, the spatial distribution characteristics of the drone aerial survey position points in step S153 specifically include the aerial survey flight curvature and the aerial survey flight angle corresponding to each flight position point.

[0041] Furthermore, step S2 includes the following steps: Step S21: performing image edge depth mining on the standard UAV aerial survey terrain image corresponding to each aerial survey time point to obtain edge feature descriptors corresponding to each UAV aerial survey terrain image, including edge contour, edge color first-order moment, edge color second-order moment, and edge color third-order moment; In an embodiment of the present invention, edge features are extracted from the drone aerial survey terrain image corresponding to each aerial survey time point. First, for each aerial survey terrain image, the Canny edge detection algorithm is used to identify the edge contour in the image. This method effectively extracts the significant edges in the image. Subsequently, the color features of the edges are calculated using image processing tools, including the first-order color moment, second-order color moment, and third-order color moment of the edge area in the image. These moments can respectively reflect the distribution, change, and fluctuation of the color of the edge area. In specific implementation, the OpenCV library in MATLAB or Python is used to perform image grayscale conversion, edge detection, and color feature calculation. After the edge contour is extracted, the color moment of each edge area is calculated by a statistical method and stored as an edge feature descriptor. Finally, the edge feature descriptor corresponding to each drone aerial survey terrain image is obtained.

[0042] Step S22: performing three-dimensional space edge matching positioning on the edge feature descriptors corresponding to each UAV aerial survey terrain image based on the UAV aerial survey flight positioning trajectory, and obtaining the UAV matching splicing positioning corresponding to each UAV aerial survey terrain image in three-dimensional space; In an embodiment of the present invention, the edge feature descriptors of each aerial survey topographic image are matched in three-dimensional space based on the flight positioning trajectory information of the drone. The specific operations include: first, determining the shooting angle and position of each image based on the flight trajectory data of the drone (including latitude, longitude and altitude information); then, using a three-dimensional matching algorithm, such as a feature-based matching method (for example, SIFT or SURF feature matching), spatially comparing and matching the image edge descriptors with other images, and locating the corresponding position of each aerial survey image in three-dimensional space by calculating the matching degree and similarity. For the matched images, by optimizing the matching parameters, it is ensured that the stitching effect between the images is accurate and has no obvious deviation. The matching process uses a sparse three-dimensional reconstruction algorithm (such as Structure from Motion) to further enhance the positioning accuracy, thereby obtaining the three-dimensional spatial positioning information corresponding to each image, ensuring that each image is accurately stitched in three-dimensional space, and ultimately obtaining the drone matching and stitching positioning corresponding to each drone aerial survey topographic image in three-dimensional space.

[0043] Step S23: Obtain the corresponding UAV flight latitude, longitude, altitude, and attitude angle by matching and stitching the UAVs corresponding to each UAV aerial survey terrain image in three-dimensional space, and perform three-dimensional space matrix transformation on the UAV flight latitude, longitude, altitude, and attitude angle to generate the corresponding aerial survey coordinate matrix and aerial survey rotation matrix for each UAV aerial survey terrain image in three-dimensional space; In an embodiment of the present invention, by obtaining the flight latitude and longitude, altitude and attitude angle corresponding to each UAV aerial survey terrain image in three-dimensional space, first, the precise flight position information is extracted through the GPS data and inertial navigation system (INS) data of the UAV. This information includes the latitude and longitude, altitude and attitude angles such as heading angle, pitch angle and roll angle of the UAV when each image is taken. Next, a three-dimensional space matrix conversion is performed to ensure that these geographic coordinates (latitude and longitude, altitude) can be converted into a unified three-dimensional space coordinate system. By applying geographic information system (GIS) tools and three-dimensional transformation matrix calculation, these flight parameters are converted into corresponding aerial survey coordinate matrices and rotation matrices. During the matrix conversion process, a conversion method between the Earth-centered Earth-fixed coordinate system (ECEF) and a local coordinate system (such as the ENU coordinate system) is used to ensure that the positioning information of each image has spatial consistency, and finally the aerial survey coordinate matrix and aerial survey rotation matrix corresponding to each UAV aerial survey terrain image in three-dimensional space are obtained.

[0044] Step S24: performing aerial survey 3D panoramic matching and stitching on the corresponding UAV aerial survey terrain standard image at each aerial survey time point based on the aerial survey coordinate matrix and the aerial survey rotation matrix, and performing gap optimization processing on the stitched 3D panoramic terrain space to generate the UAV aerial survey 3D panoramic terrain space; In an embodiment of the present invention, panoramic image stitching is performed using an aerial survey coordinate matrix and a rotation matrix. First, each image is positioned according to its three-dimensional spatial coordinates to ensure that the images are stitched at the correct angle and position. The image stitching adopts a feature matching-based stitching method, and the ORB or SIFT algorithm is used to extract key feature points in the image. The RANSAC algorithm is then used to eliminate incorrect matches to ensure the accuracy of the stitching process. The stitched image needs to be gap-optimized, and a global optimization algorithm such as the least squares method (Levenberg-Marquardt) is used for gap filling and image registration to ensure that the stitched image has no significant seams or errors, ensuring a smooth transition of the stitching effect. During the image optimization process, pixel-level image fusion is performed to ensure seamless connection. After completing the stitching of the three-dimensional panoramic image, a complete three-dimensional terrain space of the drone aerial survey is finally generated.

[0045] Step S25: performing terrain texture feature point detection on the UAV aerial survey 3D panoramic terrain space to obtain a UAV aerial survey 3D terrain texture feature point set.

[0046] In an embodiment of the present invention, texture feature point detection is performed on the spliced three-dimensional panoramic terrain space. First, texture feature points in the image are extracted through an image processing algorithm. Common methods include texture recognition based on the gray-level co-occurrence matrix (GLCM) or local binary pattern (LBP) of the image. The texture change area in the image is used to detect representative texture feature points. These feature points can effectively characterize the detailed changes in the terrain. After the detection is completed, the feature points are clustered and screened, and important texture feature points are retained for subsequent terrain analysis. In specific implementation, a three-dimensional reconstruction tool (such as COLMAP or VisualSFM) and a texture analysis algorithm can be combined to ensure that the detected texture feature points have an accurate geographical location in the three-dimensional space and can be consistently matched in multiple images, and finally a set of texture feature points of the three-dimensional terrain surveyed by the UAV is obtained.

[0047] Furthermore, step S3 includes the following steps: Step S31: performing point cloud and texture depth analysis on the standard 3D point cloud data of the UAV aerial survey terrain and the texture feature point set of the UAV aerial survey terrain, using a voxelization method to divide the corresponding standard 3D point cloud data of the UAV aerial survey terrain into uniform small cubes, and calculating the corresponding point density and distribution direction within each small cube. At the same time, using the Laplace operator to analyze the energy distribution of the texture feature points at different frequencies, the UAV aerial survey 3D point cloud and texture depth feature set are obtained; In an embodiment of the present invention, during the processing of the standard three-dimensional point cloud data and the three-dimensional terrain texture feature point set of the UAV aerial survey terrain, the point cloud data is first voxelized. This operation divides the point cloud data into uniform small cube units, determines the point density inside each cube, and traverses each small cube to count the number of points contained therein, and then calculates the point density value. At the same time, the Laplace operator is used to perform frequency domain analysis on the texture feature points. The Laplace operator is an important tool for calculating the surface features of images or point clouds. It can be used to identify texture features within different frequency ranges. Based on frequency domain analysis, the energy distribution of texture feature points at each frequency is calculated. This analysis result helps to identify texture details and then extract texture feature information. Through deep analysis of point clouds and textures, the point density and distribution direction characteristics of the three-dimensional point cloud data of the UAV aerial survey terrain and the frequency feature set of the terrain texture are generated, and finally the three-dimensional point cloud and texture depth feature set of the UAV aerial survey are obtained.

[0048] Step S32: performing geographic coordinate conversion on the standard three-dimensional point cloud data of the UAV aerial survey terrain and the three-dimensional terrain texture feature point set of the UAV aerial survey based on the UAV aerial survey three-dimensional point cloud and texture depth feature set and in combination with the geophysical model and the UAV flight attitude model, so as to convert the point cloud data and the terrain texture feature point set from the UAV coordinate system to the same geographic coordinate system, thereby obtaining the UAV aerial survey point cloud data and the UAV aerial survey terrain texture feature point set in the same three-dimensional geographic coordinate system; In an embodiment of the present invention, geographic coordinate conversion is performed based on the UAV aerial survey three-dimensional point cloud and texture depth feature set, and through the geophysical model and the UAV flight attitude model. The geophysical model provides the physical coordinate conversion relationship of the points on the earth's surface, and the flight attitude model is used to combine the flight state of the UAV with the attitude angle to ensure the accuracy of the conversion. According to the sensor data of the UAV, using the known sensor installation position and direction, combined with the aerial survey data and flight records, the point cloud data and texture feature point set in the UAV coordinate system are converted into a unified geographic coordinate system. In this process, the required transformation matrix and calculation formula are derived based on the establishment of the flight attitude data and the physical model to ensure that the UAV's point cloud data and texture feature point set can be aligned in the three-dimensional geographic coordinate system, thereby generating a data set in a unified coordinate system, and finally obtaining the UAV aerial survey point cloud data and the UAV aerial survey terrain texture feature point set in the same three-dimensional geographic coordinate system.

[0049] Step S33: Perform three-dimensional Poisson reconstruction on the terrain surveyed by the UAV based on the UAV aerial survey point cloud data and the UAV aerial survey terrain texture feature point set in the same three-dimensional geographic coordinate system to generate a three-dimensional surface model of the UAV aerial survey terrain.

[0050] In an embodiment of the present invention, three-dimensional Poisson reconstruction of the terrain is performed by utilizing data in the same three-dimensional geographic coordinate system. The Poisson reconstruction algorithm is a method for generating a three-dimensional surface model based on point cloud data. During implementation, key terrain feature points are first extracted from the point cloud data, and geometric reconstruction is performed based on these points. In the Poisson reconstruction algorithm, the normal vector of the point cloud data is first calculated. This process requires calculation based on the local characteristics of the point cloud to ensure the accuracy of the normal vector. Then, the three-dimensional surface represented by the point cloud data is reconstructed through an iterative process of solving the Poisson equation. At the same time, a set of texture feature points is used for surface texture mapping, so that the surface not only has a geometric shape, but also has realistic surface texture details, and finally a complete three-dimensional surface model of the drone aerial survey terrain is generated.

[0051] Furthermore, step S33 includes the following steps: Perform geometric connection analysis on the UAV aerial survey point cloud data and the UAV aerial survey terrain texture feature point set in the same three-dimensional geographic coordinate system to obtain the topological connection relationship between each UAV aerial survey point cloud and the geometric connection relationship between each UAV aerial survey texture feature point; In an embodiment of the present invention, the geometric relationship between each point in the three-dimensional point cloud data and its neighboring points is analyzed to construct a topological structure between the point cloud data, and the neighborhood of each point is calculated by using the K-nearest neighbor algorithm (KNN) or other geometric analysis methods, and then the connection relationship between the points is determined. This process requires adopting an appropriate neighborhood range for different point cloud densities to ensure that the connection relationship between different terrain areas can accurately reflect the actual situation. A similar geometric relationship analysis is also performed on the texture feature points to calculate the connection pattern between the texture feature point sets, and finally the topological connection relationship between each UAV aerial survey point cloud and the geometric connection relationship between each UAV aerial survey texture feature point are obtained.

[0052] Preferably, a three-dimensional Poisson reconstruction is performed on the terrain aerial survey corresponding to the drone based on the topological connection relationship between each drone aerial survey point cloud and the geometric connection relationship between each drone aerial survey texture feature point, so as to quickly construct the approximate outline corresponding to the drone aerial survey terrain according to the density distribution of each drone aerial survey point cloud, and use each drone aerial survey texture feature point to perform texture mapping and detail supplement to generate a three-dimensional surface model of the drone aerial survey terrain.

[0053] In an embodiment of the present invention, based on the previously obtained topological connection relationship and geometric connection relationship, the terrain of the UAV aerial survey is three-dimensionally reconstructed by Poisson in this step. First, the density distribution in the point cloud data is used in combination with the topological connection relationship to perform a preliminary rough reconstruction of the point cloud, and the approximate terrain contour is obtained by the Poisson equation. Then, the geometric connection relationship of the texture feature point set is used to accurately map the texture. When performing texture mapping, the spatial distribution information of the texture feature points is used to accurately fit the texture to the three-dimensional surface model, and the texture details are supplemented and repaired according to the geometric connection relationship to eliminate texture inconsistencies or dislocations. After multiple rounds of optimization and refinement, the final three-dimensional surface model is obtained. The model not only contains the precise terrain contour, but also has a realistic texture effect, and finally a three-dimensional surface model of the UAV aerial survey terrain is generated.

[0054] Furthermore, step S4 includes the following steps: Step S41: performing ground object type identification on the terrain aerial survey corresponding to the UAV to obtain the UAV aerial survey ground object type, including building, road and vegetation ground object types; In an embodiment of the present invention, ground data is collected by an unmanned aerial vehicle equipped with sensors such as a high-resolution camera and a lidar to obtain image data of the terrain surface and objects above it. The type of ground objects is identified using an image processing algorithm. Specifically, different object types in the image are identified through a pixel classification method. During the ground object classification process, a convolutional neural network (CNN) is used to train the images obtained by the unmanned aerial vehicle, and a large amount of labeled data is used to distinguish between ground object types such as buildings, roads, and vegetation. In this process, image features and remote sensing data are combined for fusion to further improve the classification accuracy. For each frame of the image, the classifier outputs the corresponding ground object type. Building recognition can rely on image texture features, and vegetation areas can be identified based on color, texture, and shape features. The recognition results of different objects are further refined through a multi-scale analysis method, and finally the type of ground objects surveyed by the unmanned aerial vehicle is obtained, which includes building, road, and vegetation ground object types.

[0055] Step S42: performing geographic target detection on the terrain aerial survey corresponding to the UAV based on the type of the UAV aerial survey ground object to obtain the corresponding terrain aerial survey geographic target; In an embodiment of the present invention, the feature type data obtained through previous analysis is compared with the drone aerial survey imagery, and further geographic target extraction is performed in combination with the feature type. In specific operations, spatial information related to target features in the aerial survey data is extracted through geographic information system (GIS) technology, and geographic targets such as building distribution, road vertical and horizontal distribution, bridges, utility poles, and flower beds are identified. To improve the accuracy of recognition, a deep learning model, such as a regional convolutional neural network (R-CNN) or a YOLO (You Only Look Once) algorithm, can be used to perform target detection and positioning on the aerial survey imagery. By combining the spatial distribution information of the aerial survey imagery and utilizing algorithms such as edge detection and texture analysis, the geometric features of the target objects, such as the outline of the building and the direction of the road, are extracted from the large-scale regional imagery. Ultimately, geographic spatial information of targets such as building distribution, road network, bridges, and utility poles is formed, ensuring the precise position and category of each target in the two-dimensional image, and ultimately obtaining the terrain aerial survey geographic targets.

[0056] Step S43: Positioning and labeling the three-dimensional surface model of the terrain surveyed by the UAV based on the terrain aerial survey geographic target to generate a three-dimensional model of the terrain surveyed by the UAV geographic target labeling.

[0057] In an embodiment of the present invention, after the geographic target detection is completed, the annotation stage of the terrain three-dimensional surface model is entered. This step first uses the acquired three-dimensional lidar data to construct a three-dimensional terrain surface model for drone aerial survey. Next, the target position information obtained through geographic target detection is used to accurately locate each target according to its spatial coordinates and annotate it into the three-dimensional model. The specific method is to connect the geographic coordinates of the two-dimensional target detection with the three-dimensional point cloud data, and use spatial interpolation and reconstruction technology to calibrate the target information into the three-dimensional model, thereby generating the specific positions of geographic targets such as buildings, roads, and telephone poles in the three-dimensional model. The annotation process relies on a high-precision geographic coordinate system and three-dimensional modeling tools, and is proofread one by one in the three-dimensional modeling software to ensure that each target is highly consistent with the terrain model. After the annotation is completed, it is visualized through virtual reality (VR) or augmented reality (AR) equipment, so that each geographic target is clearly displayed in three-dimensional space, and finally a three-dimensional model of the geographic target annotation of the drone aerial survey is generated.

[0058] Furthermore, the terrain aerial survey geographical targets in step S42 include building distribution, road vertical and horizontal distribution, bridges, electric poles and flower garden geographical targets.

[0059] Furthermore, the present invention also provides a UAV 3D aerial survey data processing system for executing the UAV 3D aerial survey data processing method described above. The UAV 3D aerial survey data processing system includes: The aerial survey terrain multi-source data acquisition module is used to equip the UAV with an optical camera, a lidar, and a global navigation satellite system receiver, and use the optical camera, lidar, and global navigation satellite system receiver to collect and pre-process the terrain multi-source data corresponding to the UAV aerial survey, thereby obtaining the UAV aerial survey terrain standard image, the UAV aerial survey terrain 3D point cloud standard data, and the UAV aerial survey flight positioning trajectory; The aerial survey terrain texture feature detection module is used to perform aerial survey 3D panoramic matching and stitching on the corresponding drone aerial survey terrain standard image at each aerial survey time point based on the drone aerial survey flight positioning trajectory to generate the drone aerial survey 3D panoramic terrain space; the module also performs terrain texture feature point detection on the drone aerial survey 3D panoramic terrain space to obtain the drone aerial survey 3D terrain texture feature point set; The terrain 3D Poisson reconstruction module is used to perform 3D Poisson reconstruction on the terrain corresponding to the drone based on the standard 3D point cloud data of the drone aerial survey terrain and the 3D terrain texture feature point set of the drone aerial survey terrain, thereby generating a 3D surface model of the drone aerial survey terrain; The aerial survey geographic target positioning and labeling module is used to obtain the corresponding terrain aerial survey geographic target through the terrain aerial survey corresponding to the UAV, and to position and label the UAV aerial survey terrain three-dimensional surface model based on the terrain aerial survey geographic target, thereby generating a UAV aerial survey geographic target labeled three-dimensional model.

[0060] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0061] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A method for processing UAV three-dimensional aerial survey data, characterized in that: The following steps are involved: Step S1: An optical camera, a laser radar, and a global navigation satellite system receiver are mounted on a UAV, and the optical camera, the laser radar, and the global navigation satellite system receiver are used to collect and preprocess terrain multi-source data for the terrain aerial survey corresponding to the UAV, thereby obtaining a standard image of the UAV aerial survey terrain, standard three-dimensional point cloud data of the UAV aerial survey terrain, and a flight positioning trajectory of the UAV aerial survey; Step S2: performing aerial survey 3D panoramic matching and stitching on the corresponding drone aerial survey terrain standard image at each aerial survey time point based on the drone aerial survey flight positioning trajectory to generate a drone aerial survey 3D panoramic terrain space; performing terrain texture feature point detection on the drone aerial survey 3D panoramic terrain space to obtain a drone aerial survey 3D terrain texture feature point set; Step S3: performing a three-dimensional Poisson reconstruction on the terrain surveyed by the UAV based on the three-dimensional point cloud standard data of the UAV aerial survey terrain and the three-dimensional terrain texture feature point set of the UAV aerial survey terrain to generate a three-dimensional surface model of the UAV aerial survey terrain; Step S4: obtaining the corresponding terrain aerial survey geographic target through the terrain aerial survey corresponding to the UAV, and positioning and labeling the UAV aerial survey terrain three-dimensional surface model based on the terrain aerial survey geographic target to generate the UAV aerial survey geographic target labeled three-dimensional model.

2. The method for processing UAV 3D aerial survey data according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: by equipping the drone with an optical camera, a laser radar, and a global navigation satellite system receiver; Step S12: using an optical camera, a laser radar, and a global navigation satellite system receiver to collect terrain multi-source data for the terrain aerial survey corresponding to the UAV, so as to extract a visible light image of the surface texture corresponding to the UAV aerial survey through the optical camera, obtain corresponding terrain three-dimensional point cloud data through the laser radar, and obtain corresponding real-time flight position point set of the UAV aerial survey through the global navigation satellite system receiver; Step S13: performing image preprocessing on the visible light image of the surface texture corresponding to the UAV aerial survey, using an adaptive filtering method to remove pixel noise corresponding to the visible light image of the surface texture, and using image enhancement technology to improve the contrast and clarity of the image, thereby obtaining a standard image of the terrain measured by the UAV aerial survey; Step S14: performing outlier filtering on the standard three-dimensional point cloud data of the drone aerial survey terrain to obtain the standard three-dimensional point cloud data of the drone aerial survey terrain; Step S15: Determine the flight trajectory of the terrain aerial survey corresponding to the UAV based on the real-time flight position point set of the UAV aerial survey to obtain the UAV aerial survey flight positioning trajectory.

3. The method for processing 3D aerial survey data of a UAV according to claim 2, characterized in that: Step S15 includes the following steps: Step S151: determining the time interval for each flight position point in the real-time flight position point set of the UAV aerial survey, and obtaining the aerial survey recording time interval corresponding to each flight position point; Step S152: performing adjacent position space dimension analysis on each flight position point in the UAV aerial survey real-time flight position point set to obtain the aerial survey space distance and aerial survey space direction between adjacent flight position points; Step S153: performing a position point spatial distribution analysis on the aerial survey spatial distance and aerial survey spatial direction between adjacent flight position points based on the aerial survey recording time interval corresponding to each flight position point, so as to obtain the spatial distribution characteristics of the UAV aerial survey position points; Step S154: Obtaining the terrain elevation, slope, and landform type corresponding to the UAV aerial survey, and performing a terrain flight constraint analysis based on the terrain elevation, slope, and landform type corresponding to the UAV aerial survey to generate the terrain flight constraints for the UAV aerial survey, including the restrictions imposed by terrain undulations and obstacle distribution on the UAV aerial survey range; Step S155: Determine the flight trajectory of the terrain aerial survey corresponding to the UAV based on the UAV aerial survey terrain flight constraints and in combination with the spatial distribution characteristics of the UAV aerial survey position points to obtain the UAV aerial survey flight positioning trajectory.

4. The method for processing UAV 3D aerial survey data according to claim 3, characterized in that: The spatial distribution characteristics of the UAV aerial survey position points in step S153 specifically include the aerial survey flight curvature and the aerial survey flight angle corresponding to each flight position point.

5. The method for processing UAV 3D aerial survey data according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: performing image edge depth mining on the standard UAV aerial survey terrain image corresponding to each aerial survey time point to obtain edge feature descriptors corresponding to each UAV aerial survey terrain image, including edge contour, edge color first-order moment, edge color second-order moment, and edge color third-order moment; Step S22: performing three-dimensional space edge matching positioning on the edge feature descriptors corresponding to each UAV aerial survey terrain image based on the UAV aerial survey flight positioning trajectory, and obtaining the UAV matching splicing positioning corresponding to each UAV aerial survey terrain image in three-dimensional space; Step S23: Obtain the corresponding UAV flight latitude, longitude, altitude, and attitude angle by matching and stitching the UAVs corresponding to each UAV aerial survey terrain image in three-dimensional space, and perform three-dimensional space matrix transformation on the UAV flight latitude, longitude, altitude, and attitude angle to generate the corresponding aerial survey coordinate matrix and aerial survey rotation matrix for each UAV aerial survey terrain image in three-dimensional space; Step S24: performing aerial survey 3D panoramic matching and stitching on the corresponding UAV aerial survey terrain standard image at each aerial survey time point based on the aerial survey coordinate matrix and the aerial survey rotation matrix, and performing gap optimization processing on the stitched 3D panoramic terrain space to generate the UAV aerial survey 3D panoramic terrain space; Step S25: performing terrain texture feature point detection on the UAV aerial survey 3D panoramic terrain space to obtain a UAV aerial survey 3D terrain texture feature point set.

6. The method for processing UAV 3D aerial survey data according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: performing point cloud and texture depth analysis on the standard 3D point cloud data of the UAV aerial survey terrain and the texture feature point set of the UAV aerial survey terrain, using a voxelization method to divide the corresponding standard 3D point cloud data of the UAV aerial survey terrain into uniform small cubes, and calculating the corresponding point density and distribution direction within each small cube. At the same time, using the Laplace operator to analyze the energy distribution of the texture feature points at different frequencies, the UAV aerial survey 3D point cloud and texture depth feature set are obtained; Step S32: performing geographic coordinate conversion on the standard three-dimensional point cloud data of the UAV aerial survey terrain and the three-dimensional terrain texture feature point set of the UAV aerial survey based on the UAV aerial survey three-dimensional point cloud and texture depth feature set and in combination with the geophysical model and the UAV flight attitude model, so as to convert the point cloud data and the terrain texture feature point set from the UAV coordinate system to the same geographic coordinate system, thereby obtaining the UAV aerial survey point cloud data and the UAV aerial survey terrain texture feature point set in the same three-dimensional geographic coordinate system; Step S33: Perform three-dimensional Poisson reconstruction on the terrain surveyed by the UAV based on the UAV aerial survey point cloud data and the UAV aerial survey terrain texture feature point set in the same three-dimensional geographic coordinate system to generate a three-dimensional surface model of the UAV aerial survey terrain.

7. The method for processing UAV 3D aerial survey data according to claim 6, characterized in that: Step S33 includes the following steps: Perform geometric connection analysis on the UAV aerial survey point cloud data and the UAV aerial survey terrain texture feature point set in the same three-dimensional geographic coordinate system to obtain the topological connection relationship between each UAV aerial survey point cloud and the geometric connection relationship between each UAV aerial survey texture feature point; Based on the topological connection relationship between each UAV aerial survey point cloud and the geometric connection relationship between each UAV aerial survey texture feature point, a three-dimensional Poisson reconstruction is performed on the terrain aerial survey corresponding to the UAV, so as to quickly construct the approximate outline corresponding to the UAV aerial survey terrain according to the density distribution of each UAV aerial survey point cloud, and use each UAV aerial survey texture feature point for texture mapping and detail supplementation to generate a three-dimensional surface model of the UAV aerial survey terrain.

8. The method for processing UAV 3D aerial survey data according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: performing ground object type identification on the terrain aerial survey corresponding to the UAV to obtain the UAV aerial survey ground object type, including building, road and vegetation ground object types; Step S42: performing geographic target detection on the terrain aerial survey corresponding to the UAV based on the type of the UAV aerial survey ground object to obtain the corresponding terrain aerial survey geographic target; Step S43: Positioning and labeling the three-dimensional surface model of the terrain surveyed by the UAV based on the terrain aerial survey geographic target to generate a three-dimensional model of the terrain surveyed by the UAV geographic target labeling.

9. The method for processing UAV 3D aerial survey data according to claim 8, characterized in that: The terrain aerial survey geographical targets described in step S42 include building distribution, road vertical and horizontal distribution, bridges, electric poles and flower garden geographical targets.

10. A UAV three-dimensional aerial survey data processing system, characterized in that: For executing the UAV 3D aerial survey data processing method according to claim 1, the UAV 3D aerial survey data processing system comprises: The aerial survey terrain multi-source data acquisition module is used to equip the UAV with an optical camera, a lidar, and a global navigation satellite system receiver, and use the optical camera, lidar, and global navigation satellite system receiver to collect and pre-process the terrain multi-source data corresponding to the UAV aerial survey, thereby obtaining the UAV aerial survey terrain standard image, the UAV aerial survey terrain 3D point cloud standard data, and the UAV aerial survey flight positioning trajectory; The aerial survey terrain texture feature detection module is used to perform aerial survey 3D panoramic matching and stitching on the corresponding drone aerial survey terrain standard image at each aerial survey time point based on the drone aerial survey flight positioning trajectory to generate the drone aerial survey 3D panoramic terrain space; the module also performs terrain texture feature point detection on the drone aerial survey 3D panoramic terrain space to obtain the drone aerial survey 3D terrain texture feature point set; The terrain 3D Poisson reconstruction module is used to perform 3D Poisson reconstruction on the terrain corresponding to the drone based on the standard 3D point cloud data of the drone aerial survey terrain and the 3D terrain texture feature point set of the drone aerial survey terrain, thereby generating a 3D surface model of the drone aerial survey terrain; The aerial survey geographic target positioning and labeling module is used to obtain the corresponding terrain aerial survey geographic target through the terrain aerial survey corresponding to the UAV, and to position and label the UAV aerial survey terrain three-dimensional surface model based on the terrain aerial survey geographic target, thereby generating a UAV aerial survey geographic target labeled three-dimensional model.

Images