Construction method and device of three-dimensional visual landscape garden and electronic equipment

Abstract

The invention discloses a three-dimensional visual landscape garden construction method and device and electronic equipment, and relates to the field of garden planning. In the method, a plurality of garden images shot by an unmanned aerial vehicle are obtained, and the garden images are shot from a plurality of preset angles when the unmanned aerial vehicle flies according to a preset route; determining feature points of the garden image; obtaining a shooting time point of each garden image and pos data of the unmanned aerial vehicle at the shooting time point; according to the pos data and the feature points, performing feature point matching on each garden image by adopting aerial triangulation operation to obtain three-dimensional point cloud data of each feature point; constructing a three-dimensional point cloud model corresponding to the landscape garden; and mapping the texture information to the three-dimensional point cloud model to obtain a three-dimensional live-action model of the landscape garden. According to the technical scheme, the unmanned aerial vehicle collects the garden data according to the preset route, large-range and multi-view garden data can be rapidly and systematically collected, and the working efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of garden planning, and specifically to a method, device and electronic equipment for constructing a three-dimensional visualized landscape. Background Art

[0002] With the development of society and the enhancement of environmental protection awareness, landscape planning has become increasingly prominent in urban planning and environmental beautification. Landscape architecture not only provides a place for people to relax and have fun, but also plays a vital role in improving the urban ecological environment and enhancing the city's image. Through well-designed garden landscapes, the quality of life of urban residents can be effectively improved. At the same time, it also has an important impact on enhancing the city's tourism appeal and economic development.

[0003] At present, the 3D visualization technology commonly used in landscape planning and design still has many shortcomings, mainly reflected in the work efficiency of model construction. In the process of landscape planning and design, the original information of the site is the premise for guiding planning and design. At present, landscape designers measure the internal environment of landscape gardens by carrying professional instruments, which consumes a lot of manpower and material resources and has a long work cycle. This technical method will lead to low work efficiency in the early stage of design.

[0004] Therefore, there is an urgent need for a method, device and electronic equipment for constructing three-dimensional visualized landscape architecture. Summary of the invention

[0005] The present application provides a method, device and electronic equipment for constructing a three-dimensional visualized landscape garden. By collecting garden data by using a drone according to a preset route, large-scale, multi-perspective garden data can be collected quickly and systematically, thereby improving work efficiency.

[0006] In a first aspect of the present application, a method for constructing a three-dimensional visualized landscape garden is provided, the method comprising: obtaining multiple garden images of a landscape garden taken by a drone, the garden images being taken from multiple preset angles when the drone flies along a preset route; determining feature points of the garden images, the feature points comprising control points and connection points, the control points being preset coordinate points, and the connection points being commonly identifiable coordinate points in the multiple garden images; obtaining the shooting time point of each of the garden images, and the POS data of the drone at the shooting time point, the POS data comprising the longitude, latitude and altitude of the drone; performing feature point matching on each of the garden images using aerial triangulation according to the POS data and the feature points, to obtain three-dimensional point cloud data of each of the feature points; constructing a three-dimensional point cloud model corresponding to the landscape garden according to the three-dimensional point cloud data; extracting texture information of the garden image according to the garden image; mapping the texture information to the three-dimensional point cloud model, to obtain a three-dimensional real-scene model of the landscape garden.

[0007] By adopting the above technical solution, the image information of the garden can be fully collected by using drones to obtain the garden images from multiple preset angles, and a rich data source can be obtained. By extracting the feature points of the garden images, including the preset control points and the connection points shared by multiple images, the geometric relationship between the garden images is established, laying the foundation for the subsequent three-dimensional reconstruction. At the same time, the shooting time point of each garden image and the corresponding drone position and posture information, namely POS data, are recorded, and the spatial position and direction of the garden image can be determined. Using POS data and feature points, feature matching between garden images is achieved through aerial triangulation (aerial triangulation), and the three-dimensional coordinates of the feature points are calculated to generate accurate three-dimensional point cloud data. On the basis of the point cloud, a three-dimensional point cloud model of the garden scene is further constructed, and then the image texture is mapped to the surface of the three-dimensional point cloud model, and finally a realistic three-dimensional real scene model is obtained. By constructing a high-precision three-dimensional real scene model, various elements of the landscape garden can be digitally recorded and displayed, and the spatial form and landscape layout of the landscape garden are fully reproduced; at the same time, the three-dimensional real scene model provides a scientific analysis tool and data support for the planning and design of the landscape garden. This method utilizes the flexible and efficient data collection capabilities of drones, combined with photogrammetry and computer vision technology, to achieve high-precision three-dimensional visualization reconstruction of landscape architecture, providing intuitive and interactive digital scenes for garden planning, design, and management, thereby improving work efficiency.

[0008] Optionally, after constructing the three-dimensional point cloud model corresponding to the landscape garden based on the three-dimensional point cloud data, the method also includes: determining whether there is a stratification phenomenon of feature points or an intersection phenomenon of feature points in the three-dimensional point cloud model; if it is determined whether there is a stratification phenomenon of feature points or an intersection phenomenon of feature points in the three-dimensional point cloud model, controlling the drone to re-acquire the garden image.

[0009] By adopting the above technical solution, by analyzing the spatial distribution characteristics of the point cloud, it is determined whether there is a stratification phenomenon or a cross phenomenon of the feature points. Both phenomena will affect the integrity and accuracy of the 3D point cloud model. By detecting stratification and crossover, the quality and reliability of the 3D point cloud model can be evaluated. When the above problems are found, the method can adaptively control the drone to re-collect data and correct and improve the 3D point cloud model.

[0010] Optionally, the determining whether there is a stratification phenomenon of feature points in the three-dimensional point cloud model specifically includes: performing stratification processing on the three-dimensional point cloud model, dividing the three-dimensional point cloud data into multiple layers according to height information; performing cluster analysis on the point cloud data in each layer to obtain multiple point cloud clusters; determining multiple target point clouds included in the target point cloud cluster, the target point cloud cluster is any one of the multiple point cloud clusters; calculating the height difference between a first target point cloud and a second target point cloud, the first target point cloud is any one of the multiple target point clouds, and the second target point cloud is any one of the multiple target point clouds except the first target point cloud; if it is determined that the height difference is greater than or equal to a preset height difference, determining that the first target point cloud and the second target point cloud are layered point clouds; determining whether the number of point clouds in the layered point clouds is greater than or equal to a preset number threshold; if it is determined that the number of point clouds in the layered point clouds is greater than or equal to a preset number threshold, determining that the stratification phenomenon exists in the three-dimensional point cloud model.

[0011] By adopting the above technical solution, the point cloud is layered according to its height information, and the three-dimensional point cloud model is divided into multiple layers, highlighting the distribution differences in the vertical direction. Then, cluster analysis is performed within each layer, and the point clouds with spatial proximity and similar attributes are clustered into clusters to form multiple point cloud clusters, which characterize the aggregation pattern in the horizontal direction. Any point cloud cluster is selected as the target cluster, and the point cloud contained therein is extracted as the target point cloud. By calculating the height difference between any two different target point clouds and setting a preset height difference, potential layered point clouds can be found. By counting the number of layered point clouds and setting a preset number threshold, it is determined whether the entire three-dimensional point cloud model has a layered phenomenon based on the situation of exceeding the preset number threshold. This method makes full use of the elevation and neighborhood information of the point cloud, effectively analyzes the spatial differentiation characteristics of the point cloud through layering and clustering, and can quantitatively evaluate the layered phenomenon.

[0012] Optionally, determining whether there is an intersection phenomenon of feature points in the three-dimensional point cloud model specifically includes: calculating the spatial distance between the first target point cloud and the second target point cloud; if it is determined that the spatial distance is less than a preset distance threshold, determining that the intersection phenomenon exists in the three-dimensional point cloud model.

[0013] By adopting the above technical solution, by calculating the spatial distance between any two target point clouds and setting a preset distance threshold, abnormally close point clouds, that is, potential intersection point clouds, can be found. When the spatial distance is less than the preset distance threshold, it means that the point clouds of different landscape elements are interspersed with each other in the three-dimensional space, and an erroneous intersection phenomenon occurs. The accuracy of the three-dimensional point cloud model can be quantitatively evaluated by measuring the degree of intersection using spatial distance. This method cleverly utilizes the spatial topological relationship of the point cloud and realizes the rapid detection of intersection phenomena through distance calculation. This method can comprehensively evaluate the quality of the three-dimensional point cloud model, provide accurate feedback for the data supplement of the drone, and improve the reliability of the three-dimensional visualization of landscape gardens. At the same time, through intelligent quality control, the workload of manual inspection and correction is reduced, and the automation level of garden digitization is improved.

[0014] Optionally, based on the POS data and the feature points, aerial triangulation is used to match the feature points of each of the garden images to obtain three-dimensional point cloud data of each of the feature points, specifically including: matching the feature points of multiple garden images to obtain the correspondence between the same feature point in the multiple garden images; constructing the correspondence between the image coordinate system and the ground coordinate system using the control point coordinates of the control point on the ground and the feature point coordinates in the garden image; obtaining the three-dimensional coordinates of each of the feature points in the three-dimensional space based on the correspondence; and densely matching the three-dimensional coordinates to generate the three-dimensional point cloud data.

[0015] By adopting the above technical solution, the same feature points are identified in different garden images through feature point matching, and the corresponding relationship between garden images is established. This step achieves accurate registration between images by measuring the similarity of feature points. Then, the coordinates of the control points on the ground and in the image are used to build the corresponding relationship between the garden image coordinate system and the ground coordinate system. According to the established corresponding relationship, the coordinates of each feature point in three-dimensional space are calculated to obtain a discrete three-dimensional point cloud. The image information and spatial information of the feature points are fully utilized, and through rigorous mathematical models and coordinate transformation, high-precision reconstruction of two-dimensional images into three-dimensional point clouds is achieved.

[0016] Optionally, extracting texture information of the garden image based on the garden image specifically includes: extracting pixel blocks corresponding to each feature point from the garden image; performing color correction on the pixel blocks to obtain corrected pixel blocks, and using the corrected pixel blocks as the texture information of the garden image.

[0017] By adopting the above technical solution, the pixel blocks corresponding to each feature point are extracted from each garden image. The feature point not only contains the location information, but the pixels around it also reflect the local texture features. Extracting pixel blocks can obtain rich texture samples. The extracted pixel blocks are color corrected. Due to the differences in lighting, exposure and other conditions between images, the color of the pixel blocks may be biased. Through color correction, these deviations can be eliminated, making the color of the pixel blocks more consistent and natural. The corrected pixel blocks serve as the texture information of the garden image, which can reflect the true color and material properties of the landscape. This method makes full use of the texture information contained in the garden image, and obtains high-quality texture samples through feature point-driven pixel block extraction.

[0018] Optionally, constructing a three-dimensional point cloud model corresponding to the landscape garden based on the three-dimensional point cloud data specifically includes: filtering and down-sampling the three-dimensional point cloud data to obtain target three-dimensional point cloud data; generating a triangular mesh surface based on the target three-dimensional point cloud data to construct the three-dimensional point cloud model.

[0019] By adopting the above technical solutions, through point cloud preprocessing and triangulation, efficient conversion from discrete point cloud to continuous surface is achieved, and a fine 3D point cloud model of landscape architecture is constructed. Filtering and downsampling ensure the data quality and computational efficiency of the 3D point cloud model, which is conducive to the rapid modeling of large-scale landscape architecture. The triangulated mesh surface provides a compact and standardized 3D representation, which is convenient for visualization rendering and spatial analysis operations.

[0020] In a second aspect of the present application, a device for constructing a three-dimensional visualized landscape garden is provided, the device comprising: an acquisition module and a processing module, the acquisition module being used to acquire multiple garden images of a landscape garden taken by a drone, the garden images being taken from multiple preset angles when the drone flies along a preset route; the processing module being used to determine feature points of the garden images, the feature points comprising control points and connection points, the control points being preset coordinate points, the connection points being coordinate points that are commonly identifiable in the multiple garden images; the acquisition module being further used to acquire the shooting time point of each of the garden images, and at the shooting time point, the The POS data of the drone includes the longitude, latitude and altitude of the drone; the processing module is further used to match the feature points of each garden image by aerial triangulation according to the POS data and the feature points, so as to obtain the three-dimensional point cloud data of each feature point; the processing module is further used to construct a three-dimensional point cloud model corresponding to the landscape garden according to the three-dimensional point cloud data; the processing module is further used to extract the texture information of the garden image according to the garden image; the processing module is further used to map the texture information to the three-dimensional point cloud model to obtain the three-dimensional real scene model of the landscape garden.

[0021] In the third aspect of the present application, an electronic device is provided, including a processor, a memory, a user interface and a network interface, the memory is used to store instructions, the user interface and the network interface are both used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes any one of the methods described above.

[0022] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores instructions, and when the instructions are executed, any of the methods described above is executed.

[0023] In summary, one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. By using drones to obtain images of landscape gardens from multiple preset angles, the image information of the gardens can be fully collected and a rich data source can be obtained. By extracting the feature points of the garden images, including the preset control points and the connection points shared by multiple images, the geometric relationship between the garden images is established, laying the foundation for the subsequent three-dimensional reconstruction. At the same time, the shooting time point of each garden image and the corresponding drone position and posture information, namely POS data, are recorded, and the spatial position and direction of the garden image can be determined. Using POS data and feature points, feature matching between garden images is achieved through aerial triangulation (aerial triangulation), and the three-dimensional coordinates of the feature points are calculated to generate accurate three-dimensional point cloud data. On the basis of the point cloud, a three-dimensional point cloud model of the garden scene is further constructed, and then the image texture is mapped to the surface of the three-dimensional point cloud model, and finally a realistic three-dimensional real scene model is obtained. By constructing a high-precision three-dimensional real scene model, various elements of the landscape garden can be digitally recorded and displayed, and the spatial form and landscape layout of the landscape garden are fully reproduced; at the same time, the three-dimensional real scene model provides scientific analysis tools and data support for the planning and design of the landscape garden. This method utilizes the flexible and efficient data collection capabilities of drones, combined with photogrammetry and computer vision technology, to achieve high-precision three-dimensional visualization reconstruction of landscape architecture, providing intuitive and interactive digital scenes for garden planning, design, and management, thereby improving work efficiency.

[0024] 2. By analyzing the spatial distribution characteristics of the 3D point cloud model, determine whether there is stratification or intersection of feature points. Both phenomena will affect the integrity and accuracy of the 3D point cloud model. By detecting stratification and intersection, the quality and reliability of the 3D point cloud model can be evaluated. When the above problems are found, the method can adaptively control the drone to re-collect data and correct and improve the 3D point cloud model.

[0025] 3. The point cloud is layered according to its height information, and the three-dimensional point cloud model is divided into multiple layers, highlighting the distribution differences in the vertical direction. Then, cluster analysis is performed within each layer, and point clouds with spatial proximity and similar attributes are clustered into clusters to form multiple point cloud clusters, which characterize the aggregation pattern in the horizontal direction. Any point cloud cluster is selected as the target cluster, and the point cloud contained therein is extracted as the target point cloud. By calculating the height difference between any two different target point clouds and setting a preset height difference, potential layered point clouds can be found. By counting the number of layered point clouds and setting a preset number threshold, it is determined whether the entire three-dimensional point cloud model has a layered phenomenon based on the situation of exceeding the preset number threshold. This method makes full use of the elevation and neighborhood information of the point cloud, effectively analyzes the spatial differentiation characteristics of the point cloud through layering and clustering, and can quantitatively evaluate the layered phenomenon. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1It is a flow chart of a method for constructing a three-dimensional visualized landscape garden disclosed in an embodiment of the present application; Figure 2 It is a module schematic diagram of a device for constructing a three-dimensional visualized landscape garden disclosed in an embodiment of the present application; Figure 3 It is a structural schematic diagram of an electronic device disclosed in an embodiment of the present application.

[0027] Explanation of the reference numerals: 201, acquisition module; 202, processing module; 300, electronic device; 301, processor; 302, communication bus; 303, user interface; 304, network interface; 305, memory. DETAILED DESCRIPTION

[0028] In order to enable technicians in this field to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0029] In the description of the embodiments of the present application, words such as "for example" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "for example" or "for example" is intended to present related concepts in a specific way.

[0030] In the description of the embodiments of the present application, the meaning of the term "multiple" refers to two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "include", "comprise", "have" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0031] This application provides a method for constructing a three-dimensional visual landscape. Figure 1 , Figure 1 1 is a flow chart of a method for constructing a three-dimensional visualized landscape architecture provided in an embodiment of the present application. The method is applied to a server and includes steps S101 to S107, which are as follows: Step S101: obtaining a plurality of garden images of a landscape garden taken by a drone, wherein the garden images are taken from a plurality of preset angles when the drone flies along a preset route.

[0032] In step S101, the server plans a preset route for the drone through dedicated software (such as Altizure). According to the area of ​​the landscape garden, the complexity of the terrain, and the required image resolution, the route parameters are designed, including the flight altitude, speed, heading overlap, and lateral overlap; the server controls the drone to fly according to the preset route and shoot the garden image from multiple preset angles. In the embodiment of the present application, preferably, the multiple preset angles are five preset angles, one perpendicular to the ground and four 45-degree angles inclined to the ground; when planning the preset route, it is also necessary to ensure that the heading overlap rate is greater than or equal to 75%, and the lateral overlap rate is greater than or equal to 65%. Vertical angle shooting can obtain orthophotos, which is convenient for extracting planar information of landscape gardens, such as roads and water systems. Shooting at an oblique angle can obtain side information of the garden, which is helpful to capture facade features such as trees and buildings. By setting the angle-free shooting angle, the three-dimensional information of the landscape garden can be fully recorded. For example, the server can instruct the drone to shoot 1 group of images at a vertical angle, shoot 4 groups of oblique images at 45-degree angles in the east, south, west, and north, and obtain a total of 5 groups of multi-view garden image data. After the drone completes the shooting mission, it sends the image data to the server in real time through a wireless transmission module (such as Wi-Fi, 4G), or stores it in an onboard storage card and transmits it to the server after the flight.

[0033] Step S102: determining feature points of the garden image, wherein the feature points include control points and connection points, wherein the control points are preset coordinate points, and the connection points are commonly identifiable coordinate points in a plurality of garden images.

[0034] In step S102, the server first determines the coordinate system of the garden image, and the server selects the WGS84 coordinate system. WGS84 is a geodetic coordinate system, which is directly compatible with GPS positioning results, and is convenient for subsequent spatial data fusion and application. Control points are markers pre-arranged at the garden site, with known high-precision three-dimensional coordinates. The control points are selected according to the image thorn points that are close to the center of the route, can map a large range and have obvious ground information as control points. On the selected control point image, the server marks at least 3 control points to form a valid control point set. Each control point should appear on at least 2 or more garden images taken at different preset angles. For example, in a landscape garden 3D modeling project, the server determined 10 garden images located at the center of the route from 500 images as control point images, and a total of 5 control points were marked on these 10 control point images. Each control point is reflected on 3-4 control point images at different angles, and each control point appears on 6 images on average.

[0035] Connection points are the same-name points that can be clearly identified on different garden images, and are used to encrypt the geometric relationship between garden images. Unlike control points, connection points do not have pre-determined three-dimensional coordinates, and their coordinates are obtained through aerial triangulation calculation. The server uses feature extraction algorithms (such as SIFT) to automatically detect significant features on garden images, such as corners and edges, and obtain a large number of candidate connection points. Then, feature matching algorithms (such as RANSAC) are used to match the same-name points between different images, eliminate incorrect matching points, and obtain reliable connection points. In order to improve the accuracy and stability of aerial triangulation, the server adds positioning constraints to control points and connection points. The constraints include origin constraints, scale constraints, direction constraints, and plane constraints. The origin constraint refers to fixing a control point as the origin of the coordinates to reduce the translational freedom; the scale constraint refers to fixing the distance between two control points to reduce the scale freedom; the direction constraint refers to fixing the direction of the line connecting the two control points to reduce the rotational freedom; the plane constraint refers to fixing the plane where the three control points are located to reduce the tilting freedom. By adding appropriate constraints, equivalent solutions in the aerial triangulation process can be eliminated, and the uniqueness and reliability of the solution results can be improved.

[0036] For example, in the above-mentioned landscape project, the server selects the control point at the entrance of the park as the origin, the line from the entrance to the visitor center as the direction constraint, and the three control points on the square in front of the visitor center as the plane constraint, thereby effectively constraining the spatial position and posture of the entire park.

[0037] Step S103: obtaining the shooting time point of each garden image and the POS data of the drone at the shooting time point, where the POS data includes the longitude, latitude and altitude of the drone.

[0038] In step S103, when the drone flies according to the preset route, the airborne control system automatically records the shooting time of each garden image, usually accurate to the millisecond level, and the shooting time will be written into the Exif (Exchangeable image file format) header of the garden image file as metadata. The server can obtain the shooting time of each garden image by reading the Exif information of the garden image file. While the drone is shooting the garden image, it also continuously records its own POS data. The POS data is obtained through the airborne GPS / INS (Inertial Navigation System) combined navigation system, including the three-dimensional spatial position (longitude, latitude, altitude) and attitude (roll angle, pitch angle, yaw angle) information of the drone. The POS data can be stored separately from the garden image data to form an independent flight log file, or it can be directly written into the Exif header of the garden image. The server uses different methods to parse and extract the data according to the storage method of the data.

[0039] Step S104: According to the POS data and the feature points, aerial triangulation is used to match the feature points of each garden image to obtain the three-dimensional point cloud data of each feature point.

[0040] In step S104, feature point matching is performed on multiple garden images to obtain the correspondence between the same feature point in the multiple garden images; the correspondence between the image coordinate system and the ground coordinate system is constructed using the control point coordinates of the control point on the ground and the feature point coordinates in the garden image; based on the correspondence, the three-dimensional coordinates of each feature point in the three-dimensional space are obtained; the three-dimensional coordinates are densely matched to generate three-dimensional point cloud data.

[0041] Specifically, the server first matches the feature points of multiple garden images to find the corresponding relationship of the same feature point on different garden images. The server uses the SIFT (Scale-Invariant Feature Transform) feature point matching algorithm to achieve feature point matching. The control points are markers pre-arranged at the garden site with known high-precision three-dimensional coordinates (X, Y, Z). At the same time, these control points are also accurately marked on the garden image to obtain their image coordinates (x, y). The server constructs the conversion relationship from the image coordinate system to the ground coordinate system based on the image coordinates and ground coordinates of the control points. The conversion relationship includes the internal orientation elements (principal point, focal length, distortion parameters) and external orientation elements (position, attitude angle) of the garden image. By combining multiple control points and multiple garden images, the server calculates these parameters through adjustment to achieve accurate conversion of the coordinate system. After having the conversion relationship from the image coordinate system to the ground coordinate system, the server converts the image coordinates of the feature points into ground coordinates to obtain the three-dimensional spatial coordinates (X, Y, Z) of each feature point. The server further performs dense matching based on the three-dimensional coordinates of the feature points. Specifically, the server can use a dense matching algorithm based on confidence propagation to perform pixel matching on multiple garden images to generate high-density three-dimensional point cloud data. Dense matching is to match all pixels of the garden image on the basis of feature point matching to obtain the three-dimensional coordinates of each pixel.

[0042] Step S105: constructing a three-dimensional point cloud model corresponding to the landscape architecture according to the three-dimensional point cloud data.

[0043] In step S105, the three-dimensional point cloud data is filtered and down-sampled to obtain target three-dimensional point cloud data; a triangular mesh surface is generated according to the target three-dimensional point cloud data to construct a three-dimensional point cloud model.

[0044] Specifically, the server first filters the 3D point cloud data to remove noise points and outliers, and improve the quality and accuracy of the point cloud. After filtering, the server downsamples the point cloud data to reduce data redundancy and improve modeling and rendering efficiency. Specifically, the server uses a statistical filtering algorithm to filter the 3D point cloud data, and the server uses a rasterization downsampling algorithm to downsample the 3D point cloud data. After filtering and downsampling, the 3D point cloud data becomes more concise and regular, but it is still composed of discrete points and lacks topological structure and surface information. In order to better express the geometric form of landscape objects, the server generates a triangular mesh surface based on the optimized 3D point cloud data, and constructs a 3D point cloud model of the landscape garden based on the generated triangular mesh surface.

[0045] After step S105, the method further includes: determining whether there is a stratification phenomenon of feature points or an intersection phenomenon of feature points in the three-dimensional point cloud model; if it is determined whether there is a stratification phenomenon of feature points or an intersection phenomenon of feature points in the three-dimensional point cloud model, controlling the drone to reacquire the garden image.

[0046] Specifically, after building the 3D point cloud model, the server will perform quality inspection and acceptance on the 3D point cloud model to ensure the integrity, accuracy and availability of the 3D point cloud model. The server mainly checks whether there are two point cloud model defects: stratification and intersection of feature points.

[0047] In a possible implementation, determining whether there is a stratification phenomenon of feature points in a three-dimensional point cloud model specifically includes: performing stratification processing on the three-dimensional point cloud model, dividing the three-dimensional point cloud data into multiple layers according to height information; performing cluster analysis on the point cloud data in each layer to obtain multiple point cloud clusters; determining multiple target point clouds included in the target point cloud cluster, the target point cloud cluster is any point cloud cluster in the multiple point cloud clusters; calculating the height difference between a first target point cloud and a second target point cloud, the first target point cloud is any target point cloud in the multiple target point clouds, and the second target point cloud is any target point cloud in the multiple target point clouds except the first target point cloud; if it is determined that the height difference is greater than or equal to a preset height difference, then determining that the first target point cloud and the second target point cloud are layered point clouds; determining whether the number of point clouds in the layered point clouds is greater than or equal to a preset number threshold; if it is determined that the number of point clouds in the layered point clouds is greater than or equal to the preset number threshold, then determining that there is a stratification phenomenon in the three-dimensional point cloud model.

[0048] Specifically, the server performs layered processing on the entire 3D point cloud model. Layered processing is to divide the point cloud data into multiple levels according to height information, and the point cloud height value in each level is within a certain range. The server implements layered processing based on the Z coordinate histogram method of the point cloud.

[0049] For example, in a 3D reconstruction project of a large park, the 3D point cloud model generated by the server contains 100 million points with a height range of 0-50 meters. The server first performs a histogram statistics on the Z coordinates of the point cloud and finds that the height distribution presents multiple obvious peaks, corresponding to different height levels of the ground, tree canopy, and building. According to the peak and valley characteristics of the histogram, the server divides the point cloud into five height layers: 0-1 meter, 1-5 meters, 5-15 meters, and 15-50 meters. The number of point clouds and height range in each layer are relatively uniform.

[0050] In each height layer, the server performs cluster analysis on the point cloud data. Cluster analysis is to divide point clouds with similar spatial distribution into a cluster, and the point clouds between different clusters are relatively separated in space. The server can use K-Means, DBSCAN algorithm, or a fast clustering algorithm based on octree and KD tree data structure. For example, in the 5-15 meter height layer of the above-mentioned park point cloud model, the server uses the DBSCAN algorithm to cluster the point cloud, sets the clustering radius to 1 meter, and the minimum number of points to 100. The clustering results show that there are 120 point cloud clusters in this layer, corresponding to different landscape elements such as tree crowns, lamp posts, pavilions, etc. The point clouds in each point cloud cluster are relatively compactly distributed, and the spacing between different point cloud clusters is greater than 1 meter, which conforms to the spatial distribution characteristics of the actual landscape. From the multiple point cloud clusters obtained in the previous step, the server arbitrarily selects one as the target point cloud cluster, and selects two point clouds from the point cloud cluster as the first target point cloud and the second target point cloud. The two target point clouds are any two point clouds in the point cloud cluster. The server calculates the height difference between the two target point clouds A and B, that is, ΔZ=|Z a -Z b |. The height difference reflects the distance of the target point cloud in the vertical direction and is an important indicator for determining whether there is a stratification phenomenon. The server compares the height difference ΔZ between the target point clouds A and B with the preset height difference. If ΔZ is greater than or equal to the threshold, the target point clouds A and B are considered to belong to different height levels and are layered point clouds; otherwise, the target point clouds A and B are considered to belong to the same height level and are not layered point clouds. The setting of the preset height difference can be adjusted according to factors such as the density of the point cloud and scene characteristics. It is usually 2-5 times the average spacing of the point cloud, and this application does not limit this.

[0051] The server counts the number of all point clouds that are determined to be layered, and compares it with the preset number threshold. If the number of layered point clouds is greater than or equal to the preset number threshold, it is considered that there is a layered phenomenon in the entire three-dimensional point cloud model, and further repair processing is required; otherwise, it is considered that there are no significant layered defects in the point cloud model, and it can directly enter the subsequent application link. The setting of the preset number threshold needs to be weighed according to factors such as the total amount of point clouds and the allowable defect ratio. It is usually taken as 0.1%-1% of the total number of point clouds, and this application does not limit this. Through the above steps, the server realizes the automatic judgment and estimation of the layered phenomenon of the three-dimensional point cloud model.

[0052] In a possible implementation, determining whether there is an intersection phenomenon of feature points in a three-dimensional point cloud model specifically includes: calculating the spatial distance between a first target point cloud and a second target point cloud; if it is determined that the spatial distance is less than a preset distance threshold, determining that there is an intersection phenomenon in the three-dimensional point cloud model.

[0053] Specifically, the server calculates the spatial distance between the first target point cloud and the second target point cloud. Spatial distance refers to the closest distance between two point clouds in three-dimensional space, which can be calculated using Euclidean distance. The server compares the calculated spatial distance with a preset distance threshold. If the spatial distance is less than the preset distance threshold, it is considered that there is an intersection between the first target point cloud and the second target point cloud, that is, the first target point cloud and the second target point cloud are interspersed or mixed with each other; otherwise, it is considered that there is no intersection between the first target point cloud and the second target point cloud, and they are relatively independent and separated in space. The setting of the preset distance threshold needs to be adjusted according to factors such as the density, scale, and scene characteristics of the point cloud. It is usually taken as 0.5-2 times the average spacing of the point cloud, and this application does not limit this.

[0054] Step S106: extracting texture information of the garden image according to the garden image.

[0055] In step S106, pixel blocks corresponding to each feature point are extracted from the garden image; color correction is performed on the pixel blocks to obtain corrected pixel blocks, and the corrected pixel blocks are used as texture information of the garden image.

[0056] Specifically, for each feature point, the server extracts a pixel block of a fixed size with the point as the center. A pixel block refers to a square or rectangular area centered on the feature point, which contains pixel information around the feature point and can express the local texture characteristics of the feature point. The size of the pixel block needs to be set according to the resolution of the image and the detail level of the texture, ranging from 16×16 to 64×64 pixels. Since garden images may be affected by factors such as lighting, weather, and shooting angle, there may be differences in color between different images, affecting the consistency and authenticity of the texture. Therefore, the server performs color correction on the extracted pixel blocks to match their colors with the standard color card or reference image. The color correction methods used by the server include white balance, color mapping, histogram matching and other correction algorithms. After color correction, the server obtains the corrected pixel block corresponding to each feature point as the texture information of the feature point. The server aggregates the texture information of all feature points and generates a complete garden image texture information dataset.

[0057] Step S107: Mapping the texture information to the three-dimensional point cloud model to obtain a three-dimensional real scene model of the landscape.

[0058] In step S107, the server maps the extracted garden image texture information with the three-dimensional point cloud model, adds a corresponding texture block to each point cloud, and generates a three-dimensional real scene model with real texture.

[0059] Reference Figure 2 The present application also provides a device for constructing a three-dimensional visualized landscape garden, which is a server. The server includes an acquisition module 201 and a processing module 202. The acquisition module 201 is used to acquire multiple garden images of the landscape garden taken by a drone. The garden images are taken from multiple preset angles when the drone flies along a preset route; the processing module 202 is used to determine the feature points of the garden images. The feature points include control points and connection points. The control points are preset coordinate points, and the connection points are commonly identifiable coordinate points in multiple garden images. The acquisition module 201 is also used to acquire the shooting time point of each garden image, and At the time of shooting, the POS data of the drone includes the longitude, latitude and altitude of the drone; the processing module 202 is also used to match the feature points of each garden image according to the POS data and the feature points, using aerial triangulation to obtain the three-dimensional point cloud data of each feature point; the processing module 202 is also used to construct a three-dimensional point cloud model corresponding to the landscape garden according to the three-dimensional point cloud data; the processing module 202 is also used to extract the texture information of the garden image according to the garden image; the processing module 202 is also used to map the texture information to the three-dimensional point cloud model to obtain the three-dimensional real scene model of the landscape garden.

[0060] In a possible implementation, after the processing module 202 constructs a three-dimensional point cloud model corresponding to the landscape garden based on the three-dimensional point cloud data, the method further includes: the processing module 202 determines whether there is a stratification phenomenon of feature points or an intersection phenomenon of feature points in the three-dimensional point cloud model; if the processing module 202 determines whether there is a stratification phenomenon of feature points or an intersection phenomenon of feature points in the three-dimensional point cloud model, the drone is controlled to reacquire the garden image.

[0061] In a possible implementation, the processing module 202 determines whether there is a hierarchical phenomenon of feature points in the three-dimensional point cloud model, specifically including: the processing module 202 performs hierarchical processing on the three-dimensional point cloud model, and divides the three-dimensional point cloud data into multiple layers according to height information; the processing module 202 performs cluster analysis on the point cloud data in each layer to obtain multiple point cloud clusters; the processing module 202 determines multiple target point clouds included in the target point cloud cluster, and the target point cloud cluster is any point cloud cluster among the multiple point cloud clusters; the processing module 202 calculates the height of the first target point cloud and the second target point cloud The processing module 202 determines that the height difference is greater than or equal to the preset height difference, the first target point cloud is any target point cloud among the multiple target point clouds, and the second target point cloud is any target point cloud among the multiple target point clouds except the first target point cloud; if the processing module 202 determines that the height difference is greater than or equal to the preset height difference, it determines that the first target point cloud and the second target point cloud are layered point clouds; the processing module 202 determines whether the number of point clouds of the layered point clouds is greater than or equal to the preset number threshold; if the processing module 202 determines that the number of point clouds of the layered point clouds is greater than or equal to the preset number threshold, it determines that there is a layered phenomenon in the three-dimensional point cloud model.

[0062] In a possible implementation, the processing module 202 determines whether there is an intersection phenomenon of feature points in the three-dimensional point cloud model, specifically including: the processing module 202 calculates the spatial distance between the first target point cloud and the second target point cloud; if the processing module 202 determines that the spatial distance is less than a preset distance threshold, it determines that there is an intersection phenomenon in the three-dimensional point cloud model.

[0063] In a possible implementation, the processing module 202 uses aerial triangulation to match the feature points of each garden image based on the POS data and the feature points to obtain three-dimensional point cloud data of each feature point, specifically including: the processing module 202 matches the feature points of multiple garden images to obtain the correspondence between the same feature point in the multiple garden images; the processing module 202 uses the control point coordinates of the control point on the ground and the feature point coordinates in the garden image to construct the correspondence between the image coordinate system and the ground coordinate system; the processing module 202 obtains the three-dimensional coordinates of each feature point in the three-dimensional space based on the correspondence; the processing module 202 performs dense matching on the three-dimensional coordinates to generate three-dimensional point cloud data.

[0064] In a possible implementation, the processing module 202 extracts texture information of the garden image based on the garden image, specifically including: the processing module 202 extracts pixel blocks corresponding to each feature point from the garden image; the processing module 202 performs color correction on the pixel blocks to obtain corrected pixel blocks, and uses the corrected pixel blocks as the texture information of the garden image.

[0065] In a possible implementation, the processing module 202 constructs a three-dimensional point cloud model corresponding to the landscape garden based on the three-dimensional point cloud data, specifically including: the processing module 202 performs filtering and down-sampling on the three-dimensional point cloud data to obtain target three-dimensional point cloud data; the processing module 202 generates a triangular mesh surface based on the target three-dimensional point cloud data to construct a three-dimensional point cloud model.

[0066] It should be noted that: when the device provided in the above embodiment realizes its function, only the division of the above functional modules is used as an example. In actual application, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0067] The present application also provides an electronic device. Figure 3 , Figure 3 The electronic device 300 may include: at least one processor 301 , at least one network interface 304 , a user interface 303 , a memory 305 , and at least one communication bus 302 .

[0068] The communication bus 302 is used to realize the connection and communication between these components.

[0069] The user interface 303 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 303 may also include a standard wired interface and a wireless interface.

[0070] The network interface 304 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0071] Among them, the processor 301 may include one or more processing cores. The processor 301 uses various interfaces and lines to connect various parts in the entire server, and executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 305, and calling data stored in the memory 305. Optionally, the processor 301 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 301 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 301, and it can be implemented separately through a chip.

[0072] Among them, the memory 305 may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read-Only Memory). Optionally, the memory 305 includes a non-transitory computer-readable storage medium. The memory 305 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 305 may optionally also be at least one storage device located away from the aforementioned processor 301. Refer to Figure 3 , the memory 305 as a computer storage medium may include an operating system, a network communication module, a user interface module and an application program of a method for constructing a three-dimensional visualized landscape.

[0073] exist Figure 3In the electronic device 300 shown, the user interface 303 is mainly used to provide an input interface for the user and obtain the data input by the user; and the processor 301 can be used to call the application program storing a method for constructing a three-dimensional visualized landscape in the memory 305, and when executed by one or more processors 301, the electronic device 300 executes one or more of the methods described in the above embodiments. It should be noted that for the aforementioned method embodiments, for the sake of simple description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for the present application.

[0074] The present application also provides a computer-readable storage medium, which stores instructions. When executed by one or more processors 301, the electronic device 300 executes one or more of the methods described in the above embodiments.

[0075] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0076] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0077] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0078] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0079] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a memory and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: various media that can store program codes, such as USB flash drives, mobile hard drives, magnetic disks or optical disks.

[0080] The above is only an exemplary embodiment of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and the truth of practice, those skilled in the art will easily think of other embodiments of the present disclosure.

[0081] This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art not described in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A method for constructing three-dimensional visual landscape architecture. It is characterized in that The method comprises: Acquire multiple images of landscape gardens taken by a drone, wherein the images of the landscape gardens are taken from multiple preset angles when the drone flies along a preset route; Determine feature points of the garden image, wherein the feature points include control points and connection points, wherein the control points are preset coordinate points, and the connection points are commonly identifiable coordinate points in a plurality of the garden images; Acquire the shooting time point of each of the garden images, and the POS data of the drone at the shooting time point, wherein the POS data includes the longitude, latitude and altitude of the drone; According to the POS data and the feature points, aerial triangulation is used to match the feature points of each of the garden images to obtain three-dimensional point cloud data of each of the feature points; Constructing a three-dimensional point cloud model corresponding to the landscape garden according to the three-dimensional point cloud data; Extracting texture information of the garden image according to the garden image; The texture information is mapped to the three-dimensional point cloud model to obtain a three-dimensional real scene model of the landscape.

2. The method according to claim 1, It is characterized in that After constructing the three-dimensional point cloud model corresponding to the landscape garden according to the three-dimensional point cloud data, the method further includes: Determine whether there is a stratification phenomenon or a crossing phenomenon of feature points in the three-dimensional point cloud model; If it is determined whether there is a layering phenomenon of feature points or an intersection phenomenon of feature points in the three-dimensional point cloud model, the drone is controlled to reacquire the garden image.

3. The method according to claim 2, It is characterized in that The determining whether there is a stratification phenomenon of feature points in the three-dimensional point cloud model specifically includes: Performing layered processing on the three-dimensional point cloud model, dividing the three-dimensional point cloud data into multiple layers according to height information; Perform cluster analysis on the point cloud data in each layer to obtain multiple point cloud clusters; Determine a plurality of target point clouds included in a target point cloud cluster, wherein the target point cloud cluster is any one of the plurality of point cloud clusters; Calculating a height difference between a first target point cloud and a second target point cloud, wherein the first target point cloud is any one of the plurality of target point clouds, and the second target point cloud is any one of the plurality of target point clouds except the first target point cloud; If it is determined that the height difference is greater than or equal to a preset height difference, determining that the first target point cloud and the second target point cloud are layered point clouds; Determining whether the number of point clouds of the layered point cloud is greater than or equal to a preset number threshold; If it is determined whether the number of point clouds of the layered point cloud is greater than or equal to a preset number threshold, it is determined that the layering phenomenon exists in the three-dimensional point cloud model.

4. The method according to claim 3, It is characterized in that The determining whether there is an intersection phenomenon of feature points in the three-dimensional point cloud model specifically includes: Calculate the spatial distance between the first target point cloud and the second target point cloud; If it is determined that the spatial distance is less than a preset distance threshold, it is determined that the intersection phenomenon exists in the three-dimensional point cloud model.

5. The method according to claim 1, It is characterized in that The method of performing feature point matching on each of the garden images by using an aerial triangulation operation based on the POS data and the feature points to obtain three-dimensional point cloud data of each of the feature points specifically includes: Performing feature point matching on the plurality of garden images to obtain a corresponding relationship between the same feature points in the plurality of garden images; Using the control point coordinates of the control point on the ground and the feature point coordinates in the garden image, a corresponding relationship between the image coordinate system and the ground coordinate system is constructed; According to the corresponding relationship, the three-dimensional coordinates of each of the feature points in the three-dimensional space are obtained; Dense matching is performed on the three-dimensional coordinates to generate the three-dimensional point cloud data.

6. The method according to claim 1, It is characterized in that The step of extracting texture information of the garden image according to the garden image specifically includes: Extracting pixel blocks corresponding to each feature point from the garden image; Color correction is performed on the pixel block to obtain a corrected pixel block, and the corrected pixel block is used as texture information of the garden image.

7. The method according to claim 1, It is characterized in that The constructing of a three-dimensional point cloud model corresponding to the landscape garden according to the three-dimensional point cloud data specifically includes: Performing filtering and down-sampling on the three-dimensional point cloud data to obtain target three-dimensional point cloud data; A triangular mesh surface is generated according to the target three-dimensional point cloud data to construct the three-dimensional point cloud model.

8. A device for constructing three-dimensional visual landscape architecture. It is characterized in that The device comprises an acquisition module (201) and a processing module (202), wherein: The acquisition module (201) is used to acquire multiple garden images of landscape gardens photographed by a drone, wherein the garden images are photographed from multiple preset angles when the drone flies along a preset route; The processing module (202) is used to determine the feature points of the garden image, wherein the feature points include control points and connection points, wherein the control points are preset coordinate points, and the connection points are coordinate points that are commonly identifiable in a plurality of the garden images; The acquisition module (201) is further used to acquire the shooting time point of each of the garden images, and the POS data of the drone at the shooting time point, wherein the POS data includes the longitude, latitude and altitude of the drone; The processing module (202) is further used to perform feature point matching on each of the garden images using an aerial triangulation operation according to the POS data and the feature points, so as to obtain three-dimensional point cloud data of each of the feature points; The processing module (202) is further used to construct a three-dimensional point cloud model corresponding to the landscape garden according to the three-dimensional point cloud data; The processing module (202) is further used to extract texture information of the garden image based on the garden image; The processing module (202) is also used to map the texture information to the three-dimensional point cloud model to obtain a three-dimensional real scene model of the landscape.

9. An electronic device, It is characterized in that The electronic device (300) comprises a processor (301), a memory (305), a user interface (303) and a network interface (304), wherein the memory (305) is used to store instructions, the user interface (303) and the network interface (304) are used to communicate with other devices, and the processor (301) is used to execute the instructions stored in the memory (305) so that the electronic device (300) executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores instructions, and when the instructions are executed, the method according to any one of claims 1 to 7 is executed.

Images