Surrounding terrain reconstruction method based on electric power facilities

Abstract

The embodiment of the invention provides a surrounding terrain reconstruction method based on an electric power facility, and the method comprises the steps: obtaining first point cloud data, carrying out the first denoising processing of the first point cloud data, and carrying out the first denoising processing of the first point cloud data, namely the point cloud data of a shielding object in the first point cloud data; obtaining coordinates of a shielding object, and obtaining first ultrasonic data according to the coordinates of the shielding object; establishing a first three-dimensional model according to the first point cloud data subjected to the first denoising processing, and establishing a second three-dimensional model according to the first ultrasonic data; calculating a coincidence position of the first three-dimensional model and the second three-dimensional model, supplementing the second three-dimensional model into the first three-dimensional model according to the coincidence position, fusing the laser radar point cloud data and the sound wave radar data, performing intelligent denoising and classification on the point cloud data, and removing the influence of a shielding object; and the terrain data of the sheltered area is obtained through the acoustic radar, and finally the two kinds of data are fused, so that the precision of terrain reconstruction is improved, and the integrity of the terrain model is enhanced.

Description

Technical Field

[0001] This invention relates to the field of three-dimensional modeling of the terrain surrounding power facilities, and more specifically, to a method for reconstructing the terrain surrounding power facilities. Background Technology

[0002] The construction, operation, and maintenance of power facilities require accurate topographic information. Accurate topographic models aid in the planning, design, construction, and subsequent inspection and maintenance of power facilities, improving their safety, reliability, and operational efficiency, and reducing accidents and malfunctions caused by topographic factors. For example, in the erection of power lines, factors such as terrain undulation and slope need to be considered to determine the optimal line route and tower locations; in the site selection and construction of substations, the flatness and drainage conditions of the surrounding terrain need to be understood. Traditional topographic surveying methods, such as manual surveying and optical surveying, suffer from low efficiency, high cost, and significant environmental limitations. In situations where the surrounding terrain of power facilities is complex and covers a large area, traditional methods struggle to quickly and accurately obtain complete topographic data. For instance, in mountainous or densely vegetated areas, manual and optical surveying are difficult, time-consuming, and easily affected by obstructions, making it impossible to obtain accurate topographic information.

[0003] While a single lidar-based point cloud data acquisition method can quickly acquire terrain data over a large area, its measurement accuracy is limited in vegetated areas or where there are obstructions, making it impossible to accurately acquire ground information. Summary of the Invention

[0004] This invention provides a method for reconstructing the surrounding terrain of power facilities, which at least solves the problem that laser point cloud measurement methods in related technologies cannot accurately reconstruct the surrounding terrain of power facilities.

[0005] According to one embodiment of the present invention, a method for reconstructing the surrounding terrain of power facilities is provided, comprising:

[0006] Acquire first point cloud data, and perform a first denoising process on the first point cloud data, wherein the first denoising process is to remove point cloud data containing occluders from the first point cloud data;

[0007] Obtain the coordinates of the obstruction, and obtain the first ultrasound data based on the coordinates of the obstruction;

[0008] A first three-dimensional model is established based on the first point cloud data after the first denoising process, and a second three-dimensional model is established based on the first ultrasound data.

[0009] Calculate the overlap position between the first 3D model and the second 3D model, and supplement the second 3D model into the first 3D model according to the overlap position.

[0010] Furthermore, the first denoising process involves removing point cloud data containing occluders from the first point cloud data, specifically including:

[0011] Feature extraction is performed on the first point cloud data;

[0012] The data from feature extraction are normalized.

[0013] The normalized data is input into a pre-trained model for filtering.

[0014] Based on the filtering results, each point cloud data in the first point cloud data is classified;

[0015] Remove point cloud data classified as non-ground points from the first point cloud data.

[0016] Furthermore, obtaining the coordinates of the obstruction and obtaining the first ultrasound data based on the coordinates of the obstruction specifically includes:

[0017] The occlusion area of ​​the occluder is obtained based on the density of the point cloud data classified as non-ground points.

[0018] Based on the occlusion area of ​​the occlusion object, obtain the coordinates of the point cloud data of non-ground points within the occlusion area;

[0019] The center coordinates are obtained based on the coordinates of the point cloud data of non-ground points within the occluded area;

[0020] The location is calculated by the acoustic radar based on the center coordinates.

[0021] Further, the step of calculating the overlap position of the first 3D model and the second 3D model, and supplementing the second 3D model into the first 3D model according to the overlap position, specifically includes:

[0022] Obtain the geometric features of the first 3D model and the second 3D model;

[0023] Based on the geometric features, a feature matching algorithm is used to perform feature matching between the first 3D model and the second 3D model;

[0024] Based on the feature point pairs after feature matching, the second 3D model is aligned to the coordinate system of the first 3D model;

[0025] The second 3D model is transformed using the iterative nearest point algorithm;

[0026] Calculate the distance from each point in the second 3D model to the surface of the first 3D model, and determine the points that are less than a preset distance threshold as points in the overlapping area;

[0027] The second 3D model is cropped based on the points in the overlapping region, and a portion of the cropped second 3D model is fitted into the first 3D model.

[0028] According to another embodiment of the present invention, a three-dimensional model construction system based on point cloud data is provided, comprising:

[0029] Point cloud data acquisition module: acquires first point cloud data, performs first denoising processing on the first point cloud data, the first denoising processing is to remove point cloud data containing occluders from the first point cloud data;

[0030] Acoustic data acquisition module: acquires the coordinates of the obstruction, and acquires first ultrasonic data based on the coordinates of the obstruction;

[0031] 3D model building module: builds a first 3D model based on the first point cloud data after the first denoising process, and builds a second 3D model based on the first ultrasound data;

[0032] 3D model compensation module: calculates the overlap position of the first 3D model and the second 3D model, and supplements the second 3D model into the first 3D model according to the overlap position.

[0033] According to yet another embodiment of the present invention, a computer-readable storage medium for reconstructing the surrounding terrain of power facilities is also provided, wherein the computer-readable storage medium stores a computer program configured to perform the steps in any of the above method embodiments when running.

[0034] According to yet another embodiment of the present invention, an electronic device for reconstructing the surrounding terrain of power facilities is also provided, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0035] This invention effectively combines the advantages of both lidar point cloud data and acoustic radar data, overcoming the limitations of a single data source and achieving a complete and accurate reconstruction of the terrain surrounding power facilities. Simultaneously, this invention utilizes a machine learning model to intelligently denoise and classify the point cloud data, removing the influence of obstructions. Then, it acquires terrain data for obstructed areas using acoustic radar. Finally, the two types of data are fused, which not only improves the accuracy of terrain reconstruction but also enhances the integrity of the terrain model, reducing measurement blind spots caused by obstructions in traditional measurement methods. This process significantly improves data processing efficiency and automation, reduces manual intervention, and provides more accurate terrain information support for the construction, operation, and maintenance of power facilities. Attached Figure Description

[0036] Figure 1 This is a flowchart of a method for reconstructing the surrounding terrain of power facilities according to an embodiment of the present invention;

[0037] Figure 2 This is a structural block diagram of a power facility-based terrain reconstruction system according to an embodiment of the present invention. Detailed Implementation

[0038] The technical solutions of the embodiments of this application will be described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0039] In the following description, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0040] Furthermore, in this application, directional terms such as "upper," "lower," "left," and "right" may be defined relative to the orientation of the components shown in the accompanying drawings. It should be understood that these directional terms can be relative concepts, used for relative description and clarification, and may change accordingly depending on the orientation of the components in the accompanying drawings.

[0041] In this application, unless otherwise expressly specified and limited, the term "connection" should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral part; it can be a direct connection or an indirect connection through an intermediate medium. Furthermore, the term "coupled" can refer to an electrical connection that enables signal transmission.

[0042] As used herein, “about,” “approximately,” or “approximately” includes the stated value and the average value within an acceptable range of deviation from the given value, wherein the acceptable range of deviation is determined by a person skilled in the art taking into account the measurement under discussion and the error associated with the measurement of the given quantity (i.e., the limitations of the measurement system).

[0043] This embodiment provides a method for reconstructing the surrounding terrain based on power facilities. Figure 1 This is a flowchart of a method for reconstructing the surrounding terrain of power facilities according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps:

[0044] Step S100: Obtain first point cloud data, and perform first denoising processing on the first point cloud data. The first denoising processing is to remove point cloud data containing occluders from the first point cloud data.

[0045] Specifically, in this embodiment, the target area is preferably scanned using a drone equipped with a LiDAR to acquire raw point cloud data containing terrain and occlusions. Simultaneously, the raw point cloud data undergoes data format conversion, transforming it into a common format such as LAS or PLY for subsequent processing. After format conversion, the raw point cloud data is preprocessed, primarily involving denoising and filtering to remove obvious noise points and outliers while preserving the main features of the terrain and occlusions. In this embodiment, denoising is performed using statistical filtering methods, calculating the voxel density or distance mean, standard deviation, and other statistical characteristics of each point to remove noise points that deviate excessively from these statistical characteristics. Simultaneously, voxel grid filtering methods are used to divide the space into regular cubic grids, processing points within each voxel to remove points that do not conform to certain rules, thus achieving the filtering effect while preserving the main features of the terrain and occlusions. Furthermore, morphological filtering methods, such as dilation and erosion, can be combined to remove points that do not conform to the terrain morphology, further optimizing data quality. Finally, the resulting first point cloud data is the data after format conversion and preprocessing. It retains the terrain features of the target area while removing most obvious noise points and outliers, laying a solid foundation for subsequent occlusion identification and removal. It is important to note that the point cloud data of occluders is not completely removed at this stage; rather, their main features are retained to enable more accurate identification and location of occluders in subsequent steps. This portion of data after the above processing constitutes the first point cloud data.

[0046] After performing the above steps, the first point cloud data undergoes a first denoising process, which includes the following steps:

[0047] Feature extraction is performed on the first point cloud data, specifically extracting multiple features for each point, including but not limited to geometric features, density features, intensity features, and structural features, to comprehensively characterize the point cloud data. Specifically, height features involve calculating the height value of each point and its height difference with neighboring points; density features involve counting the number of points within a certain neighborhood of each point and calculating the local point density; curvature features are calculated using methods such as surface fitting; and intensity features are extracted by extracting the echo intensity information of each point in the point cloud data. The extracted feature data is then normalized to eliminate differences in dimensions and orders of magnitude between different features. Commonly used methods include min-max normalization and Z-score normalization. Normalization improves the stability and accuracy of subsequent data processing. Since data normalization is already relatively well-established, this step will not be elaborated upon further.

[0048] The normalized feature data is input into a pre-trained model for filtering. This model is trained on a large amount of labeled sample data and can learn the feature differences between ground points and non-ground points in point cloud data. Through the model's calculation and judgment, each point cloud data is classified and predicted, and the probability or category label of each point belonging to ground point or non-ground point is output.

[0049] Based on the model's filtering results, each point cloud data point in the first point cloud dataset is classified into two categories: ground points and non-ground points. This determines which points belong to the ground and which belong to occlusions or other non-ground objects. Finally, the point cloud data classified as non-ground points is removed, resulting in filtered point cloud data containing only ground points. While using machine learning models for point cloud data filtering is the best approach due to technological advancements, other methods can still be employed, such as statistical filtering or voxel-based filtering. Statistical filtering involves calculating statistical characteristics for each point, such as the mean distance and standard deviation. A threshold is then set based on these characteristics to remove points that deviate excessively from the statistical features, classifying these points as noise. Voxel-based filtering, on the other hand, requires voxelizing the point cloud data, dividing the space into regular cubic grids (voxels). Within each voxel, points are statistically analyzed, and points that do not conform to certain rules are removed. Regardless of the method used, the above steps can be achieved.

[0050] Step S200: Obtain the coordinates of the obstruction, and obtain the first ultrasound data based on the coordinates of the obstruction;

[0051] Specifically, the steps include: obtaining the occlusion area of ​​the occluder based on the density of point cloud data classified as non-ground points;

[0052] Obtain the coordinates of point cloud data for non-ground points within the occlusion area based on the occlusion area of ​​the occlusion object;

[0053] Obtain the center coordinates based on the coordinates of the point cloud data of non-ground points within the occluded area;

[0054] The location is determined by the acoustic radar based on the center coordinates.

[0055] More specifically, a detailed distribution analysis is performed on the point cloud data classified as non-ground points. Using point cloud processing software or a self-written script, the density information of each non-ground point and its neighboring points is calculated. For example, the point density of a region can be assessed by calculating the number of points per unit area in the neighborhood of each point. Based on the density analysis results, point sets with significantly higher density than the surrounding areas are identified; these high-density areas typically correspond to the locations of occlusions. Density histograms or density clustering algorithms (such as DBSCAN) can be used to assist in identifying these density variation regions. Based on the density analysis and clustering results, the boundaries of the occlusion areas of occluders are determined. This can be achieved by fitting geometry to the high-density point sets, thereby accurately defining the spatial extent occupied by the occluders.

[0056] Then, within the defined occlusion area, point cloud data processing tools or programming interfaces are used to accurately extract the 3D coordinate information of all non-ground points, including the X, Y, and Z coordinates of each point, as well as any other possible attribute information (such as echo intensity, classification labels, etc.). This coordinate information not only reflects the location of the occlusion object but also provides crucial spatial references for subsequent ultrasound data acquisition. Next, geometric methods are used to calculate the center coordinates of the non-ground points within the occlusion area. For example, the average coordinates of all non-ground points can be calculated by adding the X, Y, and Z coordinates of all points and dividing by the total number of points, resulting in a coordinate point representing the center of the occlusion object. Alternatively, cluster analysis methods, such as K-means, can be used to determine the center location of the occlusion object. This divides the point cloud data within the occlusion area into multiple clusters, and then determines the center coordinates of each cluster. For a cluster formed by a single occlusion object, its center coordinates represent the center location of the occlusion object.

[0057] Finally, based on the calculated center coordinates and the geometry of the actual terrain and obstructions, the measurement position of the acoustic radar is fine-tuned and optimized. If the obstruction has a complex geometry, the angle and height of the acoustic radar may need to be adjusted to ensure comprehensive coverage of the terrain area below the obstruction. The measurement range and resolution of the acoustic radar are set appropriately according to the size and shape of the obstruction. For larger obstructions, the measurement range may need to be expanded and the resolution appropriately reduced; while for smaller obstructions, the measurement range can be reduced and the resolution increased to obtain more detailed terrain information. Simultaneously, parameters such as the acoustic radar's transmission frequency and pulse width are adjusted according to the actual application scenario and accuracy requirements to optimize measurement performance and data quality.

[0058] In the above steps, a drone equipped with a sonic radar can also be used. After determining the center coordinates of the obstruction, the drone sampling coordinates calculated based on the center coordinates are sent to the drone to complete the drone's automatic sampling.

[0059] Step S300: Establish a first three-dimensional model based on the first point cloud data after the first denoising process, and establish a second three-dimensional model based on the first ultrasound data;

[0060] The process involves building a 3D model from the first point cloud data after initial denoising. When the data volume is too large, appropriate sampling can be used. The point cloud data is then input into a selected modeling tool or library. Based on the chosen modeling method, a first 3D model is generated. Modeling methods can include voxelization and triangulation. The voxelization method first determines the voxel size based on the distribution range of the point cloud data and the required model resolution. Then, the 3D space is divided into a voxel grid, and the number of point cloud data within each voxel is counted. Based on the density or height information of the point cloud data, an attribute value is assigned to each voxel. Finally, the 3D model is constructed using the voxel grid. Voxelization tools or libraries, such as VoxelGrid, can be used. The triangulation method involves projecting the point cloud data onto a 2D plane, using the Delaunay triangulation algorithm to triangulate the 2D point set, generating a 2D triangular mesh. This 2D triangular mesh is then mapped back into 3D space to form a 3D triangular mesh. The triangular mesh is then optimized and repaired, such as removing small triangles and filling holes. The Delaunay triangulation library can be used as a triangulation generation tool. After regenerating the first 3D model, use model optimization tools or libraries to optimize the first 3D model, such as smoothing, noise removal, and hole filling, such as MeshLab, CloudCompare, etc.

[0061] For the establishment of the second 3D model, data registration is first performed. If the ultrasonic data and point cloud data are not in the same coordinate system, coordinate system registration is required to transform the ultrasonic data into the same coordinate system as the point cloud data. Then, the registered ultrasonic data is input into the selected modeling tool or library, and the second 3D model is generated according to the selected modeling method. For example, an interpolation algorithm can be used to generate regular mesh data, which is then converted into a triangular mesh model; or a surface fitting method can be used to generate a smooth surface model, which is then optimized. In this embodiment, the modeling method used is the interpolation method. Using the ultrasonic data of known points, the elevation values ​​of unknown points are estimated through an interpolation algorithm to generate a continuous terrain surface. First, a suitable interpolation algorithm needs to be selected. Based on the distribution and characteristics of the data, the interpolation parameters are determined. The interpolation algorithm is then used to interpolate the ultrasonic data, generating regular mesh data or triangular mesh data. Then, based on the interpolation results, a 3D model is constructed. Interpolation processing can be performed using the SciPy interpolation library in Python or professional interpolation software such as GS+.

[0062] Step S400: Calculate the overlap position of the first 3D model and the second 3D model, and add the second 3D model to the first 3D model according to the overlap position.

[0063] Specifically, the process includes the following steps: obtaining the geometric features of a first 3D model and a second 3D model; performing feature matching on the first 3D model and the second 3D model using a feature matching algorithm based on the geometric features; aligning the second 3D model to the coordinate system of the first 3D model based on the feature point pairs after feature matching; performing matrix transformation on the second 3D model using an iterative nearest-point algorithm; calculating the distance from each point in the second 3D model to the surface of the first 3D model, and determining points less than a preset distance threshold as points in the overlapping region; cropping the second 3D model based on the points in the overlapping region, and fitting the cropped portion of the second 3D model into the first 3D model.

[0064] More specifically, key points, such as corners, edges, and curvature extrema, are first extracted from the two 3D models. A feature descriptor is calculated for each key point, containing geometric and grayscale information about the surrounding area to distinguish different feature points. Then, a feature matching algorithm (such as brute-force matching or FLANN matching) is used to match the key points of the first 3D model with those of the second 3D model, finding possible corresponding point pairs. The RANSAC algorithm is then used to filter out correct matching pairs and remove incorrect matches. Based on the matched feature point pairs, the rotation matrix and translation vector that transform the second 3D model to the coordinate system of the first 3D model are calculated. The calculated transformation matrix is ​​applied to the second 3D model to align it to the coordinate system of the first 3D model. Then, using the results of feature matching and RANSAC filtering as the initial alignment, the rotation matrix and translation vector are continuously updated by calculating the nearest corresponding point in the first 3D model for each point in the second 3D model, minimizing the distance between point pairs. An upper limit for the number of iterations or a convergence threshold (such as root mean square error) is set, and iteration stops when the condition is met. Then, for each point in the second 3D model, calculate its distance to the surface of the first 3D model. Nearest neighbor search or KD-trees can be used to accelerate the calculation. A distance threshold is set; points smaller than this threshold are considered to be within the overlapping region. Based on the points in the overlapping region, the second 3D model is trimmed, retaining points outside the overlapping region with the first 3D model. The trimmed portion of the second 3D model is then surface-fitted to the first 3D model, using interpolation algorithms or surface fitting methods to ensure a smooth transition between the two models. Finally, the trimmed portion of the second 3D model is integrated into the first 3D model to generate a complete terrain model.

[0065] This invention effectively combines the advantages of both lidar point cloud data and acoustic radar data, overcoming the limitations of a single data source and achieving a complete and accurate reconstruction of the terrain surrounding power facilities. Simultaneously, this invention utilizes a machine learning model to intelligently denoise and classify the point cloud data, removing the influence of obstructions. Then, it acquires terrain data for obstructed areas using acoustic radar. Finally, the two types of data are fused, which not only improves the accuracy of terrain reconstruction but also enhances the integrity of the terrain model, reducing measurement blind spots caused by obstructions in traditional measurement methods. This process significantly improves data processing efficiency and automation, reduces manual intervention, and provides more accurate terrain information support for the construction, operation, and maintenance of power facilities.

[0066] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0067] This embodiment also provides a surrounding terrain reconstruction system based on power facilities, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the apparatus described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0068] Figure 2 This is a structural block diagram of a power facility-based terrain reconstruction system according to an embodiment of the present invention, such as... Figure 2 As shown, the system includes:

[0069] Point cloud data acquisition module: acquires first point cloud data, performs first denoising processing on the first point cloud data, the first denoising processing is to remove point cloud data with occlusions from the first point cloud data;

[0070] Acoustic data acquisition module: acquires the coordinates of the obstruction, and obtains the first ultrasonic data based on the coordinates of the obstruction;

[0071] 3D model building module: Builds a first 3D model based on the first point cloud data after the first denoising process, and builds a second 3D model based on the first ultrasound data;

[0072] 3D model compensation module: Calculates the overlap position of the first 3D model and the second 3D model, and supplements the second 3D model into the first 3D model according to the overlap position.

[0073] It should be noted that the above modules can be implemented by software or hardware. For the latter, they can be implemented in the following ways, but are not limited to: all the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0074] Embodiments of the present invention also provide a computer-readable storage medium for reconstructing the surrounding terrain of power facilities, wherein the computer-readable storage medium stores a computer program configured to execute the steps in any of the above method embodiments when running.

[0075] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as a USB flash drive, read-only memory (ROM), random access memory (RAM), portable hard disk, magnetic disk, or optical disk.

[0076] Embodiments of the present invention also provide an electronic device for reconstructing the surrounding terrain of power facilities, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the steps in any of the above method embodiments.

[0077] In one exemplary embodiment, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor and the input / output device is connected to the processor.

[0078] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0079] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division of a method for reconstructing the surrounding terrain of power facilities. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0080] The units described as separate components may or may not be physically separate. A component shown as a unit can be one or more physical units; that is, it can be located in one place or distributed in multiple different locations. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0081] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0082] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, in essence, or the parts that contribute to the prior art, or all or part of the technical solutions, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0083] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for reconstructing the surrounding terrain of power facilities, characterized in that, include: Acquire first point cloud data, and perform a first denoising process on the first point cloud data, wherein the first denoising process is to remove point cloud data containing occluders from the first point cloud data; Obtain the coordinates of the obstruction, and obtain the first ultrasound data based on the coordinates of the obstruction; A first three-dimensional model is established based on the first point cloud data after the first denoising process, and a second three-dimensional model is established based on the first ultrasound data. Calculate the overlap position between the first 3D model and the second 3D model, and supplement the second 3D model into the first 3D model according to the overlap position.

2. The method according to claim 1, characterized in that, The first denoising process involves removing point cloud data containing occluders from the first point cloud data, specifically including: Feature extraction is performed on the first point cloud data; The data from feature extraction are normalized. The normalized data is input into a pre-trained model for filtering. Based on the filtering results, each point cloud data in the first point cloud data is classified; Remove point cloud data classified as non-ground points from the first point cloud data.

3. The method according to claim 2, characterized in that, The step of obtaining the coordinates of the obstruction and obtaining the first ultrasound data based on the coordinates of the obstruction specifically includes: The occlusion area of ​​the occluder is obtained based on the density of the point cloud data classified as non-ground points. Based on the occlusion area of ​​the occlusion object, obtain the coordinates of the point cloud data of non-ground points within the occlusion area; The center coordinates are obtained based on the coordinates of the point cloud data of non-ground points within the occluded area; The location is calculated by the acoustic radar based on the center coordinates.

4. The method according to claim 1, characterized in that, The step of calculating the overlap position of the first 3D model and the second 3D model, and supplementing the second 3D model into the first 3D model according to the overlap position, specifically includes: Obtain the geometric features of the first 3D model and the second 3D model; Based on the geometric features, a feature matching algorithm is used to perform feature matching between the first 3D model and the second 3D model; Based on the feature point pairs after feature matching, the second 3D model is aligned to the coordinate system of the first 3D model; The second 3D model is transformed using the iterative nearest point algorithm; Calculate the distance from each point in the second 3D model to the surface of the first 3D model, and determine the points that are less than a preset distance threshold as points in the overlapping area; The second 3D model is cropped based on the points in the overlapping region, and a portion of the cropped second 3D model is fitted into the first 3D model.

5. A 3D model construction system based on point cloud data, characterized in that, include: Point cloud data acquisition module: acquires first point cloud data, performs first denoising processing on the first point cloud data, the first denoising processing is to remove point cloud data containing occluders from the first point cloud data; Acoustic data acquisition module: acquires the coordinates of the obstruction, and acquires first ultrasonic data based on the coordinates of the obstruction; 3D model building module: builds a first 3D model based on the first point cloud data after the first denoising process, and builds a second 3D model based on the first ultrasound data; 3D model compensation module: calculates the overlap position of the first 3D model and the second 3D model, and supplements the second 3D model into the first 3D model according to the overlap position.

6. A computer-readable storage medium for a method of reconstructing the surrounding terrain of power facilities, characterized in that, The computer-readable storage medium stores a computer program, wherein the computer program is configured to perform the method described in any one of claims 1 to 4 when executed.

7. An electronic device for a method of reconstructing the surrounding terrain of power facilities, comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the method described in any one of claims 1 to 4.

Images