Pole tower inclination monitoring method and system based on point cloud rapid reconstruction

Abstract

The invention relates to the technical field of pole and tower inclination monitoring, and provides a pole and tower inclination monitoring method and system based on point cloud rapid reconstruction, and the method comprises the following steps: obtaining original point cloud data of a pole and tower through a hierarchical scanning and adaptive sampling density method; performing down-sampling, noise filtering and preliminary classification on the original point cloud data to obtain tower structure point cloud data; extracting tower geometric features based on the tower structure point cloud data, and constructing a tower skeleton three-dimensional model through the tower geometric features; calculating the inclination angle of each part of the tower based on the central axis of the tower skeleton three-dimensional model; and carrying out deformation analysis and deformation prediction on the tower based on the inclination angle of each part of the tower, and generating a tower inclination monitoring report. According to the method, the whole-process deformation monitoring from data acquisition, processing, modeling to analysis and prediction is realized, and the tower inclination monitoring efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of pole and tower inclination monitoring, and particularly to a method and system for pole and tower inclination monitoring based on rapid point cloud reconstruction. Background Art

[0002] As an important part of the power transmission and communication systems, the stability of the structure of poles and towers is directly related to the safe operation and economic benefits of the entire system. With the changes in the natural environment and long-term use, poles and towers are prone to structural problems such as inclination and deformation. If these problems are not monitored and processed in a timely manner, they may lead to serious safety accidents and affect the normal supply of power and communication. Therefore, accurately and efficiently monitoring the inclination of poles and towers and evaluating their structural health status are key technical requirements for ensuring the safe operation of poles and towers.

[0003] The existing technologies usually use traditional rigid registration algorithms for pole and tower inclination monitoring. During the data acquisition process, there are often problems such as fixed scanning angles and incomplete coverage of point cloud data, resulting in incomplete point cloud data. Simple rigid registration algorithms only consider the spatial position relationship of point cloud data and ignore the complex geometric features of the pole and tower structure, resulting in low accuracy of axis extraction and large errors in calculating the inclination angle. Summary of the Invention

[0004] In view of this, the present invention proposes a method and system for pole and tower inclination monitoring based on rapid point cloud reconstruction, which solves the problems in the existing technologies that only consider the spatial position relationship of point cloud data and ignore the complex geometric features of the pole and tower structure, resulting in low accuracy of axis extraction and large errors in calculating the inclination angle.

[0005] The technical solution of the present invention is implemented as follows: In the first aspect, the present invention provides a method for pole and tower inclination monitoring based on rapid point cloud reconstruction, including the following steps:

[0006] Obtain the original point cloud data of the pole and tower through hierarchical scanning and adaptive sampling density method;

[0007] Perform downsampling, noise filtering and preliminary classification on the original point cloud data to obtain the pole and tower structure point cloud data;

[0008] Extract the geometric features of the pole and tower based on the pole and tower structure point cloud data, and construct a three-dimensional model of the pole and tower skeleton through the geometric features of the pole and tower;

[0009] Calculate the inclination angles of each part of the pole and tower based on the central axis of the three-dimensional model of the pole and tower skeleton;

[0010] Perform deformation analysis and deformation prediction on the pole and tower based on the inclination angles of each part of the pole and tower, and generate a pole and tower inclination monitoring report.

[0011] Based on the above technical solutions, preferably, the method for obtaining the original point cloud data of the pole tower by hierarchical scanning and adaptive sampling density method specifically includes:

[0012] Divide the pole tower into multiple scanning layers according to the height of the pole tower, set an overlapping area between adjacent scanning layers, and obtain the point cloud data of each scanning layer by scanning from bottom to top;

[0013] Perform data acquisition on each scanning layer using the adaptive sampling density method, and gradually reduce the sampling density from the bottom scanning layer to the top scanning layer.

[0014] Based on the above technical solutions, preferably, the method for downsampling, noise filtering and preliminary classification of the original point cloud data to obtain the pole tower structure point cloud data specifically includes:

[0015] Perform downsampling on the original point cloud data using adaptive voxel filtering, and remove noise points using the statistical outlier analysis method to obtain preprocessed point cloud data;

[0016] Based on the spatial distribution characteristics and reflection intensity characteristics of the preprocessed point cloud data, use the multi-feature fusion classification method to extract the pole tower structure point cloud data.

[0017] Based on the above technical solutions, preferably, the method for extracting the geometric features of the pole tower based on the pole tower structure point cloud data and constructing a three-dimensional model of the pole tower skeleton specifically includes:

[0018] Use the region growing algorithm to extract key geometric feature points from the pole tower structure point cloud data, and identify the main structural components of the pole tower and the connection relationship between the main structural components through curvature analysis and edge detection methods;

[0019] Use adaptive region growing parameters to dynamically adjust the threshold of region growing according to the point cloud density in different height regions of the pole tower;

[0020] Based on the key geometric feature points, use the Delaunay triangulation algorithm to construct a preliminary skeleton model of the pole tower, and use the multi-scale filtering method to perform smoothing processing on the preliminary skeleton model at different scales to eliminate small-range deviations caused by point cloud data noise, and obtain the pole tower skeleton model.

[0021] Based on the above technical solutions, preferably, the method for calculating the inclination angle of each part of the pole tower based on the central axis of the three-dimensional model of the pole tower skeleton specifically includes:

[0022] Use the improved principal component analysis method combined with the iterative closest point algorithm to extract the central axis of the pole tower skeleton model, and optimize the spatial position of the central axis through the multi-level feature fusion method to improve the axis positioning accuracy;

[0023] Based on the adaptive segmentation calculation method, the pole tower is divided into multiple calculation units, and the angle between the central axis of each calculation unit and the vertical direction is calculated to obtain the local inclination angle.

[0024] On the basis of the above technical solutions, preferably, the calculation formula of the objective function for optimizing the central axis is:

[0025]

[0026] where E(·) is the objective function, L is the central axis, is the k 1 th center of gravity point, is the weight coefficient of the k 1 th center of gravity point, is the perpendicular distance from the point to the central axis L, λ 1 is the balance coefficient, n is the number of center of gravity points, m is the number of segments of the central axis, is the slope change of the central axis in the k 2 th segment, is the inclination angle of the k 3 th segment, and μ is the smoothing coefficient of the inclination angle change;

[0027] The calculation formula of the local inclination angle is:

[0028]

[0029] where, is the inclination angle of the a th calculation unit, is the local axis direction vector of the a a th calculation unit, is the local curvature value of the a th calculation unit,

[0030] On the basis of the above technical solutions, preferably, based on the inclination angles of each part of the pole tower, deformation analysis and deformation prediction of the pole tower are carried out to generate a pole tower inclination monitoring report, which specifically includes:

[0031] Adopt a deformation feature extraction algorithm based on deep learning to extract the deformation features of the pole tower from the inclination angle data of each part of the pole tower, and combine historical monitoring data and real-time monitoring data to analyze the deformation trend of the pole tower;

[0032] The calculation formula of the deformation feature extraction algorithm is:

[0033]

[0034] Among them, F is the comprehensive deformation eigenvalue, is the inclination angle at the b-th moment, is the inclination angle at the (b - 1)-th moment, γ is the time decay coefficient, t b is the time interval between the b-th moment and the current moment, δ is the abnormal deformation weight coefficient, is the change in inclination angle of the f-th abnormal deformation point, B is the total sampling duration for deformation analysis, and p is the number of abnormal deformation points;

[0035] Predict the deformation of the pole tower through the deformation trend prediction model, and generate a pole tower inclination monitoring report according to the prediction result;

[0036] The calculation formula of the deformation trend prediction model is:

[0037]

[0038] Among them, Δφ(t) is the change in inclination angle after time t, φ 0 is the initial inclination angle, α q is the q-th amplitude coefficient, β q is the q-th decay coefficient, ω q is the q-th frequency coefficient, δ q is the q-th phase offset, Q is the number of terms of the deformation trend prediction model, and κ is the deformation trend adjustment coefficient.

[0039] In a second aspect, the present invention also provides a pole tower inclination monitoring system based on point cloud rapid reconstruction, and the system includes:

[0040] A data acquisition module, configured to acquire the original point cloud data of the pole tower through hierarchical scanning and the adaptive sampling density method;

[0041] A data processing module, configured to perform downsampling, noise filtering, and preliminary classification on the original point cloud data to obtain the pole tower structure point cloud data;

[0042] A feature extraction module, configured to extract the geometric features of the pole tower based on the pole tower structure point cloud data, and construct a three-dimensional model of the pole tower skeleton through the geometric features of the pole tower;

[0043] An inclination calculation module, configured to calculate the inclination angles of each part of the pole tower based on the central axis of the three-dimensional model of the pole tower skeleton;

[0044] A deformation prediction module, configured to perform deformation analysis and deformation prediction on the pole tower based on the inclination angles of each part of the pole tower, and generate a pole tower inclination monitoring report.

[0045] In a third aspect, the present invention also provides an electronic device, including: at least one processor, at least one memory, a communication interface, and a bus;

[0046] Among them, the processor, the memory, and the communication interface complete mutual communication through the bus. The memory stores program instructions executable by the processor, and the processor invokes the program instructions to implement the steps of a tower inclination monitoring method based on rapid point cloud reconstruction.

[0047] In a fourth aspect, the present invention also provides a computer-readable storage medium. The computer-readable storage medium stores computer instructions, and the computer instructions enable a computer to implement the steps of a tower inclination monitoring method based on rapid point cloud reconstruction.

[0048] A tower inclination monitoring method and system based on rapid point cloud reconstruction of the present invention have the following beneficial effects compared with the prior art:

[0049] (1) By using hierarchical scanning and adaptive sampling density method to obtain high-quality point cloud data, combining data processing methods such as downsampling and noise filtering to extract accurate tower structure information, and realizing precise modeling through geometric feature extraction and skeleton three-dimensional model construction. Furthermore, based on the central axis, the inclination angle is calculated and deformation analysis and prediction are carried out, and finally a monitoring report is generated, realizing the full-process deformation monitoring from data acquisition, processing, modeling to analysis and prediction, and improving the efficiency of tower inclination monitoring;

[0050] (2) By introducing an objective function that comprehensively considers the weight of the center of gravity point, the vertical distance, the change of the axis slope, and the change of the inclination angle, and combining the inclination angle calculation method based on the local axis direction vector and the reliability weight coefficient, the precise positioning of the tower central axis and the accurate calculation of the inclination angle are realized, and the reliability of inclination monitoring is improved;

[0051] (3) By introducing a deformation feature extraction algorithm that comprehensively considers time decay and abnormal deformation weight, and combining a deformation trend prediction model including multiple parameters such as amplitude, decay, frequency, and phase, the precise analysis of tower inclination deformation is realized, and the ability of the monitoring system to predict the deformation trend of the tower is improved. Description of the Drawings

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0053] Figure 1 It is a flowchart of a tower inclination monitoring method based on rapid point cloud reconstruction of the present invention;

[0054] Figure 2Structural diagram of a tower inclination monitoring system based on rapid point cloud reconstruction of the present invention. Detailed implementation manners

[0055] The following will describe clearly and completely the technical solutions in the embodiments of the present invention in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0056] Please refer to Figure 1 , the present invention provides a tower inclination monitoring method based on rapid point cloud reconstruction, including the following steps:

[0057] Obtain the original point cloud data of the tower through hierarchical scanning and adaptive sampling density method;

[0058] Perform downsampling, noise filtering and preliminary classification on the original point cloud data to obtain the tower structure point cloud data;

[0059] Extract the geometric features of the tower based on the tower structure point cloud data, and construct a three-dimensional model of the tower skeleton through the geometric features of the tower;

[0060] Calculate the inclination angles of each part of the tower based on the central axis of the three-dimensional model of the tower skeleton;

[0061] Based on the inclination angles of each part of the tower, perform deformation analysis and deformation prediction on the tower, and generate a tower inclination monitoring report.

[0062] Specifically, in this embodiment, high-quality point cloud data is obtained by using hierarchical scanning and adaptive sampling density method, accurate tower structure information is extracted by combining data processing methods such as downsampling and noise filtering, and accurate modeling is realized through geometric feature extraction and three-dimensional model construction of the skeleton. Furthermore, the inclination angle is calculated based on the central axis and deformation analysis and prediction are carried out, and finally a monitoring report is generated, realizing the full-process deformation monitoring from data acquisition, processing, modeling to analysis and prediction, and improving the efficiency of tower inclination monitoring.

[0063] The obtaining of the original point cloud data of the tower through hierarchical scanning and adaptive sampling density method specifically includes:

[0064] Divide the tower into multiple scanning layers according to the tower height, set an overlapping area between adjacent scanning layers, and use a bottom-up scanning method to obtain the point cloud data of each scanning layer;

[0065] The layer height of the scanning layer is adaptively determined according to the tower structure characteristics, and the layer height range is 3-5 meters;

[0066] The overlapping area between adjacent scanning layers accounts for 20%-30% of the layer height;

[0067] Feature points are set within the overlapping area for the registration of adjacent scanning layers;

[0068] Adaptive sampling density method is used for data acquisition of each scanning layer, and the sampling density gradually decreases from the bottom scanning layer to the top scanning layer;

[0069] The sampling density of the bottom scanning layer is 5-10 points per square centimeter;

[0070] The sampling density of the top scanning layer is 1-3 points per square centimeter;

[0071] The sampling density of the middle scanning layer decreases linearly according to the height;

[0072] The emission power and scanning angle of the laser scanner are dynamically adjusted according to weather conditions and scanning distance.

[0073] Specifically, in this embodiment, by dividing the pole tower into multiple scanning layers and setting overlapping areas, the integrity and continuity of point cloud data acquisition are realized: the layer height setting of 3-5 meters ensures the fineness of each layer of scanning, the overlapping area setting of 20%-30% ensures the effective connection of data between adjacent layers, and the feature points within the overlapping area provide a reliable reference for the registration of adjacent layers.

[0074] By adopting the adaptive sampling density method, an optimal balance between acquisition efficiency and data quality is achieved: the high-density sampling of 5-10 points per square centimeter at the bottom ensures the precise monitoring of the basic part, the lower-density sampling of 1-3 points per square centimeter at the top improves the acquisition efficiency, and the linear decrease of the density in the middle area not only ensures the data quality but also saves the acquisition time.

[0075] By dynamically adjusting the emission power and scanning angle of the laser scanner: adjusting according to weather conditions ensures stable acquisition in different environments, and adjusting according to scanning distance guarantees the quality reliability of the acquired data.

[0076] The original point cloud data is downsampled, noise filtered and preliminarily classified to obtain the pole tower structure point cloud data, which specifically includes:

[0077] Adaptive voxel filtering is used to downsample the original point cloud data, and the statistical outlier analysis method is used to remove noise points to obtain the preprocessed point cloud data;

[0078] The voxel size of voxel filtering is adaptively adjusted according to the point cloud density, and the voxel size range is 5-15 centimeters;

[0079] The statistical outlier analysis adopts the K-nearest neighbor method, the K value is taken as 20-30, and the standard deviation coefficient is set to 2.0;

[0080] Verify the data integrity of the preprocessing results to ensure that the point cloud retention rate of key structural parts is not less than 85%.

[0081] Based on the spatial distribution characteristics and reflection intensity characteristics of the preprocessed point cloud data, adopt a multi-feature fusion classification method to extract the point cloud data of the tower structure.

[0082] Extract the spatial geometric features of the preprocessed point cloud data, including normal vectors, curvatures, and local surface features.

[0083] Combine the reflection intensity values of the point cloud to construct a multi-dimensional feature vector.

[0084] Use the support vector machine algorithm for point cloud classification, and the classification accuracy is not less than 90%.

[0085] Use the region growing algorithm to optimize the classification results and eliminate isolated points and misclassified points.

[0086] Specifically, in this embodiment, through a combined processing method of adaptive voxel filtering and statistical outlier analysis, high-quality point cloud data preprocessing is achieved: an adaptive voxel size of 5 - 15 cm ensures the rationality of downsampling, a nearest neighbor analysis with a K value of 20 - 30 guarantees the accuracy of noise point removal, a standard deviation coefficient setting of 2.0 provides a suitable filtering threshold, and a key structural point cloud retention rate of more than 85% ensures the integrity of the data.

[0087] Adopting a multi-feature fusion classification method improves the extraction accuracy of the point cloud of the tower structure: combining spatial geometric features (normal vector, curvature, local surface features), fusing the point cloud reflection intensity value to construct a multi-dimensional feature vector, and the support vector machine algorithm achieves a classification accuracy of more than 90%.

[0088] Optimizing the classification results through the region growing algorithm effectively eliminates isolated points, corrects misclassified points, and improves the quality of the structural point cloud data.

[0089] The method for extracting the geometric features of the tower based on the point cloud data of the tower structure, constructing a three-dimensional model of the tower skeleton through the geometric features of the tower, specifically includes:

[0090] Use the region growing algorithm to extract key geometric feature points from the point cloud data of the tower structure, and identify the main structural components of the tower and the connection relationships of the main structural components through curvature analysis and edge detection methods.

[0091] Use adaptive region growing parameters to dynamically adjust the threshold of region growing according to the point cloud density in different height regions of the tower, ensuring that detailed features can be accurately extracted in high-density regions, while maintaining the stability of feature extraction in low-density regions.

[0092] The threshold adjustment calculation formula is:

[0093]

[0094] Among them, T 0 is the initial threshold, λ is the adjustment coefficient, and D i is the point cloud density of the i-th layer area of the pole tower, H is the height of the pole tower, and T i is the region growing threshold of the i-th layer area of the pole tower;

[0095] Based on the key geometric feature points, a preliminary skeleton model of the pole tower is constructed using the Delaunay triangulation algorithm. By adjusting the connection relationship of the triangular mesh, the redundant structure in the preliminary skeleton model is reduced, and the simplicity of the model is improved. A multi-scale filtering method is used to perform smoothing processing on the preliminary skeleton model at different scales to eliminate small-range deviations caused by point cloud data noise, and a pole tower skeleton model is obtained to ensure the overall smoothness and continuity of the skeleton model;

[0096] The smoothing coefficient of multi-scale filtering is adjusted according to the local curvature, and the calculation formula of the smoothing coefficient is:

[0097] σ j = σ 0 ×(1 + α·tanh(βC j ));

[0098] Among them, σ 0 is the basic smoothing coefficient, α is the amplitude adjustment coefficient, β is the sensitivity adjustment coefficient, C j is the curvature value of the j-th geometric feature point, tanh(βC j ) is the hyperbolic tangent activation function, and σ j is the smoothing coefficient of the j-th geometric feature point.

[0099] Specifically, in this embodiment, an accurate identification of the pole tower structure is realized through a geometric feature extraction method based on the region growing algorithm: the main structural components are identified by curvature analysis and edge detection, the connection relationship between the structural components is determined, and the adaptive region growing parameters ensure the quality of feature extraction in different height regions.

[0100] An adaptive threshold adjustment strategy is adopted to optimize the feature extraction effect in different regions: detailed features can be accurately extracted in high-density regions, the stability of feature extraction is maintained in low-density regions, and the threshold is dynamically adjusted through the calculation formula to adapt to the point cloud density characteristics in different height regions.

[0101] The method for constructing a skeleton model based on Delaunay triangulation and multi-scale filtering improves the model quality: reduces the redundant structures in the preliminary skeleton model, enhances the simplicity of the model, eliminates the deviation caused by noise through multi-scale filtering, and dynamically adjusts the smoothing coefficient according to the local curvature to ensure the overall smoothness and continuity of the skeleton model.

[0102] Based on the central axis of the three-dimensional model of the pole tower skeleton, calculate the inclination angles of each part of the pole tower, specifically including:

[0103] Adopt the improved principal component analysis method combined with the iterative closest point algorithm to extract the central axis of the pole tower skeleton model, and optimize the spatial position of the central axis through the multi-level feature fusion method to improve the axis positioning accuracy;

[0104] Perform layered slicing on the skeleton model, extract the center of gravity points for each layer of slicing, use the weighted least squares method to fit the sequence of center of gravity points, initially determine the central axis, and through the iterative optimization algorithm, combine the normal vector consistency and curvature features to finely adjust the position of the central axis. Adopt the multi-scale feature fusion method, combine the local geometric information and the global structure features to improve the robustness of axis extraction;

[0105] Based on the adaptive segmentation calculation method, divide the pole tower into multiple calculation units, calculate the angles between the central axes of each calculation unit and the vertical direction to obtain the local inclination angles;

[0106] Based on the pole tower structure characteristics and deformation degree, adaptively determine the size of the calculation unit, and the size range is dynamically adjusted according to the pole tower height and local point cloud density; within each calculation unit, use the weighted projection method to calculate the local inclination angle, where the weight coefficient is adaptively adjusted according to the point cloud density and geometric features; introduce the spatial smoothing constraint, use the inclination angle information of adjacent units, and eliminate the angle mutation between adjacent calculation units through the weighted average method; establish a confidence evaluation mechanism for the inclination angle, combine the integrity of the point cloud data and the local feature stability, and mark the suspicious calculation results.

[0107] Specifically, in this embodiment, the improved principal component analysis method and the iterative closest point algorithm are used to extract the central axis, achieving high-precision axis positioning: the multi-level feature fusion method is used to optimize the spatial position of the axis, the center of gravity point sequence is extracted by layered slicing, fine adjustment is carried out by combining the normal vector consistency and curvature features, and the multi-scale feature fusion is used to improve the robustness of axis extraction.

[0108] Adopt the adaptive segmentation calculation method to improve the accuracy of inclination angle calculation: dynamically adjust the size of the calculation unit according to the pole tower structure characteristics, adaptively adjust the weight coefficient based on the point cloud density and geometric features, calculate the local inclination angle through the weighted projection method, and introduce the spatial smoothing constraint to eliminate the angle mutation.

[0109] A perfect tilt angle confidence evaluation mechanism is established. Combining the evaluation of the integrity of point cloud data and considering the stability of local features, suspicious calculation results are marked.

[0110] The calculation formula of the objective function for optimizing the central axis is as follows:

[0111]

[0112] where E(·) is the objective function, L is the central axis, is the k 1 th centroid point, is the weight coefficient of the k 1 th centroid point, is the point to the perpendicular distance from the central axis L, λ 1 is the balance coefficient, n is the number of centroid points, m is the number of segments of the central axis, is the slope change of the central axis in the k 2 th segment, is the tilt angle of the k 3 th segment, μ is the tilt angle change smoothing coefficient;

[0113] The calculation formula of the local tilt angle is as follows:

[0114]

[0115] where, is the tilt angle of the a th calculation unit, is the local axis direction vector of the a a th calculation unit, is the local curvature value of the a th calculation unit,

[0116] Specifically, in this embodiment, by introducing the objective function, combining the weight coefficient of the centroid point with the perpendicular distance from the point to the central axis, the accuracy and smoothness of the axis fitting are effectively balanced, the positioning error is reduced, considering the slope change and tilt angle change of the central axis segments, ensuring the accurate positioning and continuity of the axis in a complex structural environment, and introducing the objective function enhances the robustness of the central axis in multi-segment changes and improves the consistency of the overall axis extraction.

[0117] By adopting the calculation method of the local axis direction vector and the vertical reference direction vector, the inclination angles of each calculation unit are accurately measured to ensure the accuracy of angle measurement. A reliability weight coefficient is introduced to dynamically adjust the sensitivity of the inclination angle according to the local curvature value, improving the response ability to structural changes. By applying the curvature sensitivity coefficient, the adaptability to different curvature regions is enhanced, ensuring the applicability and accuracy of the inclination angle calculation in various tower structures.

[0118] By combining the central axis optimization and the local inclination angle calculation, high-precision inclination monitoring data are achieved. Through the smoothing constraint and weighted average method, the angle mutation between calculation units is eliminated, ensuring the continuity of the monitoring results. A confidence evaluation mechanism for the inclination angle is established, which can effectively mark and identify suspicious calculation results, enhancing the credibility and practical value of the monitoring system.

[0119] Based on the inclination angles of each part of the tower, deformation analysis and deformation prediction of the tower are carried out, and a tower inclination monitoring report is generated, specifically including:

[0120] Adopt a deformation feature extraction algorithm based on deep learning to extract the deformation features of the tower from the inclination angle data of each part of the tower, and combine the historical monitoring data and real-time monitoring data to analyze the deformation trend of the tower;

[0121] The calculation formula of the deformation feature extraction algorithm is:

[0122]

[0123] Among them, F is the comprehensive deformation feature value, is the inclination angle at the b-th moment, is the inclination angle at the (b - 1)-th moment, γ is the time decay coefficient, t b is the time interval between the b-th moment and the current moment, δ is the abnormal deformation weight coefficient, is the inclination angle change amount of the f-th abnormal deformation point, B is the total sampling duration of the deformation analysis, and p is the number of abnormal deformation points;

[0124] Through the deformation trend prediction model, the deformation of the tower is predicted, and a tower inclination monitoring report is generated according to the prediction results, providing detailed deformation analysis and warning information;

[0125] The calculation formula of the deformation trend prediction model is:

[0126]

[0127] Among them, Δφ(t) is the inclination angle change amount after time t, φ 0 is the initial inclination angle, α q is the q-th amplitude coefficient, βq is the q-th attenuation coefficient, ω q is the q-th frequency coefficient, δ q is the q-th phase offset, Q is the number of terms of the deformation trend prediction model, and κ is the deformation trend adjustment coefficient.

[0128] Specifically, in this embodiment, the accurate extraction of the pole tower deformation characteristics is realized through the comprehensive deformation eigenvalue F, the introduction of the time attenuation coefficient ensures the reasonable weight distribution of historical data, the setting of the abnormal deformation weight coefficient enhances the sensitivity to abnormal deformation, and the effective fusion analysis of historical monitoring data and real-time monitoring data is realized.

[0129] The polynomial prediction model is adopted, which includes multiple parameters such as amplitude, attenuation, frequency and phase. The adaptive optimization of the prediction model is realized through the deformation trend adjustment coefficient, which can accurately capture the periodic characteristics of the pole tower tilt deformation and provide a quantitative prediction result of the change amount of the tilt angle.

[0130] Please refer to Figure 2 , the present invention also provides a pole tower tilt monitoring system based on point cloud rapid reconstruction, and the system includes:

[0131] A data acquisition module, which is used to obtain the original point cloud data of the pole tower through hierarchical scanning and the adaptive sampling density method;

[0132] A data processing module, which is used to perform downsampling, noise filtering and preliminary classification on the original point cloud data to obtain the pole tower structure point cloud data;

[0133] A feature extraction module, which is used to extract the geometric features of the pole tower based on the pole tower structure point cloud data, and construct a three-dimensional model of the pole tower skeleton through the geometric features of the pole tower;

[0134] An inclination calculation module, which is used to calculate the inclination angles of each part of the pole tower based on the central axis of the three-dimensional model of the pole tower skeleton;

[0135] A deformation prediction module, which is used to perform deformation analysis and deformation prediction on the pole tower based on the inclination angles of each part of the pole tower, and generate a pole tower tilt monitoring report.

[0136] Specifically, a pole tower tilt monitoring system based on point cloud rapid reconstruction in this embodiment realizes the automatic acquisition of high-quality original data through the data acquisition module; ensures the accuracy and availability of the point cloud data through the data processing module; ensures the accurate construction of the pole tower geometric features and the skeleton model through the feature extraction module; realizes the high-precision inclination angle calculation through the inclination calculation module; and provides a reliable deformation analysis function through the deformation prediction module.

[0137] The present invention also discloses an electronic device, comprising: at least one processor, at least one memory, a communication interface and a bus: wherein, the processor, the memory and the communication interface complete communication with each other through the bus; the memory stores program instructions executable by the processor, and the processor invokes the program instructions to implement a tower inclination monitoring method based on rapid point cloud reconstruction.

[0138] The present invention also discloses a computer-readable storage medium, which stores computer instructions that enable the computer to implement all or part of the steps of the tower inclination monitoring method based on rapid point cloud reconstruction according to the embodiments of the present invention. The storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory ROM, random access memory RAM, magnetic disks or optical discs that can store program codes.

[0139] The foregoing is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A tower tilt monitoring method based on point cloud rapid reconstruction, characterized in that: The following steps are involved: Obtain the original point cloud data of the tower through layered scanning and adaptive sampling density method; Downsample, filter and preliminarily classify the original point cloud data to obtain the tower structure point cloud data; Extracting the geometric features of the tower based on the tower structure point cloud data, and constructing a three-dimensional model of the tower skeleton through the geometric features of the tower; Calculating the inclination angle of each part of the tower based on the central axis of the three-dimensional model of the tower skeleton; Based on the inclination angles of various parts of the tower, deformation analysis and deformation prediction are performed on the tower to generate a tower inclination monitoring report.

2. A tower tilt monitoring method based on point cloud rapid reconstruction as claimed in claim 1, characterized in that: The original point cloud data of the tower is obtained by layered scanning and adaptive sampling density method, specifically including: The tower is divided into multiple scanning layers according to its height, and overlapping areas are set between adjacent scanning layers. The point cloud data of each scanning layer is obtained by scanning from bottom to top. An adaptive sampling density method is used to collect data for each scanning layer, and the sampling density is gradually reduced from the bottom scanning layer to the top scanning layer.

3. A tower tilt monitoring method based on point cloud rapid reconstruction as claimed in claim 2, characterized in that: The original point cloud data is downsampled, noise filtered and preliminarily classified to obtain the tower structure point cloud data, specifically including: Adaptive voxel filtering is used to downsample the original point cloud data, and the statistical outlier analysis method is used to remove noise points to obtain preprocessed point cloud data; Based on the spatial distribution characteristics and reflection intensity characteristics of the preprocessed point cloud data, a multi-feature fusion classification method is used to extract the tower structure point cloud data.

4. The method for monitoring the inclination of a tower based on rapid reconstruction of point clouds according to claim 1, characterized in that: The step of extracting the tower geometric features based on the tower structure point cloud data and constructing the tower skeleton three-dimensional model through the tower geometric features specifically includes: The key geometric feature points are extracted from the tower structure point cloud data based on the region growing algorithm, and the main structural components of the tower and the connection relationship of the main structural components are identified through curvature analysis and edge detection methods; Adaptive region growing parameters are used to dynamically adjust the region growing threshold according to the point cloud density of different tower height areas; Based on the key geometric feature points, a preliminary skeleton model of the tower is constructed using the Delaunay triangulation algorithm. The preliminary skeleton model is smoothed at different scales using a multi-scale filtering method to eliminate small-scale deviations caused by point cloud data noise, thereby obtaining a tower skeleton model.

5. The method for monitoring the inclination of a tower based on rapid reconstruction of point clouds according to claim 4, characterized in that: The calculating the inclination angles of various parts of the tower based on the central axis of the three-dimensional model of the tower skeleton specifically includes: The improved principal component analysis method combined with the iterative closest point algorithm is used to extract the central axis of the tower skeleton model. The spatial position of the central axis is optimized through the multi-level feature fusion method to improve the axis positioning accuracy. Based on the adaptive segmented calculation method, the tower is divided into multiple calculation units, and the angle between the central axis of each calculation unit and the vertical direction is calculated to obtain the local inclination angle.

6. A tower tilt monitoring method based on point cloud rapid reconstruction as claimed in claim 5, characterized in that: The objective function calculation formula for the central axis optimization is: Among them, E(·) is the objective function, L is the central axis, is the k1th centroid point, is the weight coefficient of the k1th centroid point, For point The vertical distance to the central axis L, λ1 is the balance coefficient, n is the number of center of gravity points, m is the number of central axis segments, is the slope change of the central axis in the k2th segment, is the inclination angle of the k3th segment, μ is the smoothing coefficient of the inclination angle change; The calculation formula of the local tilt angle is: in, is the tilt angle of the ath computing unit, is the local axis direction vector of the ath computational unit, is the vertical reference direction vector, γ a is the reliability weight coefficient of the ath computing unit, is the local curvature value of the ath calculation unit, is the curvature sensitivity coefficient.

7. The method for monitoring the inclination of a tower based on rapid reconstruction of point clouds according to claim 1, characterized in that: The tower is subjected to deformation analysis and deformation prediction based on the inclination angles of various parts of the tower, and a tower inclination monitoring report is generated, which specifically includes: A deformation feature extraction algorithm based on deep learning is used to extract the deformation features of the tower from the inclination angle data of each part of the tower. The deformation trend of the tower is analyzed by combining historical monitoring data and real-time monitoring data. The calculation formula of the deformation feature extraction algorithm is: Among them, F is the comprehensive deformation characteristic value, is the inclination angle at the bth moment, is the tilt angle at the b-1th moment, γ is the time attenuation coefficient, t b is the time interval between the bth moment and the current moment, δ is the abnormal deformation weight coefficient, is the change in the tilt angle of the fth abnormal deformation point, B is the total sampling time of deformation analysis, and p is the number of abnormal deformation points; The deformation trend prediction model is used to predict the tower deformation, and a tower tilt monitoring report is generated based on the prediction results. The calculation formula of the deformation trend prediction model is: Among them, Δφ(t) is the change in tilt angle after time t, φ0 is the initial tilt angle, α q is the qth amplitude coefficient, β q is the qth attenuation coefficient, ω q is the qth frequency coefficient, δ q is the qth phase offset, Q is the number of terms in the deformation trend prediction model, and κ is the deformation trend adjustment coefficient.

8. A tower tilt monitoring system based on point cloud rapid reconstruction, characterized in that: The system comprises: A data acquisition module is used to acquire the original point cloud data of the tower through layered scanning and adaptive sampling density method; The data processing module is used to downsample, filter noise and perform preliminary classification on the original point cloud data to obtain the tower structure point cloud data; A feature extraction module is used to extract the geometric features of the tower based on the tower structure point cloud data, and to construct a three-dimensional model of the tower skeleton through the geometric features of the tower; A tilt calculation module, used to calculate the tilt angle of each part of the tower based on the central axis of the three-dimensional model of the tower skeleton; The deformation prediction module is used to analyze and predict the deformation of the tower based on the inclination angles of various parts of the tower, and generate a tower inclination monitoring report.

9. An electronic device, characterized in that: include: at least one processor, at least one memory, a communication interface, and a bus; The processor, memory, and communication interface communicate with each other via the bus, the memory stores program instructions executable by the processor, and the processor calls the program instructions to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions enable a computer to implement the method according to any one of claims 1 to 7.

Images