Intelligent three-dimensional terrain modeling method and system based on surveying and mapping data

Abstract

The present invention discloses an intelligent three-dimensional terrain modeling method and system based on surveying and mapping data, which relates to the field of surveying and mapping technology. After completing local refinement and reconstruction, the method performs comprehensive calculation based on the local terrain complexity index LTCI and the terrain feature line stability index FLSI to output the overall modeling quality index GMQI, and performs a secondary comparative evaluation based on the overall modeling quality index GMQI result. When the overall modeling quality index GMQI evaluation result shows that the overall modeling quality does not meet the standard, the refinement and reconstruction is automatically triggered iteratively, and the point density is gradually optimized according to the strategy of increasing by 40% each time until the overall modeling quality meets the standard, thereby achieving global modeling quality control in the infrastructure site selection area in the mountainous area. This method not only improves the consistency and detail accuracy of the modeling, but also effectively avoids modeling deviations caused by local terrain mutations or structural instability, and improves the reliability and safety of the final three-dimensional terrain model in engineering design applications.

Description

Technical Field

[0001] The present invention relates to the field of surveying and mapping technology, and in particular to an intelligent three-dimensional terrain modeling method and system based on surveying and mapping data. Background Art

[0002] The present invention relates to the field of surveying and mapping engineering, and more particularly to three-dimensional terrain modeling technology based on remote sensing measurement data processing, and more specifically to an intelligent three-dimensional terrain modeling method based on surveying and mapping data. With the rapid development of unmanned aerial vehicle laser radar (LiDAR) technology, photogrammetry technology, and three-dimensional reconstruction algorithms, traditional surveying and mapping operations are gradually evolving towards automation and intelligence. In mountain infrastructure construction, such as dams, tunnels, and mountain road site selection, it is required to accurately and quickly establish high-precision three-dimensional terrain models in complex terrain environments to support subsequent project site selection, safety assessment, and structural design. Therefore, how to use large-scale, high-precision surveying and mapping data to intelligently and efficiently generate stable and accurate three-dimensional terrain models has become a key technical problem that urgently needs to be overcome in this field.

[0003] At present, the existing three-dimensional terrain modeling methods mainly rely on the overall deduction and surface reconstruction of point cloud data sets, and lack the active perception and utilization of the internal structural characteristics of the terrain. When faced with complex landforms such as cliffs, river valleys, and landslides, traditional methods are usually unable to effectively extract key terrain structural information such as fault lines and slope breaks, resulting in large accuracy deviations and structural distortions in the terrain models in these areas. In addition, due to the failure to dynamically respond to the complexity and stability of the terrain, there is often excessive modeling of simple areas and insufficient modeling of complex areas, resulting in a waste of modeling resources, and seriously affecting the accuracy and reliability of infrastructure site selection, and increasing the risks of subsequent engineering design and construction;

[0004] The above-mentioned current problems are mainly due to the fact that existing methods generally use point set distribution as the core to drive modeling, and lack the active extraction and in-depth analysis of key structural information such as terrain feature lines such as fault lines, slope break lines, and high-curvature continuous lines. This leads to insufficient identification of terrain change trends, cross-section continuity, and potential geological risk areas. When complex terrain areas are not accurately modeled, it may lead to errors in infrastructure site selection and bury engineering hidden dangers, such as dam foundations falling on fault zones and tunnels passing through landslide areas. In severe cases, it may even cause major accidents such as structural instability and geological disasters. Therefore, there is an urgent need for a method that can actively extract terrain feature lines and perform structural perception optimization modeling based on surveying and mapping data to improve the quality of three-dimensional terrain modeling and engineering application reliability in infrastructure site selection areas in mountainous areas. Summary of the Invention

[0005] In view of the deficiencies of the prior art, the present invention provides an intelligent three-dimensional terrain modeling method and system based on surveying and mapping data, which solves the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: comprising the following steps:

[0007] S1. Use a drone equipped with LiDAR to set up collection points to select infrastructure sites in mountainous areas, conduct aerial mapping, obtain 3D point cloud data for terrain modeling, build a mapping server, and transmit the modeling geometry data to the mapping server.

[0008] S2. Extract features from the terrain modeling 3D point cloud data in the surveying and mapping server to obtain terrain feature lines and surveying and mapping data, and pre-process the surveying and mapping data to obtain a standard surveying and mapping data set;

[0009] S3. Based on the standard surveying and mapping data set, calculate and output the local terrain complexity index (LTCI). Then, perform a preliminary comparative evaluation of the LTCI output results of all the collected points to determine the complexity of the site selection. Based on the preliminary comparative evaluation results, trigger the feature line depth analysis.

[0010] S4, the feature line depth analysis is performed by calculating and outputting a terrain feature line stability index FLSI based on a standard surveying and mapping data set, and based on the terrain feature line stability index FLSI, the stability of the terrain feature line is judged, and the unstable area is refined and reconstructed;

[0011] S5. Based on the local terrain complexity index LTCI and the terrain feature line stability index FLSI, a summary calculation is performed to output the overall modeling quality index GMQI. Based on the output results of the overall modeling quality index GMQI, a secondary comparative evaluation is performed to comprehensively analyze the complexity of the terrain and the stability of the feature lines.

[0012] Preferably, said S1 includes S11 and S12;

[0013] S11. Use a drone equipped with LiDAR, and refer to existing engineering surveying and mapping standards. Set the density of collection points based on the drone's flight altitude, speed, laser emission frequency, and image capture route. Initially set at 25 collection points per square meter, conduct aerial mapping of infrastructure sites in mountainous areas, and obtain terrain modeling 3D point cloud data for each collection point.

[0014] The terrain modeling three-dimensional point cloud data includes local point cloud three-dimensional coordinates (x, y, z), where x represents the horizontal axis coordinate, y represents the vertical axis coordinate, and z represents the vertical axis coordinate;

[0015] S12. Build a surveying and mapping server, wirelessly connect the surveying and mapping server to the drone through the 5G communication network, and transmit the real-time terrain modeling 3D point cloud data to the surveying and mapping server.

[0016] Preferably, said S2 includes S21 and S22;

[0017] S21. In the surveying and mapping server, feature extraction is performed on the local point cloud 3D coordinates (x, y, z) of the terrain modeling 3D point cloud data acquired in real time to obtain terrain feature lines and surveying and mapping data;

[0018] The surveying and mapping data includes the local point cloud three-dimensional coordinates (x, y, z), the local curvature change rate Qc, the local normal vector change gradient ▽n, the local elevation fluctuation index Bh, the point cloud density gradient ▽p and the characteristic line connection logarithm L;

[0019] The local curvature change rate Qc is obtained by using KD-terr fast domain search for each local point cloud three-dimensional coordinate (x, y, z) to search for all neighboring points of the local point cloud three-dimensional coordinate (x, y, z) within a radius r to form a domain N, and using the least squares method to perform local surface fitting in the domain N to obtain a fitting surface, and then obtaining the principal curvature K through the fitting surface, and then calculating the local principal curvature for each local point cloud three-dimensional coordinate (x, y, z) in the domain N, and then calculating the average absolute value of the difference between the local point cloud three-dimensional coordinate (x, y, z) and the principal curvature in the domain N;

[0020] The local normal vector change gradient ▽n is obtained by performing ICA principal component analysis on the domain N of the three-dimensional coordinates (x, y, z) of each local point cloud, extracting the main direction vector as the normal vector n, performing dot multiplication on each pair of normal vectors n in the domain N to calculate the angle difference, and then calculating the average change rate based on the angle difference;

[0021] The local elevation fluctuation index Bh is obtained by retrieving the three-dimensional coordinates (x, y, z) of each local point cloud, the area N within the radius r, taking the vertical axis coordinate z of the three-dimensional coordinates (x, y, z) of each local point cloud within the area N, and calculating the standard deviation of these vertical axis coordinates z;

[0022] The point cloud density gradient ▽p is obtained by calculating the ratio of the number of points in the domain N to the volume of the domain sphere to obtain the local density, and performing average difference calculation based on the local density;

[0023] Extract terrain feature lines based on the local curvature change rate Qc, local normal vector change gradient ▽n and local elevation fluctuation index Bh of each local point cloud three-dimensional coordinate (x, y, z);

[0024] The terrain characteristic lines include fault lines, slope break lines and high curvature continuous lines;

[0025] The number of feature line connection pairs L is obtained by analyzing the number of adjacent point pairs of the local point cloud three-dimensional coordinates (x, y, z) on the terrain feature line;

[0026] S22. Preprocessing the acquired surveying and mapping data in a surveying and mapping server, wherein the preprocessing includes cleaning noise and normalizing the data to obtain a standard surveying and mapping data set;

[0027] The noise cleaning is performed by using volume filtering and radius filtering on the acquired surveying and mapping data to remove outliers and clear erroneously extracted features;

[0028] The normalized data is normalized by using the mean-variance standard Z-score to eliminate the dimension effect of all parameters in the surveying and mapping data except the three-dimensional coordinates (x, y, z) of the local point cloud;

[0029] At the same time, coordinate normalization processing is used to convert the three-dimensional coordinates (x, y, z) of the local point cloud into digital parameters to eliminate the influence of dimension.

[0030] Preferably, the S3 includes S31, calculating and outputting the local terrain complexity index LTCI of all current local point cloud three-dimensional coordinates (x, y, z) based on the acquired standard surveying and mapping data set, to measure the local terrain mutation, directional instability, small elevation change amplitude and density anomaly;

[0031] The local terrain complexity index LTCI is calculated and outputted by the following algorithm formula:

[0032] ;

[0033] Where LTCI(i) represents the local terrain complexity index of the i-th collection point, Qc(i) represents the local curvature change rate of the i-th collection point, ▽n(i) represents the local normal vector change gradient of the i-th collection point, Bh(i) represents the local elevation fluctuation index of the i-th collection point, ▽p(i) represents the point cloud density gradient of the i-th collection point, Represents a small positive number that is protected against division by zero.

[0034] Preferably, the S3 further includes S32, traversing the local terrain complexity index LTCI of all collection points, and performing a preliminary comparative evaluation based on the local terrain complexity index LTCI results output by each collection point, determining the complexity of each collection point, and classifying the current mountain infrastructure site selection based on the complexity. The specific evaluation content is as follows;

[0035] When the local terrain complexity index LTCI (i) of the i-th collection point is less than 1.5, the current collection point is classified as a first-level complex point;

[0036] When 1.5≤the local terrain complexity index LTCI(i) of the i-th collection point is less than 2.5, the current collection point is classified as a secondary complex point;

[0037] When the local terrain complexity index LTCI (i) of the i-th collection point is ≥ 2.5, the current collection point is divided into a level 3 complex point;

[0038] Based on the preliminary comparison and evaluation, the total number of third-level complex points and the ratio of the total number of current collection points i are evaluated, and the trigger judgment condition Trigger is obtained to determine the triggering condition of the feature line depth analysis;

[0039] The specific judgment content of the trigger judgment condition Trigger is as follows:

[0040] ;

[0041] Among them, when the trigger judgment condition Trigger=1, the feature line depth analysis is triggered; Nhigh-LTCI∈{LTCI>2.5} represents the total number of level 3 complex points of the local terrain complexity index LTCI of the collection point; F1 represents the trigger ratio threshold, which is set by the user. When the trigger judgment condition Trigger=0, the feature line depth analysis is not triggered and standard modeling is performed.

[0042] Preferably, the S4 includes S41, after triggering the feature line depth analysis, extracting the three-dimensional coordinates (x, y, z) of the local point cloud and the standard mapping data set, analyzing the terrain feature line stability index FLSI of each pair of adjacent terrain feature lines to quantify the stability of a geological feature line in elevation change and direction change;

[0043] The terrain feature line stability index FLSI is calculated and output by the following algorithm formula:

[0044] ;

[0045] Where i represents the i-th acquisition point, j represents the j-th acquisition point, FLSI(l) represents the terrain characteristic line stability index of the l-th terrain characteristic line, represents the partial derivative, x i 、y i and z i Respectively represent the horizontal axis coordinate, vertical axis coordinate and vertical axis coordinate of the i-th acquisition point, x j 、y j and z j Respectively represent the horizontal axis coordinate, vertical axis coordinate and vertical axis coordinate of the jth acquisition point, It means that both acquisition points i and j belong to the lth terrain feature line and acquisition point i is not equal to acquisition point j, ▽n(j) represents the local normal vector change gradient of the jth acquisition point, and ▽n(i) represents the local normal vector change gradient of the ith acquisition point.

[0046] Preferably, the S4 further includes S41, based on the output result of the terrain feature line stability index FLSI, stability assessment, judging the stability of the terrain feature line, and performing detailed reconstruction on the unstable area, and the specific assessment content is as follows;

[0047] When the terrain feature line stability index FLSI ≤ 0.8, it means that the terrain feature line is stable and the model is directly built;

[0048] When the terrain feature line stability index FLSI>0.8, it means that the terrain feature line is unstable, and local refinement reconstruction is triggered;

[0049] The local refinement reconstruction is performed by restarting the UAV to perform aerial photography and mapping, reconstructing the current terrain feature lines, and initially setting the density of acquisition points to 25 per square, which is increased by 40%, and re-executing feature extraction and outputting the terrain feature line stability index FLSI and the local terrain complexity index LTCI.

[0050] Preferably, the S5 includes S51, performing comprehensive calculation based on the terrain feature line stability index FLSI obtained after local refinement and reconstruction, combined with the local terrain complexity index LTCI, to output the overall modeling quality index GMQI to measure the overall modeling quality;

[0051] The overall modeling quality index GMQI is calculated and output by the following algorithm formula:

[0052] ;

[0053] Where N represents the total number of infrastructure site selection areas in mountainous areas, R represents the infrastructure site selection area in mountainous areas, and LRCI avg (R) represents the mean of all local terrain complexity indices LTCI within the R mountain infrastructure site selection area, FLSI avg (R) represents the mean value of the FLSI of all terrain feature line stability indexes in the R mountain infrastructure site selection area.

[0054] Preferably, the S5 further includes S52, performing a secondary comparative evaluation based on the output result of the overall modeling quality index GMQI, comprehensively analyzing the complexity and characteristic line stability after local refinement and reconstruction, and evaluating the overall modeling quality. The specific evaluation content is as follows;

[0055] When the overall modeling quality index GMQI ≤ 2, it means that the overall modeling quality meets the standard, and the modeling is prompted;

[0056] When the overall modeling quality index GMQI>2, it means that the overall modeling quality does not meet the standard. At this time, the iteration triggers local refinement and reconstruction. Each iteration increases the acquisition point density by 40% based on the previous one until the overall modeling quality meets the standard and the iteration stops.

[0057] An intelligent 3D terrain modeling system based on surveying and mapping data, including a point cloud data acquisition module, a surveying and mapping data processing module, a terrain complexity analysis module, a terrain feature line stabilization module, and an overall modeling and analysis module;

[0058] The point cloud data acquisition module uses a drone equipped with LiDAR to set up collection points to select the site of infrastructure in mountainous areas, conduct aerial mapping, obtain three-dimensional point cloud data for terrain modeling, build a mapping server, and transmit the modeling geometry data to the mapping server;

[0059] The surveying and mapping data processing module extracts features from the terrain modeling three-dimensional point cloud data in the surveying and mapping server to obtain terrain feature lines and surveying and mapping data, and pre-processes the surveying and mapping data to obtain a standard surveying and mapping data set;

[0060] The terrain complexity analysis module calculates and outputs the local terrain complexity index (LTCI) based on a standard surveying and mapping dataset, and performs preliminary comparative evaluation on the output results of the local terrain complexity index (LTCI) of all collected points to determine the complexity of the site selection, and triggers feature line depth analysis based on the preliminary comparative evaluation results.

[0061] The terrain feature line stabilization module calculates and outputs a terrain feature line stability index FLSI based on the feature line depth analysis and a standard surveying and mapping data set, and judges the stability of the terrain feature line based on the terrain feature line stability index FLSI, and refines and reconstructs the unstable area;

[0062] The overall modeling and analysis module performs summary calculations based on the local terrain complexity index LTCI and the terrain feature line stability index FLSI, outputs the overall modeling quality index GMQI, and performs a secondary comparative evaluation based on the output results of the overall modeling quality index GMQI to comprehensively analyze the complexity of the terrain and the stability of the feature lines.

[0063] The present invention provides an intelligent three-dimensional terrain modeling method and system based on surveying and mapping data. It has the following beneficial effects:

[0064] (1) This method uses drones equipped with LiDAR for aerial mapping and combines it with 5G communication networks for real-time transmission to transmit terrain modeling 3D point cloud data to a mapping server. This method can quickly and efficiently complete data collection for infrastructure site selection in large mountainous areas. Furthermore, feature extraction is performed on the terrain modeling 3D point cloud data in the mapping server to extract terrain feature lines such as fault lines, slope break lines, and high curvature continuous lines. At the same time, noise is cleaned and the mapping data is normalized to generate a standard mapping data set, making the subsequent modeling data foundation more stable, reducing noise interference, and improving the data quality and modeling accuracy of 3D terrain modeling.

[0065] (2) This method calculates and outputs the local terrain complexity index (LTCI) based on a standard surveying and mapping dataset. By traversing and preliminarily comparing and evaluating the local terrain complexity index (LTCI) of each collection point and classifying it according to a three-level complexity classification standard, it can accurately identify the terrain complexity distribution in mountainous sites. When the proportion of level 3 complex points exceeds the set threshold, the feature line depth analysis is automatically triggered. By further calculating the terrain feature line stability index (FLSI), the unstable feature line area is locally refined and reconstructed, increasing the point density by 40%, and achieving local encryption and fine-grained modeling of the terrain mutation area, thereby effectively optimizing the modeling accuracy and structural reliability of the complex area.

[0066] (3) After completing the local refinement and reconstruction, this method performs a comprehensive calculation based on the local terrain complexity index (LTCI) and the terrain feature line stability index (FLSI) to output the overall modeling quality index (GMQI), and then conducts a secondary comparative evaluation based on the overall modeling quality index (GMQI). When the overall modeling quality index (GMQI) evaluation result indicates that the overall modeling quality does not meet the standard, the refinement and reconstruction is automatically triggered iteratively, and the point density is gradually optimized by increasing by 40% each time until the overall modeling quality meets the standard, thereby achieving global modeling quality control in the mountain infrastructure site selection area. This method not only improves the consistency and detail accuracy of the modeling, but also effectively avoids modeling deviations caused by local terrain mutations or structural instability, thereby improving the reliability and safety of the final three-dimensional terrain model in engineering design applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 This is a flow chart of the intelligent three-dimensional terrain modeling method based on surveying and mapping data of the present invention;

[0068] Figure 2 This is a schematic diagram of the steps of the intelligent three-dimensional terrain modeling system based on surveying and mapping data of the present invention;

[0069] Figure 3 Schematic diagram of the evaluation process of the intelligent three-dimensional terrain modeling method based on surveying and mapping data of the present invention;

[0070] Figure 4 Schematic diagram of drone aerial photography and sampling point density setting of the present invention. DETAILED DESCRIPTION

[0071] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0072] Example 1:

[0073] See also Figure 1 、 Figure 3 and Figure 4 The present invention provides an intelligent three-dimensional terrain modeling method based on surveying and mapping data. To achieve the above purpose, the present invention is implemented through the following technical solutions: comprising the following steps:

[0074] S1. Use a drone equipped with LiDAR to set up collection points to select infrastructure sites in mountainous areas, conduct aerial mapping, obtain 3D point cloud data for terrain modeling, build a mapping server, and transmit the modeling geometry data to the mapping server.

[0075] S2. Extract features from the terrain modeling 3D point cloud data in the surveying and mapping server to obtain terrain feature lines and surveying and mapping data, and pre-process the surveying and mapping data to obtain a standard surveying and mapping data set;

[0076] S3. Based on the standard surveying and mapping data set, calculate and output the local terrain complexity index (LTCI). Then, perform a preliminary comparative evaluation of the LTCI output results of all the collected points to determine the complexity of the site selection. Based on the preliminary comparative evaluation results, trigger the feature line depth analysis.

[0077] S4. Feature line depth analysis calculates and outputs the terrain feature line stability index FLSI based on the standard surveying and mapping data set, and judges the stability of the terrain feature line based on the terrain feature line stability index FLSI, and refines and reconstructs the unstable area;

[0078] S5. Based on the local terrain complexity index LTCI and the terrain feature line stability index FLSI, a summary calculation is performed to output the overall modeling quality index GMQI. Based on the output results of the overall modeling quality index GMQI, a secondary comparative evaluation is performed to comprehensively analyze the complexity of the terrain and the stability of the feature lines.

[0079] In this embodiment, the method uses a drone equipped with a LiDAR to perform aerial mapping, and combines 5G communication technology to transmit terrain modeling three-dimensional point cloud data to a mapping server in real time. Feature extraction and data preprocessing are completed on the server side to form a standard mapping data set. Subsequently, based on the standard mapping data set, the local terrain complexity index LTCI is calculated and output, and the local terrain complexity index LTCI values ​​of all acquisition points are traversed and a preliminary comparative evaluation is performed to achieve intelligent judgment of the terrain complexity of the infrastructure site selection area in the mountainous area. Based on the complexity evaluation results, the feature line depth analysis is automatically triggered. In the feature line depth analysis, the terrain feature line stability index FLSI is further calculated and output based on the standard mapping data set to determine the stability of the feature line. When an unstable feature line is detected, local refinement and reconstruction is automatically triggered to increase the density of acquisition points to optimize the modeling quality of complex terrain areas. Finally, based on the local terrain complexity index (LTCI) and the terrain feature line stability index (FLSI), a comprehensive calculation outputs the overall modeling quality index (GMQI). A secondary comparative evaluation based on the GMQI results is conducted to determine and optimize the overall modeling quality, ensuring high consistency and stability in both complex and common areas. The method of the present invention achieves the goal of intelligently and hierarchically controlling the quality of three-dimensional terrain modeling based on surveying and mapping data, without relying on traditional empirical operations. During implementation, the method significantly improves the 3D modeling accuracy of complex landforms such as cliffs, landslides, and river valleys by actively extracting terrain feature lines and performing structure-aware modeling. Furthermore, a dynamic refinement mechanism based on the dual indicators of the local complexity index (LTCI) and the feature line stability index (FLSI) avoids data redundancy and uneven quality caused by blind encryption or neglect of refinement during the modeling process. Furthermore, by comprehensively calculating the overall modeling quality index (GMQI) and iteratively optimizing it based on the GMQI, modeling quality is globally controllable and evaluable, significantly improving the consistency, accuracy, and engineering reliability of modeling in mountainous infrastructure site selection areas.

[0080] Example 2:

[0081] See also Figure 1 and Figure 4 ,Specifically: S1 includes S11 and S12;

[0082] S11. Use a drone equipped with LiDAR and refer to existing engineering surveying and mapping standards to set the density of collection points based on the drone's flight altitude, speed, laser emission frequency, and image capture line. The initial setting is 25 collection points per square. For example, if the flight altitude is 100m, the collection point density will naturally reach 25 points / m 2, conduct aerial mapping of infrastructure site selection in mountainous areas and obtain three-dimensional point cloud data for terrain modeling at each collection point;

[0083] The terrain modeling 3D point cloud data includes the local point cloud 3D coordinates (x, y, z), where x represents the horizontal axis coordinate, y represents the vertical axis coordinate, and z represents the vertical axis coordinate;

[0084] S12. Build a surveying and mapping server, wirelessly connect the surveying and mapping server to the drone through the 5G communication network, and transmit the real-time terrain modeling 3D point cloud data to the surveying and mapping server.

[0085] In this embodiment, the method utilizes a drone equipped with a LiDAR system for aerial mapping. In conjunction with existing engineering surveying standards, the method comprehensively configures the drone's flight altitude, flight speed, laser emission frequency, and flight path settings. The initial acquisition point density is set to 25 points per square meter, ensuring the target density is naturally reached at an altitude of 100 meters. This effectively covers the infrastructure site selection area in mountainous areas and acquires three-dimensional point cloud data for terrain modeling, where each point contains local three-dimensional coordinates (x, y, z). Subsequently, a mapping server system is constructed, establishing a 5G wireless communication connection between the drone and the server, enabling stable and rapid transmission of the real-time acquired three-dimensional point cloud data and ensuring the continuity and timeliness of data acquisition. Through the above-described embodiments, the present invention achieves the goal of rapidly acquiring high-precision terrain modeling data in complex mountainous terrain environments using an efficient, standardized process. Specific beneficial effects include: By rationally controlling acquisition parameters, the density of the acquired point cloud data is uniform and comprehensive, improving the integrity and accuracy of the underlying terrain data; Furthermore, through a real-time data transmission mechanism, the data transmission and processing cycle is significantly shortened, reducing packet loss and latency during data transmission. In addition, the present invention further improves the basic data quality of subsequent terrain feature extraction and intelligent three-dimensional modeling through the collaborative mechanism of high-density standardized collection and efficient real-time transmission, providing highly reliable three-dimensional terrain data support for the site selection of infrastructure in complex mountainous areas, ensuring the accuracy of modeling results and the reliability of engineering applications.

[0086] Example 3:

[0087] See also Figure 1 , specifically: S2 includes S21 and S22;

[0088] S21. In the surveying and mapping server, feature extraction is performed on the local point cloud 3D coordinates (x, y, z) of the terrain modeling 3D point cloud data acquired in real time to obtain terrain feature lines and surveying and mapping data;

[0089] The surveying and mapping data include the three-dimensional coordinates of the local point cloud (x, y, z), the local curvature change rate Qc, the local normal vector change gradient ▽n, the local elevation fluctuation index Bh, the point cloud density gradient ▽p and the characteristic line connection logarithm L;

[0090] The local curvature change rate Qc is obtained by using KD-terr fast domain search for each local point cloud three-dimensional coordinate (x, y, z) to retrieve all neighboring points of the local point cloud three-dimensional coordinate (x, y, z) within a radius r to form a domain N. The local surface is fitted using the least squares method within the domain N to obtain the fitting surface. The principal curvature K is obtained through the fitting surface. The local principal curvature is obtained for each local point cloud three-dimensional coordinate (x, y, z) in the domain N. The average absolute value of the difference between the local point cloud three-dimensional coordinate (x, y, z) and the principal curvature in the domain N is then calculated.

[0091] The local normal vector change gradient ▽n is obtained by using ICA principal component analysis on the domain N of the three-dimensional coordinates (x, y, z) of each local point cloud to extract the main direction vector as the normal vector n. The angle difference is calculated by performing dot multiplication on each pair of normal vectors n in the domain N, and the average change rate is obtained based on the angle difference.

[0092] The local elevation fluctuation index Bh is obtained by retrieving the three-dimensional coordinates (x, y, z) of each local point cloud, the area N within the radius r, taking the vertical axis coordinate z of each local point cloud three-dimensional coordinate (x, y, z) in the area N, that is, the elevation value, and calculating the standard deviation of these vertical axis coordinates z;

[0093] The point cloud density gradient ▽p is obtained by calculating the ratio of the number of points in the domain N to the volume of the domain sphere, and then performing average difference calculation based on the local density.

[0094] Extract terrain feature lines based on the local curvature change rate Qc, local normal vector change gradient ▽n and local elevation fluctuation index Bh of each local point cloud three-dimensional coordinate (x, y, z);

[0095] Topographic characteristic lines include fault lines, slope break lines, and continuous lines of high curvature;

[0096] The fault line is determined by the sudden change of the local curvature rate Qc and the terrain characteristic line of the large local normal vector change gradient ▽n;

[0097] The slope break line is determined by the sudden change of the local curvature change rate Qc in the middle amplitude combined with the change of the gradient ▽n of the local normal vector in the middle amplitude, and the characteristic line with good continuity;

[0098] High curvature continuous lines are determined by continuously distributing extremely high local curvature change rates Qc and requiring low local normal vector change gradients ▽n.

[0099] The number of feature line connection pairs L is obtained by analyzing the number of adjacent point pairs of the local point cloud three-dimensional coordinates (x, y, z) on the terrain feature line;

[0100] For example, if there are 5 local point cloud 3D coordinates (x, y, z) on a feature line, there are 4 connection pairs in total, so the number of feature line connection pairs L=4;

[0101] S22. Preprocessing the acquired surveying and mapping data in a surveying and mapping server, where the preprocessing includes cleaning noise and normalizing the data to obtain a standard surveying and mapping data set;

[0102] Noise cleaning: By using volume filtering and radius filtering on the acquired mapping data, outliers are removed and features that are incorrectly extracted are cleared;

[0103] Normalized data is performed by using the mean-variance normalized Z-score to eliminate the dimension effects of all parameters in the surveying and mapping data except the three-dimensional coordinates (x, y, z) of the local point cloud;

[0104] At the same time, coordinate normalization processing is used to convert the three-dimensional coordinates (x, y, z) of the local point cloud into digital parameters to eliminate the influence of dimension.

[0105] In this embodiment, the method extracts terrain feature lines and surveying data by performing feature extraction on the 3D coordinates (x, y, z) of local point clouds acquired in real time from terrain modeling 3D point cloud data on a surveying and mapping server. By extracting local feature parameters, three types of terrain feature lines are distinguished and extracted: breaklines, slope breaklines, and high-curvature continuous lines, based on the difference in curvature change rate and normal vector gradient amplitude. The number of feature line connections, L, is calculated based on the number of adjacent feature points, enriching the representation of terrain structure information. After feature extraction, the acquired surveying and mapping data is further preprocessed to remove outliers and erroneously extracted features through volume filtering and radius filtering, completing noise removal. Furthermore, mean-variance normalization (Z-score) is used to eliminate dimensionality effects on all parameters other than coordinates. Coordinate normalization is then used to digitally standardize the 3D coordinates (x, y, z) of the local point clouds, thereby forming a standard surveying and mapping dataset. Through the above-described embodiments, the present invention achieves the technical objectives of efficiently extracting terrain features, cleaning noise, and unifying standardized surveying and mapping datasets from large-scale surveying and mapping data. Specific benefits include: First, by comprehensively extracting multidimensional local features, it is possible to precisely and accurately describe the complex changing trends and structural information of the terrain surface; second, the preprocessing process effectively improves the cleanliness and standardization of the data, significantly reducing the interference of outliers on subsequent terrain modeling. Furthermore, by establishing a data foundation for structure-aware terrain feature lines and unifying the numerical scales of different parameters through standardization, the accuracy and stability of subsequent local complexity analysis, feature line stability assessment, and intelligent modeling processes are further improved, providing solid data support and optimization guarantees for high-precision three-dimensional terrain modeling in mountainous infrastructure site selection areas.

[0106] Example 4:

[0107] See also Figure 1 and Figure 3 Specifically: S3 includes S31, based on the acquired standard surveying and mapping data set, calculates and outputs the local terrain complexity index LTCI of the current three-dimensional coordinates (x, y, z) of all local point clouds, which measures the local terrain mutation, directional instability, small elevation change amplitude and density anomaly;

[0108] The local terrain complexity index LTCI is calculated and output by the following algorithm formula;

[0109] ;

[0110] Where LTCI(i) represents the local terrain complexity index of the i-th collection point, Qc(i) represents the local curvature change rate of the i-th collection point, ▽n(i) represents the local normal vector change gradient of the i-th collection point, Bh(i) represents the local elevation fluctuation index of the i-th collection point, ▽p(i) represents the point cloud density gradient of the i-th collection point, Represents a small positive number that is protected against division by zero.

[0111] S3 also includes S32, traversing the local terrain complexity index LTCI of all collection points, and based on the local terrain complexity index LTCI results output by each collection point, conducting a preliminary comparative evaluation to determine the complexity of each collection point, and classifying the current mountain infrastructure site selection based on the complexity. The specific evaluation contents are as follows;

[0112] When the local terrain complexity index LTCI (i) of the i-th collection point is less than 1.5, the current collection point is classified as a first-level complex point, that is, a low-complexity point;

[0113] When 1.5≤the local terrain complexity index LTCI(i) of the i-th collection point is less than 2.5, the current collection point is classified as a secondary complex point, i.e., a medium complex point;

[0114] When the local terrain complexity index LTCI (i) of the i-th collection point is ≥ 2.5, the current collection point is divided into the third level of complexity, that is, a high complexity point;

[0115] Based on the preliminary comparison and evaluation, the total number of third-level complex points and the ratio of the total number of current collection points i are evaluated, and the trigger judgment condition Trigger is obtained to determine the triggering condition of the feature line depth analysis;

[0116] The specific judgment contents of the trigger judgment condition are as follows;

[0117] ;

[0118] Among them, when the trigger judgment condition Trigger=1, the feature line depth analysis is triggered; Nhigh-LTCI∈{LTCI>2.5} represents the total number of level 3 complex points of the local terrain complexity index LTCI of the collection point; F1 represents the trigger ratio threshold, which is set by the user. When the trigger judgment condition Trigger=0, the feature line depth analysis is not triggered and standard modeling is performed;

[0119] Specific examples:

[0120] Mountain infrastructure site selection area A: contains 10,000 collection points, of which 4,000 collection points have LTCi>2.5, F1=30%, and complex point ratio = 40%>30%, triggering feature line in-depth analysis;

[0121] Mountain infrastructure site selection area B: contains 8,000 collection points, of which 1,000 collection points have LTCi>2.5, F1=30%, and the proportion of complex points = 40%<30%, which does not trigger the feature line in-depth analysis.

[0122] In this embodiment, the method calculates the 3D coordinates (x, y, z) of all current local point clouds based on a standard surveying and mapping dataset, outputting a local terrain complexity index (LTCI) for each acquisition point. The LTCI comprehensively considers four parameters: the local curvature change rate (Qc), the local normal vector change gradient (▽n), the local elevation fluctuation index (Bh), and the point cloud density gradient (▽p). Combined with a small positive integer to prevent division by zero, the LTCI is calculated using an algorithmic formula to quantitatively reflect the local terrain abruptness, surface orientation instability, minor elevation fluctuations, and density anomalies at each acquisition point. Based on the output LTCI for each acquisition point, all acquisition points are traversed for a preliminary comparative assessment. Based on a predefined grading standard, the acquisition points are classified into three levels of complexity: low, medium, and high. Furthermore, a trigger condition (Trigger) is calculated based on the ratio of the evaluated three levels of complexity points to the total number of acquisition points. When the proportion of high-complexity acquisition points exceeds a predefined trigger ratio threshold (F1), subsequent feature line depth analysis is automatically triggered. Otherwise, standard modeling is performed. Through the above-mentioned implementation mode, the present invention achieves the purpose of intelligently identifying the distribution of local terrain complexity in the site selection area of ​​mountain infrastructure based on standard surveying and mapping data. Specific beneficial effects include: on the one hand, by establishing a quantitative standard for the local terrain complexity index LTCI index, accurate distinction between terrain areas of different complexity is achieved, providing a basis for subsequent modeling strategies; on the other hand, by setting the trigger ratio threshold F1, dynamic triggering of feature line depth analysis based on the actual complexity of the terrain is achieved, avoiding invalid encryption processing of simple areas, and improving modeling efficiency and resource utilization. In addition, the present invention further improves the pertinence and necessity of detailed processing of complex terrain areas through intelligent evaluation and hierarchical control of local terrain complexity, effectively optimizes the accuracy, detail expression and stability of the overall three-dimensional terrain modeling, and significantly enhances the scientificity and reliability of terrain modeling for site selection of mountain infrastructure.

[0123] Example 5:

[0124] See also Figure 1 and Figure 3 ,Specifically: S4 includes S41, after triggering the feature line depth analysis, extracting the 3D coordinates (x, y, z) of the local point cloud and the standard mapping data set, and analyzing the terrain feature line stability index FLSI of each pair of adjacent terrain feature lines to quantify the stability of a geological feature line in elevation changes and direction changes;

[0125] The terrain feature line stability index FLSI is calculated and output by the following algorithm formula;

[0126] ;

[0127] Where i represents the i-th acquisition point, j represents the j-th acquisition point, FLSI(l) represents the terrain characteristic line stability index of the l-th terrain characteristic line, represents the partial derivative, x i 、y i and z i Respectively represent the horizontal axis coordinate, vertical axis coordinate and vertical axis coordinate of the i-th acquisition point, x j 、y j and z j Respectively represent the horizontal axis coordinate, vertical axis coordinate and vertical axis coordinate of the jth acquisition point, Indicates that both acquisition points i and j belong to the lth terrain feature line and acquisition point i is not equal to acquisition point j, ▽n(j) represents the local normal vector change gradient of the jth acquisition point, and ▽n(i) represents the local normal vector change gradient of the ith acquisition point;

[0128] represents the local elevation gradient of the i-th acquisition point;

[0129] represents the local elevation gradient of the jth acquisition point;

[0130] It represents the elevation gradient difference;

[0131] Represents the gradient difference of the normal vector, and summing the two together gives the local stability disturbance value of a single connection on the terrain feature line;

[0132] Elevation change refers to the trend of vertical height z-value changes between consecutive point pairs on a geological feature line;

[0133] Directional change refers to the continuity of the surface orientation or tilt angle along a geological feature line. Get;

[0134] Elevation continuity measures the smoothness of elevation changes between adjacent points; directional continuity measures the continuity of the terrain surface's orientation, i.e., the variation of the local normal vector. The purpose of the Feature Line Stability Index (FLSI) is to use a mathematical indicator to comprehensively represent the overall stability of a feature line in terms of its morphology and structure.

[0135] S4 also includes S41, based on the output result of the terrain feature line stability index FLSI, stability assessment, judging the stability of the terrain feature line, and refining and reconstructing the unstable area. The specific assessment contents are as follows;

[0136] When the terrain feature line stability index FLSI ≤ 0.8, it means that the terrain feature line is stable and the model is directly built;

[0137] When the terrain feature line stability index FLSI>0.8, it means that the terrain feature line is unstable, and local refinement reconstruction is triggered;

[0138] Local refinement and reconstruction is carried out by restarting the UAV for aerial mapping to reconstruct the current terrain feature lines. The density of acquisition points per square is initially set to 25, which is increased by 40%. Feature extraction is then re-executed to output the terrain feature line stability index FLSI and the local terrain complexity index LTCI.

[0139] In this embodiment, after triggering feature line depth analysis, this method analyzes the feature points of each pair of adjacent terrain feature lines based on the three-dimensional coordinates (x, y, z) of the local point cloud and a standard mapping dataset, and calculates and outputs the corresponding terrain feature line stability index (FLSI). The FLSI accumulates the difference between the local elevation gradient and the normal vector gradient for each pair of adjacent acquisition points (i, j) within the same feature line, reflecting the local stability of the terrain feature line in both elevation and orientation. This index is then normalized by the logarithm of the feature line connectivity to quantify the overall stability index. Based on the FLSI evaluation results, unstable and stable feature lines are identified, and local refinement and reconstruction are automatically triggered. During this refinement and reconstruction process, the drone is restarted for aerial mapping. The original acquisition point density in unstable feature line areas is increased by 40%, and features are re-acquired and extracted. New FLSI and local terrain complexity index (LTCI) are then re-output to optimize terrain modeling quality in unstable areas. Through the above-mentioned embodiments, the present invention achieves the technical purpose of intelligently determining and dynamically controlling the scope and granularity of refinement and reconstruction based on characteristic line stability indicators in areas with complex terrain features.

[0140] Example 6:

[0141] See also Figure 1 and Figure 3 ,Specifically: S5 includes S51, the terrain feature line stability index FLSI obtained after local refinement and reconstruction, combined with the local terrain complexity index LTCI, to perform comprehensive calculation and output the overall modeling quality index GMQI to measure the overall modeling quality;

[0142] The overall modeling quality index GMQI is calculated and output by the following algorithm formula;

[0143] ;

[0144] Where N represents the total number of infrastructure site selection areas in mountainous areas, R represents the infrastructure site selection area in mountainous areas, and LRCI avg (R) represents the mean of all local terrain complexity indices LTCI within the R mountain infrastructure site selection area, FLSIavg (R) represents the mean value of the terrain feature line stability index FLSI of all the infrastructure site selection areas in the R mountain area, and all values ​​are dimensionless after normalization.

[0145] S5 also includes S52, a secondary comparative evaluation based on the output results of the overall modeling quality index GMQI, a comprehensive analysis of the complexity and feature line stability after local refinement and reconstruction, and an evaluation of the overall modeling quality. The specific evaluation contents are as follows;

[0146] When the overall modeling quality index GMQI ≤ 2, it means that the overall modeling quality meets the standard, and the modeling is prompted;

[0147] When the overall modeling quality index GMQI>2, it means that the overall modeling quality does not meet the standard. At this time, the iteration triggers local refinement and reconstruction. Each iteration increases the acquisition point density by 40% based on the previous one until the overall modeling quality meets the standard and the iteration stops.

[0148] In this embodiment, a comprehensive calculation is performed based on the terrain feature line stability index (FLSI) and the local terrain complexity index (LTCI) obtained after local refinement and reconstruction to output an overall modeling quality indicator (GMQI) (S51). The GMQI calculates the mean of the local terrain complexity index (LTCI) and the mean of the terrain feature line stability index (FLSI) for each mountain infrastructure site selection region R, and outputs the comprehensive results according to a predefined algorithm to quantitatively measure the overall three-dimensional terrain modeling quality. Subsequently, a secondary comparative evaluation is performed based on the GMQI output results to determine whether the overall modeling quality meets the standards.

[0149] Through the above-mentioned implementation, the present invention achieves the technical purpose of realizing intelligent closed-loop evaluation and dynamic refinement optimization control based on comprehensive quality indicators in the terrain modeling process.

[0150] Example 7:

[0151] See also Figure 1 and Figure 2 , intelligent 3D terrain modeling system based on surveying and mapping data, point cloud data acquisition module, surveying and mapping data processing module, terrain complexity analysis module, terrain feature line stabilization module and overall modeling and analysis module;

[0152] The point cloud data acquisition module uses drones equipped with LiDAR to set up collection points for selecting infrastructure sites in mountainous areas, conducts aerial mapping, obtains three-dimensional point cloud data for terrain modeling, and builds a mapping server to transmit the modeling geometry data to the mapping server.

[0153] The surveying and mapping data processing module extracts features from the three-dimensional point cloud data of terrain modeling in the surveying and mapping server to obtain terrain feature lines and surveying and mapping data, and pre-processes the surveying and mapping data to obtain a standard surveying and mapping data set;

[0154] The terrain complexity analysis module calculates and outputs the local terrain complexity index (LTCI) based on a standard surveying and mapping dataset. It then traverses the output of the local terrain complexity index (LTCI) of all collected points for preliminary comparative evaluation to determine the complexity of the site selection. Based on the preliminary comparative evaluation results, it triggers a feature line depth analysis.

[0155] The terrain feature line stability module calculates and outputs the terrain feature line stability index FLSI based on the standard surveying and mapping data set through feature line depth analysis. Based on the terrain feature line stability index FLSI, the stability of the terrain feature line is judged and the unstable area is refined and reconstructed.

[0156] The overall modeling analysis module performs summary calculations based on the local terrain complexity index LTCI and the terrain feature line stability index FLSI, outputs the overall modeling quality index GMQI, and conducts a secondary comparative evaluation based on the output results of the overall modeling quality index GMQI to comprehensively analyze the complexity of the terrain and the stability of the feature lines.

[0157] While the embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes, modifications, substitutions, and alterations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. An intelligent three-dimensional terrain modeling method based on surveying and mapping data, characterized by: The following steps are involved: S1. Use a drone equipped with LiDAR to set up collection points to select infrastructure sites in mountainous areas, conduct aerial mapping, obtain 3D point cloud data for terrain modeling, build a mapping server, and transmit the modeling geometry data to the mapping server. S2. Extract features from the terrain modeling 3D point cloud data in the surveying and mapping server to obtain terrain feature lines and surveying and mapping data, and pre-process the surveying and mapping data to obtain a standard surveying and mapping data set; S3. Based on the standard surveying and mapping data set, calculate and output the local terrain complexity index (LTCI). Then, perform a preliminary comparative evaluation of the LTCI output results of all the collected points to determine the complexity of the site selection. Based on the preliminary comparative evaluation results, trigger the feature line depth analysis. The S3 includes S31, calculating and outputting the local terrain complexity index LTCI of all current local point cloud three-dimensional coordinates (x, y, z) based on the acquired standard surveying and mapping data set, to measure the local terrain mutation, directional instability, small elevation change amplitude and density anomaly; The local terrain complexity index LTCI is calculated and outputted by the following algorithm formula: ； Where LTCI(i) represents the local terrain complexity index of the i-th collection point, Qc(i) represents the local curvature change rate of the i-th collection point, ▽n(i) represents the local normal vector change gradient of the i-th collection point, Bh(i) represents the local elevation fluctuation index of the i-th collection point, ▽p(i) represents the point cloud density gradient of the i-th collection point, Represents a small positive number that prevents division by zero; S4, the feature line depth analysis is performed by calculating and outputting a terrain feature line stability index FLSI based on a standard surveying and mapping data set, and based on the terrain feature line stability index FLSI, the stability of the terrain feature line is judged, and the unstable area is refined and reconstructed; The S4 includes S41, after triggering the feature line depth analysis, extracting the three-dimensional coordinates (x, y, z) of the local point cloud and the standard mapping data set, analyzing the terrain feature line stability index FLSI of each pair of adjacent terrain feature lines to quantify the stability of a geological feature line in elevation change and direction change; The terrain feature line stability index FLSI is calculated and output by the following algorithm formula: ； Where i represents the i-th acquisition point, j represents the j-th acquisition point, FLSI(l) represents the terrain characteristic line stability index of the l-th terrain characteristic line, represents the partial derivative, x i 、y i and z i Respectively represent the horizontal axis coordinate, vertical axis coordinate and vertical axis coordinate of the i-th acquisition point, x j 、y j and z j Respectively represent the horizontal axis coordinate, vertical axis coordinate and vertical axis coordinate of the jth acquisition point, Indicates that both acquisition points i and j belong to the lth terrain feature line and acquisition point i is not equal to acquisition point j, ▽n(j) represents the local normal vector change gradient of the jth acquisition point, and ▽n(i) represents the local normal vector change gradient of the ith acquisition point; S5. Based on the local terrain complexity index LTCI and the terrain feature line stability index FLSI, a summary calculation is performed to output the overall modeling quality index GMQI. Based on the output results of the overall modeling quality index GMQI, a secondary comparative evaluation is performed to comprehensively analyze the complexity of the terrain and the stability of the feature lines.

2. The intelligent three-dimensional terrain modeling method based on surveying and mapping data according to claim 1, characterized in that: Said S1 includes S11 and S12; S11. Use a drone equipped with LiDAR, and refer to existing engineering surveying and mapping standards. Set the density of collection points based on the drone's flight altitude, speed, laser emission frequency, and image capture route. Initially set at 25 collection points per square meter, conduct aerial mapping of infrastructure sites in mountainous areas, and obtain terrain modeling 3D point cloud data for each collection point. The terrain modeling three-dimensional point cloud data includes local point cloud three-dimensional coordinates (x, y, z), where x represents the horizontal axis coordinate, y represents the vertical axis coordinate, and z represents the vertical axis coordinate; S12. Build a surveying and mapping server, wirelessly connect the surveying and mapping server to the drone through the 5G communication network, and transmit the real-time terrain modeling 3D point cloud data to the surveying and mapping server.

3. The intelligent three-dimensional terrain modeling method based on surveying and mapping data according to claim 2, characterized in that: Said S2 includes S21 and S22; S21. In the surveying and mapping server, feature extraction is performed on the local point cloud 3D coordinates (x, y, z) of the terrain modeling 3D point cloud data acquired in real time to obtain terrain feature lines and surveying and mapping data; The surveying and mapping data includes the local point cloud three-dimensional coordinates (x, y, z), the local curvature change rate Qc, the local normal vector change gradient ▽n, the local elevation fluctuation index Bh, the point cloud density gradient ▽p and the characteristic line connection logarithm L; The local curvature change rate Qc is obtained by using KD-terr fast domain search for each local point cloud three-dimensional coordinate (x, y, z) to search for all neighboring points of the local point cloud three-dimensional coordinate (x, y, z) within a radius r to form a domain N, and using the least squares method to perform local surface fitting in the domain N to obtain a fitting surface, and then obtaining the principal curvature K through the fitting surface, and then calculating the local principal curvature for each local point cloud three-dimensional coordinate (x, y, z) in the domain N, and then calculating the average absolute value of the difference between the local point cloud three-dimensional coordinate (x, y, z) and the principal curvature in the domain N; The local normal vector change gradient ▽n is obtained by performing ICA principal component analysis on the domain N of the three-dimensional coordinates (x, y, z) of each local point cloud, extracting the main direction vector as the normal vector n, performing dot multiplication on each pair of normal vectors n in the domain N to calculate the angle difference, and then calculating the average change rate based on the angle difference; The local elevation fluctuation index Bh is obtained by retrieving the three-dimensional coordinates (x, y, z) of each local point cloud, the area N within the radius r, taking the vertical axis coordinate z of the three-dimensional coordinates (x, y, z) of each local point cloud within the area N, and calculating the standard deviation of these vertical axis coordinates z; The point cloud density gradient ▽p is obtained by calculating the ratio of the number of points in the domain N to the volume of the domain sphere to obtain the local density, and performing average difference calculation based on the local density; Extract terrain feature lines based on the local curvature change rate Qc, local normal vector change gradient ▽n and local elevation fluctuation index Bh of each local point cloud three-dimensional coordinate (x, y, z); The terrain characteristic lines include fault lines, slope break lines and high curvature continuous lines; The number of feature line connection pairs L is obtained by analyzing the number of adjacent point pairs of the local point cloud three-dimensional coordinates (x, y, z) on the terrain feature line; S22. Preprocessing the acquired surveying and mapping data in a surveying and mapping server, wherein the preprocessing includes cleaning noise and normalizing the data to obtain a standard surveying and mapping data set; The noise cleaning method uses volume filtering and radius filtering on the acquired surveying and mapping data to remove outliers and clear erroneously extracted features; The normalized data is normalized by using the mean-variance standard Z-score to eliminate the dimension effect of all parameters in the surveying and mapping data except the three-dimensional coordinates (x, y, z) of the local point cloud; At the same time, coordinate normalization processing is used to convert the three-dimensional coordinates (x, y, z) of the local point cloud into digital parameters to eliminate the influence of dimension.

4. The intelligent three-dimensional terrain modeling method based on surveying and mapping data according to claim 1, characterized in that: The S3 further includes S32, traversing the local terrain complexity index LTCI of all collection points, and performing a preliminary comparative evaluation based on the local terrain complexity index LTCI results output by each collection point, determining the complexity of each collection point, and classifying the current mountain infrastructure site selection based on the complexity. The specific evaluation content is as follows; When the local terrain complexity index LTCI (i) of the i-th collection point is less than 1.5, the current collection point is classified as a first-level complex point; When 1.5≤the local terrain complexity index LTCI(i) of the i-th collection point is less than 2.5, the current collection point is classified as a secondary complex point; When the local terrain complexity index LTCI (i) of the i-th collection point is ≥ 2.5, the current collection point is divided into a level 3 complex point; Based on the preliminary comparison and evaluation, the total number of third-level complex points and the ratio of the total number of current collection points i are evaluated, and the trigger judgment condition Trigger is obtained to determine the triggering condition of the feature line depth analysis; The specific judgment content of the trigger judgment condition Trigger is as follows: ； Among them, when the trigger judgment condition Trigger=1, the feature line depth analysis is triggered; Nhigh-LTCI∈{LTCI>2.5} represents the total number of level 3 complex points of the local terrain complexity index LTCI of the collection point; F1 represents the trigger ratio threshold, which is set by the user. When the trigger judgment condition Trigger=0, the feature line depth analysis is not triggered and standard modeling is performed.

5. The intelligent three-dimensional terrain modeling method based on surveying and mapping data according to claim 1, characterized in that: The S4 further includes S41, based on the output result of the terrain feature line stability index FLSI, stability assessment, judging the stability of the terrain feature line, and refining and reconstructing the unstable area, the specific assessment content is as follows; When the terrain feature line stability index FLSI ≤ 0.8, it means that the terrain feature line is stable and the model is directly built; When the terrain feature line stability index FLSI>0.8, it means that the terrain feature line is unstable, and local refinement reconstruction is triggered; The local refinement reconstruction is performed by restarting the UAV to perform aerial photography and mapping, reconstructing the current terrain feature lines, and initially setting the density of acquisition points to 25 per square, which is increased by 40%, and re-executing feature extraction and outputting the terrain feature line stability index FLSI and the local terrain complexity index LTCI.

6. The intelligent three-dimensional terrain modeling method based on surveying and mapping data according to claim 1, characterized in that: The S5 includes S51, based on the terrain feature line stability index FLSI obtained after local refinement and reconstruction, combined with the local terrain complexity index LTCI, to perform comprehensive calculation to output the overall modeling quality index GMQI to measure the overall modeling quality; The overall modeling quality index GMQI is calculated and output by the following algorithm formula: ； Where N represents the total number of infrastructure site selection areas in mountainous areas, R represents the infrastructure site selection area in mountainous areas, and LRCI avg (R) represents the mean of all local terrain complexity indices LTCI within the R mountain infrastructure site selection area, FLSI avg (R) represents the mean value of the FLSI of all terrain feature line stability indexes in the R mountain infrastructure site selection area.

7. The intelligent three-dimensional terrain modeling method based on surveying and mapping data according to claim 6, characterized in that: The S5 also includes S52, performing a secondary comparative evaluation based on the output results of the overall modeling quality index GMQI, comprehensively analyzing the complexity and characteristic line stability after local refinement and reconstruction, and evaluating the overall modeling quality. The specific evaluation contents are as follows; When the overall modeling quality index GMQI ≤ 2, it means that the overall modeling quality meets the standard, and the modeling is prompted; When the overall modeling quality index GMQI>2, it means that the overall modeling quality does not meet the standard. At this time, the iteration triggers local refinement and reconstruction. Each iteration increases the acquisition point density by 40% based on the previous one until the overall modeling quality meets the standard and the iteration stops.

8. An intelligent three-dimensional terrain modeling system based on surveying and mapping data, applied to the intelligent three-dimensional terrain modeling method based on surveying and mapping data according to any one of claims 1 to 7, characterized in that: It includes point cloud data acquisition module, surveying and mapping data processing module, terrain complexity analysis module, terrain feature line stabilization module and overall modeling and analysis module; The point cloud data acquisition module uses a drone equipped with LiDAR to set up collection points to select the site of infrastructure in mountainous areas, conduct aerial mapping, obtain three-dimensional point cloud data for terrain modeling, build a mapping server, and transmit the modeling geometry data to the mapping server; The surveying and mapping data processing module extracts features from the terrain modeling three-dimensional point cloud data in the surveying and mapping server to obtain terrain feature lines and surveying and mapping data, and pre-processes the surveying and mapping data to obtain a standard surveying and mapping data set; The terrain complexity analysis module calculates and outputs the local terrain complexity index (LTCI) based on a standard surveying and mapping dataset, and performs preliminary comparative evaluation on the output results of the local terrain complexity index (LTCI) of all collected points to determine the complexity of the site selection, and triggers feature line depth analysis based on the preliminary comparative evaluation results. The terrain feature line stabilization module calculates and outputs a terrain feature line stability index FLSI based on the feature line depth analysis and a standard surveying and mapping data set, and judges the stability of the terrain feature line based on the terrain feature line stability index FLSI, and refines and reconstructs the unstable area; The overall modeling and analysis module performs summary calculations based on the local terrain complexity index LTCI and the terrain feature line stability index FLSI, outputs the overall modeling quality index GMQI, and performs a secondary comparative evaluation based on the output results of the overall modeling quality index GMQI to comprehensively analyze the complexity of the terrain and the stability of the feature lines.

Images