TLS observation network planning method for large-scale building

Abstract

The invention discloses a TLS observation network planning method for a large-scale building, and relates to the technical field of building simulation. Comprising the following steps: acquiring grid three-dimensional information of an observation area; taking all grid points in the screened grid as observation candidate sites, performing observation simulation on each candidate site in a final observation target point union set, and taking a building feature line length value obtained by simulation scanning as a weight value of the candidate site; scanning contents of the selected candidate sites are deleted from the screened simulation observation contents; and performing iterative selection in the remaining candidate sites until the maximum weight value of the current candidate site is lower than a set weight threshold value, obtaining a simulated overlapping degree among the remaining candidate sites, and performing station supplementation operation on the remaining candidate sites with the simulated overlapping degree lower than the set threshold value to obtain the TLS observation network. The method restrains the number of observation objects, solves the problem of low efficiency of three-dimensional simulation planning of a TLS observation network, and has the capability of processing large-scale building groups.

Description

Technical Field

[0001] The invention relates to the technical field of building simulation, and in particular to a TLS observation network planning method for large-scale buildings. Background Art

[0002] Terrestrial 3D laser scanning (TLS) is an important means of obtaining 3D morphological information of building facades. It can supplement aerial photogrammetry and aerial lidar data collection to obtain complete and high-quality 3D data of buildings. It plays an important role in applications such as 3D reconstruction of buildings, fine-scale sunshine analysis, and urban digital twins. At present, TLS observation sites are mainly determined by users' subjective decisions, which can easily lead to redundancy and voids in scanning data, seriously affecting efficiency, cost, and data quality. Therefore, scientific planning of TLS observation networks for large-scale buildings has obvious scientific and economic value.

[0003] In the prior art, the TLS observation network planning mainly adopts the model simulation method: based on the observation scene model of the small-scale area simulation planning based on the two-dimensional topographic map, the candidate station TLS scanning is simulated, the constraint equations such as observation integrity and cost are established, and the TLS observation network location is determined by greedy and other optimization methods. The main process is: obtain the two-dimensional topographic map data of the target area, establish the observation scene model of the target area, and select a series of candidate station locations in the observation scene model according to the terrain characteristics and planning requirements. For each candidate station, simulate the scanning process of TLS (terrestrial laser scanner), analyze the data quality obtained by the simulated scanning, and ensure that the scanning data obtained from each candidate station can cover the key parts of the target area. Optimization methods such as greedy algorithm, genetic algorithm, particle swarm optimization, etc. are applied to search for the TLS observation network location that meets the constraint conditions in the candidate stations. The optimization results are evaluated to check whether the constraint conditions such as observation integrity and cost are met. The coordinates of the stations in the TLS observation network are output in the form of graphics or tables.

[0004] The defects of the above-mentioned existing technologies are: the two-dimensional topographic map over-abstracts and simplifies the observation scene, resulting in serious distortion of TLS simulation. When the observation range becomes larger, such problems will become more serious, resulting in insufficient reliability of the observation network planning results. Summary of the invention

[0005] Based on this, it is necessary to provide a TLS observation network planning method for large-scale buildings to address the above technical issues.

[0006] An embodiment of the present invention provides a TLS observation network planning method for large-scale buildings, comprising:

[0007] Obtaining three-dimensional grid information of the observation area including voxel networks and building polygon vertex sets;

[0008] Construct a multi-resolution viewpoint grid based on the three-dimensional information of the grid, and determine the observation candidate area within the multi-resolution viewpoint grid; set a screening grid in the observation candidate area, and regard all grid points in the screening grid as observation candidate sites;

[0009] Extract building feature lines consisting of observation target points at the observation candidate site from the voxel grid, and use the extracted building feature lines as the observation target point set; split the observation area into multiple building cluster partitions with different building densities based on the building polygon vertex set, and obtain the simulated coverage rate of each building cluster partition;

[0010] Iterate the resolution of the screening grid of each building cluster partition until the increase ratio of the simulated coverage of each building cluster partition is lower than the set ratio threshold, and use the simulated observation content corresponding to the obtained candidate site as the screening simulated observation content, and use the obtained observation target point union as the final observation target point union;

[0011] Conduct observation simulation for each candidate site in the final observation target point set, and use the length of the building characteristic line scanned by simulation as the weight value of the candidate site; select the candidate site with the largest current weight value as the selected candidate site, and delete the scan content of the selected candidate site from the screening simulation observation content;

[0012] Iteratively select the remaining candidate sites in the final observation target point union until the maximum weight value of the current candidate site is lower than the set weight threshold, and obtain the simulated overlap between the remaining candidate sites. The remaining candidate sites whose simulated overlap is lower than the set threshold are supplemented to obtain the TLS observation network.

[0013] In addition, the three-dimensional information of the terrain and land features in the observation area is obtained through UAV aerial photogrammetry, and the three-dimensional information is expressed using voxel grids to obtain grid three-dimensional information including voxel networks and building polygon vertex sets.

[0014] In addition, the determining of the candidate area within the multi-resolution viewpoint grid specifically includes:

[0015] Produce digital elevation model DEM and digital surface model DSM based on the grid three-dimensional information of the observation area;

[0016] Obtain the areas where the difference between the digital surface model DSM in the observation area and the digital elevation model DEM in the remaining area is not 0, and mark the areas where the difference is not 0 as areas containing ponds and bushes where it is impossible to set up a station;

[0017] The digital elevation model DEM is used to obtain the slope values ​​slop of all areas in the observation area, and the areas with slope values ​​slop greater than 30 are marked as areas that affect the safety of operators;

[0018] Areas containing buildings, areas that have a serious impact on observation quality, areas where stations cannot be set up, and areas that affect operator safety are eliminated from the observation area to obtain candidate areas within the multi-resolution viewpoint grid.

[0019] In addition, the formula for obtaining the slope value slop of all areas in the observation area using the digital elevation model DEM is:

[0020]

[0021] Among them, dx is the width of the unit grid, dy is the length of the unit grid, and x i, j is a point in the digital elevation model DEM whose unit grid size is dx·dy, i is the horizontal coordinate value of the point, and j is the vertical coordinate value of the point.

[0022] In addition, the method of dividing the observation area into a plurality of building cluster partitions of building density based on the building polygon vertex set specifically includes:

[0023] Delaunay triangulation is performed on the building polygon vertex set of the three-dimensional grid information of the observation area to obtain a Delaunay triangulated network. The obtained triangles are defined as three categories: if the three vertices of a triangle belong to two or three buildings, this triangle is a first-category triangle; if a triangle is not located in any building and its three vertices are all vertices of the same building, this triangle is a second-category triangle; the remaining triangles are third-category triangles;

[0024] Construct the directional Voronoi diagram of the observation area to divide the observation area into different blocks. Each line segment in the directional Voronoi diagram is the boundary of each block.

[0025] When there are first-type triangles and second-type triangles between two buildings, the two buildings are defined as topologically adjacent;

[0026] The distances between topologically adjacent buildings are obtained, and the line segments between adjacent buildings below the set distance threshold are deleted in the directional Voronoi diagram to obtain the building cluster partitions.

[0027] In addition, the obtaining of the simulated coverage of each building cluster partition specifically includes:

[0028] The simulated observation contents of the candidate sites in each building cluster partition are combined to obtain the union of the observation target points;

[0029] The number of observation target points in the union of observation target points is divided by the number of all observation target points in the building cluster partition to obtain the simulated coverage rate of the building cluster partition.

[0030] In addition, the obtained simulated observation content corresponding to the candidate site is used as the screening simulated observation content, which specifically includes:

[0031] The first candidate site is selected according to the traversal order, and the maximum search range and the minimum search range of the candidate site are limited to complete the simulation of the scanning range;

[0032] Traverse all observation object points of the candidate site. If the observation target point is not an endpoint on the edge of the building, obtain the angle relationship between the direction vector from the candidate site to the observation target point and the edge vector where the observation target point is located, and select the observation target point that meets the preset angle condition;

[0033] Obtain the horizontal distance between the candidate site and the edge to which the point object belongs, and select the observation target points that meet the preset horizontal distance conditions;

[0034] The candidate site is taken as the starting point and the observation object point is taken as the end point. The elevation of the interpolated sight point is obtained with a fixed step length s. The elevation of the interpolated sight point is compared with the element value corresponding to the interpolated sight point in the digital surface model DSM. When the elevation of the interpolated sight point is less than the elevation of the corresponding position of the digital surface model DSM, the sight is considered to be blocked, and the observation target points whose sight is not blocked are selected.

[0035] Traverse all candidate sites and use lists to record the simulated observation contents of all candidate sites.

[0036] In addition, the formula for determining the value of the fixed step size s is:

[0037] s=int(Δ / m)

[0038] Δ=max(Δx,Δy)

[0039] Δx=x b -x a ,Δy=y b -y a

[0040] Among them, (x a ,y a ) is the coordinate of the candidate site, (x b ,y b ) is the coordinate of the observed target point, and m is the resolution of the voxel grid.

[0041] In addition, the operation of supplementing the remaining candidate sites whose simulated overlap is lower than the set threshold value specifically includes:

[0042] Construct a Delaunay triangulation network for each candidate site in the TLS observation network; the nodes in the Delaunay triangulation network are the candidate sites in the TLS observation network, and the edges of the Delaunay triangulation network represent the adjacency relationship between the candidate sites;

[0043] Based on the Delaunay triangulation, an undirected graph of the TLS observation network is constructed; the nodes of the Delaunay triangulation are the nodes of the undirected graph, and the edges of the Delaunay triangulation are the edges of the undirected graph;

[0044] The optimal registration path of the TLS observation network is generated on the undirected graph of the TLS observation network using the minimum spanning tree method;

[0045] The compatibility of the TLS observation network is evaluated. When the overlap of the scan simulations of two adjacent candidate sites on the best compatibility path is less than a certain threshold, lim When , encrypted supplementary sites are added between the two candidate sites.

[0046] In addition, the intersection of the simulated observation contents of two adjacent candidate sites divided by the minimum value of the observation target points of the two adjacent candidate sites is the simulated overlap between the remaining candidate sites.

[0047] Compared with the prior art, the TLS observation network planning method for large-scale buildings provided by the embodiment of the present invention has the following beneficial effects:

[0048] In the prior art, when planning the TLS observation network through model simulation methods, the two-dimensional topographic map over-abstracts and simplifies the observation scene, resulting in serious distortion of the TLS simulation. When the observation range becomes larger, such problems will become more serious, resulting in insufficient reliability of the observation network planning results.

[0049] The present invention sets a screening grid in the observation candidate area, regards all grid points in the screening grid as observation candidate sites, and constrains the number of candidate sites; it iterates the resolution of the screening grid for each building cluster partition to constrain the number of observation objects, which not only solves the problem of low efficiency of three-dimensional simulation planning of the TLS observation network, but also solves the problem of severe distortion of TLS simulation when the observation range becomes larger. It has the ability to handle large-scale building clusters and improves the reliability of the observation network planning results.

[0050] In addition, the remaining candidate sites of the final observation target point union are iteratively selected until the maximum weight value of the current candidate site is lower than the set weight threshold, and the simulated overlap between the remaining candidate sites is obtained. The remaining candidate sites whose simulated overlap is lower than the set threshold are supplemented to obtain the TLS observation network. This solves the problem that the TLS observation network obtained by traditional methods cannot ensure automatic alignment, and has better universality and practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A schematic diagram of a flow chart of a TLS observation network planning method for a large-scale building provided in one embodiment;

[0052] Figure 2 A schematic diagram of observation area partitioning of a TLS observation network planning method for a large-scale building provided in one embodiment. Figure 2 (a) in the figure is the directional Voronoi diagram. Figure 2 (b) in the figure is the building cluster zoning map;

[0053] Figure 3 A schematic diagram of elevations of interpolated sight-line points in a TLS observation network planning method for a large-scale building provided in one embodiment;

[0054] Figure 4 A complete elevation data map of the survey area of ​​a TLS observation network planning method for a large-scale building provided in one embodiment;

[0055] Figure 5 A filtered survey area terrain elevation data map of a TLS observation network planning method for a large-scale building provided in one embodiment;

[0056] Figure 6 A scanning planning result diagram of a TLS observation network planning method for a large-scale building provided in an embodiment. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0058] In one embodiment, a TLS observation network planning method for a large-scale building is provided, such as Figure 1 As shown, the method includes:

[0059] Step 1: Obtain the three-dimensional information of the grid including the voxel network and the building polygon vertex set of the observation area. The specific process is:

[0060] The three-dimensional information of the terrain and land features in the observation area is obtained through UAV aerial photogrammetry, and the three-dimensional information is expressed using voxel grids to obtain grid three-dimensional information including voxel networks and building polygon vertex sets.

[0061] Step 2: Construct a multi-resolution viewpoint grid based on the three-dimensional information of the grid, and determine the observation candidate area within the multi-resolution viewpoint grid. Set a screening grid in the observation candidate area, and regard all grid points in the screening grid as observation candidate sites. The specific process is:

[0062] Step 2.1: Create a digital elevation model (DEM) and a digital surface model (DSM) based on the grid three-dimensional information of the observation area.

[0063] Step 2.2: Obtain areas where the difference between the digital surface model DSM in the observation area and the digital elevation model DEM of the remaining area is non-zero, and mark the areas with non-zero difference as areas that cannot be used for station establishment, including ponds and bushes.

[0064] Step 2.3: Use the digital elevation model (DEM) to calculate the slope value slop of the remaining area obtained in step 2.3, and mark all areas with slope values ​​greater than 30, specifically marking areas located on cliffs, steep slopes, etc. that may affect the safety of operators; calculate the value of point x in the digital elevation model (DEM) with a unit grid size of dx·dy. i,j The slope formula is as shown in formula (1); all marked areas in this step are removed from the remaining area obtained in step 2.3 to obtain the candidate area for placing the TLS scanner.

[0065]

[0066] Among them, dx is the width of the unit grid, dy is the length of the unit grid, and x i,j is a point in the digital elevation model DEM whose unit grid size is dx·dy, i is the horizontal coordinate value of the point, and j is the vertical coordinate value of the point.

[0067] Step 2.4: Eliminate the areas containing buildings, areas that have a serious impact on the observation quality, areas where stations cannot be set up, and areas that affect the safety of operators from the observation area to obtain candidate areas within the multi-resolution viewpoint grid.

[0068] Step 3: Extract the building feature lines composed of observation target points of the observation candidate sites from the voxel grid, and use the extracted building feature lines as the observation target point set; the building feature lines mainly include the building bottom line, eaves line, and the inflection point wall line that can highlight the building shape.

[0069] Step 4: Divide the observation area into multiple building cluster partitions with different building densities based on the building polygon vertex set, and obtain the simulated coverage of each building cluster partition.

[0070] Specifically, the observation area partition process is as follows:

[0071] Step 4.1: Perform Delaunay triangulation on the building polygon vertex set of the three-dimensional grid information of the observation area to obtain a Delaunay triangulated network. The obtained triangles are defined as three categories: if the three vertices of a triangle belong to two or three buildings, this triangle is a first-category triangle; if a triangle is not located in any building and its three vertices are all vertices of the same building, this triangle is a second-category triangle; the remaining triangles are third-category triangles;

[0072] Step 4.2: Construct the directional Voronoi diagram of the observation area and divide the observation area into different blocks. The line segments in the directional Voronoi diagram are the boundaries of each block. The result is as follows: Figure 2 As shown in (a) in .

[0073] The specific process of dividing the observation area into different blocks is as follows:

[0074] Step 4.2.1: In the first type of triangle, connect the midpoints of two triangle sides whose endpoints belong to different buildings.

[0075] Step 4.2.2: In the second type triangle, connect the centroid of the triangle with the midpoints of the three sides of the triangle.

[0076] Step 4.3: When there are first-type triangles and second-type triangles between two buildings, the two buildings are defined as being topologically adjacent.

[0077] Step 4.4: Obtain the distances between topologically adjacent buildings, delete the line segments between adjacent buildings that are below the set distance threshold in the directional Voronoi diagram, and obtain the building cluster partitions, such as Figure 2 As shown in (b) in .

[0078] Specifically, the process of obtaining simulation coverage is as follows:

[0079] The simulated observation contents of the candidate sites in each building cluster partition are taken as the union, that is, the union of the observation target points is obtained; the number of observation target points in the union of the observation target points is divided by the number of all observation target points in the building cluster partition to obtain the simulated coverage rate of the building cluster partition.

[0080] Step 5: Iterate the resolution of the screening grid of each building cluster partition until the increase ratio of the simulated coverage of each building cluster partition is lower than the set ratio threshold, and use the simulated observation content corresponding to the obtained candidate site as the screening simulated observation content, and use the obtained union of observation target points as the final union of observation target points.

[0081] Specifically, the process of using the obtained simulated observation content corresponding to the candidate site as the screening simulated observation content is as follows:

[0082] Step 5.1: Select the first candidate site according to the traversal order, and complete the simulation of the scanning range by limiting the maximum search range and the minimum search range of the candidate site;

[0083] Step 5.2: Traverse all observation object points of the candidate site. If the observation target point is not an endpoint on the edge of the building, obtain the angle relationship between the direction vector from the candidate site to the observation target point and the edge vector where the observation target point is located, and select the observation target point that meets the preset angle condition;

[0084] Step 5.3: Obtain the horizontal distance between the candidate site and the edge to which the point object belongs, and select the observation target points that meet the preset horizontal distance conditions (when the scanner elevation angle has been determined and only the elevation angle is considered, the height difference between the observation target point and the candidate site will directly determine the shortest horizontal distance that satisfies the visibility of the two);

[0085] Step 5.4: Take the candidate site as the starting point and the observation point as the end point, and obtain the elevation of the interpolated sight line point with a fixed step length s (such as Figure 3 As shown), the elevation of the interpolated sight point is compared with the element value corresponding to the interpolated sight point in the digital surface model DSM. When the elevation of the interpolated sight point is less than the elevation of the corresponding position of the digital surface model DSM, the sight is considered to be blocked, and the observation target points whose sight is not blocked are screened. The amount of calculation will not increase exponentially with the increase of the scanning range. The value of the fixed step size s is shown in formula 2-4:

[0086] Δx=x b -x a ,Δy=y b -y a (2)

[0087] Δ=max(Δx,Δy) (3)

[0088] s=int(Δ / m) (4)

[0089] In the formula, (x a ,y a ) is the coordinate of the candidate site, (x b ,y b ) is the coordinate of the observed target point, and m is the resolution of the voxel grid.

[0090] Step 5.5: Traverse all candidate sites and use a list to record the simulated observation content of all candidate sites.

[0091] The constraints mentioned in this step mainly include the sensor parameter model that simulates the observation conditions and the environmental constraints that simulate the occlusion of terrain and objects. The main sensor parameters considered in the present invention are: scanning range, that is, the maximum measurement distance at which the target reflectivity is above 90%; angle of incidence, that is, the angle between the laser beam of the scanner and the surface normal of the observed object; elevation angle, that is, the maximum vertical angle of the scanner. The scanning range and elevation angle are determined when the instrument leaves the factory, and can be directly determined by querying the parameters. There are many factors that affect the angle of incidence. From the perspective of echo, the present invention sets the maximum angle of incidence to 70°. Environmental constraints include all objects in the real three-dimensional space that will block the scanner, including self-occlusion of the observed target, occlusion of non-observed target objects and terrain occlusion.

[0092] Step 6: Perform observation simulation on each candidate site in the final observation target point set, and use the length of the building characteristic line scanned by the simulation as the weight value of the candidate site. Select the candidate site with the largest current weight value as the selected candidate site, and delete the scan content of the selected candidate site from the screening simulation observation content.

[0093] Step 7: Iteratively select the remaining candidate sites in the final observation target point union until the weight of the candidate site with the largest current weight is lower than the set weight threshold, and obtain the simulated overlap between the remaining candidate sites.

[0094] Step 8: Perform site addition operations on the remaining candidate sites whose simulated overlap is lower than the set threshold to obtain the TLS observation network.

[0095] Specifically, step 8.1: construct a Delaunay triangulation network for each candidate site in the TLS observation network; wherein the nodes in the Delaunay triangulation network are the candidate sites in the TLS observation network, and the edges of the Delaunay triangulation network represent the adjacency relationship between the candidate sites.

[0096] Step 8.2: Based on the Delaunay triangulation, construct an undirected graph of the TLS observation network; the nodes of the Delaunay triangulation are the nodes of the undirected graph, and the edges of the Delaunay triangulation are the edges of the undirected graph. For the weight of the edge of the undirected graph, the reciprocal of the overlap of the scanned simulation data of two adjacent candidate stations on the edge is calculated and assigned. The simulation overlap between the remaining candidate stations is the intersection of the simulated observation content of the two adjacent candidate stations divided by the minimum value of the observation target points of the two adjacent candidate stations.

[0097] Step 8.3: Generate the optimal registration path of the TLS observation network on the undirected graph of the TLS observation network using the minimum spanning tree method.

[0098] Step 8.4: Evaluate the compatibility of the TLS observation network. When the overlap of the scan simulations of two adjacent candidate sites on the best compatibility path is less than a certain threshold, lim When , encrypted supplementary sites are added between the two candidate sites.

[0099] Example 1

[0100] To illustrate the effectiveness of the present invention, a verification test was conducted on a campus covering an area of ​​approximately 660,000 m 2 , including many environmental elements such as buildings, roads, water systems, and forest areas. Figure 4 Complete elevation data for the survey area; Figure 5 It is the terrain elevation data of the survey area after filtering.

[0101] The entire campus was planned, and one area was selected for field scanning to verify the results. The scanner used in the simulation was RieglVZ400, with a scanning range of 600m, an elevation angle of 60°, and a maximum horizontal incidence angle of 70°. After another 74.54 minutes, the simulated scanning of all candidate sites was obtained. Then, using the greedy algorithm based on the characteristic lines of the building, the proposed planning results were obtained after 5.43 minutes. After checking the overlap of all routes and making targeted supplementary stations, the scanning planning results and their simulations were obtained as follows Figure 6 There are 43 buildings in the 700m×1200m observation area, 278 stations in the planned TLS observation network, and the simulated coverage rate of the building facades is 74.03%.

[0102] The present invention provides a more sophisticated three-dimensional simulation space, which greatly improves the simulation coverage compared to the traditional method, thereby enhancing the reliability of the TLS observation network. The comparison is shown in Table 1.

[0103] Table 1 Comparison of simulation coverage of simulation spaces with different three-dimensionality

[0104]

[0105] The present invention has the advantage of high efficiency, and can quickly generate a TLS observation network for large-scale building facade point cloud collection. The operating efficiency comparison is shown in Table 2.

[0106] Table 2 Comparison of efficiency of different methods

[0107]

[0108] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A TLS observation network planning method for large-scale buildings, characterized in that: include: Obtaining three-dimensional grid information of the observation area including voxel networks and building polygon vertex sets; Constructing a multi-resolution viewpoint grid according to the three-dimensional information of the grid, and determining an observation candidate area within the multi-resolution viewpoint grid; Set a screening grid in the observation candidate area and regard all grid points in the screening grid as observation candidate sites; Extracting building feature lines consisting of observation target points connected at the observation candidate sites from the voxel grid, and using the extracted building feature lines as the observation target point set; Based on the building polygon vertex set, the observation area is divided into multiple building cluster partitions with different building densities, and the simulated coverage rate of each building cluster partition is obtained; Iterate the resolution of the screening grid of each building cluster partition until the increase ratio of the simulated coverage of each building cluster partition is lower than the set ratio threshold, and use the simulated observation content corresponding to the obtained candidate site as the screening simulated observation content, and use the obtained observation target point union as the final observation target point union; Conduct observation simulation for each candidate site in the final observation target point set, and use the length of the building characteristic line scanned by simulation as the weight value of the candidate site; Select the candidate site with the largest current weight value as the selected candidate site, and delete the scan content of the selected candidate site from the screening simulation observation content; Iteratively select the remaining candidate sites in the final observation target point union until the maximum weight value of the current candidate site is lower than the set weight threshold, and obtain the simulated overlap between the remaining candidate sites. The remaining candidate sites whose simulated overlap is lower than the set threshold are supplemented to obtain the TLS observation network.

2. A TLS observation network planning method for large-scale buildings as claimed in claim 1, characterized in that: The three-dimensional information of the terrain and land features in the observation area is obtained through UAV aerial photogrammetry, and the three-dimensional information is expressed using voxel grids to obtain grid three-dimensional information including voxel networks and building polygon vertex sets.

3. A TLS observation network planning method for large-scale buildings as claimed in claim 1, characterized in that: The determining of the candidate region within the multi-resolution viewpoint grid specifically includes: Produce digital elevation model DEM and digital surface model DSM based on the grid three-dimensional information of the observation area; Obtain the areas where the difference between the digital surface model DSM in the observation area and the digital elevation model DEM in the remaining area is not 0, and mark the areas where the difference is not 0 as areas containing ponds and bushes where it is impossible to set up a station; The digital elevation model DEM is used to obtain the slope values ​​slop of all areas in the observation area, and the areas with slope values ​​slop greater than 30 are marked as areas that affect the safety of operators; Areas containing buildings, areas that have a serious impact on observation quality, areas where stations cannot be set up, and areas that affect operator safety are eliminated from the observation area to obtain candidate areas within the multi-resolution viewpoint grid.

4. A TLS observation network planning method for large-scale buildings as claimed in claim 1, characterized in that: The formula for obtaining the slope value slop of all areas in the observation area using the digital elevation model DEM is: Among them, dx is the width of the unit grid, dy is the length of the unit grid, and x i, j is a point in the digital elevation model DEM whose unit grid size is dx·dy, i is the horizontal coordinate value of the point, and j is the vertical coordinate value of the point.

5. The TLS observation network planning method for large-scale buildings as claimed in claim 1, characterized in that: The method of dividing the observation area into a plurality of building cluster partitions of building density based on the building polygon vertex set specifically includes: Delaunay triangulation is performed on the building polygon vertex set of the three-dimensional grid information of the observation area to obtain a Delaunay triangulated network. The obtained triangles are defined as three categories: if the three vertices of a triangle belong to two or three buildings, this triangle is a first-category triangle; if a triangle is not located in any building and its three vertices are all vertices of the same building, this triangle is a second-category triangle; the remaining triangles are third-category triangles; Construct the directional Voronoi diagram of the observation area to divide the observation area into different blocks. Each line segment in the directional Voronoi diagram is the boundary of each block. When there are first-type triangles and second-type triangles between two buildings, the two buildings are defined as topologically adjacent; The distances between topologically adjacent buildings are obtained, and the line segments between adjacent buildings below the set distance threshold are deleted in the directional Voronoi diagram to obtain the building cluster partitions.

6. A TLS observation network planning method for large-scale buildings as claimed in claim 1, characterized in that: The step of obtaining the simulated coverage of each building cluster partition specifically includes: The simulated observation contents of the candidate sites in each building cluster partition are combined to obtain the union of the observation target points; The number of observation target points in the union of observation target points is divided by the number of all observation target points in the building cluster partition to obtain the simulated coverage rate of the building cluster partition.

7. A TLS observation network planning method for large-scale buildings as claimed in claim 1, characterized in that: The step of using the obtained simulated observation content corresponding to the candidate site as the screening simulated observation content specifically includes: The first candidate site is selected according to the traversal order, and the maximum search range and the minimum search range of the candidate site are limited to complete the simulation of the scanning range; Traverse all observation object points of the candidate site. If the observation target point is not an endpoint on the edge of the building, obtain the angle relationship between the direction vector from the candidate site to the observation target point and the edge vector where the observation target point is located, and select the observation target point that meets the preset angle condition; Obtain the horizontal distance between the candidate site and the edge to which the point object belongs, and select the observation target points that meet the preset horizontal distance conditions; The candidate site is taken as the starting point and the observation object point is taken as the end point. The elevation of the interpolated sight point is obtained with a fixed step length s. The elevation of the interpolated sight point is compared with the element value corresponding to the interpolated sight point in the digital surface model DSM. When the elevation of the interpolated sight point is less than the elevation of the corresponding position of the digital surface model DSM, the sight is considered to be blocked, and the observation target points whose sight is not blocked are selected. Traverse all candidate sites and use lists to record the simulated observation contents of all candidate sites.

8. A TLS observation network planning method for large-scale buildings as claimed in claim 7, characterized in that: The formula for determining the value of the fixed step size s is: s=int(Δ / m) Δ=max(Δx,Δy) Δx=x b -x a ,Δy=y b -y a Among them, (x a ,y a ) is the coordinate of the candidate site, (x b ,y b ) is the coordinate of the observed target point, and m is the resolution of the voxel grid.

9. A TLS observation network planning method for large-scale buildings as claimed in claim 1, characterized in that: The operation of supplementing the remaining candidate sites whose simulated overlap is lower than the set threshold value specifically includes: Construct a Delaunay triangulation network for each candidate site in the TLS observation network; the nodes in the Delaunay triangulation network are the candidate sites in the TLS observation network, and the edges of the Delaunay triangulation network represent the adjacency relationship between the candidate sites; Based on the Delaunay triangulation, an undirected graph of the TLS observation network is constructed; the nodes of the Delaunay triangulation are the nodes of the undirected graph, and the edges of the Delaunay triangulation are the edges of the undirected graph; The optimal registration path of the TLS observation network is generated on the undirected graph of the TLS observation network using the minimum spanning tree method; The compatibility of the TLS observation network is evaluated. When the overlap of the scan simulations of two adjacent candidate sites on the best compatibility path is less than a certain threshold, lim When , encrypted supplementary sites are added between the two candidate sites.

10. A TLS observation network planning method for large-scale buildings as claimed in claim 1, characterized in that: The intersection of the simulated observation contents of two adjacent candidate sites divided by the minimum value of the observation target points of the two adjacent candidate sites is the simulated overlap between the remaining candidate sites.

Images