Ground laser scanner positioning network determination method for pre-assembled structure

Abstract

The invention provides a ground laser scanner positioning network determination method for a pre-assembled structure. Generating a planar model of a target structure and a peripheral area in the NetLogo platform, and marking a restricted area and a shielding area in the peripheral area to obtain a multi-agent model; the method comprises the following steps: discretizing the perimeter of a splicing interface and a non-splicing interface of a target structure according to different densities, obtaining a visual area of each discrete point according to a TLS constraint parameter, and superposing the visual areas of all discrete points on the perimeter to obtain a perimeter visual area global domain, namely a TLS site area global domain; the method comprises the following steps: calculating an initial site position set by adopting a genetic algorithm according to the whole site region, then establishing a TLS site network, and optimizing the site network by adopting a registration evaluation method to obtain a TLS positioning network. According to the method, the emphasis on the splicing interface is considered when the whole region of the TLS site region is established, the quality of the point cloud data obtained by the TLS positioning network can be improved, and the accuracy, rapidness and high efficiency of structure splicing are ensured.

Description

Technical Field

[0001] The invention relates to the technical field of scanning planning of a terrestrial laser scanner, and in particular to a method for determining a terrestrial laser scanner positioning network for a pre-assembled structure. Background Art

[0002] In recent years, prefabricated structures have been increasingly widely used in my country. Physical pre-assembly is an effective measure to ensure that the spatially related components of prefabricated structures can be accurately installed on site, but it has the disadvantages of high frame and labor costs, large site occupation and low efficiency. With the development and application of computer technology and three-dimensional measurement technology, physical pre-assembly has gradually been replaced by virtual pre-assembly. Virtual pre-assembly is a method of simulating the assembly of component splicing control points obtained by three-dimensional measurement in special software to achieve assembly accuracy. The terrestrial laser scanner (TLS) can quickly obtain the three-dimensional point cloud data of the structure. It has the advantages of high data accuracy, little external influence, and strong operability. It has become the preferred three-dimensional measurement equipment for virtual pre-assembly.

[0003] At present, the decision of TLS site layout mainly depends on the experience of operators, that is, the selection of TLS site location is highly subjective, and it is difficult to ensure the integrity of the final point cloud. Due to the lack of a systematic and reasonable decision-making method, it is still impossible to intelligently select the best layout location of TLS sites based on the site environment and target structure. At the same time, the point cloud acquisition of pre-assembled structures needs to meet the requirements of virtual pre-assembly work for point cloud data quality. Therefore, a systematic method is urgently needed to solve this problem.

[0004] Some engineers and technicians have transformed the scanning planning problem into the problem of finding the minimum number of predefined viewpoints, which provide complete coverage of the scanning target while meeting the data quality requirements. However, when measuring the external structure, many solutions do not consider that when placing TLS sites, there are obstructions that affect the line of sight of TLS, so that the final solution cannot obtain the expected results. Traditional site measurement also focuses less on special areas. For example, in the work of obtaining point cloud data for virtual pre-assembly of pre-assembled structures, it is necessary to obtain more accurate point cloud data for the assembly interface area to complete the extraction of characteristic assembly points. The lack of focused measurement of the assembly interface will result in the point cloud quality not meeting the virtual pre-assembly requirements. In addition, when aligning the point cloud, a certain degree of overlap is required between each TLS site to ensure the alignment effect. If only the minimum number of predefined sites is considered, the overlap between some adjacent sites is insufficient, resulting in the inability to automatically align the sites, making the structure assembly accuracy low. Repeated calculations will increase the workload of site calculations and the efficiency is low. However, too many TLS sites will reduce scanning efficiency and greatly increase measurement costs. Summary of the invention

[0005] In view of the technical problems existing in the prior art, the present invention provides a method for determining a terrestrial laser scanner positioning network suitable for pre-assembled structures, which has stronger application promotion, higher scanning efficiency in actual work, better data quality, and controllable measurement cost.

[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions.

[0007] A method for determining a terrestrial laser scanner positioning network for a pre-assembled structure comprises the following steps:

[0008] Step S1: Establish a structural information model in BIM software according to the target structural design plan;

[0009] Step S2: flatten the structural information model and import it into the NetLogo integrated platform to obtain a plane model of the target structure; generate a surrounding area model outside the plane model based on the photographs of the surrounding area of ​​the target structure obtained by shooting, mark the area where the laser scanner cannot be placed as a restricted area in the surrounding area model, and mark the area that will block the laser scanner's line of sight when measuring the target structure as an occlusion area, and obtain a multi-agent model; the multi-agent includes a target structure agent, an occlusion area agent, and a restricted area agent;

[0010] Step S3: discretize the perimeter of the target structure intelligent body into a point set, wherein the density of discrete points on the perimeter corresponding to the target structure assembly interface is greater than the density of discrete points on the perimeter corresponding to the non-assembly interface; take any discrete point on the perimeter as the starting point, and according to the constraint parameters of the laser scanner, search for a laser scanner site that can observe the starting point in the surrounding area model, and use the obtained site set as the visible area of ​​the starting point; superimpose the visible areas of all starting points on the perimeter to obtain the entire visible area of ​​the perimeter; based on the principle of visual heterogeneity, use the entire visible area of ​​the perimeter as the entire site area of ​​the terrestrial laser scanner;

[0011] Step S4: using a genetic algorithm to calculate the initial site position set of the laser scanner based on the entire site area of ​​the laser scanner obtained in S3 and the blocked area agent and restricted area agent obtained in S2;

[0012] Step S5: According to the initial site position set of the laser scanner, a laser scanner site network is established using the Delaunay triangulation method; the site network is optimized using a registration evaluation method to obtain a laser scanner positioning network.

[0013] It should be noted that in step S2, the photograph of the surrounding area of ​​the target structure is taken by a drone or satellite, or by a total station positioning plus camera photography method.

[0014] It should be noted that in step S2, the restricted area includes narrow spaces in the surrounding area of ​​the target structure, unstable ground, and natural areas where equipment cannot be installed; among them, natural areas where equipment cannot be installed include rivers and canyons.

[0015] It should be noted that, in step S3, the constraint parameters of the laser scanner are working distance, incident angle, and field of view angle.

[0016] It should be noted that in step S4, the fitness function used by the genetic algorithm is expressed as follows according to formula (1):

[0017] (1)

[0018] Where: F(I i ) is individual I i The fitness score of R(I i ) is individual I i The fitness score of the non-penalty term is based on its performance without violating any constraints; P(I i ) is a penalty term used to reduce the fitness score of individuals that violate the constraints;

[0019] The individual refers to any set of laser scanner sites in the population; the population refers to any set of laser scanner sites in the genetic algorithm calculation process;

[0020] Among them, R(I i ) is expressed as follows according to formula (2):

[0021] (2)

[0022] Where: For individual I i The entropy of opinions; For individual I i The size of the positioning network; For individual I i The degree of connectivity between the sites;

[0023] Among them, the penalty term P(I i ) can be expressed as follows according to formula (3):

[0024] (3)

[0025] Where: n is individual I i The total number of sites j in the is the weight associated with site j; is the indicator function, if individual I i If the middle station j is placed in the blocked area or restricted area, then =1, otherwise =0.

[0026] It should be noted that in step S4, when the genetic algorithm is used to calculate the initial site position set of the laser scanner, the selection operation adopts the tournament selection method, and the crossover operation adopts the single-point crossover method.

[0027] It should be noted that the step S5 further includes:

[0028] S51, based on the initial site location set of the laser scanner, a site network is established using the Delaunay triangulation method, where nodes represent sites and edges represent adjacency relationships between sites;

[0029] S52, assigning a weight to each edge in the site network to obtain an undirected weighted graph; the weight of each edge is the overlap of the corresponding visible areas between adjacent sites;

[0030] S53, generating an initial registration path using a maximum spanning tree algorithm according to the undirected weighted graph;

[0031] S54, checking the overlap of each edge in the initial registration path, if the overlap is lower than a threshold, adding new sites between adjacent sites until the overlap of all adjacent sites meets the threshold requirement, thereby obtaining a preferred registration path;

[0032] S55. Obtain a laser scanner positioning network according to the preferred registration path and corresponding sites.

[0033] The threshold is an overlap threshold determined according to an automatic registration algorithm. The automatic registration algorithm refers to a method for automatic registration between point cloud data of adjacent sites, and generally can be 2% to 10%.

[0034] The basic principle of the present invention is as follows: the present invention establishes a multi-agent model including a target structure, an occluded area, and a restricted area in the NetLogo integrated platform, and then discretizes the perimeter of the target structure assembly interface and the non-assembly interface perimeter according to different point densities. After obtaining the corresponding visible area for each discrete point, it is superimposed to obtain the entire visible area of ​​the perimeter of the target structure agent (i.e., the entire candidate area of ​​the site), and then based on the principle of visual interchange, the entire visible area of ​​the perimeter is used as the entire site area of ​​the laser scanner, and then the site position set is calculated by a genetic algorithm based on the entire site area, and the best ground laser is determined based on the registration evaluation method. Scanner (TLS) positioning network; the present invention considers the emphasis on the assembly interface area to meet the requirements of point cloud data quality during virtual pre-assembly, and generates a reasonable TLS positioning network on this basis, so that the generated TLS positioning network can obtain a large amount of high-quality point cloud data in the shortest working time during virtual pre-assembly, reduce the time required for measurement work, and thus reduce measurement costs; at the same time, improve the quality of point cloud data, reduce point cloud data redundancy, ensure the integrity of point cloud data collection, and achieve faster and more effective post-processing, ensure the accuracy of structure assembly, have the advantages of fast and efficient, and greatly reduce the workload of site calculation.

[0035] Compared with the prior art, the present invention has the following advantages:

[0036] 1) By dividing and marking the areas where the line of sight is blocked and where TLS cannot be placed, the resulting TLS positioning network is more in line with engineering practice;

[0037] 2) By using the NetLogo integrated platform to build a multi-agent model, the computational efficiency is improved while making the computational process more intuitive and accurate;

[0038] 3) By focusing on the target structure assembly interface, the quality of the point cloud data obtained by the TLS positioning network is improved to meet the requirements for point cloud data quality during virtual pre-assembly work;

[0039] 4) The site network is optimized through the registration evaluation method, so that the optimized positioning network gets rid of the problem of existing methods that only pursue the minimum number of sites and ignore the need for a certain degree of overlap between point cloud data of adjacent sites to ensure the registration effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 The present invention is a flowchart of a method for determining a terrestrial laser scanner positioning network.

[0041] Figure 2 Schematic diagram of the incident angle of the terrestrial laser scanner.

[0042] Figure 3This is a schematic diagram of the target structure plane model in the NetLogo integration platform.

[0043] Figure 4 A schematic diagram of the visible area around the target structure in the NetLogo integration platform.

[0044] Figure 5 Schematic diagram of a set of terrestrial laser scanner site locations.

[0045] Figure 3-5 In the figure, black is the background color of the NetLogo integration platform, the edge line with a small triangle represents the boundary of the blocked area, the edge line with a small diamond represents the boundary of the virtual pre-assembly interface, the edge line with a small circle represents the boundary of the restricted area, the ordinary edge line represents the boundary of the non-pre-assembly interface of the target structure, the gray shaded area represents the visible area of ​​the discrete points on the perimeter of the target structure, and the small white dots represent the ground laser scanner sites.

[0046] Figure 6 This is the flow chart of the genetic algorithm.

[0047] Figure 7 Schematic diagram of the Delaunay triangulation site network.

[0048] Figure 8 Schematic diagram of the undirected weighted graph of the site network.

[0049] Fig. 9 Schematic diagram of the initial registration path.

[0050] Fig.10 Schematic diagram of the optimized registration path. DETAILED DESCRIPTION

[0051] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0052] The present invention provides a method for designing a terrestrial laser scanner positioning network for pre-assembled structures, and uses on-site photos and design schemes of known structures to build a structural information model of the construction site in the BIM software Revit. The structural information model is flattened and imported into the NetLogo integrated platform, and parameter constraints are obtained according to the specifications of the terrestrial laser scanner used. Restricted areas where the terrestrial laser scanner cannot be placed and blocked areas that will affect the measurement line of sight are defined according to the structural information model and the construction site photos; a multi-agent model is generated by setting the model structure and model parameters; in the multi-agent model, a TLS site location set is generated according to the emphasis on the assembly interface, and a genetic algorithm is used to calculate the site network of the terrestrial laser scanner, and the optimal location set is obtained from the site location set. Finally, the positioning network is optimized based on the registration evaluation method to obtain a point cloud data scanning positioning network that meets the registration standard.

[0053] Example 1: Please refer to Figure 1-10 .

[0054] The embodiment of the present invention discloses a method for designing a terrestrial laser scanner positioning network for a pre-assembled structure, comprising the following steps:

[0055] Step S1: According to the design plan of the target structure, a structural information model of the target structure is built in the BIM software.

[0056] The design plan (design drawings) of the target structure is generally provided by the construction unit or design unit, and the BIM software generally uses Revit software.

[0057] Step S2: Flatten the structural information model and import it into the NetLogo integrated platform to obtain a plane model of the target structure; generate a surrounding area model outside the plane model of the target structure based on the photographs of the surrounding area of ​​the target structure taken, mark the area where the laser scanner cannot be placed as a restricted area in the surrounding area model, and mark the area that will block the laser scanner's line of sight when measuring the target structure as an occlusion area, to obtain a multi-agent model; the multi-agent includes a target structure agent, an occlusion area agent and a restricted area agent.

[0058] The photos of the surrounding area of ​​the construction site can be taken by drone or satellite, or by total station positioning plus traditional photo shooting methods, such as high-definition camera shooting, to determine all the features of the surrounding area and their positional relationship relative to the target structure, including structures, facilities, landforms, etc. The surrounding area generally refers to the ground area outside the perimeter of the target structure and with a radius not less than the working distance of the ground laser scanner.

[0059] The areas in the surrounding area where terrestrial laser scanners cannot be placed are restricted areas for terrestrial laser scanners, including narrow spaces (space dimensions are smaller than the dimensions required for terrestrial laser scanner installation), unstable ground (i.e., ground that cannot provide a stable measurement environment for terrestrial laser scanners), and natural areas where equipment cannot be set up; among them, natural areas where equipment cannot be set up include rivers and canyons.

[0060] When building a multi-agent model in the NetLogo integrated platform, you need to first set the background and size of the multi-agent model according to the size of the target structure and the surrounding area, and construct a plane layout of the target structure plane model and its surrounding area model. The target structure agent can be obtained from the plane model of the target structure. Then, based on the photographs taken, mark the occlusion area and restricted area in the surrounding area, adjust and determine the shape and position of each agent, create an occlusion area agent and a restricted area agent respectively, and select colors and attributes for each agent in the NetLogo integrated platform to achieve visual distinction between the agents, such as Figure 3 .

[0061] Step S3: discretize the perimeter of the target structure intelligent body into a point set, wherein the density of discrete points on the perimeter corresponding to the target structure assembly interface is greater than the density of discrete points on the perimeter corresponding to the non-assembly interface; take any discrete point on the perimeter as the starting point, and according to the constraint parameters of the terrestrial laser scanner, search for the position of the laser scanner site that can observe the starting point in the surrounding area model, and use the obtained site set as the visible area of ​​the starting point; superimpose the visible areas of all starting points on the perimeter to obtain the entire visible area of ​​the perimeter; based on the principle of visual heterogeneity, use the entire visible area as the entire site area of ​​the terrestrial laser scanner.

[0062] The constraint parameters of the terrestrial laser scanner are the working distance, incident angle, and field of view angle of the terrestrial laser scanner;

[0063] Working distance: the distance that a terrestrial laser scanner device can scan;

[0064] Incident angle: The angle between the laser beam emitted by the terrestrial laser scanner and the corresponding normal line of a surface of the scanned target structure; Figure 2 ;

[0065] Field of view (FOV): The viewing angle of a terrestrial laser scanner;

[0066] In this embodiment, a Trimble X7 terrestrial laser scanner is taken as an example, whose working distance is 0.6m-80m, the incident angle is 40°, and the field of view is 360°×282°.

[0067] The visible area of ​​any starting point on the perimeter of the target structure is marked with the same reference color. Color coding is used to indicate the overlap of the visible areas corresponding to all starting points on the perimeter of the target structure. The depth of the color corresponds to the degree of overlap. The lighter the color, the higher the overlap of the visible area, and the darker the color, the lower the overlap of the visible area. The lighter the color, the superposition of the visible areas of multiple perimeter points, which means that the visibility of the station locations in these areas is higher (more points on the perimeter are observed), and the darker the color, the superposition of the visible areas of fewer perimeter points, which means that the visibility of these areas is lower (fewer points on the perimeter are observed). Figure 4. This color coding method allows intuitive identification of which site locations can observe a larger range of target structures. Ultimately, this process helps us quantify the amount of observation information for each potential terrestrial laser scanner site location. Because it is necessary to focus on the assembly interface area, when looking for the visible area outside the perimeter corresponding to the assembly interface, by encrypting the observed points (i.e., discrete points) on the perimeter and increasing the amount of information contained in the observed points at the site location (observation point), the weight of the observed point information contained in the site in the visible area outside the perimeter corresponding to the assembly interface can be increased relative to the visible area outside the perimeter of the non-assembly interface.

[0068] Since the density of discrete points on the perimeter corresponding to the target structure assembly interface is greater than the density of discrete points on the perimeter corresponding to the non-assembly interface, that is, after the discrete points on the perimeter of the assembly interface are encrypted (densified), after the discrete points on the perimeter corresponding to the assembly interface generate a visible area, the amount of information contained in the ground laser scanner site position set composed of this part of the area is greater than the amount of information contained in the corresponding perimeter of the non-assembly interface or the corresponding site position set before encryption. Then, when calculating the ground laser scanner site at the adjacent position of the perimeter corresponding to the assembly interface and the perimeter corresponding to the non-assembly interface, based on the above encryption work, when calculating the position of the ground laser scanner site in the subsequent calculation, it will naturally be closer to the perimeter corresponding to the assembly interface, so that the ground laser scanner can obtain more and more accurate point cloud data about the assembly interface during measurement, thereby completing the extraction of characteristic assembly points. The characteristic assembly point is the splicing control point during virtual pre-assembly. The assembly control point is a key factor in ensuring the accuracy of engineering assembly, improving construction efficiency and ensuring structural safety. The more accurate the characteristic assembly point, the higher the assembly accuracy.

[0069] Step S4: Based on the entire site area of ​​the terrestrial laser scanner obtained in S3 and the blocked area agent and restricted area agent obtained in S2, a genetic algorithm is used to calculate the initial site position set of the terrestrial laser scanner.

[0070] See also Figure 6 , the steps of calculating the site location set by the genetic algorithm can be refined into the following process:

[0071] 1) Initialization: Randomly generate a set of terrestrial laser scanner sites as the initialization population;

[0072] The initialization population includes multiple randomly generated laser scanner site sets; site set I i That is, individual I in the population i ; Individual I i The number of sites included in is determined according to the size of the target structure. i The number of sites included in is the same. In this embodiment, the initial number of sites is 12.

[0073] 2) Individual evaluation: calculate the fitness of each individual in the population;

[0074] The fitness function used by the genetic algorithm is expressed as follows:

[0075] (1)

[0076] Where: F(I i ) is individual I i The fitness score of R(I i ) is individual I i The fitness score of the non-penalty term is based on its performance without violating any constraints; P(I i ) is a penalty term used to reduce the fitness score of individuals that violate the constraints;

[0077] Among them, R(I i ) is defined as the weighted sum of three parameters, expressed as follows according to formula (2):

[0078] (2)

[0079] Where: For individual I i The entropy of opinions; For individual I i The size of the positioning network; For individual I i The degree of connectivity between sites in the network; the opinion entropy Γ quantifies a possible solution (i.e., individual I i ) provides information; the positioning network size Φ quantifies a possible solution (i.e., individual I i ) of the site locations; the connectivity Δ between the site locations measures a possible solution (i.e., individual I i ) between the site locations; , , The parameters are , , The weight of .

[0080] Determine the weight , , The principle is to ensure a balance between the amount of information obtained and the number of sites and to ensure that the visible area has a certain degree of overlap during subsequent registration. , , They are 0.5, 0.4 and 0.1 respectively.

[0081] Individual Ii The proportion of the perimeter of the target structure scanned by the site set is determined by the ratio of the number of discrete points on the perimeter of the target structure scanned by the site set to the total number of discrete points on the perimeter; It is determined based on the ratio of the number of inactivated sites to the total number of sites in the site set; It is determined by the connectivity of the site set, where connectivity is 1 and disconnection is 0.

[0082] During calculation, an individual is a coding string (including sites and their order). During the calculation process, an explicit or implicit activation mechanism is usually added to the coding string. When the site is in the entire site area but not in the blocked area agent and the restricted area agent, it is dominant, and in other cases it is recessive. Explicit means activated, and recessive means inactivated. The purpose of setting recessive sites is to increase the genetic diversity of individuals and facilitate crossover and mutation operations.

[0083] Among them, the penalty term P(I i ) can be expressed as follows according to formula (3):

[0084] (3)

[0085] Where: n is individual I i The total number of sites j in ; is the weight associated with site j, which can be set according to the importance of the blocked area or restricted area. The larger the weight, the more severe the penalty for violating the constraint; is the indicator function, if individual I i If the middle station j is placed in the blocked area or restricted area, then =1, otherwise =0; In this embodiment, when a site (individual I i ) is in a blocked area or restricted area, take =∞, thereby eliminating such sites.

[0086] The fitness function is used to quantify the quality of individuals (site sets). The penalty term is proposed because the sites in restricted or blocked areas are eliminated, thereby completing the screening of individuals (site sets).

[0087] 3) Selection operation: select individuals according to their fitness values ​​to reproduce and form a new generation, using the tournament selection method;

[0088] 4) Crossover operation: Generate new individuals through crossover operation. Select single-point crossover, that is, randomly select a crossover point in the encoding string of two individuals (position sets), and then exchange the sequence after the crossover point to generate a new individual;

[0089] 5) Mutation operation: Mutation operation is applied to increase the diversity of the population. Mutation can randomly change certain genes of individuals, which is manifested as changing the position of a certain site in the position set, the structure of the position set, etc. After selection, crossover and mutation operations, the fitness is calculated again to eliminate some individuals to generate a new generation of population;

[0090] 6) Repeat the steps of selection, crossover, mutation, evaluation, elimination of some individuals and generation of a new generation of population until the preset iteration termination condition is reached.

[0091] The iteration termination condition is: when the fitness of the best individual and the fitness of the group no longer increase, the iteration terminates. In this embodiment, when the number of iterations is less than 3000, the fitness no longer increases. After the algorithm stops, the most fit individual is output as the site location set, such as Figure 5 shown.

[0092] The entire area of ​​the terrestrial laser scanner site is the definition domain of the genetic algorithm calculation, and the occluded area agent and the restricted area agent are the constraints of the calculation.

[0093] Step S5: According to the initial site position set of the laser scanner, a laser scanner site network is established using the Delaunay triangulation method; the site network is optimized using a registration evaluation method to obtain a ground laser scanner positioning network;

[0094] In this embodiment, the positioning network is optimized based on the registration evaluation method, which can ensure that the overlap between the scanning areas of adjacent sites meets the basic requirements of the automatic registration algorithm. The automatic registration algorithm refers to the automatic registration between the point cloud data of adjacent sites. When measuring with a terrestrial laser scanner, there is generally a matching field software to achieve automatic registration.

[0095] Step S5 can be further refined as follows:

[0096] 1) Based on the initial site location set of the terrestrial laser scanner, the Delaunay triangulation method is used to establish a site network, where nodes represent sites and edges represent the adjacency relationship between sites, such as Figure 7 ;

[0097] 2) Assign weights to the edges in the site network. The weight of each edge is the overlap of the corresponding visible areas between adjacent sites, and an undirected weighted graph is obtained. This ensures that the total weight of the registration path is maximized, thereby optimizing the registration process.

[0098] 3) Based on the undirected weighted graph, the maximum spanning tree algorithm is used to generate the initial registration path, such as Fig. 9 ; The resulting registration path will connect all sites and ensure that the point cloud data of the entire site network can be effectively registered through the resulting registration path;

[0099] 4) Check the overlap of each edge in the registration path. If the overlap is lower than the preset threshold, add a new site between adjacent sites. Since the automatic registration algorithm will automatically realize the automatic registration of the visible areas of adjacent sites to ensure the overlap of the site point cloud data, the preset overlap threshold of the non-assembly interface site can be taken as >0% (generally 2%~10%, 5% in this embodiment); then for the assembly interface, since the present invention increases the weight of the observed point information contained in the site outside the perimeter corresponding to the assembly interface, the preset overlap threshold of the assembly interface can be taken as the same overlap threshold as the non-assembly interface site. The larger the overlap threshold, the more ground laser scanner sites there are, so it is necessary to select according to the actual project situation. If the overlap between adjacent scanning positions is lower than the preset threshold, successful registration may not be possible. In this case, it is necessary to add new scanning positions between these adjacent scanning positions to increase their overlap. Until the overlap of all adjacent scanning positions meets the threshold requirement, the optimal registration path is finally obtained, such as Fig.10 .

[0100] 5) According to the preferred registration path and the obtained sites, a laser scanner positioning network is obtained; the nodes in the preferred registration path are the ground laser scanner sites, and the edges are the connections between adjacent ground laser scanner sites, that is, the positioning network of the ground laser scanner is obtained.

[0101] When adding a new site between adjacent sites with insufficient overlap, first try to take the midpoint of the two adjacent sites as the new site; then check whether the new site is feasible and safe (whether it is in the restricted area and the blocked area), and whether the overlap with the visible area of ​​the adjacent site meets the requirements; if the conditions are met, the new site will be adopted; if not, search for new sites in the adjacent ground until the requirements are met.

[0102] In this embodiment, there are 3 new sites, totaling 15 sites, and the number of new sites is small, which is basically in line with expectations. When the number of new sites exceeds expectations, a small number of sites can be added when selecting the initial number of sites (generating the initialization population) so that the total number of sites recalculated meets expectations, thereby reducing measurement costs.

[0103] The above are only preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications made without departing from the principle of the present invention should be regarded as the protection scope of the present invention.

Claims

1. A method for determining the positioning of a terrestrial laser scanner for a pre-assembled structure, characterized in that: The steps include: Step S1: Establish a structural information model in BIM software according to the design plan of the target structure; Step S2: flatten the structural information model and import it into the NetLogo integrated platform to obtain a plane model of the target structure; generate a surrounding area model outside the plane model based on the photographs of the surrounding area of ​​the target structure obtained by shooting, mark the area where the laser scanner cannot be placed as a restricted area in the surrounding area model, and mark the area that will block the laser scanner's line of sight when measuring the target structure as an occlusion area, and obtain a multi-agent model; the multi-agent includes a target structure agent, an occlusion area agent, and a restricted area agent; Step S3: discretize the perimeter of the target structure intelligent body into a point set, wherein the density of discrete points on the perimeter corresponding to the target structure assembly interface is greater than the density of discrete points on the perimeter corresponding to the non-assembly interface; take any discrete point on the perimeter as the starting point, and according to the constraint parameters of the laser scanner, search for a laser scanner site that can observe the starting point in the surrounding area model, and use the obtained site set as the visible area of ​​the starting point; Superimposing the visible areas of all starting points on the perimeter to obtain the entire visible area of ​​the perimeter; Based on the principle of visual difference, the entire visible area of ​​the perimeter is used as the entire site area of ​​the terrestrial laser scanner; Step S4: using a genetic algorithm to calculate the initial site position set of the laser scanner based on the entire site area of ​​the laser scanner obtained in S3 and the blocked area agent and restricted area agent obtained in S2; Step S5: According to the initial site position set of the laser scanner, a laser scanner site network is established using the Delaunay triangulation method; the site network is optimized using a registration evaluation method to obtain a laser scanner positioning network.

2. The method according to claim 1, characterized in that: In step S2, the photograph of the surrounding area of ​​the target structure is taken by using a drone or satellite, or by using a total station positioning plus camera photography.

3. The method according to claim 1, characterized in that: In step S2, the restricted area includes narrow spaces in the surrounding area of ​​the target structure, unstable ground, and natural areas where equipment cannot be installed; wherein the natural areas where equipment cannot be installed include rivers and canyons.

4. The method according to claim 1, characterized in that: The constraint parameters of the laser scanner are working distance, incident angle, and field of view angle.

5. The method according to claim 1, characterized in that: In step S4, the fitness function used by the genetic algorithm is expressed as follows according to formula (1): （1） Where: F(I i ) is individual I i The fitness score of R(I i ) is individual I i The fitness score when the penalty term is not triggered; P(I i ) is a penalty term; The individual refers to any set of laser scanner sites in the population; the population refers to any set of laser scanner sites in the genetic algorithm calculation process; Among them, R(I i ) is expressed as follows according to formula (2): （2） Where: For individual I i The entropy of opinions; For individual I i The size of the positioning network; For individual I i The degree of connectivity between the sites; Among them, P(I i ) is expressed as follows according to formula (3): （3） Where: n is individual I i The total number of sites j in ; is the weight associated with site j; is the indicator function, if individual I i If the middle station j is placed in the blocked area or restricted area, then =1, otherwise =0.

6. The method according to claim 1, characterized in that: In step S4, when the genetic algorithm is used to calculate the initial site position set of the laser scanner, the selection operation adopts the tournament selection method, and the crossover operation adopts the single point crossover.

7. The method according to claim 1, characterized in that: The step S5 further comprises: S51, based on the initial site location set of the laser scanner, a site network is established using the Delaunay triangulation method, where nodes represent sites and edges represent adjacency relationships between sites; S52, assigning a weight to each edge in the site network to obtain an undirected weighted graph; the weight of each edge is the overlap of the corresponding visible areas between adjacent sites; S53, generating an initial registration path using a maximum spanning tree algorithm according to the undirected weighted graph; S54, checking the overlap of each edge in the initial registration path, if the overlap is lower than a threshold, adding new sites between adjacent sites until the overlap of all adjacent sites meets the threshold requirement, thereby obtaining a preferred registration path; S55. Obtain a laser scanner positioning network according to the preferred registration path and corresponding sites.

Images