Manufacturing method and device of bridge live-action three-dimensional point cloud data set and medium

Abstract

The invention belongs to the field of point cloud data, and discloses a bridge live-action three-dimensional point cloud data set manufacturing method and device and a medium, and the method comprises the steps: obtaining a data set manufacturing demand; when the data set manufacturing demand is the point cloud data set manufacturing demand of the bridge components, obtaining a plurality of bridge components in the bridge model; extracting component parameters of each bridge component, and constructing a point cloud data set of the bridge components based on the component parameters of each bridge component and the parameters of the bridge model; and / or when the data set manufacturing demand is the point cloud data set manufacturing demand of the bridge model, obtaining entity objects corresponding to bridge components in the bridge model; and performing triangularization and laser simulation scanning on the surface of the entity object to obtain point cloud space information of the entity object, and constructing a point cloud data set of the bridge model according to the point cloud space information of the entity object. By means of the data set manufacturing method, the data set meeting the deep learning requirement can be efficiently generated.

Description

Technical Field

[0001] The present invention belongs to the technical field of point cloud data, and in particular relates to a method, device and medium for producing a three-dimensional point cloud data set of a real bridge scene. Background Art

[0002] Identifying key bridge components and constructing 3D scenes are crucial for bridge construction and subsequent dynamic monitoring. Deep learning methods for addressing these challenges require sequential, real-world 3D observations. However, bridges vary in structure, have long construction cycles, and sample sizes are severely limited.

[0003] The existing methods for acquiring small amounts of simulated data mainly rely on surface sampling, which cannot simulate the actual laser scanning process and cannot fully estimate the impact of factors such as scanning intervals and occlusions. The results obtained are very different from the actual data. Summary of the Invention

[0004] The purpose of the present invention is to provide a method, device and medium for producing a three-dimensional point cloud dataset of a real bridge scene, so as to solve the problem that the existing method for obtaining a small amount of simulated data mainly relies on surface sampling, cannot simulate the actual laser scanning process, cannot fully estimate the influence of factors such as scanning interval and occlusion, and the obtained results are very different from the actual data.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions: In a first aspect, the present invention provides a method for producing a real-scene three-dimensional point cloud dataset of a bridge, the method comprising: Acquiring data set production requirements, the data set production requirements including: point cloud data set production requirements for a bridge model and / or point cloud data set production requirements for bridge components; When the dataset preparation requirement is a requirement for preparing a point cloud dataset of bridge components, obtaining several bridge components in the bridge model; extracting component parameters of each bridge component, and constructing a point cloud dataset of the bridge components based on the component parameters of each bridge component and parameters of the bridge model; and / or When the dataset production requirement is a point cloud dataset production requirement for a bridge model, the entity objects corresponding to each bridge component in the bridge model are obtained; the surfaces of the entity objects are triangulated and laser simulated scanned to obtain point cloud spatial information of the entity objects, and the point cloud dataset of the bridge model is constructed based on the point cloud spatial information of the entity objects.

[0006] Preferably, constructing a point cloud dataset of the bridge components based on the component parameters of each bridge component and the parameters of the bridge model includes: Automatically correcting component parameters based on the proportions of each bridge component to obtain corrected component parameters of each bridge component; Converting the bridge model corresponding to the corrected component parameters of each bridge component into a first mesh model, wherein the first mesh model includes a plurality of triangular faces; Based on the uniform sampling method, determine the sampling points of each triangle; Based on the sampling points of each triangular face and the corrected component parameters of each bridge component, a point cloud dataset of each bridge component is constructed.

[0007] Preferably, the uniform sampling method is: Calculate the area of ​​each triangle of the first mesh model and the total area of ​​all triangles of the first mesh model; Determining the number of sampling points that need to be generated on each triangular face of the first mesh model according to the proportion of the area of ​​each triangular face of the first mesh model to the total area of ​​all triangular faces of the first mesh model; Construct the barycentric coordinate system of each triangle face of the first mesh model; In the barycentric coordinate system of each triangular face of the first mesh model, uniformly distributed sampling points are generated according to the vertices of the triangular face and the number of sampling points to be generated on the triangular face.

[0008] Preferably, before constructing the point cloud dataset of each bridge component based on the sampling points of each triangular face and the corrected component parameters of each bridge component, the method also includes: adding random noise to the sampling points of each triangular face, and the random noise includes: Gaussian noise, random block noise and / or random block excavation.

[0009] Preferably, when the random noise is random block noise and / or random block excavation, the random block is a cubic block and / or a spherical block, wherein, when the random block is a spherical block, spherical block noise and / or spherical block excavation is added to the sampling point of each triangular surface based on the spherical space constraint; when the random block is a cubic block, cubic block noise and / or cubic block excavation is added to the sampling point of each triangular surface based on the cubic space constraint.

[0010] Preferably, the spherical space constraint is that the noised or removed sampling points are evenly distributed within the spherical radius; The cubic space constraint is that the sampling points to be noised or mined are located within the range of the cube and are distributed according to a preset distribution rule.

[0011] Preferably, triangulation and laser simulation scanning are performed on the surface of the physical object to obtain point cloud spatial information of the physical object, including: Reading the faces of the physical object, triangulating the faces of the physical object to obtain a number of triangular facets, and constructing labels for each triangular facet based on the bridge components corresponding to the physical object; Convert the bridge model into a second mesh model, load the second mesh model, and use a simulated drone to perform laser simulation scanning on the second mesh model; Obtain the laser emission direction, drone position, time interval between laser signal emission and reception, and scanning timestamp of the simulated drone during the laser simulation scanning process; According to the laser emission direction of the simulated drone, the position of the drone and the time interval between the emission and reception of the laser signal during the laser simulation scanning process, the point cloud coordinates of each triangular facet are determined, and the point cloud coordinates of each triangular facet of the physical object, the label of each triangular facet and the point cloud coordinates of each triangular facet are used as the point cloud spatial information of the physical object.

[0012] Preferably, a simulated drone is used to perform laser simulation scanning on the second mesh model, including: The center line of the bridge deck model is constructed based on the bridge model, and the scanning route is generated based on the center line of the model; The simulated UAV performs flight scanning according to a scanning route, and emits a scanning laser within a preset laser emission range and in a preset laser emission direction to perform laser simulation scanning on the second mesh model.

[0013] In a second aspect, the present invention provides a device for producing a real-scene 3D point cloud dataset of a bridge, for implementing the above-mentioned method for producing a real-scene 3D point cloud dataset of a bridge, the device comprising: An acquisition module is used to acquire a data set production requirement, wherein the data set production requirement includes: a point cloud data set production requirement of a bridge model and / or a point cloud data set production requirement of a bridge component; a production module, configured to obtain a plurality of bridge components in a bridge model when the data set production requirement is a point cloud data set production requirement for bridge components; extract component parameters of each bridge component, and construct a point cloud data set of the bridge components based on the component parameters of each bridge component and parameters of the bridge model; and / or When the dataset production requirement is a point cloud dataset production requirement for a bridge model, the entity objects corresponding to each bridge component in the bridge model are obtained; the surfaces of the entity objects are triangulated and laser simulated scanned to obtain point cloud spatial information of the entity objects, and the point cloud dataset of the bridge model is constructed based on the point cloud spatial information of the entity objects.

[0014] In a third aspect, the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned method for producing a real-scene three-dimensional point cloud dataset of a bridge when executing the computer program.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for producing a real-scene three-dimensional point cloud dataset of a bridge.

[0016] Beneficial effects: 1. The present invention can produce two different point cloud datasets according to needs. Specifically, the present invention extracts the component parameters of each bridge component and constructs a point cloud dataset of the bridge component based on the component parameters of each bridge component and the parameters of the bridge model. At the same time, the present invention also obtains the physical objects corresponding to each bridge component in the bridge model. The surfaces of the physical objects are triangulated and laser simulated to scan the physical objects to obtain the point cloud spatial information of the physical objects. Based on the point cloud spatial information of the physical objects, a point cloud dataset of the bridge model is constructed. Therefore, during the training process, the present invention can select different point cloud datasets according to different training tasks. For example, the point cloud dataset of the bridge component is selected for the classification task, and the point cloud dataset of the bridge model is selected for the segmentation task. 2. The dataset preparation method of the present invention can efficiently generate a dataset that meets the needs of deep learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following detailed description, they are used to explain the embodiments of the present invention, but do not constitute a limitation of the embodiments of the present invention. In the accompanying drawings: Figure 1 This is a flow chart of a method for producing a real-scene three-dimensional point cloud dataset of a bridge provided by one embodiment of the present invention; Figure 2 This is a block diagram of a device for producing a real-scene three-dimensional point cloud dataset of a bridge provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the present invention will be briefly introduced below in conjunction with the drawings and the description of the embodiments or the prior art. Obviously, the following description of the structure of the drawings is only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention.

[0019] Example 1 Figure 1 This is a flow chart of a method for producing a real-scene three-dimensional point cloud dataset of a bridge provided by one embodiment of the present invention. Figure 1 As shown, this embodiment provides a method for producing a bridge real-scene 3D point cloud dataset, the method comprising: Step S10: Obtaining dataset production requirements, which include: requirements for producing a point cloud dataset of a bridge model and / or requirements for producing a point cloud dataset of bridge components. In this embodiment, the point cloud dataset of the bridge model is mainly used for segmentation tasks. The purpose of the segmentation task is to accurately separate and mark the boundaries and areas of each part of the bridge. Therefore, the segmentation task requires a complete bridge model and its accurate geometric information to ensure that the model can identify and segment each component. The point cloud dataset of the bridge component is mainly used for classification tasks. Classification tasks usually only require the dataset to contain an image or model of a single bridge component. Such tasks mainly focus on feature recognition or parameter prediction of specific components. Therefore, the production of the dataset can focus on a local detail of the bridge. Step S20: when the dataset preparation requirement is a requirement for preparing a point cloud dataset of bridge components, obtaining a plurality of bridge components in the bridge model; extracting component parameters of each bridge component, and constructing a point cloud dataset of the bridge components based on the component parameters of each bridge component and the parameters of the bridge model; and / or When the dataset production requirement is a point cloud dataset production requirement for a bridge model, the entity objects corresponding to each bridge component in the bridge model are obtained; the surfaces of the entity objects are triangulated and laser simulated scanned to obtain point cloud spatial information of the entity objects, and the point cloud dataset of the bridge model is constructed based on the point cloud spatial information of the entity objects.

[0020] The present invention can produce two different point cloud datasets according to needs, namely, by extracting the component parameters of each bridge component, constructing a point cloud dataset of the bridge component based on the component parameters of each bridge component and the parameters of the bridge model; at the same time, the entity objects corresponding to each bridge component in the bridge model are obtained; the surface of the entity object is triangulated and laser simulated scanned to obtain the point cloud spatial information of the entity object, and the point cloud dataset of the bridge model is constructed according to the point cloud spatial information of the entity object; therefore, during the training process, the present invention can select different point cloud datasets according to different training tasks, for example: the classification task selects the point cloud dataset of the bridge component, and the segmentation task selects the point cloud dataset of the bridge model; and the dataset production method of the present invention can efficiently generate a dataset that meets the needs of deep learning.

[0021] As a further optimization of this embodiment, a point cloud dataset of the bridge components is constructed based on the component parameters of each bridge component and the parameters of the bridge model, including: Step A10: Automatically correct the component parameters based on the proportions of each bridge component to obtain the corrected component parameters of each bridge component; by automatically correcting the component parameters, a model library of different types and parameters of a certain bridge structure can be batch produced to meet the sufficient diversity of sample data for deep learning; taking bridge piers as an example: produce a certain category of bridge piers as benchmark piers, limit the transformation range of their model parameters to 70% to 130%, and then adjust the parameters such as pier cap height, top height, long side radius, wide side radius, pier base width, pier base length, pier top length, pier top width, pier body connection width, pier body connection length, and pier body height.

[0022] In this example, the bridge model uses a BIM model (Building Information Modeling, a conventional parametric model of a bridge). This model provides detailed geometric information about the bridge and allows for flexible adjustment of the bridge structure's parametric data to meet the needs of different tasks. This BIM model allows for efficient generation of datasets that meet the requirements of deep learning.

[0023] In this embodiment, the specific steps for automatically correcting component parameters are as follows: according to the proportions of each bridge component, a parameter adjustment range is set. For example, if the length of a component needs to be adjusted, a minimum and maximum value can be set as the adjustment range; this ensures that the corrected component parameters of the bridge component meet the actual requirements; Then, the time operator is introduced as the default seed value, and a different random number is generated for each component parameter, and this random number is within the parameter adjustment range; that is, the current system time is used as the seed value for random number generation to ensure that the random number sequence generated each time the program is run is different. For each parameter, the time operator is used as the seed to generate a random number within the set range; in this embodiment, the generated random number will be a number between 0.7 and 1.3, and the component parameter is adjusted by multiplying the random number.

[0024] Step A20: converting the bridge model corresponding to the corrected component parameters of each bridge component into a first mesh model, wherein the first mesh model includes a plurality of triangular faces; The mesh model is a grid model, which is a model that uses a series of polygons of similar size and shape to approximately represent a three-dimensional object. The first mesh model uses triangles, and the first mesh model has a plurality of triangular faces.

[0025] After converting the bridge model to the first mesh model, the first mesh model and the corresponding corrected component parameters of each bridge component are automatically recorded in a txt file. The parameters in the txt file correspond one-to-one with the saved model. After saving, the model is restored to the sample model to ensure that parameter distortion after repeated adjustments is not present. The sample model serves as a benchmark for a single category. Because the parameters are automatically adjusted using random numbers, if the model is completely distorted after hundreds or thousands of random adjustments, a sample recovery step is required after each automated parameter adjustment to ensure that the model after each parameter adjustment is similar to the actual structure. Sample models also come in multiple categories. For example, the most basic bridge structures include piers, beams, and abutments. Piers can be straight piers, hyperbolic piers, and can be differentiated by whether they have sluice holes or not. Therefore, a benchmark sample is required for each category.

[0026] Step A30: Determine the sampling points of each triangle based on the uniform sampling method; In this embodiment, the uniform sampling method is: Step a301: Calculate the area of ​​each triangle of the first mesh model and the total area of ​​all triangles of the first mesh model; the calculation formula for the area of ​​each triangle is: (1); In formula (1), is the area of ​​the triangle, A, B, and C are the three vertices of the triangle, Represents a vector With vector The modulus of the product of .

[0027] Step a302: Determine the number of sampling points that need to be generated on each triangular face of the first mesh model based on the proportion of the area of ​​each triangular face of the first mesh model to the total area of ​​all triangular faces of the first mesh model. In this embodiment, the larger the area of ​​the triangular face, the more sampling points thereon.

[0028] Step a303: Construct the barycentric coordinate system of each triangular face of the first mesh model.

[0029] Step a304: in the barycentric coordinate system of each triangular face of the first mesh model, generate uniformly distributed sampling points according to the vertices of the triangular face and the number of sampling points to be generated on the triangular face.

[0030] In this embodiment, for any triangular face, the three vertices of the triangular face are 、 、 , then the sampling point can be expressed as P, and the calculation formula of the sampling point is: ( (1- )) +( ) (2); In formula (2), 、 is a random number in the interval [0,1], if 1, then map them to the interior of the triangle, a set 、 A sampling point can be generated.

[0031] As a further optimization of this embodiment, after generating the sampling points on each triangular face, the method further includes: adding random noise to the sampling points of each triangular face, wherein the random noise includes: Gaussian noise, random block noise and / or random block mining; this embodiment introduces random block noise and random block mining to more realistically simulate the problems encountered in the actual acquisition process, such as tree occlusion and missing sampling points.

[0032] Among them, Gaussian noise is a type of noise that obeys Gaussian distribution. An offset value is added to each sampling point. The offset value The calculation formula is: (3); In formula (3), 、 are the parameters of the distribution, which are the expectation and variance of the Gaussian distribution, e is the natural base, and x is the independent variable.

[0033] In which, when the random noise is random block noise and / or random block excavation, the random block is a cubic block and / or a spherical block. When the random block is a spherical block, spherical block noise and / or spherical block excavation is added to the sampling point of each triangular surface based on the spherical space constraint; when the random block is a cubic block, cubic block noise and / or cubic block excavation is added to the sampling point of each triangular surface based on the cubic space constraint.

[0034] In this embodiment, the spherical spatial constraint is that the noised or removed sampling points are evenly distributed within the spherical radius; the cubic spatial constraint is that the noised or removed sampling points are located within the range of the cube and are distributed according to a preset distribution rule.

[0035] In this embodiment, when random block noise and random block excavation are introduced, the size and shape of these random noise and excavated block areas are adjusted according to the scale of the model, thereby achieving different noise addition and missing effects. Specifically, random block noise and excavation mainly adopt two geometric shapes: cubic blocks and spherical blocks. When adding noise, a certain number of points will be added to the randomly generated block area, and when excavating, the points in the area will be deleted. In this way, various occlusions and missing conditions that may occur in the real world can be effectively simulated. The center points of these random blocks are arbitrarily selected points, but in order to ensure that they are in a reasonable position, they must be constrained to a certain distance range within the boundary of the model point cloud, namely the spherical space constraint and the cubic space constraint.

[0036] The expression of the cubic space constraint is: (4); In formula (4), is the center point of the randomly selected noise block. Assuming that the noise block is a cube, the side length is defined as , for any point in the block Satisfies formula (4).

[0037] The expression of the spherical space constraint is: (5); In formula (5), is the center point of the randomly selected noise block. Assuming that the noise block is spherical, the radius is defined as , for any point in the block, formula (5) is satisfied.

[0038] For example, for a cubic block, all points added or removed must lie within the block's boundaries and adhere to certain distribution rules; for a spherical block, all points must be evenly distributed within the radius. This approach to noise addition and data removal, based on cubic and spherical spatial constraints, not only increases the model's robustness but also better simulates real-world uncertainty.

[0039] Step A40: constructing a point cloud dataset of each bridge component based on the sampling points of each triangular face and the corrected component parameters of each bridge component.

[0040] As a further optimization of this embodiment, the surface of the entity object is triangulated and laser simulated scanned to obtain point cloud spatial information of the entity object, including: Step B10: Read the surface of the physical object, triangulate the surface of the physical object to obtain a number of triangular facets, and construct a label for each triangular facet based on the bridge component corresponding to the physical object; In this embodiment, the entity objects in the BIM model are obtained according to different bridge components, and the faces of the entity objects are read to obtain the geometric information of the entity objects and save the geometric information; then the faces of the entity objects are triangulated, and the faces are initially segmented first, and each partial triangle after segmentation is evaluated to check whether its flatness meets the conditions, and the parts that do not meet the conditions are further subdivided and the above steps are repeated until all triangles meet the conditions, and then all the finally segmented parts are converted into triangular facets; and the triangular facets are named after the bridge component type or with self-made numbers to construct labels for the triangular facets.

[0041] Step B20: converting the bridge model into a second mesh model, loading the second mesh model, and performing laser simulation scanning on the second mesh model using a simulated drone; In this embodiment, a simulated drone is used to perform laser simulation scanning on the second mesh model, including: The centerline of the model is constructed based on the bridge deck of the bridge model, and the scanning route is generated based on the model centerline. That is, the bridge surface is divided into objects, and an object similar to the model centerline is drawn on the connecting line of its two end points and saved separately. Then, the flight route of the simulated drone is designed based on the centerline. The flight route is the scanning route. When designing the simulated drone flight route, the height from the highest point of the bridge to the bridge deck is included in the design of two routes that are parallel to the bridge and slightly higher than the bridge. The simulated UAV performs flight scanning according to a scanning route, and emits a scanning laser within a preset laser emission range and in a preset laser emission direction to perform laser simulation scanning on the second mesh model.

[0042] After loading the mesh model and starting the simulation, ToF (time-of-flight) is used to calculate the distance of the scanning point. The scanning point is a point on the triangular patch. Starting from the simulation starting position, the drone will continuously emit simulated laser signals with a range and fixed direction, and calculate the intersection of each geometric entity in the scene with the laser. At the intersection, the reflected light is calculated and the new light is recursively traced. When the reflected light of the geometric entity intersects with the current position of the drone, the laser ranging process is completed.

[0043] Step B30: Obtain the laser emission direction of the simulated UAV, the UAV position, the time interval between the emission and reception of the laser signal, and the scanning timestamp during the laser simulation scanning process.

[0044] Step B40: Determine the point cloud coordinates of each triangular facet based on the laser emission direction of the simulated drone, the drone position, and the time interval between the emission and reception of the laser signal during the laser simulation scanning process, and use the point cloud coordinates of each triangular facet of the physical object, the label of each triangular facet, and the point cloud coordinates of each triangular facet as the point cloud spatial information of the physical object.

[0045] In this embodiment, by simulating the laser emission direction of the drone, the drone's position, and the time interval between the emission and reception of the laser signal, the point cloud coordinates of each triangular facet can be calculated. The point cloud coordinate calculation expression of the triangular facet is: (6); In formula (6), is the distance from the scanning point to the simulated drone, is the pitch angle (vertical angle) of the laser, is the yaw angle (horizontal angle) of the laser, is the three-dimensional coordinate of the scanning point in the simulated drone coordinate system ( ).

[0046] Among them, the distance from the scanning point to the simulated drone The calculation expression is: (7); (8); In equations (7) and (8), c is the speed of light, is the time interval between transmitting and receiving laser signals, 、 are the transmitting time and receiving time respectively.

[0047] In the point cloud data generated by the present invention, the point cloud data of a single structure can be used for the classification of bridge components, and combined with the model parameters recorded in the output process, the parameter fitting of the bridge components can be completed; the point cloud data of the entire bridge is accompanied by category labels, which can complete the semantic segmentation of the entire bridge.

[0048] Example 2 Figure 2 This is a block diagram of a device for producing a real-scene three-dimensional point cloud data set of a bridge provided by an embodiment of the present invention. Figure 2 As shown, this embodiment provides a device for producing a real-scene 3D point cloud dataset of a bridge, which is used to implement the method for producing a real-scene 3D point cloud dataset of a bridge in Example 1. The device includes: An acquisition module is used to acquire a data set production requirement, wherein the data set production requirement includes: a point cloud data set production requirement of a bridge model and / or a point cloud data set production requirement of a bridge component; a production module, configured to obtain a plurality of bridge components in a bridge model when the data set production requirement is a point cloud data set production requirement for bridge components; extract component parameters of each bridge component, and construct a point cloud data set of the bridge components based on the component parameters of each bridge component and parameters of the bridge model; and / or When the dataset production requirement is a point cloud dataset production requirement for a bridge model, the entity objects corresponding to each bridge component in the bridge model are obtained; the surfaces of the entity objects are triangulated and laser simulated scanned to obtain point cloud spatial information of the entity objects, and the point cloud dataset of the bridge model is constructed based on the point cloud spatial information of the entity objects.

[0049] This embodiment also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for producing a real-scene three-dimensional point cloud dataset of a bridge in embodiment one is implemented.

[0050] This embodiment further provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the method for producing a real-scene three-dimensional point cloud dataset of a bridge in the first embodiment is implemented.

[0051] The present invention can produce two different point cloud datasets according to needs, namely, by extracting the component parameters of each bridge component, constructing a point cloud dataset of the bridge component based on the component parameters of each bridge component and the parameters of the bridge model; at the same time, the entity objects corresponding to each bridge component in the bridge model are obtained; the surface of the entity object is triangulated and laser simulated scanned to obtain the point cloud spatial information of the entity object, and the point cloud dataset of the bridge model is constructed according to the point cloud spatial information of the entity object; therefore, during the training process, the present invention can select different point cloud datasets according to different training tasks, for example: the classification task selects the point cloud dataset of the bridge component, and the segmentation task selects the point cloud dataset of the bridge model; and the dataset production method of the present invention can efficiently generate a dataset that meets the needs of deep learning.

[0052] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0053] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0054] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A method for preparing a bridge real-scene three-dimensional point cloud data set, characterized in that: The method comprises: Acquiring data set production requirements, the data set production requirements including: point cloud data set production requirements of a bridge model and / or point cloud data set production requirements of bridge components; When the data set production requirement is a point cloud data set production requirement for bridge components, obtaining a number of bridge components in the bridge model; extracting component parameters of each bridge component, and constructing a point cloud data set of the bridge component based on the component parameters of each bridge component and parameters of the bridge model; and / or When the dataset production requirement is a point cloud dataset production requirement for a bridge model, the entity objects corresponding to the various bridge components in the bridge model are obtained; the surfaces of the entity objects are triangulated and laser simulated scanned to obtain point cloud spatial information of the entity objects, and the point cloud dataset of the bridge model is constructed according to the point cloud spatial information of the entity objects.

2. The method for producing a bridge real-scene three-dimensional point cloud data set according to claim 1, characterized in that: A point cloud dataset of bridge components is constructed based on the component parameters of each bridge component and the parameters of the bridge model, including: Automatically correcting the component parameters based on the proportions of the bridge components to obtain corrected component parameters of the bridge components; Converting the bridge model corresponding to the modified component parameters of each bridge component into a first mesh model, wherein the first mesh model includes a plurality of triangular faces; Based on the uniform sampling method, determine the sampling points of each triangle; Based on the sampling points of each triangular face and the corrected component parameters of each bridge component, a point cloud dataset of each bridge component is constructed.

3. The method for producing a bridge real-scene three-dimensional point cloud dataset according to claim 2, characterized in that: The uniform sampling method is: Calculate the area of ​​each triangle of the first mesh model and the total area of ​​all triangles of the first mesh model; Determine the number of sampling points that need to be generated on each triangular face of the first mesh model according to the proportion of the area of ​​each triangular face of the first mesh model to the total area of ​​all triangular faces of the first mesh model; Construct the barycentric coordinate system of each triangle face of the first mesh model; In the barycentric coordinate system of each triangular face of the first mesh model, uniformly distributed sampling points are generated according to the vertices of the triangular face and the number of sampling points to be generated on the triangular face.

4. The method for producing a bridge real-scene three-dimensional point cloud dataset according to claim 2, characterized in that: Before constructing a point cloud data set of each bridge component based on the sampling points of each triangular face and the corrected component parameters of each bridge component, the method also includes: adding random noise to the sampling points of each triangular face, and the random noise includes: Gaussian noise, random block noise and / or random block excavation.

5. The method for producing a bridge real-scene three-dimensional point cloud data set according to claim 4, characterized in that: When the random noise is random block noise and / or random block excavation, the random block is a cubic block and / or a spherical block, wherein, when the random block is a spherical block, spherical block noise and / or spherical block excavation is added to the sampling points of each triangular surface based on the spherical space constraint; when the random block is a cubic block, cubic block noise and / or cubic block excavation is added to the sampling points of each triangular surface based on the cubic space constraint.

6. The method for preparing a bridge real-scene three-dimensional point cloud data set according to claim 5, characterized in that: The spherical space constraint is that the sampling points to be noised or mined are evenly distributed within the spherical radius; The cubic space constraint is that the sampling points to be noised or mined are located within the range of the cube and are distributed according to a preset distribution rule.

7. The method for producing a bridge real-scene three-dimensional point cloud data set according to claim 1, characterized in that: Triangulate the surface of the entity object and perform laser simulation scanning to obtain the point cloud spatial information of the entity object, including: Read the surface of the entity object, triangulate the surface of the entity object to obtain a plurality of triangular face patches, and construct a label of each triangular face patch based on the bridge component corresponding to the entity object; Convert the bridge model into a second mesh model, load the second mesh model, and use a simulated drone to perform laser simulation scanning on the second mesh model; Obtain the laser emission direction of the simulated drone, the drone position, the time interval between the emission and reception of the laser signal, and the scanning timestamp during the laser simulation scanning process; According to the laser emission direction of the simulated drone, the position of the drone and the time interval between the emission and reception of the laser signal during the laser simulation scanning process, the point cloud coordinates of each triangular facet are determined, and the point cloud coordinates of each triangular facet of the physical object, the label of each triangular facet and the point cloud coordinates of each triangular facet are used as the point cloud spatial information of the physical object.

8. The method for producing a bridge real-scene three-dimensional point cloud data set according to claim 7, characterized in that: Use a simulated drone to perform laser simulation scanning on the second mesh model, including: The center line of the model is constructed based on the bridge deck of the bridge model, and the scanning route is generated based on the center line of the model; The simulated UAV performs flight scanning according to a scanning route, and emits scanning lasers in a preset laser emission range and a preset laser emission direction to perform laser simulation scanning on the second mesh model.

9. A device for producing a real-scene three-dimensional point cloud data set of a bridge, used to implement the method for producing a real-scene three-dimensional point cloud data set of a bridge as claimed in any one of claims 1 to 8, characterized in that: The device comprises: An acquisition module is used to acquire a data set production requirement, wherein the data set production requirement includes: a point cloud data set production requirement of a bridge model and / or a point cloud data set production requirement of a bridge component; A production module, used for obtaining a number of bridge components in the bridge model when the data set production requirement is a point cloud data set production requirement for bridge components; extracting component parameters of each bridge component, and constructing a point cloud data set of the bridge component based on the component parameters of each bridge component and the parameters of the bridge model; and / or When the dataset production requirement is a point cloud dataset production requirement for a bridge model, the entity objects corresponding to the various bridge components in the bridge model are obtained; the surfaces of the entity objects are triangulated and laser simulated scanned to obtain point cloud spatial information of the entity objects, and the point cloud dataset of the bridge model is constructed according to the point cloud spatial information of the entity objects.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for producing a real-scene three-dimensional point cloud data set of a bridge described in any one of claims 1 to 8 is implemented.

Images