Three-dimensional reconstruction method and device of grid structure, computer equipment and storage medium

By identifying and matching the two-dimensional mapping of the grid structure, the problems of grid structure accuracy loss and high computational complexity in three-dimensional reconstruction technology are solved, and more efficient and accurate three-dimensional model construction is achieved, providing high-quality data for subsequent analysis .

CN118918162AActive Publication Date: 2024-11-08CHINA CONSTR SCI & IND CORP LTD

Patent Information

Application Number
CN202410603708.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-05-15
Publication Date
2024-11-08
Estimated Expiration
2044-05-15

AI Technical Summary

Technical Problem

The existing three-dimensional reconstruction technology has problems of accuracy loss and high computational complexity when dealing with grid structures, and lacks targeted optimization.

Method used

By obtaining the original image of the mesh structure, identifying the member boundary of the mesh rod and the center position of the mesh ball, determining the two-dimensional map, matching the original image with the two-dimensional map, building a three-dimensional model, and correcting it through multi-view image and optimize model parameters to improve reconstruction accuracy and speed.

Benefits of technology

It improves the accuracy and computing speed of three-dimensional reconstruction, reduces error accumulation, provides a more accurate and high-quality three-dimensional model, and provides solid data support for subsequent structural analysis and design.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118918162A_ABST
    Figure CN118918162A_ABST
Patent Text Reader

Abstract

The invention relates to the field of buildings, and discloses a three-dimensional reconstruction method and device for a grid structure, computer equipment and a storage medium, and the method comprises the steps: obtaining an original image of the grid structure which comprises grid balls and grid rods; identifying a rod piece boundary of a net rack rod in the original image, and determining a central position of a net rack ball; determining two-dimensional mapping of the grid structure based on the rod piece boundary of the grid rod and the central position of the grid ball; and matching the original image with the two-dimensional mapping to obtain a three-dimensional model of the grid structure. According to the method, the problems of precision loss and slow operation when the grid structure is processed by the existing three-dimensional reconstruction technology are solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of architecture, and in particular to a three-dimensional reconstruction method, device, computer equipment and storage medium of a grid structure. Background Art

[0002] With the continuous advancement of computer vision and graphics processing technology, 3D reconstruction technology has played an important role in many fields. As a common form of building structure, the demand for 3D reconstruction of grid structures is growing. When dealing with grid structures, traditional 3D reconstruction methods often face problems such as difficulty in data acquisition, low reconstruction accuracy, high computational complexity, and low efficiency. At present, the data description methods for 3D reconstruction are mainly divided into four types: voxel, point cloud, mesh, and implicit surface.

[0003] However, these methods have simplified the original surface to varying degrees, resulting in a loss of accuracy. This loss of accuracy may be acceptable in some scenarios where accuracy is not high, but in the 3D reconstruction of grid structures, this loss of accuracy may affect subsequent key links such as structural analysis and safety assessment. Therefore, the current 3D reconstruction technology has problems such as loss of accuracy, high computational complexity, and lack of targeted optimization when dealing with grid structures. Summary of the invention

[0004] In view of this, embodiments of the present invention provide a three-dimensional reconstruction method, apparatus, computer equipment and storage medium for a grid structure to solve the problems of accuracy loss and slow calculation in existing three-dimensional reconstruction technology when processing grid structures.

[0005] In a first aspect, an embodiment of the present invention provides a three-dimensional reconstruction method of a grid structure, the method comprising:

[0006] Acquire an original image of a grid structure, wherein the grid structure comprises: grid balls and grid rods;

[0007] Identifying the rod boundary of the grid rod in the original image and determining the center position of the grid ball;

[0008] Determining a two-dimensional mapping of the grid structure based on the rod boundaries of the grid rods and the center position of the grid ball;

[0009] The original image is matched with the two-dimensional mapping to obtain a three-dimensional model of the grid structure.

[0010] In an optional implementation manner of the present application, the identifying the rod boundary of the grid rod in the original image and determining the center position of the grid ball includes:

[0011] Eliminating irrelevant information in the original image to obtain a preprocessed image;

[0012] Performing edge detection on the preprocessed image to obtain the rod boundary of the grid rod and the spherical boundary of the grid ball;

[0013] The center position of the grid ball is determined according to the rod boundary and the spherical surface boundary.

[0014] In an optional embodiment of the present application, the determining of the two-dimensional mapping of the grid structure based on the rod boundaries of the grid rods and the center position of the grid ball includes:

[0015] Determining four corner nodes based on the original image of the grid structure;

[0016] Constructing a three-dimensional coordinate system according to the four corner nodes, and determining the position information of a collection device using the three-dimensional coordinate system, wherein the collection device refers to a device that collects the original image;

[0017] Extracting spatial geometric parameters of the grid ball, and constructing point cloud data of the grid ball according to the spatial geometric parameters;

[0018] The point cloud inertia principal axis of the grid ball is calculated according to the point cloud data, and the two-dimensional mapping of the grid structure is determined according to the posture information and the point cloud inertia principal axis.

[0019] In an optional implementation manner of the present application, matching the original image with the two-dimensional mapping to obtain the three-dimensional model of the grid structure includes:

[0020] Matching the original image with the two-dimensional map to obtain a corresponding relationship;

[0021] Adjusting the coordinate information of the grid ball in the two-dimensional mapping based on the corresponding relationship to obtain an adjusted two-dimensional mapping;

[0022] Constructing an original three-dimensional model of the grid structure according to the adjusted two-dimensional mapping;

[0023] The original three-dimensional model is modified to obtain a target three-dimensional model of the grid structure.

[0024] In an optional implementation manner of the present application, the modifying the original three-dimensional model to obtain the target three-dimensional model of the grid structure includes:

[0025] Acquire multi-view images of the grid rods collected by a collection device in different positions;

[0026] Identifying multi-view rod boundaries of grid rods in the multi-view image;

[0027] Comparing the rod boundary of the grid rod and the multi-view rod boundary, determining the correction coefficient of the acquisition device at different positions, and adjusting the original three-dimensional model of the grid structure according to the correction coefficient to obtain multiple candidate three-dimensional models;

[0028] The target three-dimensional model is determined from the candidate three-dimensional models.

[0029] In an optional implementation manner of the present application, determining the target three-dimensional model from the candidate three-dimensional models includes:

[0030] Calculating the reconstruction accuracy of the plurality of candidate three-dimensional models;

[0031] Determining the optimal posture of the acquisition device based on the reconstruction accuracy;

[0032] The model parameters of the three-dimensional model are adjusted according to the optimal posture to obtain a target three-dimensional model.

[0033] In an optional implementation manner of the present application, after obtaining the target three-dimensional model, the method further includes:

[0034] Disconnecting the coordinate link between the grid rods and the grid balls in the target three-dimensional model to obtain first position data of the grid rods and second position data of the grid balls;

[0035] Modeling the net frame rods based on the first position data to obtain a first model, and modeling the net frame ball based on the second position data to obtain a second model;

[0036] The rod end coordinates in the target three-dimensional model are adjusted according to the first model and the second model to obtain an optimized target three-dimensional model.

[0037] In a second aspect, an embodiment of the present invention provides a three-dimensional reconstruction device for a grid structure, the device comprising:

[0038] An acquisition module, used to acquire an original image of a grid structure, wherein the grid structure comprises: grid balls and grid rods;

[0039] A recognition module, used for recognizing the rod boundary of the grid rod in the original image and determining the center position of the grid ball;

[0040] A determination module, configured to determine a two-dimensional mapping of the grid structure based on the rod boundaries of the grid rods and the center position of the grid ball;

[0041] A matching module is used to match the original image with the two-dimensional mapping to obtain a three-dimensional model of the grid structure.

[0042] In a third aspect, an embodiment of the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the three-dimensional reconstruction method of the grid structure of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0043] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the three-dimensional reconstruction method of the grid structure of the first aspect or any corresponding embodiment thereof.

[0044] The method provided in the embodiment of the present application has the following beneficial effects:

[0045] The method provided in the embodiment of the present application provides key basic data for the subsequent three-dimensional reconstruction process by acquiring the original image of the grid structure, ensuring the accuracy and flexibility of the reconstruction and making the method widely applicable.

[0046] The method provided in the embodiment of the present application can accurately identify the rod boundary of the grid rod and the spherical boundary of the grid ball by eliminating irrelevant information in the original image and performing edge detection on the pre-processed image, thereby improving the recognition accuracy. Through the recognition of the spherical boundary, the center position of the grid ball can be accurately determined. Extracting the accurate rod boundary and the center position of the grid ball provides reliable data support for the subsequent three-dimensional reconstruction steps. Through accurate recognition and positioning, error accumulation can be reduced, thereby improving the quality and accuracy of the final three-dimensional model.

[0047] The method provided in the embodiment of the present application provides reliable reference points for two-dimensional mapping by determining the four corner nodes; determines the position of the acquisition device by constructing a three-dimensional coordinate system to ensure accurate mapping of the three-dimensional structure and the two-dimensional plane; accurately captures the shape and position of the grid ball by constructing point cloud data; and determines the two-dimensional mapping by comprehensively considering the overall characteristics of the grid, which provides a solid foundation for subsequent three-dimensional reconstruction and improves the calculation speed during the three-dimensional reconstruction process.

[0048] The method provided in the embodiment of the present application can accurately establish the corresponding relationship between two-dimensional and three-dimensional by matching the original image with the two-dimensional mapping. This matching process ensures that each point in the three-dimensional model corresponds to the position in the two-dimensional mapping, thereby improving the accuracy and operation speed of three-dimensional reconstruction. By adjusting the coordinate information of the grid ball in the two-dimensional mapping through the corresponding relationship, a more accurate adjusted two-dimensional mapping is obtained, which can correct the errors that may exist in the initial two-dimensional mapping and further optimize the construction basis of the three-dimensional model. The original three-dimensional model of the grid structure is constructed according to the adjusted two-dimensional mapping, which provides a basis for subsequent model correction and optimization. By correcting the original three-dimensional model, it is helpful to obtain a target three-dimensional model that is closer to the actual grid structure.

[0049] The method provided in the embodiment of the present application improves the accuracy of the three-dimensional model by disconnecting the coordinate link between the grid rods and the grid balls, and independently analyzing and processing the position data of each component; the grid rods and the grid balls are modeled separately using these data to obtain a more refined model; the three-dimensional model is optimized by adjusting the rod end coordinates in the model to make it more realistic in structure and visual sense, and the calculation speed is faster, thereby providing high-quality data support for subsequent design and analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0051] Figure 1 is a schematic flow chart of a three-dimensional reconstruction method of a grid structure according to some embodiments of the present invention;

[0052] Figure 2 is a schematic diagram of edge detection of an original image according to an embodiment of the present invention;

[0053] Figure 3 is a schematic diagram of four corner nodes in a grid structure according to an embodiment of the present invention;

[0054] Figure 4 is a schematic diagram of a three-dimensional coordinate system according to an embodiment of the present invention;

[0055] Figure 5 is a schematic diagram of matching an original image with a two-dimensional mapping according to an embodiment of the present invention;

[0056] Figure 6 is a schematic diagram of marking positions for recognition errors according to an embodiment of the present invention;

[0057] Figure 7 is a structural block diagram of a three-dimensional reconstruction device of a grid structure according to an embodiment of the present invention;

[0058] Figure 8 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0060] According to an embodiment of the present invention, a three-dimensional reconstruction method, apparatus, computer equipment and storage medium for a grid structure are provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0061] In this embodiment, a three-dimensional reconstruction method of a grid structure is provided. Figure 1 is a flow chart of a three-dimensional reconstruction method of a grid structure according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0062] Step S11, obtaining an original image of a grid structure, wherein the grid structure includes grid balls and grid rods.

[0063] In the embodiment of the present application, the grid structure is a spatial grid structure composed of a plurality of grid balls and grid rods connected to each other, and this structure is widely used in the fields of buildings, bridges, engineering facilities, etc. The grid ball is a key connection point in the grid structure, usually made of steel, and its shape is similar to a sphere, and there are threaded holes on the sphere for connecting the grid rods. The grid rod is a rod-shaped component connecting the grid balls, which is used to form the basic skeleton of the grid. In addition, obtaining the original image of the grid structure requires using an image acquisition device (such as a camera) to shoot the grid structure. The original image shot should contain as many details of the grid structure as possible, especially the shape and position relationship of the grid balls and the grid rods.

[0064] The method provided in the embodiment of the present application provides key basic data for the subsequent three-dimensional reconstruction process by acquiring the original image of the grid structure, ensuring the accuracy and flexibility of the reconstruction and making the method widely applicable.

[0065] Step S12, identifying the rod boundary of the grid rod in the original image and determining the center position of the grid ball.

[0066] In the embodiment of the present application, step S12 includes the following steps A1-A3:

[0067] Step A1, remove irrelevant information in the original image to obtain a preprocessed image.

[0068] In the embodiment of the present application, in addition to the grid structure (including grid balls and grid rods), the acquired original image also contains a lot of irrelevant information, such as background, shadows caused by changes in illumination, construction equipment, personnel, etc. This information may interfere with subsequent edge detection, feature extraction, and three-dimensional reconstruction, so it is necessary to remove it in the preprocessing stage. Image processing techniques such as filtering, threshold segmentation, morphological operations, etc. can be used to eliminate irrelevant information to eliminate noise and background and highlight the characteristics of the grid structure. For example, deep learning techniques such as convolutional networks can be used to identify the boundaries of grid balls and rods, and further delete irrelevant information outside the grid balls.

[0069] Step A2: perform edge detection on the preprocessed image to obtain the rod boundary of the grid rod and the spherical boundary of the grid ball.

[0070] In the embodiments of the present application, Figure 2 As shown, edge detection technology is used to identify the outline of objects in the image or the boundaries between different objects. This technology can achieve this goal by capturing areas with obvious brightness changes in the image, and is used to identify the rod boundaries of the grid rods and the spherical boundaries of the grid balls. The rod boundaries, that is, the outline or edge of the grid rods in the image, are crucial for subsequent three-dimensional reconstruction. Similarly, the spherical boundary of the grid ball, that is, the outer contour of the grid ball in the image, plays a decisive role in determining the center position of the grid ball. In order to accurately extract these key boundaries, algorithms such as Canny edge detection and Sobel edge detection can be used. These algorithms can accurately identify edges based on the grayscale gradient changes in the image.

[0071] Step A3, determining the center position of the grid ball according to the rod boundary and the spherical surface boundary.

[0072] In the embodiment of the present application, the center position refers to the geometric center point of the grid ball in the image, which is an important parameter for locating the grid ball in three-dimensional reconstruction. The center position of the grid ball is determined by using the rod boundary and spherical boundary identified in the previous step. The image processing technology involved includes: morphological operation, contour analysis after edge detection, or analysis based on regional characteristics. Among them, morphological operation can strengthen the boundary and make the boundary of the grid ball clearer. By performing contour analysis on these boundaries, the center position of the grid ball can be determined. In addition, analyzing the grayscale, texture and other characteristics of the image area can also help locate the center.

[0073] The method provided in the embodiment of the present application can accurately identify the rod boundary of the grid rod and the spherical boundary of the grid ball by eliminating irrelevant information in the original image and performing edge detection on the pre-processed image, thereby improving the recognition accuracy. Through the recognition of the spherical boundary, the center position of the grid ball can be accurately determined. Extracting the accurate rod boundary and the center position of the grid ball provides reliable data support for the subsequent three-dimensional reconstruction steps. Through accurate recognition and positioning, error accumulation can be reduced, thereby improving the quality and accuracy of the final three-dimensional model and improving the operation speed during the three-dimensional reconstruction process.

[0074] Step S13, determining a two-dimensional mapping of the grid structure based on the rod boundaries of the grid rods and the center position of the grid balls.

[0075] In the embodiment of the present application, step S13 includes the following steps B1-B4:

[0076] Step B1, determining the four corner nodes based on the original image of the grid structure.

[0077] In the embodiments of the present application, Figure 3 As shown, the four-corner nodes refer to the specific points at the four corners of the grid structure, which are also the support points of the grid or the significant feature points on the structure. These points are used as reference points for establishing a three-dimensional coordinate system in subsequent steps. The nodes at the four corners of the grid structure are identified and located from the original image. This can be achieved through edge detection, feature point extraction or template matching methods using image processing technology. After the nodes are determined, they can be used as control points in the subsequent three-dimensional reconstruction process.

[0078] Step B2, constructing a three-dimensional coordinate system based on the four corner nodes, and using the three-dimensional coordinate system to determine the position and posture information of the acquisition device, wherein the acquisition device refers to the device that acquires the original image.

[0079] In the embodiments of the present application, Figure 4As shown, based on the determined four corner nodes, a three-dimensional coordinate system can be constructed, and the coordinate system includes the principal axis of inertia 1, the principal axis of inertia 2, and the principal axis of inertia 3. In this coordinate system, each node is assigned a three-dimensional coordinate (x, y, z). This process is to accurately convert the two-dimensional image information into three-dimensional space information in the future. The pose information includes the position and orientation of the acquisition device (such as a camera) in space. By comparing the position of the four corner nodes in the image with the position of the four corner nodes in the actual three-dimensional coordinate system, the position and shooting angle of the camera when taking this photo can be calculated, which can be achieved through perspective transformation and camera calibration.

[0080] Step B3, extracting the spatial geometric parameters of the grid ball, and constructing the point cloud data of the grid ball according to the spatial geometric parameters.

[0081] In the embodiment of the present application, the spatial geometric parameters of the grid ball refer to the parameters that describe the position, shape and size of the grid ball in three-dimensional space, which may include the coordinates of the center of the sphere, the radius, etc. Point cloud data is a data format that represents the surface shape of an object in three-dimensional space, and is composed of a large number of points, each of which contains its coordinate information in three-dimensional space. Based on the spatial geometric parameters of the grid ball, point cloud data representing the surface of the grid ball can be generated. Specifically, a series of points can be generated on the surface of the grid ball according to the coordinates of the center of the sphere and the radius of the grid ball through a certain algorithm (such as a spherical sampling algorithm), and these points constitute the point cloud data of the grid ball.

[0082] Step B4, calculating the point cloud inertia principal axis of the grid ball according to the point cloud data, and determining the two-dimensional mapping of the grid structure according to the posture information and the point cloud inertia principal axis.

[0083] In an embodiment of the present application, the principal axis of inertia of the point cloud is determined by calculating the mass distribution of the point cloud (or the point cloud density distribution), which can reflect the main direction and distribution characteristics of the point cloud shape. In physics, the principal axis of inertia of an object refers to the axis that passes through the center of mass of the object and has the maximum and minimum moment of inertia relative to the mass distribution of the object. When calculating the principal axis of inertia of the point cloud, the mass of each point can be regarded as the same, and these axes are found by calculating the covariance matrix. Two-dimensional mapping refers to the representation of a three-dimensional object or scene projected onto a two-dimensional plane, and a two-dimensional mapping of the grid structure is generated based on the principal axis of inertia of the grid ball point cloud and the pose information of the camera. This mapping will be used to match the two-dimensional image with the three-dimensional model in subsequent steps.

[0084] The method provided in the embodiment of the present application provides a reliable reference point for two-dimensional mapping by determining the four corner nodes; determines the position of the acquisition device by constructing a three-dimensional coordinate system to ensure accurate mapping of the three-dimensional structure and the two-dimensional plane; accurately captures the shape and position of the grid ball by constructing point cloud data; and determines the two-dimensional mapping by comprehensively considering the overall characteristics of the grid, providing a solid foundation for subsequent three-dimensional reconstruction.

[0085] Step S14, matching the original image with the two-dimensional mapping to obtain a three-dimensional model of the grid structure.

[0086] In the embodiment of the present application, step S14 includes the following steps C1-C4:

[0087] Step C1, matching the original image with the two-dimensional mapping to obtain a corresponding relationship.

[0088] In the embodiment of the present application, the comparison and matching process between the original image and the two-dimensional mapping is as follows: Figure 5 As shown in the figure, first, the algorithm extracts key features from the original image, such as the shape, size and position of the grid balls, as well as the lines and intersections of the grid rods. Second, these features are compared with elements in the 2D map, which can be achieved through traditional template matching methods (comparing a part of the image to a predefined template) or using more advanced object detection techniques (such as deep learning-based models). After finding a match, the correspondence between the original image and the 2D map is established. It means that it has been determined which point in the 2D map represents which grid ball or grid rod in the original image.

[0089] Step C2, adjusting the coordinate information of the grid ball in the two-dimensional mapping based on the corresponding relationship to obtain an adjusted two-dimensional mapping.

[0090] In an embodiment of the present application, the original image is compared with the two-dimensional mapping to find out the differences between the two, wherein the differences may manifest as position offset, size inconsistency, etc.; based on the above differences, the adjustment amount required for the grid ball coordinates in the two-dimensional mapping is calculated; according to the calculated adjustment amount, the grid ball coordinates in the two-dimensional mapping are fine-tuned; after the adjustment, the original image is compared with the adjusted two-dimensional mapping again to ensure a higher degree of match between the two.

[0091] Step C3, constructing an original three-dimensional model of the grid structure according to the adjusted two-dimensional mapping.

[0092] In an embodiment of the present application, the coordinates in the two-dimensional mapping are converted into coordinates in the three-dimensional space using the determined three-dimensional coordinate system and the position information of the acquisition device. The three-dimensional structure can be restored from the two-dimensional image using the techniques of geometry and computer vision, such as the inverse process of perspective projection. A model of the grid structure is constructed in the three-dimensional space based on the three-dimensional coordinates obtained by the conversion. The grid balls are placed at the calculated three-dimensional coordinate positions, and the grid rods are reconstructed in the three-dimensional space based on their two-dimensional boundary information. After the three-dimensional model is initially constructed, verification and optimization can be performed, that is, checking whether the geometric relationship of the model is correct, such as whether the connection relationship between the grid balls and the grid rods conforms to the actual structure, and whether there are any geometric errors or inconsistencies.

[0093] Step C4, correcting the original three-dimensional model to obtain a target three-dimensional model of the grid structure.

[0094] In the embodiment of the present application, step C4 specifically includes the following steps C41-C44:

[0095] Step C41, collecting multi-view images of the grid rods in different positions and postures by the collecting device.

[0096] In an embodiment of the present application, ensure that the acquisition device (such as a digital camera or a professional 3D scanning device) is ready, and check its settings to ensure image quality; adjust the position and posture of the acquisition device to obtain different viewing angles of the grid rods, which may include changing the height, angle and distance of the camera; in each different posture, use the acquisition device to photograph or scan and save the grid rods.

[0097] Step C42, identifying the multi-view rod boundary of the grid rod in the multi-view image.

[0098] In the embodiment of the present application, similar to the previous steps, the collected multi-view images are cleaned to remove blurred, repeated or invalid images; image processing technology is used to extract key features of the truss rods, such as edges, corners, etc.; the extracted features are labeled to facilitate subsequent three-dimensional reconstruction and model correction.

[0099] Step C43, comparing the rod boundary of the grid rod and the multi-view rod boundary, determining the correction coefficient of the acquisition device in different positions, and adjusting the original three-dimensional model of the grid structure according to the correction coefficient to obtain multiple candidate three-dimensional models.

[0100] In the embodiment of the present application, the original three-dimensional model is aligned with the extracted features to ensure the accuracy of the model, and the error position is obtained and marked to obtain the marked position, wherein the marked position is as follows: Figure 6As shown; based on the extracted features and annotation data, the original three-dimensional model is preliminarily corrected and adjusted; the error between the preliminarily corrected model and the actual grid structure is analyzed to determine the further correction direction.

[0101] Step C44, determining a target three-dimensional model from the candidate three-dimensional models.

[0102] In an embodiment of the present application, determining a target three-dimensional model from candidate three-dimensional models includes: calculating the reconstruction accuracy of multiple candidate three-dimensional models; determining the optimal posture of the acquisition device based on the reconstruction accuracy; and adjusting the model parameters of the three-dimensional model according to the optimal posture to obtain the target three-dimensional model.

[0103] In an embodiment of the present application, first, the reconstruction accuracy of multiple candidate three-dimensional models is calculated, wherein the reconstruction accuracy is a key indicator for measuring the similarity between a three-dimensional model and an actual scene or object. By comparing the candidate three-dimensional model with the actual image or data, the accuracy of each model can be evaluated. Based on the reconstruction accuracy, the optimal posture of the acquisition device can be determined. Secondly, by analyzing the influence of the data collected under different postures on the reconstruction accuracy, a posture with the highest reconstruction accuracy, i.e., the optimal posture, can be determined. After determining the optimal posture, the model parameters of the three-dimensional model are adjusted according to this posture, including adjusting the position, rotation angle, scaling, etc. of the model, so that the three-dimensional model is more consistent with the geometric shape and style type of the actual scene or object. Finally, a target three-dimensional model with a high degree of similarity to the scene is obtained.

[0104] The method provided in the embodiment of the present application can accurately establish the corresponding relationship between two-dimensional and three-dimensional by matching the original image with the two-dimensional mapping. This matching process ensures that each point in the three-dimensional model corresponds to the position in the two-dimensional mapping, thereby improving the accuracy and operation speed of three-dimensional reconstruction. By adjusting the coordinate information of the grid ball in the two-dimensional mapping through the corresponding relationship, a more accurate adjusted two-dimensional mapping is obtained, which can correct the errors that may exist in the initial two-dimensional mapping and further optimize the construction basis of the three-dimensional model. The original three-dimensional model of the grid structure is constructed according to the adjusted two-dimensional mapping, which provides a basis for subsequent model correction and optimization. By correcting the original three-dimensional model, it is helpful to obtain a target three-dimensional model that is closer to the actual grid structure.

[0105] In the embodiment of the present application, after obtaining the target three-dimensional model, the method further includes the following steps D1-D3:

[0106] Step D1, disconnecting the coordinate link between the grid rods and the grid balls in the target three-dimensional model, and obtaining the first position data of the grid rods and the second position data of the grid balls.

[0107] In an embodiment of the present application, in a traditional three-dimensional model, the endpoints of the grid rods are usually directly connected to the center coordinates of the grid ball, indicating that the rod end coordinates are the spherical coordinates. However, in the actual installation process, due to factors such as construction errors and material deformation, the rod end coordinates may not completely coincide with the spherical coordinates. Therefore, disconnecting the coordinate link can allow the actual position of each component to be located more accurately. In the three-dimensional model, the coordinate link between the grid rods and the grid ball is disconnected to make them independent entities. The first position data of the grid rods after the link is disconnected (i.e., the rod end coordinates) and the second position data of the grid ball (i.e., the spherical center coordinates) are recorded separately.

[0108] Step D2, modeling the grid rods based on the first position data to obtain a first model, and modeling the grid ball based on the second position data to obtain a second model.

[0109] In the embodiment of the present application, based on the first position data (coordinates of the rod end), a modeling technique (such as a line segment, a cylinder, etc.) is used to represent the grid rod. The parameters of the model may include the length, diameter, position, etc. of the rod. Based on the second position data (coordinates of the sphere center), a sphere is used to represent the grid ball. The parameters of the model include the three-dimensional coordinates of the sphere center and the radius of the sphere.

[0110] Step D3, adjusting the rod end coordinates in the target three-dimensional model according to the first model and the second model to obtain an optimized target three-dimensional model.

[0111] In the embodiment of the present application, according to the positional relationship between the first model (grid rod model) and the second model (grid ball model), the rod end coordinates of the grid rods in the target three-dimensional model are adjusted to ensure that they match the positions of the grid balls. After adjusting the rod end coordinates, the entire three-dimensional model is further optimized, such as adjusting the details of the model, improving the lighting and materials, etc., to improve the visualization effect of the model, and finally obtaining the optimized target three-dimensional model.

[0112] The method provided in the embodiment of the present application improves the accuracy of the three-dimensional model by disconnecting the coordinate link between the grid rods and the grid balls, and independently analyzing and processing the position data of each component; the grid rods and the grid balls are modeled separately using these data to obtain a more refined model; the three-dimensional model is optimized by adjusting the rod end coordinates in the model to make it more realistic in structure and visual sense, and the calculation speed is faster, thereby providing high-quality data support for subsequent design and analysis.

[0113] In this embodiment, a three-dimensional reconstruction device of a grid structure is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0114] This embodiment provides a three-dimensional reconstruction device of a grid structure, such as Figure 7 As shown, including:

[0115] An acquisition module 71 is used to acquire an original image of a grid structure, wherein the grid structure includes: grid balls and grid rods;

[0116] The recognition module 72 is used to recognize the rod boundary of the grid rod in the original image and determine the center position of the grid ball;

[0117] A determination module 73, for determining a two-dimensional mapping of the grid structure based on the rod boundaries of the grid rods and the center position of the grid balls;

[0118] The matching module 74 is used to match the original image with the two-dimensional mapping to obtain a three-dimensional model of the grid structure.

[0119] In an optional embodiment of the present application, the recognition module 72 is used to eliminate irrelevant information in the original image to obtain a preprocessed image; perform edge detection on the preprocessed image to obtain the rod boundaries of the grid rods and the spherical boundaries of the grid ball; and determine the center position of the grid ball based on the rod boundaries and the spherical boundaries.

[0120] In an optional embodiment of the present application, a determination module 73 is used to determine the four corner nodes based on the original image of the grid structure; construct a three-dimensional coordinate system based on the four corner nodes, and use the three-dimensional coordinate system to determine the posture information of the acquisition device, wherein the acquisition device refers to the device that acquires the original image; extract the spatial geometric parameters of the grid ball, and construct point cloud data of the grid ball based on the spatial geometric parameters; calculate the point cloud inertia principal axis of the grid ball based on the point cloud data, and determine the two-dimensional mapping of the grid structure based on the posture information and the point cloud inertia principal axis.

[0121] In an optional embodiment of the present application, the matching module 74 is used to match the original image with the two-dimensional mapping to obtain a corresponding relationship; adjust the coordinate information of the grid ball in the two-dimensional mapping based on the corresponding relationship to obtain an adjusted two-dimensional mapping; construct the original three-dimensional model of the grid structure according to the adjusted two-dimensional mapping; and correct the original three-dimensional model to obtain a target three-dimensional model of the grid structure.

[0122] In an optional embodiment of the present application, the matching module 74 is used to obtain multi-perspective images of truss rods acquired by an acquisition device in different postures; identify multi-perspective rod boundaries of the truss rods in the multi-perspective images; compare the rod boundaries of the truss rods and the multi-perspective rod boundaries to determine the correction coefficients of the acquisition device in different postures, and adjust the original three-dimensional model of the truss structure according to the correction coefficients to obtain multiple candidate three-dimensional models; and determine the target three-dimensional model from the candidate three-dimensional models.

[0123] In an optional embodiment of the present application, the matching module 74 is used to calculate the reconstruction accuracy of multiple candidate three-dimensional models; determine the optimal posture of the acquisition device based on the reconstruction accuracy; adjust the model parameters of the three-dimensional model according to the optimal posture to obtain the target three-dimensional model.

[0124] In an optional embodiment of the present application, the device also includes: an optimization module, which is used to disconnect the coordinate link between the grid rods and the grid balls in the target three-dimensional model to obtain first position data of the grid rods and second position data of the grid balls; modeling the grid rods based on the first position data to obtain a first model, and modeling the grid balls based on the second position data to obtain a second model; adjusting the rod end coordinates in the target three-dimensional model according to the first model and the second model to obtain an optimized target three-dimensional model.

[0125] See also Figure 8 , Figure 8 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 8 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system).

[0126] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0127] The memory 20 stores instructions executable by at least one processor 10, so that at least one processor 10 executes the method shown in the above embodiment.

[0128] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created by the use of a computer device based on the presentation of a small program landing page, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0129] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0130] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0131] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0132] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A three-dimensional reconstruction method of a grid structure, characterized in that: The method comprises: Acquire an original image of a grid structure, wherein the grid structure comprises: grid balls and grid rods; Identifying the rod boundary of the grid rod in the original image and determining the center position of the grid ball; Determining a two-dimensional mapping of the grid structure based on the rod boundaries of the grid rods and the center position of the grid ball; The original image is matched with the two-dimensional mapping to obtain a three-dimensional model of the grid structure.

2. The method according to claim 1, characterized in that The identifying the rod boundary of the grid rod in the original image and determining the center position of the grid ball includes: Eliminating irrelevant information in the original image to obtain a preprocessed image; Performing edge detection on the preprocessed image to obtain the rod boundary of the grid rod and the spherical boundary of the grid ball; The center position of the grid ball is determined according to the rod boundary and the spherical surface boundary.

3. The method according to claim 1, characterized in that The determining of the two-dimensional mapping of the grid structure based on the rod member boundaries of the grid rods and the center position of the grid ball comprises: Determining four corner nodes based on the original image of the grid structure; Constructing a three-dimensional coordinate system according to the four corner nodes, and determining the position information of a collection device using the three-dimensional coordinate system, wherein the collection device refers to a device that collects the original image; Extracting spatial geometric parameters of the grid ball, and constructing point cloud data of the grid ball according to the spatial geometric parameters; The point cloud inertia principal axis of the grid ball is calculated according to the point cloud data, and the two-dimensional mapping of the grid structure is determined according to the posture information and the point cloud inertia principal axis.

4. The method according to claim 1, characterized in that: The matching of the original image with the two-dimensional mapping to obtain the three-dimensional model of the grid structure includes: Matching the original image with the two-dimensional map to obtain a corresponding relationship; Adjusting the coordinate information of the grid ball in the two-dimensional mapping based on the corresponding relationship to obtain an adjusted two-dimensional mapping; Constructing an original three-dimensional model of the grid structure according to the adjusted two-dimensional mapping; The original three-dimensional model is modified to obtain a target three-dimensional model of the grid structure.

5. The method according to claim 4, characterized in that The method of modifying the original three-dimensional model to obtain a target three-dimensional model of the grid structure includes: Acquire multi-view images of the grid rods collected by a collection device in different positions; Identifying multi-view rod boundaries of grid rods in the multi-view image; Comparing the rod boundary of the grid rod and the multi-view rod boundary, determining the correction coefficient of the acquisition device at different positions, and adjusting the original three-dimensional model of the grid structure according to the correction coefficient to obtain multiple candidate three-dimensional models; The target three-dimensional model is determined from the candidate three-dimensional models.

6. The method according to claim 5, characterized in that The determining the target three-dimensional model from the candidate three-dimensional models comprises: Calculating the reconstruction accuracy of the plurality of candidate three-dimensional models; Determining the optimal posture of the acquisition device based on the reconstruction accuracy; The model parameters of the three-dimensional model are adjusted according to the optimal posture to obtain the target three-dimensional model.

7. The method according to claim 6, characterized in that After obtaining the target three-dimensional model, the method further includes: Disconnecting the coordinate link between the grid rods and the grid balls in the target three-dimensional model to obtain first position data of the grid rods and second position data of the grid balls; Modeling the net frame rods based on the first position data to obtain a first model, and modeling the net frame ball based on the second position data to obtain a second model; The rod end coordinates in the target three-dimensional model are adjusted according to the first model and the second model to obtain an optimized target three-dimensional model.

8. A three-dimensional reconstruction device of a grid structure, characterized in that: The device comprises: An acquisition module, used to acquire an original image of a grid structure, wherein the grid structure comprises: grid balls and grid rods; A recognition module, used for recognizing the rod boundary of the grid rod in the original image and determining the center position of the grid ball; A determination module, configured to determine a two-dimensional mapping of the grid structure based on the rod boundaries of the grid rods and the center position of the grid ball; A matching module is used to match the original image with the two-dimensional mapping to obtain a three-dimensional model of the grid structure.

9. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method according to any one of claims 1 to 7 by executing the computer instructions.

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

Citation Information

Patent Citations

  • Camera three-dimensional truss calibrating target for automotive visual detection system

    CN104101299A

  • Installation structure of plant roof grid structure project and construction method thereof

    CN108590033A

  • Method for extracting welding ball from space truss based on point cloud and processing terminal

    CN115564761A

  • Three-dimensional target object mapping image determination method and device, equipment and storage medium

    CN116205978A

  • Three-dimensional reconstruction method, device and equipment based on neural radiation field and storage medium

    CN116310076A

Cited By

  • Space truss internal force monitoring method, device and equipment based on three-dimensional reconstruction

    CN119321840A

  • Method, device and equipment for monitoring internal force of space grid based on three-dimensional reconstruction

    CN119321840B

  • Texture reconstruction method for steel structure circular tube penetration welding joint

    CN122492907A