Construction method, device and storage medium of industrial digital twin scene

Abstract

The present application provides a method, device and storage medium for constructing an industrial digital twin scene. The method includes: obtaining point cloud data obtained by scanning an industrial scene using a multi-site scanning method, splicing the point cloud data using a hybrid splicing method to obtain a three-dimensional point cloud of the industrial scene; converting the three-dimensional point cloud using a K-nearest neighbor method to obtain a smooth and continuous point cloud, and converting the smooth and continuous point cloud into a point cloud mesh model using Poisson reconstruction technology; identifying and processing the features of the point cloud mesh model to determine the shape type of the point cloud mesh model, simplifying the point cloud mesh model according to the shape type using a QEM algorithm to obtain a lightweight point cloud model; and constructing an industrial digital twin scene using a lightweight point cloud model. This method solves the problem that the prior art ignores the lightweight processing of the mesh model of the point cloud reconstruction, resulting in poor operating efficiency of the digital twin scene.

Description

Technical Field

[0001] The present application relates to the technical field of constructing industrial digital twin scenarios, and specifically, to a method for constructing an industrial digital twin scenario, a method for monitoring an industrial digital twin scenario, a device for constructing an industrial digital twin scenario, and a computer-readable storage medium. Background Art

[0002] Digital twins are the key to the transformation and upgrading of modern manufacturing. They achieve a deep integration of the physical world and the digital world by creating a digital mirror in a virtual space that completely corresponds to the physical entity. This process is highly dependent on the accuracy of the three-dimensional model.

[0003] In line with the needs of Industry 4.0 and smart manufacturing, the manufacturing industry is undergoing a transformation from automation to intelligence. As one of the key technologies of smart manufacturing, digital twins can help enterprises achieve more efficient management and control throughout the entire life cycle of design, production, operation and maintenance, and improve production flexibility and resource utilization. It is not only a response to the current needs of manufacturing transformation and upgrading, but also an important way to explore intelligent, efficient and sustainable manufacturing models. The construction of industrial digital twin scenarios is the basis for the construction of digital twin smart factories. In order for digital twin models to accurately reflect the behavior and performance of physical objects, it is necessary to build highly accurate three-dimensional geometric models and complex physical models.

[0004] Most existing point cloud acquisition and splicing technologies are limited to specific single application environments. They are characterized by limited point cloud data acquisition scale and relatively single object types for point cloud mesh reconstruction, which are not suitable for complex industrial site layouts and dense equipment scenarios. In addition, existing technologies ignore the lightweight processing of point cloud reconstruction mesh models, which greatly affects the operating efficiency of digital twin scenarios, thereby hindering the actual deployment and application of industrial digital twin technology. Summary of the invention

[0005] The main purpose of this application is to provide a method for constructing an industrial digital twin scene, a method for monitoring an industrial digital twin scene, a device for constructing an industrial digital twin scene, and a computer-readable storage medium, so as to at least solve the problem that the prior art ignores the lightweight processing of the mesh model of point cloud reconstruction, resulting in poor operating efficiency of the digital twin scene.

[0006] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a method for constructing an industrial digital twin scene is provided, including: obtaining point cloud data obtained by scanning an industrial scene using a multi-site scanning method, and splicing the point cloud data using a hybrid stitching method to obtain a three-dimensional point cloud of the industrial scene; converting the three-dimensional point cloud using a K-nearest neighbor method to obtain a smooth and continuous point cloud, and converting the smooth and continuous point cloud into a point cloud mesh model using a Poisson reconstruction technique; identifying features of the point cloud mesh model to determine a shape type of the point cloud mesh model, and simplifying the point cloud mesh model according to the shape type using a QEM algorithm to obtain a lightweight point cloud model; and using the lightweight point cloud model to construct an industrial digital twin scene.

[0007] Optionally, after converting the three-dimensional point cloud using the K nearest neighbor method to obtain a smooth and continuous point cloud, the method further includes: determining the spatial domain weight of each point in the smooth and continuous point cloud based on the spatial distance between each point and the surrounding points, and determining the attribute domain weight of each point based on the similarity of the attribute values ​​of each point and the surrounding points; convolving the spatial domain weight with the attribute domain weight to obtain a target weight; and weighted averaging the attribute values ​​of each point in the smooth and continuous point cloud and the surrounding points based on the target weight to update the attribute values ​​of each point in the smooth and continuous point cloud to obtain an updated smooth and continuous point cloud.

[0008] Optionally, point cloud data obtained by scanning an industrial scene using a multi-site scanning method is obtained, and the point cloud data is stitched using a hybrid stitching method to obtain a three-dimensional point cloud of the industrial scene, including: aligning and overlapping a preset number of public targets of two adjacent sites through target stitching to spatially stitch the point cloud data to obtain spatially stitched point cloud data; and performing view stitching on the spatially stitched point cloud data from a first coordinate axis, a second coordinate axis, and a third coordinate axis, respectively, to obtain the three-dimensional point cloud of the industrial scene, wherein the first coordinate axis, the second coordinate axis, and the third coordinate axis are perpendicular to each other.

[0009] Optionally, before using the K nearest neighbor method to convert the three-dimensional point cloud to obtain a smooth and continuous point cloud, the method also includes: respectively setting the maximum and minimum values ​​in the first coordinate axis, the second coordinate axis, and the third coordinate axis to determine the spatial range, and deleting the point clouds that are not within the spatial range according to the coordinates of each point cloud in the three-dimensional point cloud to obtain a processed three-dimensional point cloud, wherein the first coordinate axis, the second coordinate axis, and the third coordinate axis are perpendicular to each other.

[0010] Optionally, after constructing the industrial digital twin scene using the lightweight point cloud model, the method further includes: using Level of Detail technology to set a multi-level precision display form of the lightweight point cloud model according to different distances between the industrial digital twin scene and the observer.

[0011] Optionally, the smooth continuous point cloud is converted into a point cloud mesh model using Poisson reconstruction technology, including: solving the Poisson equation according to the smooth continuous point cloud to obtain a scalar field; extracting an isosurface from the scalar field, and determining the point cloud mesh model according to the isosurface.

[0012] Optionally, after using the QEM algorithm to simplify the point cloud mesh model according to the shape type to obtain a lightweight point cloud model, the method also includes: an optimization step: determining the error value of each triangle in the lightweight point cloud model, selecting and improving the triangle with the largest error value; executing the optimization step multiple times to iteratively optimize the lightweight point cloud model, and after executing the optimization step a preset number of times, obtaining the optimized lightweight point cloud model.

[0013] According to another aspect of the present application, a method for monitoring an industrial digital twin scene is provided, comprising: constructing an industrial digital twin scene using any one of the methods for constructing an industrial digital twin scene; and monitoring an industrial site using the industrial digital twin scene.

[0014] According to another aspect of the present application, a device for constructing an industrial digital twin scene is provided, including: a stitching processing unit, used to obtain point cloud data obtained by scanning an industrial scene using a multi-site scanning method, and stitching the point cloud data using a hybrid stitching method to obtain a three-dimensional point cloud of the industrial scene; a conversion processing unit, used to convert the three-dimensional point cloud using a K-nearest neighbor method to obtain a smooth and continuous point cloud, and convert the smooth and continuous point cloud into a point cloud mesh model using a Poisson reconstruction technique; a simplification processing unit, used to identify and process features of the point cloud mesh model, determine a shape type of the point cloud mesh model, and simplify the point cloud mesh model according to the shape type using a QEM algorithm to obtain a lightweight point cloud model; and a construction unit, used to construct an industrial digital twin scene using the lightweight point cloud model.

[0015] According to another aspect of the present application, a computer-readable storage medium is provided, which includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the methods for constructing an industrial digital twin scenario.

[0016] By applying the technical solution of the present application, point cloud data obtained by scanning an industrial scene using a multi-site scanning method is acquired, and a hybrid stitching method is used to stitch the point cloud data to obtain a three-dimensional point cloud of the industrial scene; the three-dimensional point cloud is converted using the K nearest neighbor method to obtain a smooth and continuous point cloud, and the smooth and continuous point cloud is converted into a point cloud mesh model using Poisson reconstruction technology; the features of the point cloud mesh model are identified and processed to determine the shape type of the point cloud mesh model, and the point cloud mesh model is simplified according to the shape type using the QEM algorithm to obtain a lightweight point cloud model; and the lightweight point cloud model is used to construct an industrial digital twin scene. According to the shape type identification of the point cloud mesh model, the most suitable mesh simplification parameters are matched according to the shape type, and the QEM algorithm is used to simplify the point cloud mesh model, which solves the model deformation problem caused by the use of a single mesh simplification parameter to the greatest extent, while ensuring the accuracy of the mesh simplification and improving the efficiency of the simplification processing; and through point cloud splicing, point cloud conversion, point cloud mesh model recognition and mesh model lightweight technology, the workload of building industrial digital twin scenes is greatly reduced, ensuring the efficient and smooth operation of the scene; It solves the problem that the existing technology ignores the lightweight processing of the mesh model reconstructed by the point cloud, resulting in poor operating efficiency of the digital twin scene. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings constituting part of the present application are used to provide a further understanding of the present application. The exemplary embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0018] Figure 1 A hardware structure block diagram of a mobile terminal for executing a method for constructing an industrial digital twin scenario provided in an embodiment of the present application is shown;

[0019] Figure 2 A schematic diagram of a process for constructing an industrial digital twin scenario provided according to an embodiment of the present application is shown;

[0020] Figure 3 A schematic diagram of a process for constructing a specific industrial digital twin scenario provided in accordance with an embodiment of the present application is shown;

[0021] Figure 4 A schematic diagram of the process of point cloud data stitching provided according to an embodiment of the present application is shown;

[0022] Figure 5 A schematic diagram of a point cloud lightweight process according to an embodiment of the present application is shown;

[0023] Figure 6 A schematic diagram of a scattered and disordered point cloud provided according to an embodiment of the present application is shown;

[0024] Figure 7 A schematic diagram of a smooth continuous point cloud provided according to an embodiment of the present application is shown;

[0025] Figure 8 A schematic diagram of a process for lightweighting a point cloud mesh model provided according to an embodiment of the present application is shown;

[0026] Fig. 9 A structural block diagram of a device for constructing an industrial digital twin scenario provided according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0027] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0028] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0029] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0030] For the convenience of description, some nouns or terms involved in the embodiments of the present application are explained below:

[0031] Lidar: A system that emits laser beams and receives echoes to obtain three-dimensional information about a target;

[0032] Point cloud: A collection of points distributed in three-dimensional space in a discrete and irregular manner;

[0033] LiDAR point cloud: point cloud obtained by LiDAR scanning;

[0034] Point cloud density: the average number of laser points in the point cloud per unit area, with the elevation direction as the normal direction;

[0035] Digital twin: Use physical sensors to apply equipment and production line information in a virtual environment to achieve interaction between physical entities and virtual models, and provide support for factory online management, real-time monitoring and other technologies;

[0036] Physical entity: refers to actual objects such as factory buildings, production equipment, and products;

[0037] Digital Twin: A digital representation of a physical entity;

[0038] QEM: Quadtree Error Metric Algorithm, Geometric Mesh Simplification and Optimization Technique.

[0039] As introduced in the background technology, the prior art ignores the lightweight processing of the mesh model for point cloud reconstruction, resulting in poor operating efficiency of the digital twin scene. In order to solve the problem that the prior art ignores the lightweight processing of the mesh model for point cloud reconstruction, resulting in poor operating efficiency of the digital twin scene, the embodiments of the present application provide a method for constructing an industrial digital twin scene, a method for monitoring an industrial digital twin scene, a device for constructing an industrial digital twin scene, and a computer-readable storage medium.

[0040] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0041] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Taking running on a mobile terminal as an example, Figure 1 1 is a hardware structure block diagram of a mobile terminal of a method for constructing an industrial digital twin scene in an embodiment of the present invention. Figure 1 As shown, the mobile terminal may include one or more ( Figure 1 Only one is shown in the figure) a processor 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA) and a memory 104 for storing data, wherein the mobile terminal may also include a transmission device 106 and an input / output device 108 for communication functions. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the mobile terminal. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0042] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as the computer program corresponding to the method for constructing the industrial digital twin scene in the embodiment of the present invention. The processor 102 executes various functional applications and data processing by running the computer program stored in the memory 104, that is, to implement the above method. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the mobile terminal 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 a combination thereof. The transmission device 106 is used to receive or send data via a network. The above-mentioned specific network examples may include a wireless network provided by a communication provider of the mobile terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 may be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0043] In this embodiment, a method for constructing an industrial digital twin scenario running on a mobile terminal, a computer terminal or a similar computing device is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the 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.

[0044] Figure 2 : is a flowchart of a method for constructing an industrial digital twin scenario according to an embodiment of the present application. Figure 2 As shown, the method comprises the following steps:

[0045] Step S201, obtaining point cloud data obtained by scanning an industrial scene using a multi-site scanning method, and splicing the point cloud data using a hybrid splicing method to obtain a three-dimensional point cloud of the industrial scene;

[0046] Specifically, the complex scenes in industrial environments contain many devices, and the interweaving of the devices makes the laser signal of the laser scanner vulnerable to occlusion, which directly affects the integrity of the point cloud data. In order to ensure that complete scene information is collected, a multi-site scanning method is adopted to scan the scene.

[0047] Among them, the hybrid stitching method is a hybrid stitching method of target stitching and view stitching;

[0048] Through the hybrid stitching method of target stitching and view stitching, multiple site point cloud data obtained from scanning at different locations are integrated into the same coordinate system to construct a complete three-dimensional point cloud of the industrial scene.

[0049] Step S202, using the K nearest neighbor method to convert the three-dimensional point cloud to obtain a smooth continuous point cloud, and using the Poisson reconstruction technology to convert the smooth continuous point cloud into a point cloud mesh model;

[0050] Specifically, the K nearest neighbor method is used to set the point cloud neighborhood to collect the point set around each point, and the standard deviation analysis is performed on the position of the point cloud within the given neighborhood. Each point is judged by comparing with the set threshold, and the points that do not meet the requirements are filtered out based on the above judgment, and the points that meet the statistical characteristics are retained, which can effectively remove outliers and abnormal points in the point cloud and eliminate redundant points in overly dense areas.

[0051] Step S203, identifying and processing the features of the point cloud mesh model, determining the shape type of the point cloud mesh model, and simplifying the point cloud mesh model according to the shape type using a QEM algorithm to obtain a lightweight point cloud model;

[0052] Specifically, mesh simplification mainly uses the QEM algorithm, which is an optimization algorithm based on quadratic metric error. It evaluates and optimizes the quadratic error metric of the mesh surface, collapses a certain edge, and combines the two points of the edge into one point to merge the patches, thereby reducing the number of polygons in the mesh model while trying to maintain the main features of the original geometric shape.

[0053] When simplifying curved meshes such as cylinders and cones, the main focus is on maintaining the smoothness and shape characteristics of the surface. The QEM algorithm calculates an error value for each triangle or vertex, and decides which triangles can be merged or deleted based on the error threshold to maintain the key feature of curvature.

[0054] When simplifying planar meshes such as cubes and tetrahedrons, the focus is on maintaining the flatness of the plane and the accuracy of the boundaries. Therefore, when calculating errors, more emphasis is placed on the consistency of the plane's normal vectors and the alignment of the boundaries to reduce the "staircase effect".

[0055] Step S204: Use a lightweight point cloud model to build an industrial digital twin scene.

[0056] Specifically, building an industrial digital twin scene is a process of reproducing the physical industrial environment with high precision in the virtual space. It combines CAD to accurately restore the layout structure of the industrial site in the three-dimensional engine, including production equipment, production line layout, logistics channels, etc., to ensure that the digital twin model is highly consistent with the actual scene in terms of geometric dimensions and spatial relationships.

[0057] Through this embodiment, point cloud data obtained by scanning an industrial scene using a multi-site scanning method is acquired, and a hybrid stitching method is used to stitch the point cloud data to obtain a three-dimensional point cloud of the industrial scene; the three-dimensional point cloud is converted using the K nearest neighbor method to obtain a smooth and continuous point cloud, and the Poisson reconstruction technology is used to convert the smooth and continuous point cloud into a point cloud mesh model; the features of the point cloud mesh model are identified and processed to determine the shape type of the point cloud mesh model, and the QEM algorithm is used to simplify the point cloud mesh model according to the shape type to obtain a lightweight point cloud model; and the lightweight point cloud model is used to construct an industrial digital twin scene. According to the shape type identification of the point cloud mesh model, the most suitable mesh simplification parameters are matched according to the shape type, and the QEM algorithm is used to simplify the point cloud mesh model, which solves the model deformation problem caused by the use of a single mesh simplification parameter to the greatest extent, while ensuring the accuracy of the mesh simplification and improving the efficiency of the simplification processing; and through point cloud splicing, point cloud conversion, point cloud mesh model recognition and mesh model lightweight technology, the workload of building industrial digital twin scenes is greatly reduced, ensuring the efficient and smooth operation of the scene; It solves the problem that the existing technology ignores the lightweight processing of the mesh model reconstructed by the point cloud, resulting in poor operating efficiency of the digital twin scene.

[0058] In the specific implementation process, after the K nearest neighbor method is used to convert the three-dimensional point cloud to obtain a smooth and continuous point cloud, the method also includes: determining the spatial domain weight of each point according to the spatial distance between each point in the smooth and continuous point cloud and the surrounding points, and determining the attribute domain weight of each point according to the similarity of the attribute values ​​of each point and the surrounding points; convolving the spatial domain weight with the attribute domain weight to obtain the target weight; and weighted averaging the attribute values ​​of each point in the smooth and continuous point cloud and the surrounding points according to the target weight to update the attribute values ​​of each point in the smooth and continuous point cloud to obtain an updated smooth and continuous point cloud.

[0059] This method calculates the weight of the spatial domain based on the spatial distance between each point and the surrounding points, and then calculates the weight of the attribute domain based on the similarity of the attribute values, and convolves the two weights to obtain the final weight. Using the weights calculated above, the attribute values ​​of the center point and the points in its neighborhood are weighted averaged to obtain the new attribute values ​​after filtering. Not only does it efficiently remove the noise on the surface of the three-dimensional model, it also excellently preserves the key structural features in the point cloud data, preventing the loss of details caused by over-smoothing, thereby maintaining the integrity and authenticity of the point cloud data.

[0060] Specifically, point cloud data obtained by scanning an industrial scene using a multi-site scanning method is obtained, and the point cloud data is stitched using a hybrid stitching method to obtain a three-dimensional point cloud of the industrial scene, including: aligning and overlapping a preset number of public targets of two adjacent sites through target stitching to spatially stitch the point cloud data to obtain spatially stitched point cloud data; and performing view stitching on the spatially stitched point cloud data from a first coordinate axis, a second coordinate axis, and a third coordinate axis, respectively, to obtain a three-dimensional point cloud of the industrial scene, wherein the first coordinate axis, the second coordinate axis, and the third coordinate axis are perpendicular to each other.

[0061] Among them, the preset number can be set to 3, the first coordinate axis is the X axis, the second coordinate axis is the Y axis, and the third coordinate axis is the Z axis.

[0062] This method first uses target stitching to align and overlap the three common targets in two adjacent sites to achieve spatial stitching of point clouds, ensuring that the overlap of the point clouds of the two sites meets the restriction requirements, generally not less than 75%. Then, through view stitching, point clouds are aligned in three axes: X-axis, Y-axis, and Z-axis to ensure that the point clouds in the horizontal and vertical directions are aligned correctly, and the overall error between the stitching sites and the error in each direction are strictly controlled within 3mm. Finally, the point cloud stitching of all sites is completed to obtain complete industrial scene point cloud data.

[0063] More specifically, before using the K nearest neighbor method to convert the three-dimensional point cloud to obtain a smooth and continuous point cloud, the method also includes: respectively setting the maximum and minimum values ​​in the first coordinate axis, the second coordinate axis, and the third coordinate axis to determine the spatial range, and deleting the point clouds that are not within the spatial range according to the coordinates of each point cloud in the three-dimensional point cloud to obtain a processed three-dimensional point cloud, wherein the first coordinate axis, the second coordinate axis, and the third coordinate axis are perpendicular to each other.

[0064] This method sets the filtering range according to the target scene of the point cloud scanning, sets the maximum and minimum values ​​in the three axes of X, Y, and Z, uses the algorithm to traverse each point in the entire point cloud data set, checks the coordinate value of each point in space, retains the points whose coordinate axis values ​​fall within the range, filters the points whose coordinate axis values ​​fall outside the range, narrows the scope of subsequent point cloud data processing, and reduces the burden of subsequent point cloud data processing.

[0065] Furthermore, after constructing the industrial digital twin scene using the lightweight point cloud model, the method also includes: using Level of Detail technology to set a multi-level precision display form of the lightweight point cloud model according to different distances between the industrial digital twin scene and the observer.

[0066] This method uses Level of Detail (LOD) technology to set the model's multi-level precision display form for various application scenarios and different distances of the observer. When the object is far away from the visual range, it is presented using a simplified model, which effectively reduces the pressure of graphics rendering and ensures a smooth visual experience. As the object approaches the observation focus, it seamlessly transitions to a high-precision model with higher resolution and more levels of detail.

[0067] Furthermore, a Poisson reconstruction technique is used to convert the smooth continuous point cloud into a point cloud mesh model, including: solving the Poisson equation according to the smooth continuous point cloud to obtain a scalar field; extracting an isosurface from the scalar field, and determining the point cloud mesh model according to the isosurface.

[0068] This method constructs an indicator function based on point cloud data to distinguish the internal and external areas of the object. This function will be applied to the finite difference method as a boundary condition to solve the Poisson equation to obtain a smooth scalar field. The isosurface is extracted from this continuous scalar field. Subsequently, the isosurface extraction operation is performed on the scalar field to improve and close the surface model, and finally a smooth and uninterrupted three-dimensional mesh model is obtained.

[0069] Specifically, after using the QEM algorithm to simplify the point cloud mesh model according to the shape type to obtain a lightweight point cloud model, the method also includes: an optimization step: determining the error value of each triangle in the lightweight point cloud model, selecting and improving the triangle with the largest error value; executing the optimization step multiple times to iteratively optimize the lightweight point cloud model, and after executing the preset number of optimization steps, an optimized lightweight point cloud model is obtained.

[0070] The preset number of times can be set to 50, 100, or 200, depending on the iteration requirements;

[0071] The method can also determine whether the iterative optimization is completed by setting an error threshold. When the error values ​​of each triangle in the lightweight point cloud model are calculated to be less than the error threshold, the optimized lightweight point cloud model is obtained.

[0072] Specifically, after mesh simplification, an iterative optimization process is performed to continuously select and improve the triangles with the largest errors, gradually simplifying the mesh while striving to maintain the accuracy of the overall shape. The errors of all triangles are recalculated in each round, and the optimization strategy is adjusted according to the new mesh layout. The model that has completed mesh simplification is checked to ensure that the model has not deformed. If deformation occurs, the processing threshold needs to be adjusted and the original model needs to be meshed again.

[0073] An embodiment of the present application also includes a method for monitoring an industrial digital twin scene, comprising: constructing an industrial digital twin scene using any of the above-mentioned industrial digital twin scene construction methods; and monitoring the industrial site using the above-mentioned industrial digital twin scene.

[0074] In order to enable those skilled in the art to more clearly understand the technical solution of the present application, the implementation process of the method for constructing the industrial digital twin scene of the present application will be described in detail below in combination with specific embodiments.

[0075] This embodiment relates to a specific method for constructing an industrial digital twin scene. Through methods such as point cloud stitching, point cloud denoising, point cloud mesh reconstruction, and mesh model lightweighting, the workload of building an industrial digital twin scene is greatly reduced, ensuring efficient and smooth operation of the scene. In addition, this embodiment adopts model lightweighting technology, which adds model classification based on feature recognition before the previous lightweight processing, and performs different QEM processing on curved mesh bodies and plane mesh bodies, which solves the model deformation problem caused by the use of a single mesh simplification parameter to the greatest extent, while ensuring the accuracy of mesh simplification and improving the efficiency of simplification processing. Figure 3 As shown, the following steps are included:

[0076] Step S1: scene scanning, collecting industrial scene point cloud data;

[0077] Step S2: point cloud stitching, integration and matching of multi-site point cloud data;

[0078] Step S3: Lightweight the point cloud, purify the point cloud data, and improve the point cloud quality;

[0079] Step S4: point cloud mesh reconstruction, point cloud data is converted into a mesh model;

[0080] Step S5: Lighten the mesh model and optimize the mesh model to reduce the load;

[0081] Step S6: Construction of industrial digital twin scene.

[0082] The complex scene of the industrial environment in step S1 contains many devices. The interweaving of the devices makes the laser signal of the laser scanner easily blocked, which directly affects the integrity of the point cloud data. To ensure that the complete scene information is collected, such as Figure 4As shown in the figure, a multi-site scanning method is used to scan the scene. The sampling interval between each site is kept at about 5m, and the sites can form a complete closed loop. In the acquisition scene, a target is set every 3m in the continuous plane area, and targets are added in areas that are easily blocked to ensure that each site can scan 5 targets, the number of common reference targets between two adjacent sites is ≥3, and each target can be collected. Through such scene scanning measures, the integrity and accuracy of the point cloud scanning data can be maximized, thereby obtaining the best scanning results.

[0083] Step S2 is to integrate multiple site point cloud data obtained from scanning at different locations into the same coordinate system through a hybrid stitching method of target stitching and view stitching, thereby constructing a complete three-dimensional point cloud of the industrial scene.

[0084] like Figure 4 As shown in the figure, firstly, through target stitching, the three public targets in the two adjacent sites are aligned and overlapped to realize the spatial stitching of the point cloud, ensuring that the overlap of the point cloud stitching of the two sites meets the restriction requirements, generally not less than 75%. Then, through view stitching, the point cloud is aligned in three axes: X-axis, Y-axis, and Z-axis to ensure that the point cloud is aligned correctly in the horizontal and vertical directions, and the overall error between the stitching sites and the error in each direction are strictly controlled within 3mm. Finally, the point cloud stitching of all sites is completed to obtain the complete industrial scene point cloud data.

[0085] Step S3 aims to remove non-target noise points generated during the scanning process, including abnormal points, isolated points and irrelevant background points. These noise points may interfere with the accuracy and efficiency of point cloud mesh reconstruction. In order to ensure the accuracy and clarity of the point cloud, Figure 5 As shown, the following steps are used to purify the point cloud data:

[0086] 1) Through filtering: Set the filtering range according to the target scene of the point cloud scanning, set the maximum and minimum values ​​in the three axes of X, Y, and Z, use the algorithm to traverse each point in the entire point cloud data set, check the coordinate value of each point in space, retain the points whose coordinate axis values ​​fall within the range, filter out the points whose coordinate axis values ​​fall outside the range, narrow the scope of subsequent point cloud data processing, and reduce the burden of subsequent point cloud data processing.

[0087] 2) Statistical filtering: Use the K nearest neighbor method to set the point cloud neighborhood to collect the point set around each point, perform standard deviation analysis on the position of the point cloud within the given neighborhood, and judge each point by comparing it with the set threshold. Based on the above judgment, filter out points that do not meet the requirements and retain points that meet the statistical characteristics, effectively removing outliers and abnormal points in the point cloud, and eliminating redundant points in overly dense areas.

[0088] 3) Bilateral filtering: Calculate the weight of the spatial domain based on the spatial distance between each point and the surrounding points, and then calculate the weight of the attribute domain based on the similarity of the attribute values, and convolve the two weights to get the final weight. Using the weights calculated above, perform weighted averaging on the attribute values ​​of the center point and the points in its neighborhood to get the new attribute values ​​after filtering. It not only efficiently removes the noise on the surface of the 3D model, but also excellently preserves the key structural features in the point cloud data, preventing the loss of details caused by over-smoothing, thereby maintaining the integrity and authenticity of the point cloud data.

[0089] The above step S3 can effectively improve the quality of point cloud data by performing point cloud lightweighting through the above steps, and convert the scattered and disordered point cloud before processing into a smooth and continuous point cloud, laying a solid foundation for subsequent point cloud mesh reconstruction, analysis and application. The schematic diagram of the scattered and disordered point cloud is as follows: Figure 6 As shown, the schematic diagram of smooth continuous point cloud is as follows Figure 7 shown.

[0090] Step S4 converts the 3D point cloud into a connected, smooth mesh model. The specific steps of point cloud mesh reconstruction are as follows:

[0091] 1) Poisson reconstruction: An indicator function is constructed based on the point cloud data to distinguish the internal and external areas of the object. This function will be used as a boundary condition in the finite difference method to solve the Poisson equation to obtain a smooth scalar field. The isosurface is extracted from this continuous scalar field. Subsequently, the isosurface extraction operation is performed on the scalar field to improve and close the surface model, and finally a smooth and uninterrupted three-dimensional mesh model is obtained.

[0092] 2) Post-processing: Smooth the surface of the reconstructed mesh model by artificial methods, remove the bumps and depressions in the model, fill the holes in the polygons, improve the closure of the model, and finally fit the model surface to improve the quality of the polygonal model. Then, reconstruct the solid model based on the polygonal model through 3D geometric modeling methods.

[0093] The three-dimensional model obtained by point cloud mesh reconstruction in step S5 has the problem of too many triangular faces. A high-faceted three-dimensional model will increase the load on the GPU and CPU and affect the operating efficiency of the twin scene. Therefore, it is necessary to lightweight the reconstructed three-dimensional model. While ensuring that the model accuracy is as unaffected as possible, reduce the model size, computational complexity and memory usage, so that the scene rendering is faster and smoother.

[0094] In step S5, the model is lightweighted and hidden culling technology is used to perform visibility analysis from multiple different perspectives to ensure accurate identification of all visual elements. Subsequently, the minimum enclosed space required to be displayed in the scene is defined and maintained, thereby eliminating elements in a state of visual occlusion to achieve efficient optimization and presentation of the scene.

[0095] The schematic diagram of the lightweight process of point cloud mesh model is as follows Figure 8 As shown in the figure, with the help of feature recognition technology, the geometric shapes of the model, such as cubes, cylinders, cones, tetrahedrons and other basic elements, are accurately identified and labeled accordingly. Based on these labels, a classified mesh optimization strategy is adopted to ensure that each geometric element can match the most suitable mesh simplification parameters, which solves the model deformation problem caused by the use of a single mesh simplification parameter to the greatest extent, and improves the efficiency of the simplification process while ensuring the accuracy of mesh simplification.

[0096] The mesh simplification in this embodiment mainly uses the QEM algorithm, which is an optimization algorithm based on quadratic metric error. It evaluates and optimizes the quadratic error metric of the mesh surface, collapses a certain edge, and combines the two points of the edge into one point to merge the patches, thereby reducing the number of polygons in the mesh model while trying to maintain the main features of the original geometric shape.

[0097] When simplifying curved meshes such as cylinders and cones, the main focus is on maintaining the smoothness and shape characteristics of the surface. The QEM algorithm calculates an error value for each triangle or vertex, and decides which triangles can be merged or deleted based on the error threshold to maintain the key feature of curvature.

[0098] When simplifying planar meshes such as cubes and tetrahedrons, the focus is on maintaining the flatness of the plane and the accuracy of the boundaries. Therefore, when calculating errors, more emphasis is placed on the consistency of the plane's normal vectors and the alignment of the boundaries to reduce the "staircase effect".

[0099] After the mesh is simplified, an iterative optimization process is performed to continuously select and improve the triangles with the largest errors, gradually simplifying the mesh while trying to maintain the accuracy of the overall shape. Each round recalculates the errors of all triangles and adjusts the optimization strategy based on the new mesh layout.

[0100] Check the model after mesh simplification to ensure that the model is not deformed. If deformation occurs, adjust the processing threshold and re-simplify the original model.

[0101] Using Level of Detail (LOD) technology, we set up multi-level precision display forms of the model for various application scenarios and different distances of the observer. When the object is far away from the visual range, it is presented with a simplified model, which effectively reduces the pressure of graphics rendering and ensures a smooth visual experience. As the object approaches the observation focus, it seamlessly transitions to a high-precision model with higher resolution and more levels of detail.

[0102] Through this series of model lightweight processing, we successfully created a lightweight model that meets the needs of industrial digital twin scenarios, achieving efficient and accurate model display.

[0103] Building the industrial digital twin scene in step S6 is a process of reproducing the physical industrial environment in the virtual space with high precision. It combines CAD to accurately restore the layout structure of the industrial site in the three-dimensional engine, including production equipment, production line layout, logistics channels, etc., to ensure that the digital twin model is highly consistent with the actual scene in terms of geometric dimensions and spatial relationships.

[0104] Most existing point cloud scanning and acquisition technologies are limited to specific single scenes, and the scale of point cloud data acquisition is limited. When faced with complex industrial site layouts and dense equipment, there is a lack of effective point cloud acquisition and stitching methods to obtain complete scene point cloud data. Based on this, this embodiment adopts a point cloud acquisition method that sets point cloud targets, and ensures the integrity of point cloud stitching through a hybrid stitching method of target stitching and view stitching; in addition, in order to solve the operation bottleneck problem caused by large-scale industrial three-dimensional models, this embodiment also adopts model lightweight technology, and adds feature recognition-based model classification before the previous lightweight processing, and performs different QEM processing on curved mesh bodies and plane mesh bodies, which solves the model deformation problem caused by the use of a single mesh simplification parameter to the greatest extent, while ensuring the accuracy of mesh simplification and improving the efficiency of simplification processing.

[0105] This embodiment realizes point cloud data processing in complex industrial environments, covering multiple key links such as comprehensive collection, precise splicing, point cloud lightweighting and high-quality reconstruction. In addition, on the basis of realizing laser point cloud mesh reconstruction, the model lightweight processing process is introduced to effectively solve the problem that the large number of triangular faces of industrial models hinders the actual deployment and application of industrial digital twins.

[0106] The embodiment of the present application also provides a device for constructing an industrial digital twin scene. It should be noted that the device for constructing an industrial digital twin scene of the embodiment of the present application can be used to execute the method for constructing an industrial digital twin scene provided in the embodiment of the present application. The device is used to implement the above-mentioned embodiments and preferred implementations, and those that have been described will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware for a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.

[0107] The following introduces the device for constructing the industrial digital twin scene provided in the embodiment of the present application.

[0108] Fig. 9 Schematic diagram of a construction device for an industrial digital twin scene according to an embodiment of the present application. Fig. 9 As shown, the device comprises:

[0109] A stitching processing unit 91 is used to obtain point cloud data obtained by scanning the industrial scene using a multi-site scanning method, and stitch the point cloud data using a hybrid stitching method to obtain a three-dimensional point cloud of the industrial scene;

[0110] Specifically, the complex scenes in industrial environments contain many devices, and the interweaving of the devices makes the laser signal of the laser scanner vulnerable to occlusion, which directly affects the integrity of the point cloud data. In order to ensure that complete scene information is collected, a multi-site scanning method is adopted to scan the scene.

[0111] A conversion processing unit 92 is used to convert the three-dimensional point cloud using a K-nearest neighbor method to obtain a smooth and continuous point cloud, and convert the smooth and continuous point cloud into a point cloud mesh model using a Poisson reconstruction technique;

[0112] Specifically, the K nearest neighbor method is used to set the point cloud neighborhood to collect the point set around each point, and the standard deviation analysis is performed on the position of the point cloud within the given neighborhood. Each point is judged by comparing with the set threshold, and the points that do not meet the requirements are filtered out based on the above judgment, and the points that meet the statistical characteristics are retained, which can effectively remove outliers and abnormal points in the point cloud and eliminate redundant points in overly dense areas.

[0113] The simplification processing unit 93 is used to identify the features of the point cloud mesh model, determine the shape type of the point cloud mesh model, and use the QEM algorithm to simplify the point cloud mesh model according to the shape type to obtain a lightweight point cloud model;

[0114] Specifically, mesh simplification mainly uses the QEM algorithm, which is an optimization algorithm based on quadratic metric error. It evaluates and optimizes the quadratic error metric of the mesh surface, collapses a certain edge, and combines the two points of the edge into one point to merge the patches, thereby reducing the number of polygons in the mesh model while trying to maintain the main features of the original geometric shape.

[0115] The construction unit 94 is used to construct an industrial digital twin scene using a lightweight point cloud model.

[0116] Specifically, building an industrial digital twin scene is a process of reproducing the physical industrial environment with high precision in the virtual space. It combines CAD to accurately restore the layout structure of the industrial site in the three-dimensional engine, including production equipment, production line layout, logistics channels, etc., to ensure that the digital twin model is highly consistent with the actual scene in terms of geometric dimensions and spatial relationships.

[0117] In this embodiment, the stitching processing unit is used to obtain point cloud data obtained by scanning an industrial scene using a multi-site scanning method, and use a hybrid stitching method to stitch the point cloud data to obtain a three-dimensional point cloud of the industrial scene; the conversion processing unit is used to convert the three-dimensional point cloud using a K-nearest neighbor method to obtain a smooth and continuous point cloud, and use Poisson reconstruction technology to convert the smooth and continuous point cloud into a point cloud mesh model; the simplification processing unit is used to identify and process the features of the point cloud mesh model, determine the shape type of the point cloud mesh model, and use the QEM algorithm to simplify the point cloud mesh model according to the shape type to obtain a lightweight point cloud model; the construction unit is used to construct an industrial digital twin scene using a lightweight point cloud model. According to the shape type identification of the point cloud mesh model, the most suitable mesh simplification parameters are matched according to the shape type, and the QEM algorithm is used to simplify the point cloud mesh model, which solves the model deformation problem caused by the use of a single mesh simplification parameter to the greatest extent, while ensuring the accuracy of the mesh simplification and improving the efficiency of the simplification processing; and through point cloud splicing, point cloud conversion, point cloud mesh model recognition and mesh model lightweight technology, the workload of building industrial digital twin scenes is greatly reduced, ensuring the efficient and smooth operation of the scene; It solves the problem that the existing technology ignores the lightweight processing of the mesh model reconstructed by the point cloud, resulting in poor operating efficiency of the digital twin scene.

[0118] As an optional scheme, the device also includes a determination unit, a convolution unit and an update unit. The determination unit is used to determine the spatial domain weight of each point in the smooth continuous point cloud according to the spatial distance between each point and the surrounding points after converting the three-dimensional point cloud using the K nearest neighbor method to obtain a smooth and continuous point cloud, and determine the attribute domain weight of each point according to the similarity of the attribute values ​​of each point with the surrounding points; the convolution unit is used to convolve the spatial domain weight with the attribute domain weight to obtain a target weight; the update unit is used to perform a weighted average of the attribute values ​​of each point in the smooth and continuous point cloud and the surrounding points according to the target weight to update the attribute value of each point in the smooth and continuous point cloud to obtain an updated smooth and continuous point cloud.

[0119] Specifically, the weight of the spatial domain is calculated based on the spatial distance between each point and the surrounding points, and the weight of the attribute domain is calculated based on the similarity of the attribute values. The two weights are convolved to obtain the final weight. Using the weights calculated above, the attribute values ​​of the center point and the points in its neighborhood are weighted averaged to obtain the new attribute values ​​after filtering. Not only does it efficiently remove the noise on the surface of the three-dimensional model, it also excellently preserves the key structural features in the point cloud data, preventing the loss of details caused by over-smoothing, thereby maintaining the integrity and authenticity of the point cloud data.

[0120] An optional solution, the stitching processing unit includes a first stitching module and a second stitching module; the first stitching module is used to align and overlap a preset number of public targets of two adjacent sites through target stitching, so as to spatially stitch the point cloud data and obtain the point cloud data after spatial stitching; the second stitching module is used to perform view stitching on the point cloud data after spatial stitching from the first coordinate axis, the second coordinate axis, and the third coordinate axis, respectively, to obtain a three-dimensional point cloud of the industrial scene, wherein the first coordinate axis, the second coordinate axis, and the third coordinate axis are perpendicular to each other.

[0121] Among them, the preset number can be set to 3, the first coordinate axis is the X axis, the second coordinate axis is the Y axis, and the third coordinate axis is the Z axis.

[0122] An optional solution, the device also includes a deletion processing unit, which is used to set the maximum and minimum values ​​in the first coordinate axis, the second coordinate axis, and the third coordinate axis to determine the spatial range before converting the three-dimensional point cloud using the K nearest neighbor method to obtain a smooth and continuous point cloud, and delete the point clouds that are not within the spatial range according to the coordinates of each point cloud in the three-dimensional point cloud to obtain a processed three-dimensional point cloud, wherein the first coordinate axis, the second coordinate axis, and the third coordinate axis are perpendicular to each other.

[0123] Specifically, the filtering range is set according to the target scene of the point cloud scanning, the maximum and minimum values ​​in the three axes of X, Y, and Z are set, and the algorithm is used to traverse each point in the entire point cloud data set, check the coordinate value of each point in space, retain the points whose coordinate axis values ​​fall within the range, filter out the points whose coordinate axis values ​​fall outside the range, narrow the scope of subsequent point cloud data processing, and reduce the burden of subsequent point cloud data processing.

[0124] An optional solution is that the device also includes a setting unit, which is used to use Level of Detail technology to set the multi-level precision display form of the lightweight point cloud model according to different distances between the industrial digital twin scene and the observer after constructing the industrial digital twin scene using the lightweight point cloud model.

[0125] Specifically, the Level of Detail (LOD) technology is used to set the multi-level precision display form of the model for various application scenarios and different distances of the observer. When the object is far away from the visual range, a simplified model is used to present it, which effectively reduces the pressure of graphics rendering and ensures a smooth visual experience.

[0126] An optional solution, the conversion processing unit includes a solution module and a determination module, the solution module is used to solve the Poisson equation according to the smooth continuous point cloud to obtain a scalar field; the determination module is used to extract an isosurface from the scalar field and determine the point cloud mesh model according to the isosurface.

[0127] Specifically, an indicator function is constructed based on the point cloud data to distinguish the internal and external areas of the object. This function will be used as a boundary condition in the finite difference method to solve the Poisson equation to obtain a smooth scalar field. The isosurface is extracted from this continuous scalar field. Subsequently, the isosurface extraction operation is performed on the scalar field to improve and close the surface model, and finally a smooth and uninterrupted three-dimensional mesh model is obtained.

[0128] An optional solution, the device also includes an optimization unit and an execution unit, the optimization unit is used to simplify the point cloud mesh model according to the shape type using the QEM algorithm to obtain a lightweight point cloud model, and then execute the optimization steps: determine the error value of each triangle in the lightweight point cloud model, select and improve the triangle with the largest error value; the execution unit is used to execute the optimization steps multiple times to iteratively optimize the lightweight point cloud model, and after executing the preset number of optimization steps, obtain the optimized lightweight point cloud model.

[0129] Specifically, whether the iterative optimization is completed can also be determined by setting an error threshold. When the error values ​​of each triangle in the lightweight point cloud model are calculated to be less than the error threshold, the optimized lightweight point cloud model is obtained.

[0130] The construction device of the industrial digital twin scene includes a processor and a memory. The above-mentioned splicing processing unit, conversion processing unit, simplification processing unit, construction unit, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize the corresponding functions. The above-mentioned modules are all located in the same processor; or, the above-mentioned modules are located in different processors in any combination.

[0131] The processor contains a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the kernel parameters can be adjusted to solve the problem that the existing technology ignores the lightweight processing of the mesh model reconstructed from the point cloud, resulting in poor operating efficiency of the digital twin scene.

[0132] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0133] An embodiment of the present invention provides a computer-readable storage medium, which includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the method for constructing the industrial digital twin scenario.

[0134] An embodiment of the present invention provides a processor, which is used to run a program, wherein the method for constructing the industrial digital twin scenario is executed when the program is running.

[0135] An embodiment of the present invention provides a device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, at least the steps of the method for constructing the industrial digital twin scenario are implemented.

[0136] The devices in this article can be servers, PCs, PADs, mobile phones, etc.

[0137] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that initializes at least the steps of the method for constructing the industrial digital twin scene.

[0138] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

[0139] 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 adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0140] 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 box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes 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 generate 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.

[0141] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0142] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0143] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0144] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0145] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0146] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0147] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for constructing an industrial digital twin scene, characterized in that: include: Acquire point cloud data obtained by scanning an industrial scene using a multi-site scanning method, and perform splicing processing on the point cloud data using a hybrid splicing method to obtain a three-dimensional point cloud of the industrial scene; The three-dimensional point cloud is converted by using a K-nearest neighbor method to obtain a smooth and continuous point cloud, and the smooth and continuous point cloud is converted into a point cloud mesh model by using a Poisson reconstruction technique; Identify and process the features of the point cloud mesh model, determine the shape type of the point cloud mesh model, and use a QEM algorithm to simplify the point cloud mesh model according to the shape type to obtain a lightweight point cloud model; Using the lightweight point cloud model to build an industrial digital twin scene; Wherein, after converting the three-dimensional point cloud using the K nearest neighbor method to obtain a smooth and continuous point cloud, the method further includes: Determine the spatial domain weight of each point according to the spatial distance between each point and surrounding points in the smooth continuous point cloud, and determine the attribute domain weight of each point according to the similarity of the attribute values ​​of each point and the surrounding points; Convolving the spatial domain weight with the attribute domain weight to obtain a target weight; The attribute values ​​of each point in the smooth and continuous point cloud and the surrounding points are weighted averaged according to the target weight to update the attribute value of each point in the smooth and continuous point cloud to obtain an updated smooth and continuous point cloud.

2. The method according to claim 1, characterized in that Acquiring point cloud data obtained by scanning an industrial scene using a multi-site scanning method, and splicing the point cloud data using a hybrid splicing method to obtain a three-dimensional point cloud of the industrial scene, including: By target stitching, a preset number of public targets of two adjacent sites are aligned and overlapped to spatially stitch the point cloud data to obtain spatially stitched point cloud data; The spatially stitched point cloud data is respectively stitched along a first coordinate axis, a second coordinate axis, and a third coordinate axis to obtain the three-dimensional point cloud of the industrial scene, wherein the first coordinate axis, the second coordinate axis, and the third coordinate axis are perpendicular to each other.

3. The method according to claim 1, characterized in that Before converting the three-dimensional point cloud using the K nearest neighbor method to obtain a smooth and continuous point cloud, the method further includes: The maximum and minimum values ​​along the first coordinate axis, the second coordinate axis, and the third coordinate axis are respectively set to determine the spatial range, and the point clouds that are not within the spatial range are deleted according to the coordinates of each point cloud in the three-dimensional point cloud to obtain a processed three-dimensional point cloud, wherein the first coordinate axis, the second coordinate axis, and the third coordinate axis are perpendicular to each other.

4. The method according to claim 1, characterized in that: After constructing the industrial digital twin scene using the lightweight point cloud model, the method further includes: The Level of Detail technology is adopted to set a multi-level precision display form of the lightweight point cloud model according to different distances between the industrial digital twin scene and the observer.

5. The method according to any one of claims 1 to 4, characterized in that The smooth continuous point cloud is converted into a point cloud mesh model using Poisson reconstruction technology, including: Solving Poisson's equation according to the smooth continuous point cloud to obtain a scalar field; An isosurface is extracted from the scalar field, and the point cloud mesh model is determined according to the isosurface.

6. The method according to any one of claims 1 to 4, characterized in that After simplifying the point cloud mesh model according to the shape type using the QEM algorithm to obtain a lightweight point cloud model, the method further includes: Optimization step: determining the error value of each triangle in the lightweight point cloud model, selecting and improving the triangle with the largest error value; The optimization step is performed multiple times to iteratively optimize the lightweight point cloud model. After performing the optimization step a preset number of times, the optimized lightweight point cloud model is obtained.

7. A monitoring method for an industrial digital twin scene, characterized in that: include: An industrial digital twin scene is constructed by using the method for constructing an industrial digital twin scene as described in any one of claims 1 to 6; The industrial digital twin scenario is used to monitor the industrial site.

8. A device for constructing an industrial digital twin scene, characterized in that: include: A splicing processing unit, used to obtain point cloud data obtained by scanning the industrial scene using a multi-site scanning method, and splice the point cloud data using a hybrid splicing method to obtain a three-dimensional point cloud of the industrial scene; A conversion processing unit, used to convert the three-dimensional point cloud using a K-nearest neighbor method to obtain a smooth and continuous point cloud, and convert the smooth and continuous point cloud into a point cloud mesh model using a Poisson reconstruction technique; A simplification processing unit, used to identify the features of the point cloud mesh model, determine the shape type of the point cloud mesh model, and use a QEM algorithm to simplify the point cloud mesh model according to the shape type to obtain a lightweight point cloud model; A construction unit, used to construct an industrial digital twin scene using the lightweight point cloud model; The device also includes a determination unit, a convolution unit and an update unit. The determination unit is used to determine the spatial domain weight of each point in the smooth and continuous point cloud according to the spatial distance between each point and the surrounding points, and determine the attribute domain weight of each point according to the similarity of the attribute values ​​of each point and the surrounding points; the convolution unit is used to convolve the spatial domain weight with the attribute domain weight to obtain a target weight; the update unit is used to perform weighted averaging of the attribute values ​​of each point in the smooth and continuous point cloud and the surrounding points according to the target weight, so as to update the attribute value of each point in the smooth and continuous point cloud and obtain an updated smooth and continuous point cloud.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is run, the device where the computer-readable storage medium is located is controlled to execute the method for constructing an industrial digital twin scenario as described in any one of claims 1 to 6.

Images