Implementation method and device for virtual-real mapping of equipment measuring points
By segmenting and converting the target point cloud input neural network model of the target device, efficient virtual and real mapping of the device measurement points is achieved, solving the problems of low efficiency and low accuracy in traditional methods.
Patent Information
- Application Number
- CN202411978319.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-31
- Publication Date
- 2025-04-08
AI Technical Summary
Traditional methods are inefficient and accurate in the collection and modeling of position information of equipment measurement points, and are prone to loss of measurement points, affecting the accuracy of virtual and real mapping of measurement points.
By obtaining the target point cloud of the target device and inputting the trained neural network model, point cloud segmentation is performed to obtain the sensor point cloud, and then converting the measured point coordinates into the three-dimensional coordinates of the target device twin model to render the measured point.
The efficiency and accuracy of virtual and real mapping of measurement points is improved, and the labor investment is reduced and the error rate is reduced compared with manual annotation.
Smart Images
Figure CN119784953A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of digital twin and data processing technology, and in particular to a method and device for realizing virtual-real mapping of equipment measurement points. Background Art
[0002] In the process of building digital twins of traditional power plant equipment, the main equipment includes boilers, steam turbines, generators, etc., as well as their corresponding auxiliary equipment, and usually needs to collect and mark the location information of equipment measurement points. These measurement points refer to sensor devices installed on the equipment, which are used to monitor the operating status of the equipment in real time, including but not limited to parameters such as temperature and air pressure. Currently, commonly used methods include manual labeling and on-site collection of equipment point clouds, and manual extraction of measurement point location information.
[0003] However, due to the large number of equipment measurement points and the complex structure of the host equipment, manual labeling requires a lot of manpower and is prone to errors; on-site collection of equipment point cloud data requires time and resource investment, and there is a risk of losing measurement points when extracting measurement point location information, which will affect the accuracy of the virtual-reality mapping of the measurement points, and thus affect the effect of the measurement point sensors in the digital twin application of power plants.
[0004] Therefore, it is necessary to propose a more efficient and accurate method for collecting and modeling equipment measurement point location information to address the problems existing in traditional methods. Summary of the invention
[0005] The present disclosure provides a method and a device for realizing virtual-real mapping of equipment measuring points, so as to at least solve the problem of low efficiency and low accuracy in realizing virtual-real mapping of measuring points in the related art.
[0006] The first aspect of the present application proposes a method for implementing virtual-to-real mapping of device measurement points, including: obtaining a target point cloud corresponding to a target device, where at least one sensor is installed on the target device; inputting the target point cloud into a trained neural network model to obtain a point category corresponding to each point in the target point cloud output by the neural network model; performing point cloud segmentation on the target point cloud according to the point category corresponding to each point in the target point cloud to obtain a sensor point cloud obtained after the segmentation; obtaining sensor measurement point coordinates corresponding to the sensor point cloud, and converting the sensor measurement point coordinates into three-dimensional coordinates corresponding to a target device twin model based on conversion parameters, and performing measurement point rendering on the target device twin model based on the three-dimensional coordinates.
[0007] According to one embodiment of the present application, the neural network model adopts the PointNet++ neural network model, and the training process of the PointNet++ neural network model includes: obtaining training point cloud samples, wherein each point in the training point cloud samples carries a category label; iteratively training the initial PointNet++ neural network model according to the training point cloud samples until the training is completed, thereby obtaining a trained PointNet++ neural network model.
[0008] According to one embodiment of the present application, obtaining a target point cloud corresponding to a target device includes: performing point cloud data collection on the target device to obtain an initial point cloud obtained by the collection; filtering and denoising the initial point cloud to obtain a target point cloud obtained after filtering and denoising.
[0009] According to an embodiment of the present application, obtaining the sensor point coordinates corresponding to the sensor point cloud includes: constructing an envelope box based on the sensor point cloud; obtaining the center point coordinates of the envelope box; and using the center point coordinates as the sensor point coordinates corresponding to the sensor point cloud.
[0010] According to an embodiment of the present application, constructing an envelope box based on a sensor point cloud includes: acquiring reference coordinates of an envelope box corresponding to the sensor point cloud; and constructing the envelope box according to the reference coordinates of the envelope box.
[0011] According to one embodiment of the present application, a method for obtaining conversion parameters includes: selecting multiple same-name point pairs of a target point cloud and a target device twin model, wherein each same-name point pair includes a target point cloud data point and a target device twin model data point; and calculating the conversion parameters based on the coordinate positions corresponding to the multiple same-name point pairs according to the seven-parameter conversion method.
[0012] According to one embodiment of the present application, a method for constructing a twin model of a target device includes: collecting basic feature data corresponding to the target device; and modeling based on the basic feature data to obtain a twin model of the target device.
[0013] The second aspect of the present application proposes an implementation device for virtual-to-real mapping of device measurement points, including: a first acquisition module, used to obtain a target point cloud corresponding to a target device, where at least one sensor is installed; a second acquisition module, used to input the target point cloud into a trained neural network model, and obtain the point category corresponding to each point in the target point cloud output by the neural network model; a point cloud segmentation module, used to perform point cloud segmentation on the target point cloud according to the point category corresponding to each point in the target point cloud to obtain a sensor point cloud obtained after the segmentation is completed; a measurement point rendering module, used to obtain the sensor measurement point coordinates corresponding to the sensor point cloud, and convert the sensor measurement point coordinates into three-dimensional coordinates corresponding to a target device twin model based on conversion parameters, and perform measurement point rendering on the target device twin model based on the three-dimensional coordinates.
[0014] According to one embodiment of the present application, the neural network model in the second acquisition module adopts the PointNet++ neural network model, and the implementation device of the virtual-real mapping of the equipment measurement points also includes a model training module for acquiring training point cloud samples, wherein each point in the training point cloud samples carries a category label; the initial PointNet++ neural network model is iteratively trained according to the training point cloud samples until the training is completed to obtain a trained PointNet++ neural network model.
[0015] According to one embodiment of the present application, the first acquisition module is further used to: collect point cloud data from the target device to obtain the collected initial point cloud; filter and reduce noise on the initial point cloud to obtain the target point cloud obtained after filtering and reducing noise.
[0016] According to an embodiment of the present application, the measuring point rendering module is further used to: construct an envelope box based on the sensor point cloud; obtain the center point coordinates of the envelope box; and use the center point coordinates as the sensor measuring point coordinates corresponding to the sensor point cloud.
[0017] According to an embodiment of the present application, the measurement point rendering module is further used to: obtain the envelope box reference coordinates corresponding to the sensor point cloud; and construct the envelope box according to the envelope box reference coordinates.
[0018] According to one embodiment of the present application, the device for implementing the virtual-to-real mapping of equipment measurement points also includes a conversion parameter acquisition module, which is used to: select multiple point pairs with the same name of the target point cloud and the target equipment twin model, wherein each point pair with the same name includes a target point cloud data point and a target equipment twin model data point; and calculate the conversion parameters according to the seven-parameter conversion method in combination with the coordinate positions corresponding to the multiple point pairs with the same name.
[0019] According to one embodiment of the present application, the device for implementing the virtual-real mapping of equipment measurement points also includes a twin model construction module, which is used to collect basic feature data corresponding to the target equipment; modeling is performed based on the basic feature data to obtain a twin model of the target equipment.
[0020] The third aspect embodiment of the present application proposes an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to implement the method for implementing virtual-to-real mapping of device measurement points as described in the first aspect embodiment of the present application.
[0021] The fourth aspect embodiment of the present application proposes a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to implement the method for implementing the virtual-real mapping of device measurement points as described in the first aspect embodiment of the present application.
[0022] The fifth aspect embodiment of the present application proposes a computer program product, including a computer program, which, when executed by a processor, implements the method for implementing the virtual-real mapping of device measurement points as described in the first aspect embodiment of the present application.
[0023] The technical solution provided by the embodiments of the present disclosure brings at least the following beneficial effects: the present application combines the target point cloud corresponding to the target device with the neural network model to realize the virtual-reality mapping of the measurement points. Compared with manual marking of the measurement points on the device twin model, the efficiency of the virtual-reality mapping of the measurement points is improved, and the accuracy of the virtual-reality mapping of the measurement points is improved.
[0024] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:
[0026] Figure 1 It is a schematic diagram of an exemplary implementation method of a method for implementing virtual-real mapping of device measurement points shown in an embodiment of the present application.
[0027] Figure 2 Schematic diagram of a PointNet++ neural network model shown in one embodiment of the present application.
[0028] Figure 3 It is a schematic diagram of an exemplary implementation method of a method for implementing virtual-real mapping of device measurement points shown in an embodiment of the present application.
[0029] Figure 4 The above figure is a rendering of a power plant boiler measuring point (T1~T6) shown in an embodiment of the present application.
[0030] Figure 5 It is a schematic diagram of a device for implementing virtual-real mapping of equipment measurement points shown in an embodiment of the present application.
[0031] Figure 6 It is a schematic diagram of a device for implementing virtual-real mapping of equipment measurement points shown in an embodiment of the present application.
[0032] Figure 7 It is a schematic diagram of an electronic device shown in one embodiment of the present application. DETAILED DESCRIPTION
[0033] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.
[0034] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.) and signals involved in this application are all authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.
[0035] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.
[0036] Figure 1 is a schematic diagram of an exemplary implementation method of a method for implementing virtual-real mapping of a device measurement point shown in the present application, such as Figure 1 As shown, the method for implementing the virtual-real mapping of the device measurement point includes the following steps:
[0037] S101, obtaining a target point cloud corresponding to a target device, where at least one sensor is installed on the target device.
[0038] The target device is a device for which virtual-to-real mapping of device measurement points is to be performed. For example, the target device may be a boiler, a steam turbine, a generator, or other device equipped with measurement point sensors.
[0039] Exemplarily, a laser radar is used to scan a target device, and a target point cloud is acquired by reflecting and receiving a laser beam. Optionally, the target point cloud may be in a laser radar data exchange format (LAS).
[0040] It is not difficult to understand that since the target device is installed with a sensor, the collected target point cloud contains point cloud data corresponding to the sensor in addition to the point cloud data corresponding to the target device, so that the point cloud can be classified in the subsequent process.
[0041] S102, inputting the target point cloud into the trained neural network model, and obtaining the point category corresponding to each point in the target point cloud output by the neural network model.
[0042] In this application, a deep learning algorithm is used to extract features from the target point cloud to obtain the point category corresponding to each point in the target point cloud. The point category is used to characterize whether the point is a point corresponding to the sensor, so the point category can be simply divided into sensor category and other category.
[0043] In some implementations, in order to realize the specific category corresponding to each sensor when mapping the virtual and real measurement points, the point category can also be divided into multiple sensor categories and other categories, for example, it can be divided into temperature sensor category, pressure sensor category, humidity sensor category and other categories.
[0044] Optionally, the neural network model can be PointNet, PointNet++, PointCNN, etc.
[0045] For example, the neural network model uses the PointNet++ neural network model as an example. Figure 2 is a schematic diagram of the PointNet++ neural network model shown in this application, such as Figure 2 As shown in the figure, the PointNet++ neural network model learns the features of hierarchical point sets, and after sampling, grouping, and feature learning processes, it outputs the point category corresponding to each point in the target point cloud.
[0046] The following is a brief introduction to the architecture and implementation of the PointNet++ neural network model:
[0047] The PointNet++ neural network model is a deep learning model for processing point cloud data. It is improved on the basis of PointNet to improve performance and efficiency. The core idea of PointNet++ is to capture the local structure and global structure information of point cloud data through multi-level local and global feature aggregation, thereby improving its representation ability. The PointNet++ neural network model consists of many set abstraction layers. Each set abstraction layer consists of three key layers: sampling layer, grouping layer, and PointNet layer.
[0048] Among them, the sampling layer: the farthest point sampling (FPS) method is used to sample the input points and select N center points. The FPS algorithm randomly selects a point, then selects the point farthest from this point, adds it to the result set, and iterates this process until the number of points in the result set reaches a given value.
[0049] Among them, the grouping layer: within a certain range of the center point extracted by the sampling layer, find the nearest K neighboring points to form a group and generate N local areas; the K values of different groups are different to ensure a fixed regional scale.
[0050] Among them, PointNet layer: N local features are extracted through N local areas.
[0051] Through the synergy of the above three key components, PointNet++ can effectively capture the local and global structural information of point cloud data. The PointNet++ neural network model adopts a layered propagation strategy based on distance interpolation and cross-layer skip links to obtain the point features of all original points in the target point cloud. Among them, the k-nearest neighbor method (KNN) is used to reversely weighted average to achieve feature propagation, and then the point category corresponding to each point in the calibrated point cloud is obtained through a fully connected network.
[0052] The training process of the PointNet++ neural network model includes: obtaining training point cloud samples, wherein each point in the training point cloud samples carries a category label; iteratively training the initial PointNet++ neural network model according to the training point cloud samples until the training is completed, thereby obtaining a trained PointNet++ neural network model.
[0053] S103, performing point cloud segmentation on the target point cloud according to the point category corresponding to each point in the target point cloud to obtain a sensor point cloud obtained after the segmentation is completed.
[0054] For example, if 10 sensors are installed on the target device, 10 sensor point clouds can be obtained after the target point cloud is segmented according to the point category corresponding to each point in the target point cloud.
[0055] S104, obtaining the sensor measurement point coordinates corresponding to the sensor point cloud, and converting the sensor measurement point coordinates into three-dimensional coordinates corresponding to the target device twin model based on the conversion parameters, and rendering the measurement points on the target device twin model based on the three-dimensional coordinates.
[0056] This application uses digital twin technology. A twin model refers to a physical object presented in a digital way in a virtual space, that is, a virtual model is created for a physical object in a digital way, and its behavioral characteristics in a real environment are simulated. Such a twin model can build a virtual environment that is highly similar to the actual system, and analyze and predict the performance and behavior of the actual system by simulating various conditions and parameter changes in it.
[0057] In the present application, a method for constructing a twin model of a target device includes: collecting basic feature data corresponding to the target device; including the target device model, size, design parameters, etc., and modeling based on the basic feature data to obtain a twin model of the target device.
[0058] In the present application, optionally, for any sensor point cloud obtained as described above, coordinates corresponding to a point may be randomly selected from the sensor point cloud as the sensor measurement point coordinates corresponding to the sensor point cloud.
[0059] In the present application, optionally, for any sensor point cloud obtained as described above, an envelope box may be constructed based on the sensor point cloud, and the coordinates of the center point of the envelope box may be used as the coordinates of the sensor measurement point corresponding to the sensor point cloud.
[0060] After obtaining the sensor measurement point coordinates as mentioned above, in this application, the sensor measurement point coordinates are converted into three-dimensional coordinates corresponding to the target device twin model based on the conversion parameters, and the measurement point rendering is performed on the target device twin model based on the three-dimensional coordinates to realize the measurement point mapping in the virtual three-dimensional space (that is, what is ultimately achieved is to mark the sensor on the target device twin model).
[0061] The embodiment of the present application proposes a method for realizing virtual-to-real mapping of device measuring points, by obtaining a target point cloud corresponding to a target device, where at least one sensor is installed; inputting the target point cloud into a trained neural network model, and obtaining the point category corresponding to each point in the target point cloud output by the neural network model; performing point cloud segmentation on the target point cloud according to the point category corresponding to each point in the target point cloud to obtain the sensor point cloud obtained after the segmentation; obtaining the sensor measuring point coordinates corresponding to the sensor point cloud, and converting the sensor measuring point coordinates into three-dimensional coordinates corresponding to the target device twin model based on conversion parameters, and rendering the measuring points on the target device twin model based on the three-dimensional coordinates. The present application combines the target point cloud corresponding to the target device with the neural network model to realize virtual-to-real mapping of measuring points, and improves the efficiency of virtual-to-real mapping of measuring points and the accuracy of virtual-to-real mapping of measuring points compared to manual marking of measuring points on the device twin model.
[0062] Figure 3 is a schematic diagram of an exemplary implementation method of a method for implementing virtual-real mapping of a device measurement point shown in the present application, such as Figure 3 As shown, the method for implementing the virtual-real mapping of the device measurement point includes the following steps:
[0063] S301, collecting point cloud data from a target device to obtain an initial point cloud.
[0064] S302, filtering and denoising the initial point cloud to obtain a target point cloud obtained after filtering and denoising.
[0065] The initial point cloud is filtered and denoised to effectively reduce noise and unnecessary details in the data, thereby improving the efficiency and accuracy of subsequent processing.
[0066] Optionally, common point cloud filtering methods include Gaussian filtering, mean filtering, median filtering, etc. When using the point cloud filtering method, appropriate parameters can be selected to adjust the filtering effect according to the requirements of the application scenario.
[0067] S303, inputting the target point cloud into the trained neural network model, and obtaining the point category corresponding to each point in the target point cloud output by the neural network model.
[0068] S304 , performing point cloud segmentation on the target point cloud according to the point category corresponding to each point in the target point cloud to obtain a sensor point cloud obtained after the segmentation is completed.
[0069] Regarding the specific implementation of steps S303 to S304, reference may be made to the specific introduction of the relevant parts in the above embodiments, which will not be elaborated here.
[0070] S305, constructing an envelope box based on the sensor point cloud.
[0071] For any sensor point cloud obtained above, obtain the reference coordinates of the envelope box corresponding to the sensor point cloud.
[0072] Optionally, in this application, the envelope box reference coordinates are the minimum XYZ coordinates (X min , Y min , Z m in) and maximum XYZ coordinates (Xmax, Ymax, Zmax).
[0073] After the envelope box reference coordinates are obtained as described above, the envelope box is constructed according to the envelope box reference coordinates.
[0074] S306, obtaining the center point coordinates of the envelope box.
[0075] Calculate the coordinates (X, Y, Z) of the center point of the envelope box. The calculation formula is X = (X min +X max ) / 2, Y=(Y min +Y max ) / 2, Z=(Z min +Z max ) / 2.
[0076] S307 , using the center point coordinates as the sensor measurement point coordinates corresponding to the sensor point cloud.
[0077] S308, converting the sensor measurement point coordinates into three-dimensional coordinates corresponding to the target device twin model based on the conversion parameters.
[0078] A method for obtaining conversion parameters includes: selecting multiple same-name point pairs of a target point cloud and a target device twin model, wherein each same-name point pair includes a target point cloud data point and a target device twin model data point; and calculating the conversion parameters based on the coordinate positions corresponding to the multiple same-name point pairs according to a seven-parameter conversion method.
[0079] The transformation parameters include translation parameters, rotation parameters and scale factors.
[0080] Among them, the translation parameter (T x 、T y 、T z ): Represents the translation along the X-axis, Y-axis, and Z-axis respectively, and is used to specify the position offset of one coordinate system relative to another coordinate system.
[0081] Among them, the rotation parameter (ω x ,ω y ,ω z ): Represents the rotation angles around the X-axis, Y-axis, and Z-axis respectively, and is used to specify the rotation transformation of one coordinate system relative to another coordinate system.
[0082] Among them, the scale factor is used to represent the scale transformation of one coordinate system relative to another coordinate system, usually expressed as a proportional factor.
[0083] S309, performing measurement point rendering on the target device twin model based on the three-dimensional coordinates.
[0084] Figure 4 This is a rendering of a power plant boiler measuring point (T1-T6) shown in this application, as shown in the figure above. Figure 4 As shown, measurement point rendering is performed on the target device twin model based on three-dimensional coordinates to achieve virtual-real mapping of device measurement points.
[0085] This application combines the target point cloud corresponding to the target device with the neural network model to realize the virtual-reality mapping of the measurement points. Compared with manual marking of the measurement points on the device twin model, it improves the efficiency of the virtual-reality mapping of the measurement points and improves the accuracy of the virtual-reality mapping of the measurement points.
[0086] Figure 5 is a schematic diagram of a device for implementing virtual-real mapping of a device measurement point shown in this application, such as Figure 5 As shown, the device 500 for implementing virtual-real mapping of the device measuring point includes a first acquisition module 501, a second acquisition module 502, a point cloud segmentation module 503 and a measuring point rendering module 504, wherein:
[0087] The first acquisition module 501 is used to acquire a target point cloud corresponding to a target device, where at least one sensor is installed.
[0088] The second acquisition module 502 is used to input the target point cloud into the trained neural network model, and obtain the point category corresponding to each point in the target point cloud output by the neural network model.
[0089] The point cloud segmentation module 503 is used to perform point cloud segmentation on the target point cloud according to the point category corresponding to each point in the target point cloud to obtain a sensor point cloud obtained after the segmentation is completed.
[0090] The measurement point rendering module 504 is used to obtain the sensor measurement point coordinates corresponding to the sensor point cloud, and convert the sensor measurement point coordinates into three-dimensional coordinates corresponding to the target device twin model based on the conversion parameters, and perform measurement point rendering on the target device twin model based on the three-dimensional coordinates.
[0091] This device combines the target point cloud corresponding to the target device with the neural network model to realize the virtual-reality mapping of the measurement points. Compared with manual marking of the measurement points on the equipment twin model, it improves the efficiency of the virtual-reality mapping of the measurement points and improves the accuracy of the virtual-reality mapping of the measurement points.
[0092] Figure 6 is a schematic diagram of a device for implementing virtual-real mapping of a device measurement point shown in this application, such as Figure 6 As shown, the device 600 for implementing virtual-real mapping of the device measuring point includes a first acquisition module 601, a second acquisition module 602, a point cloud segmentation module 603, a measuring point rendering module 604, a model training module 605, a conversion parameter acquisition module 606 and a twin model construction module 607, wherein:
[0093] The first acquisition module 601 is used to acquire a target point cloud corresponding to a target device, where at least one sensor is installed.
[0094] The second acquisition module 602 is used to input the target point cloud into the trained neural network model, and obtain the point category corresponding to each point in the target point cloud output by the neural network model.
[0095] The point cloud segmentation module 603 is used to perform point cloud segmentation on the target point cloud according to the point category corresponding to each point in the target point cloud to obtain a sensor point cloud obtained after the segmentation is completed.
[0096] The measurement point rendering module 604 is used to obtain the sensor measurement point coordinates corresponding to the sensor point cloud, and convert the sensor measurement point coordinates into three-dimensional coordinates corresponding to the target device twin model based on the conversion parameters, and perform measurement point rendering on the target device twin model based on the three-dimensional coordinates.
[0097] Furthermore, the neural network model in the second acquisition module 602 adopts the PointNet++ neural network model, and the implementation device 600 of the virtual-real mapping of the equipment measurement point also includes a model training module 605, and the model training module 605 is used to: obtain training point cloud samples, wherein each point in the training point cloud samples carries a category label; iteratively train the initial PointNet++ neural network model according to the training point cloud samples until the training is completed, thereby obtaining a trained PointNet++ neural network model.
[0098] Furthermore, the first acquisition module 601 is also used to: collect point cloud data from the target device to obtain an initial point cloud obtained by the collection; filter and reduce noise on the initial point cloud to obtain a target point cloud obtained after filtering and reducing noise.
[0099] Furthermore, the measuring point rendering module 604 is further used to: construct an envelope box based on the sensor point cloud; obtain the center point coordinates of the envelope box; and use the center point coordinates as the sensor measuring point coordinates corresponding to the sensor point cloud.
[0100] Furthermore, the measurement point rendering module 604 is further used to: obtain the reference coordinates of the envelope box corresponding to the sensor point cloud; and construct the envelope box according to the reference coordinates of the envelope box.
[0101] Furthermore, the implementation device 600 of the virtual-to-real mapping of the equipment measurement points also includes a conversion parameter acquisition module 606, which is used to: select multiple point pairs with the same name of the target point cloud and the target equipment twin model, wherein each point pair with the same name includes a target point cloud data point and a target equipment twin model data point; and calculate the conversion parameters according to the seven-parameter conversion method in combination with the coordinate positions corresponding to multiple point pairs with the same name.
[0102] Furthermore, the device 600 for implementing the virtual-real mapping of equipment measurement points also includes a twin model construction module 607, which is used to collect basic feature data corresponding to the target equipment; modeling is performed based on the basic feature data to obtain a twin model of the target equipment.
[0103] This device combines the target point cloud corresponding to the target device with the neural network model to realize the virtual-reality mapping of the measurement points. Compared with manual marking of the measurement points on the equipment twin model, it improves the efficiency of the virtual-reality mapping of the measurement points and improves the accuracy of the virtual-reality mapping of the measurement points.
[0104] In order to implement the above embodiment, the present application embodiment also proposes an electronic device 700, such as Figure 7 As shown, the electronic device 700 includes: a processor 701 and a memory 702 that is communicatively connected to the processor, the memory 702 stores instructions that can be executed by at least one processor, and the instructions are executed by at least one processor 701 to implement the method for implementing virtual-real mapping of device measurement points as shown in the above embodiment.
[0105] In order to implement the above embodiment, the embodiment of the present application also proposes a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to implement the method for implementing the virtual-real mapping of device measurement points as shown in the above embodiment.
[0106] In order to implement the above embodiment, the embodiment of the present application also proposes a computer program product, including a computer program, which, when executed by a processor, implements the method for implementing the virtual-real mapping of device measurement points as shown in the above embodiment.
[0107] In the description of the present application, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present application.
[0108] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.
[0109] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.
[0110] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.
[0111] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.
Claims
1. A method for implementing virtual-real mapping of equipment measurement points, characterized in that: include: Acquire a target point cloud corresponding to a target device, where at least one sensor is installed on the target device; Input the target point cloud into a trained neural network model, and obtain the point category corresponding to each point in the target point cloud output by the neural network model; Performing point cloud segmentation on the target point cloud according to the point category corresponding to each point in the target point cloud to obtain a sensor point cloud obtained after the segmentation; The sensor point coordinates corresponding to the sensor point cloud are obtained, and the sensor point coordinates are converted into three-dimensional coordinates corresponding to the target device twin model based on the conversion parameters, and the point rendering is performed on the target device twin model based on the three-dimensional coordinates.
2. The method according to claim 1, characterized in that The neural network model adopts the PointNet++ neural network model, and the training process of the PointNet++ neural network model includes: Obtain training point cloud samples, where each point in the training point cloud samples carries a category label; The initial PointNet++ neural network model is iteratively trained according to the training point cloud samples until the training is completed to obtain a trained PointNet++ neural network model.
3. The method according to claim 1 or 2, characterized in that: The step of obtaining a target point cloud corresponding to the target device includes: Performing point cloud data collection on the target device to obtain an initial point cloud obtained by the collection; The initial point cloud is filtered and denoised to obtain the target point cloud obtained after filtering and denoising.
4. The method according to claim 3, characterized in that The obtaining the sensor point coordinates corresponding to the sensor point cloud includes: constructing an envelope box based on the sensor point cloud; Obtaining the center point coordinates of the envelope box; The center point coordinates are used as the sensor measurement point coordinates corresponding to the sensor point cloud.
5. The method according to claim 4, characterized in that The constructing an envelope box based on the sensor point cloud comprises: Obtaining reference coordinates of an envelope box corresponding to the sensor point cloud; An envelope box is constructed according to the envelope box reference coordinates.
6. The method according to claim 5, characterized in that The method for obtaining the conversion parameters includes: Select multiple same-name point pairs of the target point cloud and the target device twin model, where each same-name point pair includes a target point cloud data point and a target device twin model data point; According to the seven-parameter conversion method, the conversion parameters are calculated by combining the coordinate positions corresponding to the multiple pairs of points with the same name.
7. The method according to claim 6, characterized in that The method for constructing the target device twin model includes: Collecting basic feature data corresponding to the target device; Modeling is performed according to the basic feature data to obtain the twin model of the target device.
8. A device for implementing virtual-real mapping of equipment measurement points, characterized in that: include: A first acquisition module is used to acquire a target point cloud corresponding to a target device, where at least one sensor is installed; A second acquisition module is used to input the target point cloud into a trained neural network model to obtain a point category corresponding to each point in the target point cloud output by the neural network model; A point cloud segmentation module, used for performing point cloud segmentation on the target point cloud according to the point category corresponding to each point in the target point cloud to obtain a sensor point cloud obtained after the segmentation is completed; A measurement point rendering module is used to obtain the sensor measurement point coordinates corresponding to the sensor point cloud, and convert the sensor measurement point coordinates into three-dimensional coordinates corresponding to the target device twin model based on conversion parameters, and perform measurement point rendering on the target device twin model based on the three-dimensional coordinates.
9. An electronic device, comprising: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.
10. A computer program product comprising a computer program, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 7.