A multi-source fusion structural deformation measurement and calibration method and related device

Abstract

This application discloses a multi-source fusion structural deformation measurement and calibration method and related apparatus, relating to the field of structural deformation measurement and calibration. The method includes acquiring first deformation information after deformation collected by fiber Bragg grating sensors deployed on the test piece; acquiring initial state point cloud information of the test piece before deformation scanned by a binocular vision system; inputting the initial state point cloud information and the first deformation information into a Gaussian process regression model, and then inputting the non-fiber Bragg grating monitoring point information from the initial state point cloud into the Gaussian process regression model to obtain the deformation information corresponding to the non-fiber Bragg grating monitoring points; inputting the multi-source fused deformation information, composed of the first deformation information and the deformation information corresponding to the non-fiber Bragg grating monitoring points, into a trained NDT-CNN-SA network model to obtain calibrated deformation information. This invention enables calibration based on multi-source fusion structural deformation measurement data, improving the accuracy of structural deformation data.

Description

Technical Field

[0001] This application relates to the field of structural deformation measurement and calibration, and in particular to a multi-source fusion structural deformation measurement and calibration method and related apparatus. Background Technology

[0002] With the increasing digitalization of aerospace products, the precision requirements for real-time monitoring of structural deformation data are also rising. High-precision structural deformation data not only ensures the basic performance of aerospace products but also improves assembly accuracy, optimizes assembly processes, and enables the construction of more accurate digital twins. Typical aerospace vehicles, such as satellites and drones, are characterized by long operating times, high workloads, and high speeds. During prolonged operation, their structures are prone to deformation. Prolonged structural deformation can lead to fatigue damage or fracture, affecting product efficiency and accuracy, reducing safety and stability, and even causing irreversible consequences.

[0003] Currently, structural deformation measurement generally employs methods such as laser tracking, fiber optic gratings, and vision measurement systems. Vision measurement systems utilize cameras to acquire images of large components, deriving the pose of the measured image in the vision measurement coordinate system through parallax and model parameters. This allows for the visualization of high-density point clouds to represent the component's surface information. However, the deformation of the measured component can only be obtained through repeated scans and data processing, making real-time observation of the component's surface deformation impossible in terms of time and spatial breadth. Fiber optic gratings offer direct measurement by attaching sensors to the component surface, enabling deformation surface reconstruction. The deformation response at the grating attachment point is the most accurate, but the attachment location lacks precise positional information in the model, and the points calculated through interpolation between the micro-element arcs of the deformation curve differ from the actual deformation of the component. Laser tracking systems offer unique advantages in measuring field coordinate unification due to their high measurement point accuracy. However, given their measurement principle of acquiring only one spatial point coordinate per measurement, they are less efficient for measuring large and complex surface features and do not provide comprehensive data.

[0004] Based on this, in order to solve the problem of low accuracy of fiber optic monitoring point cloud data, structural deformation measurement technology based on multi-source fusion has also become a research hotspot. This method can obtain more accurate and complete point cloud data by adding high-precision optical equipment. However, it is also limited by the error between the fiber bonding position and the theoretical position, as well as the accuracy of the reconstruction algorithm or the accuracy of the added prediction algorithm. This method can obtain the most accurate real-time structural deformation data to date, but the accuracy of the data obtained still cannot match the accuracy of high-precision optical equipment. Summary of the Invention

[0005] The purpose of this application is to provide a multi-source fusion structural deformation measurement and calibration method and related device, which can perform calibration based on multi-source fusion structural deformation measurement data and improve the accuracy of structural deformation data.

[0006] To achieve the above objectives, this application provides the following solution:

[0007] Firstly, this application provides a multi-source fusion structural deformation measurement and calibration method, including:

[0008] Construct a global coordinate system;

[0009] The first deformation information after deformation is acquired by a fiber Bragg grating sensor deployed on the test piece; the fiber Bragg grating sensor is deployed on the test piece based on the global coordinate system.

[0010] The initial point cloud information of the test part before deformation is obtained by scanning with a binocular vision system; the binocular vision system is deployed based on the global coordinate system.

[0011] After inputting the initial state point cloud information and the first deformation information into the Gaussian process regression model, the non-fiber grating monitoring point information in the initial state point cloud is input into the Gaussian process regression model to obtain the deformation information corresponding to the non-fiber grating monitoring points; the first deformation information and the deformation information corresponding to the non-fiber grating monitoring points constitute multi-source fused deformation information;

[0012] The multi-source fused deformation information is input into the trained NDT-CNN-SA network model to obtain calibrated deformation information; the NDT-CNN-SA network model includes an NDT module, a CNN model, a self-attention mechanism module, and a perceptron connected in series.

[0013] Secondly, this application provides a multi-source fusion structural deformation measurement and calibration method apparatus, comprising:

[0014] The coordinate system construction module is used to construct the global coordinate system;

[0015] The first data acquisition module is used to acquire the first deformation information after deformation collected by the fiber optic grating sensor deployed on the test piece; the fiber optic grating sensor is deployed on the test piece based on the global coordinate system.

[0016] The second data acquisition module is used to acquire the initial state point cloud information of the test piece before deformation, scanned by the binocular vision system; the binocular vision system is deployed based on the global coordinate system;

[0017] The Gaussian regression analysis module is used to input the initial state point cloud information and the first deformation information into the Gaussian process regression model, and then input the non-fiber grating monitoring point information in the initial state point cloud into the Gaussian process regression model to obtain the deformation information corresponding to the non-fiber grating monitoring points; the first deformation information and the deformation information corresponding to the non-fiber grating monitoring points constitute multi-source fused deformation information;

[0018] The deformation calibration module is used to input the multi-source fused deformation information into the trained NDT-CNN-SA network model to obtain calibrated deformation information; the NDT-CNN-SA network model includes an NDT module, a CNN model, a self-attention mechanism module, and a perceptron connected in series.

[0019] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described multi-source fusion structural deformation measurement and calibration method.

[0020] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described multi-source fusion structural deformation measurement and calibration method.

[0021] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the above-described multi-source fusion structural deformation measurement and calibration method.

[0022] According to the specific embodiments provided in this application, the following technical effects are disclosed:

[0023] This application provides a multi-source fusion structural deformation measurement calibration method and related device. In order to improve the accuracy of multi-source fusion real-time structural deformation data and fill the technical gap of no calibration method for multi-source fusion deformation measurement, a neural network framework based on NDT-CNN-SA is proposed. The high-precision point cloud data obtained by binocular scanner is used to calibrate the multi-source fusion deformation measurement structure, improve data fidelity, and realize higher precision multi-source fusion deformation measurement and monitoring. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1This is an application environment diagram of a multi-source fusion structural deformation measurement and calibration method according to an embodiment of this application;

[0026] Figure 2 A schematic flowchart of a multi-source fusion structural deformation measurement and calibration method provided in an embodiment of this application;

[0027] Figure 3 This is a schematic diagram of the technical route of a multi-source fusion structural deformation measurement and calibration method provided in an embodiment of this application;

[0028] Figure 4 This is a schematic diagram of the structure of an NDT-CNN-SA network model provided in an embodiment of this application;

[0029] Figure 5 This is a schematic diagram of the structure of an NDT module provided in an embodiment of this application;

[0030] Figure 6 This is a schematic diagram of the structure of a CNN module provided in an embodiment of this application;

[0031] Figure 7 This is a schematic diagram of the structure of a self-attention mechanism module provided in an embodiment of this application;

[0032] Figure 8 This is a schematic diagram of the structure of a sensor provided in one embodiment of this application;

[0033] Figure 9 This is a schematic diagram of the functional modules of a multi-source fusion structural deformation measurement and calibration device provided in another embodiment of this application.

[0034] Figure 10 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0035] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0036] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0037] The multi-source fusion structural deformation measurement and calibration method provided in this application is specifically a multi-source fusion structural deformation measurement and calibration method based on an NDT-CNN-SA network, which can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be set up independently, integrated into server 104, or placed in the cloud or on another server. Terminal 102 can send global coordinate system information, the first deformation information collected by fiber optic sensors deployed on the test piece after deformation, and the initial state point cloud information of the test piece before deformation scanned by a binocular vision system to server 104. After receiving the relevant data, server 104 inputs the initial state point cloud information and the first deformation information into a Gaussian process regression model, and then inputs the non-fiber optic monitoring point information from the initial state point cloud into the Gaussian process regression model to obtain the deformation information corresponding to the non-fiber optic monitoring points. The first deformation information and the deformation information corresponding to the non-fiber optic monitoring points constitute multi-source fused deformation information. The multi-source fused deformation information is input into the trained NDT-CNN-SA network model to obtain calibrated deformation information. Server 104 can feed back the calibrated deformation information to terminal 102. Furthermore, in some embodiments, the multi-source fusion structural deformation measurement calibration method can also be implemented separately by server 104 or terminal 102. For example, terminal 102 can directly perform deformation information calibration processing on the global coordinate system information, the first deformation information collected by the fiber optic grating sensors deployed on the test piece, and the initial state point cloud information of the test piece before deformation scanned by the binocular vision system. Alternatively, server 104 can obtain the global coordinate system information, the first deformation information collected by the fiber optic grating sensors deployed on the test piece, and the initial state point cloud information of the test piece before deformation scanned by the binocular vision system from the data storage system, and perform deformation information calibration processing.

[0038] The terminal 102 can be, but is not limited to, various desktop computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, and smart in-vehicle devices. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted devices. The server 104 can be implemented using a standalone server or a server cluster composed of multiple servers, or it can be a cloud server.

[0039] In one exemplary embodiment, such as Figure 2 and Figure 3 As shown, a multi-source fusion structural deformation measurement and calibration method is provided. This method is executed by a computer device, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps 201 to 205.

[0040] in:

[0041] Step 201: Construct a global coordinate system.

[0042] To achieve synchronous collaboration between various digital devices and tooling, a laser tracker is needed to establish accurate relative positional relationships between the values ​​of each digital device and tooling. If the area of ​​deformation measured on the component is relatively small, a global coordinate system can be established without using a laser tracker.

[0043] Step 202: Obtain the first deformation information after deformation collected by the fiber optic grating sensor deployed on the test piece; the fiber optic grating sensor is deployed on the test piece based on the global coordinate system.

[0044] The fiber optic positions are pasted according to the theoretical design scheme based on the global coordinate system. At this time, the initial coordinates of the fiber optic monitoring position points are known.

[0045] During the monitoring process, the fiber optic sensor directly contacts the surface of the workpiece to obtain deformation information in real time. The deformation information obtained by the fiber optic sensor is represented by the change in wavelength, and then converted into coordinate form.

[0046] Step 203: Obtain the initial point cloud information of the test piece before deformation, scanned by the binocular vision system; the binocular vision system is deployed based on the global coordinate system.

[0047] Based on the global coordinate system, a binocular scanning point cloud is used to perform visual scanning of the product in its initial state and obtain the initial state point cloud information of the test part.

[0048] Step 204: After inputting the initial state point cloud information and the first deformation information into the Gaussian process regression model, input the non-fiber grating monitoring point information in the initial state point cloud into the Gaussian process regression model to obtain the deformation information corresponding to the non-fiber grating monitoring points; the first deformation information and the deformation information corresponding to the non-fiber grating monitoring points constitute multi-source fused deformation information.

[0049] Based on data with a unified data source format, Gaussian process regression analysis is used to fuse low-precision deformation data from fiber optic grating monitoring and initial state point cloud data from binocular scanning. Gaussian process regression training is performed using the coordinates of points before and after deformation at the fiber optic grating monitoring location to determine hyperparameters and the Gaussian process model as prior input. Then, fusion prediction is performed on the undeformed points (points without fiber optic grating sensors) in the binocular scanning zero-load state (initial state before deformation). Different kernel functions are selected to obtain the deformation data of points without fiber optic grating sensors.

[0050] Step 205: Input the multi-source fused deformation information into the trained NDT-CNN-SA network model to obtain the calibrated deformation information; such as Figure 4 As shown, the NDT-CNN-SA network model body includes an NDT module, a CNN model, a self-attention mechanism module, and a perceptron connected in series.

[0051] By inputting the multi-source fused deformation information from Gaussian process regression into the trained NDT-CNN-SA network model, the model can output real-time calibration with an accuracy close to that of binocular visual scanning measurement point clouds, thereby achieving multi-source fused deformation measurement calibration.

[0052] The NDT module divides the multi-source fused deformed point cloud into grids of a specified size according to their spatial distribution, and calculates the multidimensional normal distribution parameters for each grid. The normal distribution N(u, C) of each NDT cell consists of an average vector and a covariance matrix C, defined as follows:

[0053]

[0054] Where x k =1, ..., n are the three-dimensional points in each cell, i.e., each point in the multi-source fused deformed point cloud. For example... Figure 5 The diagram illustrates the visualization of the NDT principle, which involves meshing a multi-source fused deformable point cloud and calculating the mean matrix and covariance matrix of the points within the mesh.

[0055] like Figure 6 As shown, the CNN module is a feature extraction module. The mean vector and covariance matrix data obtained from the NDT module are input into this module. This module uses three one-dimensional convolutional layers to extract features from the point cloud. To avoid offsets and biases, a batch normalization layer (BN) is added to normalize the output of the one-dimensional convolutional layers, making the outputs have a similar distribution. Finally, an activation function is used for non-linear transformation, and the normalized results are processed to retain positive values, enhancing the network's non-linear expressive power. Figure 6 As shown, the CNN model includes multiple CNN units connected in series; each CNN unit includes a first convolutional layer, a first batch normalization layer and a first activation layer connected in series.

[0056] The self-attention mechanism is a deep learning mechanism for processing sequential data. Its core idea is to calculate the correlation between each location and other locations, thereby generating a weight for each location. This weight adaptively adjusts the importance of each input feature, thus improving the overall performance of the network. For example... Figure 7 A schematic diagram of the self-attention mechanism module is shown. Figure 7 In the middle, Fp This represents the feature vector output by the CNN feature extraction module; F represents p After a transpose transformation along a specified dimension, during forward propagation, the attention scores, used to represent the importance of each sequence, are... Calculated weight init and bias init This represents the initial weights and biases. Figure 7 In the diagram, `weight` and `bias` without subscripts represent the weights and biases, respectively. Softmax is applied to normalize the attention scores, yielding the attention weights. This involves combining the elements of the feature vector within the range [0,1] along a specified dimension into a single value of 1. The calculation formula is as follows: Matmul is matrix multiplication.

[0057] Multilayer perceptrons (MLPs) apply a series of linear and nonlinear transformations to the input features, mapping the output of the layer to the desired output size. This allows them to learn more complex and abstract representations in the feature space of the input data, resulting in better predictive performance. For example... Figure 8 As shown, the perceptron includes multiple sensing units connected in series, a third convolutional layer, and a splicing layer; each sensing unit includes a second convolutional layer, a second batch normalization layer, and a second activation layer connected in series.

[0058] Implementing steps 201 to 205 above provides a calibration method for the multi-source fusion structural deformation monitoring system, filling the gap in the lack of calibration methods for multi-source fusion structural deformation measurement. Compared with the previous use of multi-source fusion systems alone, this calibration method compares the calibrated deformation point cloud with the actual scanned deformation point cloud. Under different loads, the standard deviation of the surface is below 0.1, and the proportion of point clouds with deviation values ​​within the ±0.1 confidence interval can reach 97%. Its calibration accuracy is higher than that of multi-source fusion measurement.

[0059] In another exemplary embodiment of this application, the present invention incorporates an NDT-CNN-SA network model after the Gaussian process regression model. This network is a single-input, single-output calibration network. The data input to the constructed network includes a large amount of multi-source fused deformation sample information with different deformation amounts and deformed point clouds monitored by a binocular vision scanning device after deformation (deformation monitoring accuracy is very high). Both are simultaneous. The multi-source fused deformation sample information has relatively low accuracy and is used as the point cloud to be calibrated. The point cloud information measured by the binocular vision scanning device is used as the high-precision labeled point cloud. The network uses these two types of data as input and output data to train the constructed neural network. Therefore, as... Figure 3As shown, before performing the step "inputting the multi-source fused deformation information into the trained NDT-CNN-SA network model", the following steps are included:

[0060] (1) Obtain multi-source fusion deformation sample information and corresponding scanned deformation sample information of the tested component after deformation obtained by scanning with a binocular vision system.

[0061] (2) The NDT-CNN-SA network model is trained using multi-source fusion deformation sample information as input and scanned deformation sample information as label.

[0062] (3) When the model loss error converges or the number of training iterations reaches the preset number of iterations, the trained NDT-CNN-SA network model is obtained.

[0063] This application also provides an application scenario in which the above-mentioned multi-source fusion structural deformation measurement and calibration method is applied. Specifically, the multi-source fusion structural deformation measurement and calibration method provided in this embodiment can be applied in the digital assembly process. As the state of the assembly process changes, the product will deform and needs to be monitored. In order to improve the accuracy of monitoring, the monitoring is calibrated, especially for large composite material components. This invention is mainly applied in scenarios where composite material components are prone to deformation during the assembly process.

[0064] Based on the same inventive concept, this application also provides a multi-source fusion structural deformation measurement and calibration device for implementing the multi-source fusion structural deformation measurement and calibration method described above. The solution provided by this device is similar to the implementation described in the above method. Therefore, the specific limitations in one or more embodiments of the multi-source fusion structural deformation measurement and calibration device provided below can be found in the limitations of the multi-source fusion structural deformation measurement and calibration method described above, and will not be repeated here.

[0065] In one exemplary embodiment, such as Figure 9 As shown, a multi-source fusion structural deformation measurement and calibration device is provided, comprising:

[0066] The coordinate system construction module M1 is used to construct the global coordinate system.

[0067] The first data acquisition module M2 is used to acquire the first deformation information after deformation collected by the fiber optic grating sensor deployed on the test piece; the fiber optic grating sensor is deployed on the test piece based on the global coordinate system.

[0068] The second data acquisition module M3 is used to acquire the initial state point cloud information of the test piece before deformation, which is scanned by the binocular vision system; the binocular vision system is deployed based on the global coordinate system.

[0069] The Gaussian regression analysis module M4 is used to input the initial state point cloud information and the first deformation information into the Gaussian process regression model, and then input the non-fiber grating monitoring point information in the initial state point cloud into the Gaussian process regression model to obtain the deformation information corresponding to the non-fiber grating monitoring points; the first deformation information and the deformation information corresponding to the non-fiber grating monitoring points constitute multi-source fused deformation information.

[0070] The deformation calibration module M5 is used to input the multi-source fused deformation information into the trained NDT-CNN-SA network model to obtain the calibrated deformation information; the NDT-CNN-SA network model includes an NDT module, a CNN model, a self-attention mechanism module, and a perceptron connected in series.

[0071] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 10 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores multi-source fusion structural deformation measurement and calibration data. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a multi-source fusion structural deformation measurement and calibration method.

[0072] Those skilled in the art will understand that Figure 10 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0073] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0074] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0075] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0076] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0077] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0078] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0079] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0080] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A multi-source fusion structural deformation measurement and calibration method, characterized in that, include: Construct a global coordinate system; Acquire the first deformation information after deformation collected by the fiber Bragg grating sensor deployed on the test piece; The fiber optic grating sensor is deployed on the device under test based on the global coordinate system; Acquire the initial point cloud information of the test part before deformation, scanned by a binocular vision system; The binocular vision system is deployed based on the global coordinate system; After inputting the initial state point cloud information and the first deformation information into the Gaussian process regression model, the non-fiber grating monitoring point information in the initial state point cloud is input into the Gaussian process regression model to obtain the deformation information corresponding to the non-fiber grating monitoring points. The first deformation information and the deformation information corresponding to the non-fiber grating monitoring point constitute multi-source fused deformation information; The multi-source fused deformation information is input into the trained NDT-CNN-SA network model to obtain calibrated deformation information; the NDT-CNN-SA network model includes a cascaded NDT module, a CNN model, a self-attention mechanism module, and a perceptron. The output of the NDT module is a mean vector and a covariance matrix; the expressions for the mean vector and the covariance matrix are as follows: in, Let represent the mean vector and covariance matrix of an NDT unit, respectively; , ..., , for each three-dimensional point in the cell; The CNN model includes multiple CNN units connected in series; each CNN unit includes a first convolutional layer, a first batch normalization layer and a first activation layer connected in series. The perceptron includes multiple sensing units connected in series, a third convolutional layer, and a splicing layer; each sensing unit includes a second convolutional layer, a second batch normalization layer, and a second activation layer connected in series.

2. The multi-source fusion structural deformation measurement and calibration method according to claim 1, characterized in that, Before performing the step "inputting the multi-source fused deformation information into the trained NDT-CNN-SA network model", the following steps are included: Acquire multi-source fusion deformation sample information and corresponding scanned deformation sample information of the tested component after deformation obtained by binocular vision system scanning; The NDT-CNN-SA network model is trained using multi-source fused deformation sample information as input and scanned deformation sample information as labels. When the model loss error converges or the number of training iterations reaches the preset number of iterations, the trained NDT-CNN-SA network model is obtained.

3. A multi-source fusion structural deformation measurement and calibration device, characterized in that, include: The coordinate system construction module is used to construct the global coordinate system; The first data acquisition module is used to acquire the first deformation information after deformation collected by the fiber optic grating sensor deployed on the test piece. The fiber optic grating sensor is deployed on the device under test based on the global coordinate system; The second data acquisition module is used to acquire the initial state point cloud information of the test piece before deformation, which is scanned by the binocular vision system. The binocular vision system is deployed based on the global coordinate system; The Gaussian regression analysis module is used to input the initial state point cloud information and the first deformation information into the Gaussian process regression model, and then input the non-fiber grating monitoring point information in the initial state point cloud into the Gaussian process regression model to obtain the deformation information corresponding to the non-fiber grating monitoring points. The first deformation information and the deformation information corresponding to the non-fiber grating monitoring point constitute multi-source fused deformation information; The deformation calibration module is used to input the multi-source fused deformation information into the trained NDT-CNN-SA network model to obtain the calibrated deformation information; the NDT-CNN-SA network model includes an NDT module, a CNN model, a self-attention mechanism module, and a perceptron connected in series. The output of the NDT module is a mean vector and a covariance matrix; the expressions for the mean vector and the covariance matrix are as follows: in, Let represent the mean vector and covariance matrix of an NDT unit, respectively; , ..., , for each three-dimensional point in the cell; The CNN model includes multiple CNN units connected in series; each CNN unit includes a first convolutional layer, a first batch normalization layer and a first activation layer connected in series. The perceptron includes multiple sensing units connected in series, a third convolutional layer, and a splicing layer; each sensing unit includes a second convolutional layer, a second batch normalization layer, and a second activation layer connected in series.

4. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the multi-source fusion structural deformation measurement and calibration method according to any one of claims 1-2.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the multi-source fusion structural deformation measurement and calibration method as described in any one of claims 1-2.

6. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the multi-source fusion structural deformation measurement and calibration method as described in any one of claims 1-2.

Images