A material mechanics multi-source heterogeneous data synchronous detection method and system

Abstract

The application discloses a material mechanics multi-source heterogeneous data synchronous detection method and system, and belongs to the field of material mechanics testing. The method comprises the following steps: preparing a speckle on the surface of a sample and recording an initial gauge length, installing a universal testing machine and establishing a 3D-DIC coordinate system; synchronously collecting a load-time sequence and a speckle image sequence during a loading process; extracting three-dimensional coordinates of displacement reference points from the images, calculating real space distances and absolute displacements, and constructing an optical measurement displacement-time sequence; identifying multiple characteristic time points of the load and the optical measurement sequence respectively, recording anchor points and extracting corresponding time point sets; calculating time deviations of the two systems according to the time point sets, taking the time of the testing machine as a reference, performing time domain translation and resampling on the optical measurement sequence; performing multiple time point alignment error quantification verification on the resampled data, and outputting a synchronous fusion sequence when the verification is satisfied. The application can effectively eliminate the time phase difference and space perspective error between the joint testing system, and ensures the measurement accuracy of the data.

Description

Technical Field

[0001] This invention relates to the field of materials mechanics testing, specifically to a method and system for synchronous detection of multi-source heterogeneous data in materials mechanics. Background Technology

[0002] With the rapid iteration of high-end equipment manufacturing and new materials industries, aerospace, new energy, and other fields have placed extremely high demands on the mechanical performance evaluation of complex components and cutting-edge alloy materials. Against this backdrop, combining a universal testing machine with digital image correlation (DIC) technology to construct a multi-device coordinated testing system has become the industry's standard paradigm for obtaining the true mechanical response of materials. The universal testing machine primarily provides high-precision global load data, while the DIC system provides full-field displacement and strain data of the specimen surface through machine vision algorithms.

[0003] However, existing joint testing systems face two serious technical bottlenecks at the data cross-fusion level: First, achieving absolute time-domain synchronization of wideband heterogeneous data is extremely difficult. The underlying control unit of the universal testing machine and the industrial camera of the DIC belong to two completely independent hardware architectures. Not only is there clock drift, but the sampling frequency also varies greatly. For example, the testing machine normally operates at a high frequency of 100Hz or even higher, while the DIC camera, limited by image transmission and computing power, usually maintains a low frequency of around 15Hz. Traditional solutions heavily rely on external expensive hardware triggers such as BNC trigger lines, high-speed data acquisition cards, and dedicated synchronization controllers for forced alignment. This hard synchronization method not only significantly increases the manufacturing cost of the testing system, but also often results in incompatible communication protocols between different manufacturers' equipment. More critically, in industrial testing environments with strong electromagnetic interference, hardware triggers are prone to millisecond-level signal delays and jitter, causing the generated stress-strain curves to lag and data misalignment in the critical yielding stage, leading to the failure of calculations for core mechanical parameters such as the elastic modulus.

[0004] Second, three-dimensional spatial calibration is highly sensitive to errors caused by misalignment of the physical axes. When using 3D-DIC to extract axial strain, existing commercial testing methods typically rely on forcibly projecting the spatial displacement vector onto a single coordinate axis (such as the Y-axis) of the DIC global coordinate system. However, in real engineering testing environments, due to camera mounting eccentricity or slight deformation of the testing machine frame under stress, a slight spatial misalignment inevitably exists between the DIC coordinate system established by the stereo calibration of the binocular camera and the actual physical tensile axis of the universal testing machine. This fundamental misalignment will cause the projected displacement component to be smaller than the actual physical tensile displacement, resulting in irreversible measurement errors. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes a method and system for synchronous detection of multi-source heterogeneous data in materials mechanics. This invention effectively eliminates time phase differences and spatial perspective errors between joint testing systems, ensuring the accuracy of data measurement.

[0006] The technical solution adopted in this invention is: A method for synchronous detection of multi-source heterogeneous data in materials mechanics employs a joint testing system, which includes a universal testing machine, a DIC non-contact optical measurement system, and a computer terminal. The universal testing machine is equipped with a high-frequency load sensor. The DIC non-contact optical measurement system includes a binocular industrial camera, which is positioned on one side of the universal testing machine. Both the universal testing machine and the binocular industrial camera are connected to the computer terminal. The method includes the following steps: S1. Create a speckle pattern on the outer surface of the sample to be tested, measure and record the initial gauge length of the displacement reference point in the speckle pattern; mount the sample to be tested on the fixture of the universal testing machine, and establish a 3D-DIC global spatial coordinate system; S2. The universal testing machine is controlled by a computer terminal to load the sample. During the loading process, the universal testing machine is used to collect load-time data sequences in real time. At the same time, a binocular industrial camera is used to collect image sequences of speckle patterns on the surface of the sample to be tested and the corresponding timestamps are recorded. S3. Extract the three-dimensional physical coordinates of the displacement reference points in S1 at each time stamp through the image sequence, and calculate the real spatial distance between the displacement reference points; then combine the initial gauge length to calculate the real spatial absolute displacement sequence, and then construct the optical displacement-time data sequence with timestamps. S4. Perform multi-feature time point identification on the load-time data sequence obtained in S2 and the optical displacement-time data sequence obtained in S3, and record the feature time anchor points; then extract the time point set corresponding to the testing machine sequence and the time point set corresponding to the optical sequence, respectively. S5. Based on the time point set corresponding to the testing machine sequence and the time point set corresponding to the optical measurement sequence obtained in S4, calculate the time deviation between the universal testing machine and the DIC non-contact optical measurement system; using the time axis of the universal testing machine as the reference, perform time-domain translation and resampling calculations on the optical measurement displacement-time data sequence with time deviation. S6. Perform multi-time-point alignment error quantization verification on the data resampled in S5. If the conditions are not met, return to S5 and re-execute the adaptive optimization of the time deviation. If the conditions are met, output the time-dimensional synchronized fusion sequence to complete the multi-source heterogeneous data synchronization detection.

[0007] This invention also provides a material mechanics multi-source heterogeneous data synchronous detection system for implementing the method described above, comprising: The loading execution module is used to control the universal testing machine to execute a variable load path that includes preloading, unloading, and formal loading features; The heterogeneous acquisition module is used to acquire the high-frequency load sequence of the universal testing machine and the low-frequency image sequence of the DIC non-contact optical measurement system, respectively. The spatial coordinate calculation module is used to track the three-dimensional physical coordinates of the speckle and calculate the real spatial absolute displacement sequence unaffected by camera pose based on the three-dimensional Euclidean distance. The feature recognition module is used to extract multiple feature extreme value time anchor points in high-frequency load sequences and real space absolute displacement sequences by using discrete first-order difference and threshold judgment. The time-domain interpolation synchronization module is used to calculate the time deviation based on multiple feature extreme time anchor points, and to complete time difference compensation and upsampling resampling to generate a fused sequence.

[0008] The beneficial technical effects of the present invention are as follows:

[0009] This invention endows traditional testing machines and DIC non-contact optical measurement systems with adaptive and high-precision temporal alignment capabilities through data processing and feature matching methods, namely, three-dimensional Euclidean distance calculation, discrete first-order difference feature anchor point extraction, PCHIP conformal interpolation resampling, and closed-loop adaptive optimization. This eliminates time phase differences and spatial perspective errors between different platforms, ensuring the accuracy of data measurement. From a principle perspective, three-dimensional Euclidean distance calculation completely eliminates spatial geometric projection errors caused by coordinate system misalignment; discrete first-order difference accurately captures the inflection point of the actual physical response as an absolute time reference; and PCHIP interpolation, by unifying the high and low frequency time axes, avoids numerical overshoot at extreme values ​​in conventional algorithms, achieving lossless data fidelity and high-precision spatiotemporal alignment.

[0010] Compared to traditional hardware-triggered testing systems, this invention addresses the communication jitter and signal delay defects inherent in hardware triggering from an algorithmic perspective. By introducing quantitative evaluation constraints such as root mean square synchronization error and maximum synchronization deviation, it rigorously converges the time misalignment between heterogeneous systems to the tolerance limit of the high-frequency sampling period, mathematically eliminating data misalignment in heterogeneous systems and accurately restoring the true constitutive relationship and mechanical response characteristics of materials. This invention can not only serve as a benchmark performance indicator for the research and development of new materials such as aerospace alloys and new energy composite materials, but also provide absolutely reliable underlying physical data support for CAE simulation design of high-end equipment, structural fatigue life prediction, and digital twin models. This invention replaces expensive hardware triggering with a data-driven method, saving equipment costs while ensuring the accuracy of data measurement. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating a method for synchronous detection of multi-source heterogeneous data in materials mechanics according to the present invention. Figure 2 This is a schematic diagram of the combined testing system used in the method of the present invention; wherein: 1 is a universal testing machine; 2 is the test sample; 3 is a binocular industrial camera; Figure 3 This is a schematic diagram of the speckle pattern on the outer surface of the sample to be tested. Figure 4 This is a flowchart of the steps for extracting multiple feature time anchors in the method of the present invention; Figure 5 This is a block diagram of the module logic of the material mechanics multi-source heterogeneous data synchronous detection system of the present invention; Figure 6 This is an overview diagram of the multi-source heterogeneous data processing method of the present invention; wherein: ① Part shows the matching method and effect of optical displacement and machine load curves, P1-P4 represent the starting time point, the pre-loading to unloading time point, the unloading to formal loading time point, and the final fracture time point, respectively; ② Part shows the 3D-DIC calculated displacement field distribution corresponding to different time points, as well as the displacement evolution process of the upper and lower displacement reference points selected in S1; ③ Part shows the reason why the three-dimensional coordinate change leads to the inaccuracy of two-dimensional displacement; ④ Part shows the difference between the displacement calculation method proposed in this invention and the traditional displacement calculation method; Figure 7 This is a diagram illustrating the effect of time-axis synchronization and alignment of heterogeneous data after time-domain shifting and resampling in this invention. Figure 8 This is a schematic diagram illustrating the principle of the PCHIP algorithm for time-domain resampling and time difference compensation of high and low frequency data in the method of this invention. Figure 9 This is the material load-displacement relationship diagram obtained in the present invention. Detailed Implementation

[0012] Currently, the testing of material mechanical properties largely relies on imported DIC synchronous triggering hardware, which results in high equipment procurement costs and large synchronization errors under industrial electromagnetic environments, hindering the localization of high-end material testing equipment. To address this, this invention eliminates camera coordinate system misalignment errors through three-dimensional Euclidean distance and achieves autonomous synchronization of load-displacement data based on feature anchor points and time-domain interpolation. It can be integrated with a universal testing machine to build a low-cost, high-precision intelligent mechanical testing platform, solving the problems of hardware dependence, poor synchronization accuracy, and measurement distortion in existing testing equipment.

[0013] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0014] Example 1: A method for synchronous detection of multi-source heterogeneous data in materials mechanics, which employs a joint testing system, such as... Figure 2As shown, the combined testing system includes a universal testing machine 1, a DIC non-contact optical measurement system, and a computer terminal. The universal testing machine 1 is equipped with a high-frequency load sensor, used to apply load to the test sample 2, or the test piece, and to acquire high-frequency load data sequences in real time. The DIC non-contact optical measurement system includes a binocular industrial camera 3, arranged on one side of the universal testing machine 1, used to synchronously acquire low-frequency image sequences of speckle patterns on the test piece surface. Both the universal testing machine 1 and the binocular industrial camera 3 are connected to the computer terminal. The computer terminal is used to control the loading of the testing machine, receive the acquired data, and execute the data matching method of this invention.

[0015] like Figure 1 As shown, the method of the present invention specifically includes the following steps: S1. Create a speckle pattern on the outer surface of the sample to be tested, such as... Figure 3 As shown. Complete speckle spots near the upper and lower edges of the speckle pattern are selected as the upper displacement reference point and the lower displacement reference point, respectively, and denoted as... , Measure and record the initial gauge length of the specimen. ; in , , , , , These are the initial x, y, and z coordinates of the upper and lower displacement reference points, respectively. The subscript number 1 represents the upper reference point of the specimen, and the subscript number 2 represents the lower reference point of the specimen.

[0016] S2. Install the test sample onto the fixture of the universal testing machine and adjust the position of the test sample so that its stress axis is strictly aligned with the axis of the testing machine.

[0017] S3. Use a calibration board to perform stereo vision calibration of the binocular industrial camera and establish a 3D-DIC global spatial coordinate system. Adjust the placement of the binocular industrial camera to enable it to capture clear full-field speckle images of the test sample.

[0018] S4. Control the universal testing machine through the computer terminal to execute a variable load path or indicative loading path containing characteristic inflection points. The variable load path includes, in sequence, a preloading section, an unloading section, and a formal loading to the fracture section of the specimen, so as to deliberately create the macroscopic inflection point characteristics of "peak-trough" at the physical source.

[0019] S5. During the loading process in S4, a universal testing machine is used to acquire discrete high-frequency load-time data sequences in real time. Its sampling frequency is ; For the time axis of the testing machine, for Load data corresponding to each moment. Low-frequency image sequences of speckle patterns on the test sample surface are simultaneously acquired using a binocular industrial camera, with a sampling frequency of [missing information]. , Record the corresponding timestamp , where the subscript i represents the i-th data point of the signal collected by the testing machine, and j represents the j-th time point of the image collected by the binocular industrial camera.

[0020] S6. Perform 3D-DIC spatial feature point motion behavior calculation on the acquired low-frequency image sequence, track and extract the three-dimensional physical coordinates of the reference point in the 3D-DIC global spatial coordinate system at each time stamp, i.e., the timestamps of the upper displacement reference point and the lower displacement reference point. The three-dimensional physical coordinates below, denoted in chronological order, are as follows: and The subscript j represents the j-th time point.

[0021] S7. Calculate the real-space absolute displacement sequence unaffected by camera pose and coordinate system deflection. For example... Figure 6 As shown in the figure, this diagram provides an overview of the data processing method and error prevention principle of the present invention. Specifically: Figure 6 Figure ① shows the time-domain matching effect of optical displacement and machine load curves, where P1 to P4 represent the four key feature time points extracted. Figure 6 Figure ② shows the 3D-DIC solution displacement field distribution at different time points during the loading process, as well as the displacement evolution trajectories of the selected upper and lower reference points; Figure 6 ③ The surface morphology of the specimen was compared between the uniform deformation stage and the large deformation stage, revealing that large deformation of the material can induce out-of-plane spatial deformation. Figure 6 Section ④ provides a direct comparison of the differences between the two-point distance calculated in this invention based on three-dimensional space and the traditional two-point distance calculated based on two-dimensional space.

[0022] Combination Figure 6 ③ and Figure 6 As can be seen from section ④, in existing technologies, the difference between two points on a single coordinate axis is usually taken directly (e.g., Projecting the image onto a two-dimensional plane as displacement, once the Y-axis of the camera coordinate system has a spatial angle with the actual stretching axis... Or the sample undergoes changes in the thickness direction. Figure 6 The spatial deformation shown in ③ will have its two-dimensional measured value reduced to the actual physical displacement due to the principle of geometric projection. times.

[0023] This invention employs a real spatial distance calculation formula, using the three-dimensional physical coordinates obtained in S6 to calculate the real spatial distance between the upper and lower displacement reference points at various time points. : in, , , , , , These are the x, y, and z coordinates of the j-th time point on the upper and lower displacement reference points, respectively.

[0024] Use the spatial distance of the first frame image as the initial gauge length. The real-space absolute displacement sequence, unaffected by camera pose and coordinate system deflection, is calculated. This allows for the construction of a low-frequency optical displacement-time data sequence with timestamps. This effectively avoids the geometric projection defects caused by the camera's perspective.

[0025] S8. Load-time data sequence obtained from S5 And the optical displacement-time data sequence obtained by S7 Multiple feature time points are identified separately, that is, multiple feature time points are extracted to establish a set of time anchor points.

[0026] Combination Figure 4 and Figure 7 By setting the noise floor threshold And calculate the sliding step size as Discrete first-order difference, identify and record the following four feature anchor time points: Start time: The data value first crosses the noise floor threshold. The moment; Preload to unload moment: the peak moment when the discrete first-order difference changes from positive to negative; The moment when unloading transitions to formal loading: the trough moment when the discrete first-order difference changes from negative to positive; Fracture time: The slope of the load data is less than the preset negative failure threshold. The slope of the displacement data is greater than the preset damage threshold. The sudden moment.

[0027] Subsequently, the time point sets corresponding to the testing machine sequence were extracted. and the set of time points corresponding to the optical measurement sequence. .

[0028] S9, Combination Figure 8 and Figure 9 Based on the set of time points corresponding to the testing machine sequence The set of time points corresponding to the optical measurement sequence Calculate the time deviation between the testing machine and the DIC non-contact optical measurement system. Using the testing machine's time axis Based on this, a piecewise cubic Hermite conformal interpolation algorithm was used to process optical displacement-time data sequences with time deviations. Perform a time-domain translation, with the translation amount being... And upsampling and resampling calculations.

[0029] First, we introduce the time deviation. Global temporal shift compensation is performed on the low-frequency optical measurement time axis to obtain the shifted and aligned optical measurement timestamp: For any time i on the high-frequency time axis of the testing machine Find the translational photometry time interval in which it is located. And calculate the time step s: Finally, construct a cubic Hermite conformal interpolation polynomial and calculate... upsampling shift corresponding to time step : In the formula, The length of the local time interval; and These are the first derivatives at adjacent nodes, constrained by monotonicity constraints. , , , This is a standard cubic Hermitian basis function. This step ensures the smoothness of time difference compensation while avoiding unrealistic numerical overshoot at physical extrema.

[0030] S10. Verify the multiple time anchor alignment error of the data after resampling in S9, and extract the four feature anchor times from the interpolated and shifted photometric sequence. Calculate the anchor point time corresponding to its sequence in the testing machine. Absolute alignment error between : in, These correspond to the start time, the preload to unload time, the unload to formal load time, and the break time, respectively.

[0031] A global quantization alignment error evaluation model is established to calculate the root mean square synchronization error of alignment at four time points. : and maximum synchronization deviation :

[0032] The acceptable standard for system-level quantization alignment is set as follows: the maximum synchronization deviation is less than or equal to the high-frequency sampling period of the universal testing machine, i.e. And the root mean square synchronization error satisfies .

[0033] If the quantization alignment criteria are not met, an exception feedback mechanism is triggered, and the time deviation is re-executed. Adaptive optimization, that is, based on the current computation... As the initial solution, the sampling period of the low-frequency optical measurement system is used. Construct a local search interval for the search radius and set the optimization iteration step size. Within the local search interval, [the algorithm] optimizes [the search]. Perform iterative updates. In each iteration, utilize the updated... The optical displacement sequence was resampled and shifted in the time domain using PCHIP, and the alignment error of the feature anchor points was recalculated. and objective function This continues until the root mean square error meets the aforementioned requirements.

[0034] When the quantization alignment criteria are met, the four feature anchor points are determined to be successfully aligned on the global time axis, and a high-precision fusion sequence with point-to-point correspondence and absolute synchronization in the time dimension is output. .

[0035] S11. Based on the obtained absolutely synchronized fusion sequence Calculate the engineering stress of the sample to be tested: In the formula This represents the initial cross-sectional area. And the engineering strain: ,in This is the initial gauge length.

[0036] Based on the above method, a high-precision stress-strain curve without hysteresis error can be obtained, and the true elastic modulus E and yield strength characteristics can be extracted through linear regression.

[0037] Example 2: like Figure 5 As shown, a material mechanics multi-source heterogeneous data synchronous detection system is used to implement the method described above, including: The loading execution module is used to control the universal testing machine to execute a variable load path that includes preloading, unloading, and formal loading features.

[0038] The heterogeneous acquisition module includes a high-frequency load sequence acquisition unit and a low-frequency speckle image sequence acquisition unit, which are used to acquire the high-frequency load sequence of the universal testing machine and the low-frequency image sequence of the DIC non-contact optical measurement system, respectively.

[0039] The spatial coordinate calculation module is used to track the 3D physical coordinates of the speckle and calculate the true spatial absolute displacement sequence unaffected by camera pose based on 3D Euclidean distance. Specifically, it extracts the 3D coordinates of reference points based on 3D-DIC and calculates the absolute spatial displacement using Euclidean distance.

[0040] The feature recognition module is used to extract multiple feature extreme value time anchor points in high-frequency load sequences and real space absolute displacement sequences by using discrete first-order difference and threshold judgment.

[0041] The time-domain interpolation synchronization module is used to calculate the time deviation based on multiple feature extreme time anchor points and to complete time difference compensation and upsampling resampling. That is, PCHIP conformal interpolation is used to perform resampling fusion to generate a fused sequence.

[0042] For any parts not mentioned above, existing technologies can be adopted or referenced.

[0043] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for synchronous detection of multi-source heterogeneous data in materials mechanics, characterized in that: A joint testing system is adopted, which includes a universal testing machine, a DIC non-contact optical measurement system, and a computer terminal. The universal testing machine is equipped with a high-frequency load sensor. The DIC non-contact optical measurement system includes a binocular industrial camera, which is arranged on one side of the universal testing machine. Both the universal testing machine and the binocular industrial camera are connected to the computer terminal. The method includes the following steps: S1. Create a speckle pattern on the outer surface of the sample to be tested, measure and record the initial gauge length of the displacement reference point in the speckle pattern; mount the sample to be tested on the fixture of the universal testing machine, and establish a 3D-DIC global spatial coordinate system; S2. The universal testing machine is controlled by a computer terminal to load the sample. During the loading process, the universal testing machine is used to collect load-time data sequences in real time. At the same time, a binocular industrial camera is used to collect image sequences of speckle patterns on the surface of the sample to be tested and the corresponding timestamps are recorded. S3. Extract the three-dimensional physical coordinates of the displacement reference points in S1 at each time stamp through the image sequence, and calculate the real spatial distance between the displacement reference points; then combine the initial gauge length to calculate the real spatial absolute displacement sequence, and then construct the optical displacement-time data sequence with timestamps. S4. Perform multi-feature time point identification on the load-time data sequence obtained in S2 and the optical displacement-time data sequence obtained in S3, and record the feature time anchor points; then extract the time point set corresponding to the testing machine sequence and the time point set corresponding to the optical sequence, respectively. S5. Based on the time point set corresponding to the testing machine sequence and the time point set corresponding to the optical measurement sequence obtained in S4, calculate the time deviation between the universal testing machine and the DIC non-contact optical measurement system; using the time axis of the universal testing machine as the reference, perform time-domain translation and resampling calculations on the optical measurement displacement-time data sequence with time deviation. S6. Perform multi-time-point alignment error quantization verification on the data resampled in S5. If the conditions are not met, return to S5 and re-execute the adaptive optimization of the time deviation. If the conditions are met, output the time-dimensional synchronized fusion sequence to complete the multi-source heterogeneous data synchronization detection.

2. The method for synchronous detection of multi-source heterogeneous data in materials mechanics according to claim 1, characterized in that, S1 includes the following steps: S11. Create a speckle pattern on the outer surface of the sample to be tested. Select complete speckle spots at the upper and lower edges of the speckle pattern as the upper displacement reference point and the lower displacement reference point, respectively, and denot them as follows: , Measure and record the initial gauge length. ; in , , , , , These are the initial x, y, and z coordinates of the upper and lower displacement reference points, respectively. S12. Install the test sample onto the fixture of the universal testing machine, and adjust the position of the test sample so that its stress axis is strictly aligned with the axis of the testing machine; S13. Use the calibration board to complete the stereo vision calibration of the binocular industrial camera and establish the 3D-DIC global spatial coordinate system; adjust the placement of the binocular industrial camera so that it can capture clear speckle images of the test sample across the entire field.

3. The method for synchronous detection of multi-source heterogeneous data in materials mechanics according to claim 2, characterized in that, S2 includes the following steps: S21. Control the universal testing machine through a computer terminal to execute a variable load path containing characteristic inflection points. The variable load path includes, in sequence, a preloading section, an unloading section, and a formal loading to the specimen fracture section. S22. During the loading process in S21, a universal testing machine is used to collect discrete load-time data sequences in real time. Its sampling frequency is ; For the time axis of the testing machine, for Load data corresponding to each moment; simultaneously, image sequences of speckle patterns on the surface of the test sample are acquired using a binocular industrial camera, with a sampling frequency of [missing information]. , Record the corresponding timestamp ; where the subscript i represents the i-th data point of the signal collected by the testing machine, and j represents the j-th time point of the image collected by the binocular industrial camera.

4. The method for synchronous detection of multi-source heterogeneous data in materials mechanics according to claim 3, characterized in that, S3 includes the following steps: S31. Perform 3D-DIC spatial feature point motion behavior calculation on the acquired image sequence, and extract the timestamps of the upper and lower displacement reference points. The three-dimensional physical coordinates below, denoted in chronological order, are as follows: and ; S32. Based on the three-dimensional physical coordinates obtained in S31, calculate the true spatial distance between the upper displacement reference point and the lower displacement reference point at each time step. : in, , , , , , These are the x, y, and z coordinates of the j-th time point on the upper and lower displacement reference points, respectively. Use the spatial distance of the first frame image as the initial gauge length. The real-space absolute displacement sequence, unaffected by camera pose and coordinate system deflection, is calculated. This allows for the construction of a time-stamped optical displacement-time data sequence. .

5. The method for synchronous detection of multi-source heterogeneous data in materials mechanics according to claim 4, characterized in that, S4 includes the following steps: For load-time data series and optical displacement-time data series Perform multi-feature time point identification separately; By setting the background noise threshold And calculate the sliding step size as Discrete first-order difference, identify and record the following four feature anchor time points: Start time: Characterized by the value first crossing the noise floor threshold. The moment; Preloading to unloading moment: characterized by the peak moment when the discrete first-order difference changes from positive to negative; The moment when unloading transitions to formal loading is characterized by the trough when the discrete first-order difference changes from negative to positive. Fracture time: characterized by a load data slope less than a preset negative failure threshold. The slope of the displacement data is greater than the preset damage threshold. The sudden moment; Subsequently, the time point sets corresponding to the testing machine sequence were extracted. and the set of time points corresponding to the optical measurement sequence. .

6. The method for synchronous detection of multi-source heterogeneous data in materials mechanics according to claim 5, characterized in that, S5 includes the following steps: Based on the set of time points corresponding to the testing machine sequence The set of time points corresponding to the optical measurement sequence Calculate the time deviation between the testing machine and the DIC non-contact optical measurement system. ; Using the testing machine time axis Based on this, a piecewise cubic Hermite conformal interpolation algorithm was used to process optical displacement-time data sequences with time deviations. Perform a time-domain translation, with the translation amount being... And upsampling and resampling calculations.

7. The method for synchronous detection of multi-source heterogeneous data in materials mechanics according to claim 6, characterized in that, The specific process of time-domain shifting and upsampling resampling is as follows: First, we introduce the time deviation. Global temporal shift compensation is performed on the low-frequency optical measurement time axis to obtain the shifted and aligned optical measurement timestamp: For any time i on the high-frequency time axis of the testing machine Find the translational photometry time interval in which it is located. And calculate the time step s: Finally, construct a cubic Hermite conformal interpolation polynomial and calculate... upsampling shift corresponding to time step : In the formula, The length of the local time interval; and These are the first derivatives at adjacent nodes, constrained by monotonicity constraints. , , , These are standard cubic Hermitian basis functions.

8. The method for synchronous detection of multi-source heterogeneous data in materials mechanics according to claim 7, characterized in that, S6 includes the following steps: The resampled data underwent multi-timepoint alignment error quantization verification, and the times of four feature anchor points in the interpolated and shifted optical measurement sequence were extracted respectively. Calculate the anchor point time corresponding to its sequence in the testing machine. Absolute alignment error between : in, The root mean square synchronization error of the alignment at the four time points—start time, pre-loading to unloading time, unloading to formal loading time, and break time—is calculated, corresponding to the start time, pre-loading to unloading time, unloading to formal loading time, and break time, respectively. : and maximum synchronization deviation : Let the standard for quantization alignment be: the maximum synchronization deviation is less than or equal to the sampling period of the universal testing machine, i.e. And the root mean square synchronization error satisfies ; If the quantization alignment criteria are not met, the time skewness will be re-executed. Adaptive optimization, that is, based on the current computation... As the initial solution, the sampling period of the low-frequency optical measurement system is used. Construct a local search interval for the search radius and set the optimization iteration step size. Within the local search interval, [the algorithm] optimizes [the search]. Perform iterative updates; in each iteration, utilize the updated... The optical displacement sequence was resampled and shifted in the time domain using PCHIP, and the alignment error of the feature anchor points was recalculated. and objective function until the quantization alignment criteria are met; If the quantization alignment criteria are met, the four feature anchor points are considered to be successfully aligned on the global time axis, and a point-to-point, synchronized fused sequence is output in the time dimension. .

9. A method for synchronous detection of multi-source heterogeneous data in materials mechanics according to claim 8, characterized in that, It also includes the following steps: Based on the obtained fusion sequence Calculate the engineering stress of the sample to be tested: In the formula This represents the initial cross-sectional area; And engineering strain: ,in This is the initial gauge length.

10. A material mechanics multi-source heterogeneous data synchronous detection system, used to implement the method as described in any one of claims 1-9, characterized in that, include: The loading execution module is used to control the universal testing machine to execute a variable load path that includes preloading, unloading, and formal loading features; The heterogeneous acquisition module is used to acquire the high-frequency load sequence of the universal testing machine and the low-frequency image sequence of the DIC non-contact optical measurement system, respectively. The spatial coordinate calculation module is used to track the three-dimensional physical coordinates of the speckle and calculate the real spatial absolute displacement sequence unaffected by camera pose based on the three-dimensional Euclidean distance. The feature recognition module is used to extract multiple feature extreme value time anchor points in high-frequency load sequences and real space absolute displacement sequences by using discrete first-order difference and threshold judgment. The time-domain interpolation synchronization module is used to calculate the time deviation based on multiple feature extreme time anchor points, and to complete time difference compensation and upsampling resampling to generate a fused sequence.

Images