Digital image correlation measurement method and system combining single-pixel imaging and polarization suppression

Abstract

The application discloses a kind of digital image correlation measurement method and system combined with single-pixel imaging and polarization suppression, belong to optical measurement and experimental mechanics field.The present application constructs the imaging module including light source, orthogonal linear polarizer, digital micromirror array and single-pixel detector for the problem that traditional DIC method is difficult to effectively suppress interference light under complex illumination.Measuring, Hadamard coding mode is projected to sample surface by digital micromirror array, and single-pixel detector is synchronously collected after the light signal filtered by orthogonal polarizer, and differential acquisition is carried out to positive and negative mode;After gray scale image is reconstructed using differential signal, PINN-DIC neural network is input to calculate the full-field displacement field and strain field of sample surface.The present application effectively suppresses interference light at image acquisition source by the synergistic effect of orthogonal polarization and single-pixel differential acquisition, significantly improves the anti-interference ability, precision and stability of DIC measurement under complex illumination environment.

Description

Technical Field

[0001] This invention belongs to the field of optical measurement and experimental mechanics technology, specifically relating to a digital image correlation measurement method and system that combines single-pixel imaging and polarization suppression. It is suitable for high-precision measurement of the full-field displacement and strain of material or structural surfaces under complex lighting environments such as strong light and stray light. Background Technology

[0002] Digital Image Correlation (DIC) is a non-contact, full-field measurement technique based on image grayscale correlation, enabling quantitative measurement of displacement and strain fields on the surface of materials or structures. This method typically involves preparing random speckle patterns on the sample surface and acquiring corresponding digital image sequences before and after loading. Correlation algorithms are then used to match and analyze the grayscale changes in the speckle regions of the images, thereby retrieving deformation information of the sample under different loading states. Due to its advantages such as high measurement accuracy, strong adaptability, and non-contact operation, DIC technology has been widely applied in fields such as materials mechanics, structural engineering, biomechanics, and experimental mechanics.

[0003] In existing technologies, DIC systems typically rely on camera imaging systems based on CCD or CMOS sensors to acquire images. These imaging systems use arrayed photodetectors to perform pixel-by-pixel integration imaging of incident light. Under stable lighting conditions and with minimal environmental interference, they can obtain speckle images with good grayscale contrast and spatial resolution, thus meeting the basic requirements of digital image correlation algorithms for input image quality. However, in practical engineering tests and complex experimental environments, imaging conditions are often difficult to maintain ideally. The presence of ambient light, stray light, and strong external directional light sources can easily interfere with the imaging process.

[0004] When the sample surface is exposed to strong light or there are additional interfering light sources, the interfering light enters the camera imaging system through reflection or scattering, and superimposes on the same imaging plane with the effective speckle reflection light from the sample surface. Because the camera imaging system has a uniform response to all incident light signals, it cannot distinguish between the target light signal and the interfering light signal. The interfering light directly causes local gray-level saturation, decreased contrast, or gray-level distribution distortion in the image, thereby destroying the randomness and stability of the speckle pattern. These problems significantly reduce the correlation between images, thus affecting the accuracy of the DIC algorithm in inverting displacement and strain. In severe cases, it can even lead to the failure of correlation calculations, making it difficult to obtain effective full-field measurement results.

[0005] To address the impact of optical interference on the accuracy of DIC measurements, various improvement schemes have been proposed in existing technologies. For example, filtering, denoising, or enhancing the acquired image can reduce the influence of interfering light on the image's grayscale distribution; or introducing optical devices such as polarizers and filters into the imaging optical path can suppress specific types of interfering light to a certain extent. However, these methods are mostly compensation or correction measures after image acquisition. Interfering light has already caused irreversible damage to image quality during the imaging stage, making it difficult for subsequent processing to fully recover speckle characteristic information. Furthermore, when the sample surface is relatively rough or has significant volume scattering characteristics, the interfering light often exhibits partial depolarization characteristics after scattering, making it difficult to effectively isolate the interfering light simply by relying on polarization suppression.

[0006] In the prior art, the closest implementation to this invention is a camera-based digital image correlation measurement system. This system acquires surface images of a sample under different loading states using a camera and performs correlation calculations on the images using the DIC algorithm to obtain the displacement and strain fields of the sample surface. In some applications, this system can also be combined with polarization suppression or illumination condition control to improve image quality and measurement stability. However, this type of system still relies on an array-based imaging mechanism. During the acquisition process, the imaging system cannot distinguish between modulated light signals and unmodulated interference light. When the intensity of the interference light is high or the spatial distribution is uneven, it is still difficult to avoid the interference light adversely affecting the image quality and correlation calculation results, thus limiting its application in complex illumination environments. Chinese patent CN120707663A discloses a single-pixel imaging reconstruction method based on compressed sensing, but it does not involve DIC measurement or consider its combination with polarization technology.

[0007] Therefore, existing camera-based DIC measurement schemes still suffer from limited imaging anti-interference capabilities and insufficient measurement stability under complex optical interference conditions, and urgently need further improvement to meet the practical needs of high-precision full-field displacement and strain measurement in complex optical environments. Summary of the Invention

[0008] To address the aforementioned technical problems, this invention provides a digital image correlation measurement method and system that combines single-pixel imaging and polarization suppression, aiming to achieve stable and high-precision full-field displacement and strain measurement under complex lighting conditions.

[0009] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: In a first aspect, the present invention provides a digital image correlation measurement method combining single-pixel imaging and polarization suppression, comprising the following steps: Step S1: Construct a measurement system, which includes an optical imaging module, a mechanical loading module, and a data processing module; the optical imaging module includes a light source, a first linear polarizer, a digital micromirror array, a second linear polarizer, and a single-pixel detector; the first linear polarizer is placed in front of the light source, and the second linear polarizer is placed in front of the single-pixel detector, with the transmission axis of the second linear polarizer being orthogonal to the transmission axis of the first linear polarizer; Step S2: Control the digital micromirror array to generate a preset coding pattern sequence and project it onto the surface of the test sample in sequence; The single-pixel detector synchronously collects the light signal intensity after being reflected from the sample surface and transmitted through the second linear polarizer. For each coding pattern, the light signals corresponding to its positive mode and negative mode are collected respectively to obtain a pair of differential signals. Step S3: Using the differential signal obtained in step S2, combined with the preset encoding mode sequence, the grayscale image of the sample surface is calculated by a single-pixel imaging reconstruction algorithm. Step S4: Input the grayscale image reconstructed in step S3 into the pre-constructed PINN-DIC network. The network takes the image pixel coordinates as input and the displacement field as output, and fits the displacement field function through a neural network. After the network training converges, the full-field displacement distribution of the sample surface is obtained from the output layer, and then the corresponding strain field is calculated.

[0010] Furthermore, the encoding mode described in step S2 is the Hadamard encoding mode.

[0011] Furthermore, the order of the Hadamard encoding pattern is 256×256 pixels.

[0012] Furthermore, the differential signal acquisition and processing process described in step S2 is as follows:

[0013] in For the k-th projection mode, For a valid signal, To interfere with the light component.

[0014] Furthermore, the PINN-DIC network described in step S4 includes an input layer, a hidden layer, and an output layer; the input layer has 2 neurons corresponding to pixel coordinates (x, y); the output layer has 2 neurons corresponding to displacement components (u, v); the hidden layer has 4 layers, each containing 50 neurons, and the activation function is the tanh function.

[0015] Furthermore, the loss function of the PINN-DIC network described in step S4 is:

[0016] in The predicted deformed image is obtained by interpolating and resampling the reference image based on the displacement field predicted by the network. These are the deformed images that were actually captured.

[0017] Another aspect of the present invention provides a digital image correlation measurement system that combines single-pixel imaging and polarization suppression to implement the above-described method, comprising: Optical imaging module, including: A light source, used to produce a beam of light; A first linear polarizer is disposed in the light output path of the light source to modulate the illumination light into linearly polarized light; A digital micromirror array is used to receive linearly polarized light modulated by a first linear polarizer, and modulate it into a structured light field according to a preset encoding mode before projecting it onto the surface of the test sample. The second linear polarizer is placed in the optical path for receiving optical signals. Its transmission axis is orthogonal to that of the first linear polarizer. It is used to filter out components in the incident light that maintain the same polarization state as the illumination light. A single-pixel detector, located at the rear end of the second linear polarizer, is used to receive the optical signal filtered by the second linear polarizer and convert it into an electrical signal; The mechanical loading module is used to apply a controllable mechanical load to the test sample; The data processing module includes: The signal acquisition and control unit is connected to the digital micromirror array and the single-pixel detector, and is used to control the digital micromirror array to generate a preset encoding pattern sequence and synchronously acquire the electrical signal output by the single-pixel detector. The image reconstruction unit is used to reconstruct a grayscale image of the sample surface based on the acquired signal and a preset encoding mode using a single-pixel imaging reconstruction algorithm. The deformation calculation unit, with a built-in PINN-DIC network, is used to input the image sequence output by the image reconstruction unit into the network to calculate the full-field displacement field and strain field of the sample surface.

[0018] Furthermore, the mechanical loading module is a tensile testing machine, which clamps both ends of the specimen with fixtures to achieve quasi-static tensile loading on the specimen.

[0019] Furthermore, the preset encoding mode is the Hadamard encoding mode, with an order of 256×256 pixels.

[0020] Furthermore, the angle between the transmission axes of the first linear polarizer and the second linear polarizer is 90°.

[0021] The mechanism of action of this invention can be summarized as a "three-layer protection" mechanism: The first layer is the optical layer: it attenuates directional interference and some scattered interference through orthogonal polarizers, reducing the interference energy entering the detector by more than 50%. Specifically, the first linear polarizer defines the known linear polarization state of the interfering light, and the second linear polarizer is orthogonal to it, effectively blocking polarization-dependent interference components.

[0022] The second layer is the acquisition layer: through the differential acquisition characteristics of single-pixel imaging, residual uniform or slowly varying interference light is canceled out, further improving the signal-to-noise ratio. For each coding mode, the optical signals corresponding to its positive and negative modes are acquired separately and differential operations are performed, which can effectively eliminate the influence of global interference components.

[0023] The third layer is the computational layer: through the physical constraints of the PINN-DIC network, it performs fault-tolerant processing of high-frequency local interference, ultimately achieving stable measurement under strong light interference. This network takes image coordinates as input and displacement field as output, and is trained and optimized by the mean square error between the predicted image and the actual deformed image.

[0024] Compared with the prior art, the beneficial technical effects of the present invention are: First, it significantly improves imaging quality under complex lighting conditions. Traditional cameras cannot distinguish between effective signals and interfering light under strong light interference, easily leading to image saturation and decreased contrast. This invention combines orthogonal polarizers with single-pixel differential acquisition to effectively attenuate interfering light energy at the image acquisition source, canceling residual interference, resulting in reconstructed images with uniform grayscale and clear speckle, thus significantly improving imaging quality.

[0025] Second, it achieves high-precision and high-stability deformation measurement. In existing technologies, interference light can disrupt image correlation, leading to a decrease in DIC calculation accuracy or even failure. This invention provides a high-quality input image and combines it with a PINN-DIC network for displacement field fitting. Even under strong light interference, it can still obtain a continuous and smooth full-field displacement distribution. The measured values ​​are in high agreement with the theoretical values, and the calculation stability and accuracy are superior to traditional methods.

[0026] Third, it overcomes the limitations of single interference suppression methods. Post-processing denoising cannot eliminate the irreversible effects of the imaging stage, and single polarization suppression has limited effectiveness against depolarization interference. This invention combines orthogonal polarization with single-pixel differential acquisition, which can simultaneously suppress directional and scattering interference, making it suitable for complex scenarios such as rough surfaces and expanding the application scope of DIC technology.

[0027] Fourth, it balances measurement accuracy and acquisition efficiency. By optimizing the Hadamard encoding order, this invention can complete a single measurement within 60 seconds at a resolution of 256×256, which not only meets the DIC's requirements for speckle detail resolution but also takes into account the timeliness of practical applications.

[0028] In summary, this invention significantly improves the anti-interference capability, measurement accuracy, and applicability of digital image correlation measurement technology in complex lighting environments, and has good engineering application value. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the experimental apparatus according to an embodiment of the present invention.

[0030] In the figure: 1. Light source; 2. First linear polarizer; 3. Digital micromirror array; 4. Second linear polarizer; 5. Single pixel detector; 6. Standard specimen; 7. Tensile testing machine. Detailed Implementation

[0031] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0032] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0033] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0034] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0035] Example 1 This embodiment provides a DIC measurement method that combines single-pixel imaging technology with orthogonal polarization suppression, which is used to improve the measurement accuracy of full-field displacement and strain under complex lighting conditions.

[0036] The method includes the following steps: Step S1, System Setup and Optical Path Calibration: Construct the measurement system, including an optical imaging module, a mechanical loading module, and a data processing module. The optical imaging module includes a light source 1, a first linear polarizer 2, a digital micromirror array 3, a second linear polarizer 4, and a single-pixel detector 5. The first linear polarizer 2 is placed in front of the light source 1 to generate linearly polarized light to illuminate the surface of the standard sample 6; the second linear polarizer 4 is placed in front of the single-pixel detector 5, with its transmission axis orthogonal to the transmission axis of the first linear polarizer 2. The optical path is calibrated to ensure that the illumination and detection optical paths meet the orthogonal polarization configuration.

[0037] Step S2, Projection and Signal Acquisition of Encoded Patterns: The digital micromirror array 3 is controlled to generate a preset Hadamard encoding pattern sequence and project it sequentially onto the surface of the standard specimen 6. The standard specimen 6 is under load under the action of the tensile testing machine 7. The single-pixel detector 5 synchronously acquires the intensity of the light signal reflected from the surface of the standard specimen 6 and transmitted through the second linear polarizer 4. For each encoded pattern, the light signals corresponding to its positive and negative modes are acquired respectively to obtain a pair of differential signals.

[0038] Step S3, Single-pixel image reconstruction: Using the differential signal obtained in step S2, combined with a preset Hadamard coding pattern sequence, a high signal-to-noise ratio grayscale image of the surface of standard sample 6 is calculated using a single-pixel imaging reconstruction algorithm. Due to the use of differential acquisition, globally uniform or slowly varying residual interference light is further canceled out during the differential process, thereby effectively suppressing its impact on imaging quality.

[0039] Step S4, Deformation field calculation based on PINN-DIC: The high-quality grayscale image reconstructed in step S3 is input into the pre-constructed PINN-DIC network. This network takes the image pixel coordinates (x, y) as input and the displacement field (u, v) as output, fitting the displacement field function through a neural network. The objective function for network training is the mean square error between the predicted deformation image and the actual deformation image. After network training converges, the full-field displacement distribution of the surface of standard specimen 6 is obtained from the output layer, and the corresponding strain field is then calculated.

[0040] In this embodiment, the digital micromirror array 3 uses a Texas Instruments DLP Discovery 4100 with a resolution of 768×1024. The projection mode is a 256×256-order Hadamard encoding mode with a sampling rate of 40%, requiring a total of 26,384 modes to be projected. Each mode is sampled once in both forward and reverse states, with a single exposure time of 950 μs and a total acquisition time of approximately 53 seconds. The single-pixel detector 5 uses a Thorlabs PDA100A2, along with a preamplifier and a data acquisition card (DAQ) for signal acquisition.

[0041] In this embodiment, the first linear polarizer 2 (DHC GCL-050003) is placed in front of the light source 1, causing the incident light to become horizontally linearly polarized light. The second linear polarizer 4 (DHC GCL-050004) is placed in front of the single-pixel detector 5, and its polarization angle is precisely adjusted to 90° using a graduated polarizer holder, forming an orthogonal configuration with the first linear polarizer 2. The standard sample 6 is made of PE (polyethylene), and its surface has a certain roughness, which has a partial depolarization effect on the incident linearly polarized light. After being scattered by the standard sample 6, part of the originally horizontally linearly polarized effective signal light is transformed into other polarization states. Among them, the component that still maintains horizontal polarization is completely blocked by the second linear polarizer 4, while about 50% of the energy of the depolarized component (including the effective signal) can pass through the second linear polarizer 4 and enter the detector. The horizontally polarized component in the interference light is also blocked by the second linear polarizer 4, while the depolarized component is halved.

[0042] In this embodiment, the mathematical expression for the differential signal can be represented as:

[0043] in For the k-th projection mode, For a valid signal, As the interfering light component, its contribution is significantly reduced after differential and polarization suppression. Combined with the pre-attenuation of the interfering light by orthogonal polarizers, the signal-to-noise ratio of the signal entering the detector is further improved.

[0044] In this embodiment, the PINN-DIC network structure comprises six layers: one input layer, four hidden layers, and one output layer. The input layer has two neurons, corresponding to pixel coordinates (x, y). The output layer has two neurons, corresponding to displacement components (u, v). Each of the four hidden layers contains 50 neurons, and the activation function is the tanh function. The loss function is defined as:

[0045] in The predicted deformed image is obtained by interpolating and resampling the reference image based on the displacement field predicted by the network. The images shown are actual deformed images. The network was trained using the Adam optimizer with a learning rate of 0.001 for 2000 iterations. After training, the network can output displacement values ​​at any coordinate point, forming a continuous full-field displacement distribution, which can then be used to calculate the strain field through numerical differentiation.

[0046] Example 2 This embodiment optimizes the encoding strategy for single-pixel imaging based on Embodiment 1.

[0047] Considering the impact of Hadamard coding order on imaging quality and acquisition efficiency, this embodiment compares the measurement effects under different coding orders through experiments. The experiments show that when the Hadamard coding order is below 128×128, the spatial resolution of the reconstructed image is insufficient, resulting in the loss of some speckle detail information, leading to poorer subsequent DIC matching and increased displacement measurement error. When the coding order is above 512×512, the number of required projection patterns increases dramatically, resulting in excessively long total acquisition time (exceeding 200 seconds), making it difficult to meet the needs of dynamic measurements or applications requiring high acquisition efficiency.

[0048] Therefore, as a preferred solution, this embodiment employs a 256×256 order Hadamard coding pattern. This scheme can ensure that the reconstructed image contains sufficient speckle details to meet the high-precision matching requirements of DIC, while keeping the total acquisition time of a single measurement within 60 seconds, achieving a good balance between measurement accuracy and efficiency.

[0049] Example 3 This embodiment provides a system for implementing the above measurement method, such as... Figure 1 As shown. The system includes: Optical imaging module, including: Light source 1, used to generate an illumination beam; The first linear polarizer 2 is disposed in the light output path of the light source 1 and is used to modulate the illumination light into linearly polarized light; The digital micromirror array 3 is used to receive linearly polarized light modulated by the first linear polarizer 2, and modulate it into a structured light field according to a preset encoding mode before projecting it onto the surface of the standard sample 6. The second linear polarizer 4 is disposed on the optical path for receiving optical signals. Its transmission axis is orthogonal to the transmission axis of the first linear polarizer 2. It is used to filter out the components in the incident light that maintain the same polarization state as the illumination light. A single-pixel detector 5 is located at the rear end of the second linear polarizer 4 and is used to receive the optical signal filtered by the second linear polarizer 4 and convert it into an electrical signal.

[0050] The mechanical loading module is used to apply a controllable mechanical load to the standard specimen 6. In this embodiment, a tensile testing machine 7 is used, specifically an MTS electro-hydraulic servo material testing system. The standard specimen 6 is clamped at both ends by a fixture, and the upper clamp is precisely controlled to generate an upward step displacement (e.g., Δx = 100 μm each time) to achieve quasi-static tensile loading on the standard specimen 6, ensuring the stability of the loading process and the uniformity of the force on the specimen.

[0051] The data processing module includes: The signal acquisition and control unit is connected to the digital micromirror array 3 and the single-pixel detector 5. It is used to control the digital micromirror array 3 to generate a preset encoding pattern sequence and synchronously acquire the electrical signal output by the single-pixel detector 5. The image reconstruction unit is used to reconstruct the grayscale image of the surface of the standard sample 6 based on the acquired signal and the preset encoding mode using a single-pixel imaging reconstruction algorithm. The deformation calculation unit, with a built-in PINN-DIC network, is used to input the image sequence output by the image reconstruction unit into the network to calculate the full-field displacement field and strain field on the surface of the standard specimen 6.

[0052] The system's workflow is as follows: The data processing module controls the digital micromirror array 3 to generate a Hadamard-coded pattern sequence. Linearly polarized light, modulated by the first linear polarizer 2, illuminates the digital micromirror array 3 and is modulated into a structured light field, which is then projected onto the surface of the standard specimen 6. Simultaneously, the tensile testing machine 7 loads the standard specimen 6. The light signal reflected from the surface of the standard specimen 6 (containing effective signals and residual interference) passes through the second linear polarizer 4, is received by the single-pixel detector 5, and converted into an electrical signal. The data processing module acquires the electrical signal, combines it with the coding pattern to perform image reconstruction and PINN-DIC calculation, and finally outputs the full-field displacement and strain measurement results.

[0053] Example 4 To verify the technical effect of the present invention, a comparative experiment was conducted under strong light interference conditions in this embodiment.

[0054] The experiment was conducted with two control groups: the first group used a traditional camera-based DIC measurement system (control group); the second group used the measurement method and system described in Example 1 of this invention (experimental group). Both groups applied the same quasi-static tensile load to the same PE material standard specimen 6. During the experiment, an additional strong light source was used to apply directional light interference to the surface of the standard specimen 6.

[0055] Experimental results show that under strong light interference, the control group exhibits significant local saturation and contrast reduction in the acquired images. DIC-related calculations fail in the interference area, resulting in incomplete displacement field data, and the displacement measurement results in the calculable area fluctuate considerably. In contrast, the experimental group, under the same interference conditions, effectively attenuated directional interference light through an orthogonal polarization system composed of the first linear polarizer 2 and the second linear polarizer 4. Combined with the single-pixel imaging differential acquisition characteristics achieved by the digital micromirror array 3 and the single-pixel detector 5, residual interference was further suppressed. The reconstructed image showed uniform grayscale distribution, clear speckle features, and high consistency with the image acquired under interference-free conditions. PINN-DIC calculation based on the reconstructed image converged stably, and the obtained full-field displacement distribution was smooth and continuous. The measured values ​​highly matched the theoretical calculated values ​​(uniform tensile strain), verifying the high accuracy and robustness of the method under strong light interference.

[0056] In summary, this invention combines single-pixel imaging technology with orthogonal polarization suppression, effectively reducing the influence of interference light at the physical level during image acquisition. This provides high-quality input images for the DIC algorithm. Furthermore, by combining the PINN-DIC method, it achieves high-precision and high-stability measurement of full-field displacement and strain under complex lighting conditions. Compared with traditional technologies, this invention significantly improves the anti-interference capability and measurement reliability of the measurement system under complex lighting conditions, and has promising prospects for engineering applications.

[0057] 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.

[0058] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention, and no reference numerals in the claims should be construed as limiting the scope of the claims.

[0059] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A digital image correlation measurement method combining single-pixel imaging and polarization suppression, characterized in that, Includes the following steps: Step S1: Construct a measurement system, which includes an optical imaging module, a mechanical loading module, and a data processing module; The optical imaging module includes a light source, a first linear polarizer, a digital micromirror array, a second linear polarizer, and a single-pixel detector; the first linear polarizer is placed in front of the light source, the second linear polarizer is placed in front of the single-pixel detector, and the transmission axis of the second linear polarizer is orthogonal to the transmission axis of the first linear polarizer. Step S2: Control the digital micromirror array to generate a preset coding pattern sequence and project it onto the surface of the test sample in sequence; The single-pixel detector synchronously collects the light signal intensity after being reflected from the sample surface and transmitted through the second linear polarizer. For each coding pattern, the light signals corresponding to its positive mode and negative mode are collected respectively to obtain a pair of differential signals. Step S3: Using the differential signal obtained in step S2, combined with the preset encoding mode sequence, the grayscale image of the sample surface is calculated by a single-pixel imaging reconstruction algorithm. Step S4: Input the grayscale image reconstructed in step S3 into the pre-constructed PINN-DIC network. The network takes the image pixel coordinates as input and the displacement field as output, and fits the displacement field function through a neural network. After the network training converges, the full-field displacement distribution of the sample surface is obtained from the output layer, and then the corresponding strain field is calculated.

2. The digital image correlation measurement method combining single-pixel imaging and polarization suppression according to claim 1, characterized in that, The encoding mode described in step S2 is the Hadamard encoding mode.

3. The digital image correlation measurement method combining single-pixel imaging and polarization suppression according to claim 2, characterized in that, The order of the Hadamard coding pattern is 256×256 pixels.

4. The digital image correlation measurement method combining single-pixel imaging and polarization suppression according to claim 1, characterized in that, The differential signal acquisition and processing process described in step S2 is as follows: in For the k-th projection mode, For a valid signal, To interfere with the light component.

5. The digital image correlation measurement method combining single-pixel imaging and polarization suppression according to claim 1, characterized in that, The PINN-DIC network described in step S4 includes an input layer, a hidden layer, and an output layer; the input layer has 2 neurons, corresponding to pixel coordinates (x, y); the output layer has 2 neurons, corresponding to displacement components (u, v); the hidden layer has 4 layers, each containing 50 neurons, and the activation function is the tanh function.

6. The digital image correlation measurement method combining single-pixel imaging and polarization suppression according to claim 1 or 5, characterized in that, The loss function of the PINN-DIC network in step S4 is: in The predicted deformed image is obtained by interpolating and resampling the reference image based on the displacement field predicted by the network. These are the deformed images that were actually captured.

7. A digital image correlation measurement system combining single-pixel imaging and polarization suppression, implementing the method of any one of claims 1-6, characterized in that, include: An optical imaging module includes: a light source for generating an illumination beam; A first linear polarizer is disposed in the light output path of the light source to modulate the illumination light into linearly polarized light; A digital micromirror array is used to receive linearly polarized light modulated by a first linear polarizer, and modulate it into a structured light field according to a preset encoding mode before projecting it onto the surface of the test sample. The second linear polarizer is placed in the optical path for receiving optical signals. Its transmission axis is orthogonal to that of the first linear polarizer. It is used to filter out components in the incident light that maintain the same polarization state as the illumination light. A single-pixel detector, located at the rear end of the second linear polarizer, is used to receive the optical signal filtered by the second linear polarizer and convert it into an electrical signal; The mechanical loading module is used to apply a controllable mechanical load to the test sample; The data processing module includes: The signal acquisition and control unit is connected to the digital micromirror array and the single-pixel detector, and is used to control the digital micromirror array to generate a preset encoding pattern sequence and synchronously acquire the electrical signal output by the single-pixel detector. The image reconstruction unit is used to reconstruct a grayscale image of the sample surface based on the acquired signal and a preset encoding mode using a single-pixel imaging reconstruction algorithm. The deformation calculation unit, with a built-in PINN-DIC network, is used to input the image sequence output by the image reconstruction unit into the network to calculate the full-field displacement field and strain field of the sample surface.

8. The digital image correlation measurement system combining single-pixel imaging and polarization suppression according to claim 7, characterized in that, The mechanical loading module is a tensile testing machine, which clamps both ends of the specimen to achieve quasi-static tensile loading on the specimen.

9. The digital image correlation measurement system combining single-pixel imaging and polarization suppression according to claim 7, characterized in that, The preset encoding mode is the Hadamard encoding mode, with an order of 256×256 pixels.

10. The digital image correlation measurement system combining single-pixel imaging and polarization suppression according to claim 7, characterized in that, The angle between the transmission axes of the first linear polarizer and the second linear polarizer is 90°.

Images