Adaptive matrix block weighted distributed optical fiber strain measurement method

Abstract

The present invention relates to a distributed optical fiber strain measurement method, and specifically to an adaptive matrix block weighted distributed optical fiber strain measurement method, comprising the following steps: S1, respectively collecting a reference signal and a test signal; S2, performing frequency domain windowing processing on the reference signal and the test signal, and performing local cross-correlation analysis on each windowed signal to obtain a one-dimensional cross-correlation result; S3, integrating the one-dimensional cross-correlation results of all windows, and reconstructing them into a two-dimensional image signal over the optical fiber distance; S4, denoising the image by an adaptive matrix block weighted processing method. S5, outputting the strain measurement result. The present invention can suppress noise interference in long-distance, high-spatial-resolution, and large-scale distributed strain sensing measurement signals, thereby improving the accuracy and reliability of the signal. It also has broad application potential in the fields of intelligent optical fiber sensing, structural health monitoring, temperature sensing, etc., and can improve measurement accuracy and data quality.

Description

Technical Field

[0001] The present invention relates to a distributed optical fiber strain measurement method, and in particular to an adaptive matrix block weighted distributed optical fiber strain measurement method. Background Art

[0002] In distributed fiber-optic sensing technology, optical fibers serve not only as sensing media but also as transmission media for measurement signals. By exploiting the propagation properties of light waves within optical fibers, such as Raman scattering, Rayleigh scattering, and Brillouin scattering, real-time monitoring of the external environment can be achieved, with the monitoring range extending along the length of the fiber. This technology offers significant advantages, including strong immunity to electromagnetic interference, simple structure, high spatial resolution, and long sensing distance. Leveraging these features, distributed fiber-optic sensing has been gradually applied in a variety of fields, such as bridge safety monitoring, civil engineering inspection, underground tunnel fire alarms, and geological surveys, generating significant social benefits. Optical frequency domain reflectometry (OFDR), a representative example of this technology, boasts advantages such as portability, compact size, high sensitivity, strong immunity to electromagnetic interference, and excellent spatial resolution. It can continuously measure changes in external physical quantities such as strain, vibration, and temperature along the fiber. In recent years, with the development of OFDR technology, its application areas have continued to expand, extending to emerging fields such as shape sensing and acoustic sensing.

[0003] The operating principle of OFDR is as follows: a linearly swept light beam emitted by a tunable laser source is split into two beams by a coupler. One beam enters the optical fiber under test and, after returning as Rayleigh scattered light from the fiber under test, undergoes beat frequency interference with the other reference beam. By collecting the beat frequency signal and performing fast Fourier transform processing, distance domain information along the optical fiber can be obtained. During the OFDR measurement process, it is necessary to first collect a reference signal free of external interference, then collect a test signal affected by external factors, and calculate external changes through cross-correlation operations. A prominent feature of the OFDR system is its high spatial resolution, which can reach the millimeter level, making it widely used in high-precision monitoring fields such as aerospace. However, as the sensing distance and spatial resolution increase, large strain stretching can cause multiple peaks or false peaks in the cross-correlation results, thereby affecting the accuracy of the measurement results. This is mainly due to the accumulation of long-distance noise and the reduction in correlation caused by strain stretching. Therefore, how to effectively achieve long-distance, high-spatial-resolution large strain measurement has become a key research issue that needs to be urgently addressed.

[0004] Based on this, the present invention provides an adaptive matrix block weighted distributed optical fiber strain measurement method to solve the above problems. Summary of the Invention

[0005] In order to make up for the shortcomings of the existing technology, the present invention proposes a method based on adaptive matrix block weighting to improve the quality of long-distance, high spatial resolution and large-scale distributed strain sensing measurement signals. By adaptively adjusting the size of the matrix block and the weighting parameters, the local area of ​​the signal is adaptively weighted, and the influence of noise and the cross-correlation error caused by strain stretching are dealt with at the same time.

[0006] The technical solution of the present invention is:

[0007] The adaptive matrix block weighted distributed optical fiber strain measurement method comprises the following steps:

[0008] S1, collects reference signal and test signal respectively;

[0009] S2, perform frequency domain windowing on the reference signal and the test signal, and perform local cross-correlation analysis on each windowed signal to obtain a one-dimensional cross-correlation result;

[0010] S3, integrates the one-dimensional cross-correlation results of all sub-windows and reconstructs them into a two-dimensional image signal over the fiber distance;

[0011] S4, denoising the image by using an adaptive matrix block weighting processing method;

[0012] S5, output the strain measurement results.

[0013] In order to better implement the present invention, a further technical solution is:

[0014] S1 is specifically:

[0015] The reference signal is a signal that does not contain strain information, and the test signal is a signal that contains strain information.

[0016] S2 is specifically,

[0017] S21, the reference signal and the test signal are mapped to the range domain through fast Fourier transform, and the signal is divided into N equal parts using a window size C, where the window size C determines the spatial resolution of the system;

[0018] S22, performing a fast inverse Fourier transform on the local distance domain information of the first reference signal and the measurement signal;

[0019] S23, performing cross-correlation calculation on the reference signal and the measurement signal after the fast inverse Fourier transform to obtain a one-dimensional cross-correlation result.

[0020] S3 is specifically:

[0021] Repeat S22-S23 to obtain the cross-correlation result of each corresponding position of the optical fiber, and reconstruct all the obtained one-dimensional cross-correlation results into a two-dimensional image signal over the optical fiber distance.

[0022] S4 is specifically:

[0023] S41, calculate the local variance using a 3×3 matrix block for each two-dimensional data: , the noise intensity of the local matrix block is judged according to the variance. This operation is used to preliminarily estimate the intensity of the change in the area around each pixel, so as to adaptively determine the size of the matrix block involved in the weighted calculation;

[0024] in, is the pixel value of the k-th pixel in the window, is the average value of all pixels in the window, n is the total number of pixels in the window, is the variance of the current window, indicating the intensity of local changes;

[0025] S42, calculate the average value of all matrix block variances of the entire two-dimensional image: ;

[0026] in, is the average value of all matrix block variances, M is the total number of image rows, and N is the total number of image columns;

[0027] S43, since the variance can reflect the strength of the noise, an exponential mapping function is introduced to convert the local variance of each pixel into the matrix block size of the weighted calculation, so as to facilitate the adaptive adjustment of the weighted matrix block size: ;

[0028] in, Represents the radius of the matrix block involved in weight calculation, 、 is the set minimum and maximum matrix block radius values, k is the slope coefficient of the mapping function, the larger the value, the more sensitive the matrix block size is to the variance response. is the variance of the local matrix block, is the average value of all matrix block variances;

[0029] S44, calculate the similarity weight between the current pixel and each pixel in the neighborhood, using an adaptive weighting function: ; Normalize the obtained weights: ;

[0030] in, is the pixel value at the center of the matrix block, is the pixel value of other points in the matrix block, Is the value that controls the degree of weight decay, weight Adaptive dynamic adjustment based on the similarity between pixels, is the normalized weight value.

[0031] S45, performing weighted averaging processing on the pixels in the matrix block using the calculated similarity weights to obtain denoised pixel values. The weighted averaging formula is: ;

[0032] in, is the normalized weight value, is the pixel value of other points in the matrix block, The new pixel value is the weighted average of the pixel values ​​in the neighborhood. This new pixel value can effectively remove noise while preserving signal details.

[0033] S4 is specifically:

[0034] In S43, the matrix block size is adaptively determined, and the pixel value of the matrix block is: ;

[0035] Where r is the radius of the matrix block, i and j are the row and column coordinates of the pixel in the image, The pixel values ​​of other points in the matrix block.

[0036] S5 is specifically:

[0037] The strain information along the sensing fiber is analyzed by determining the appropriate spectral shift between the reference signal and the measured signal through cross-correlation analysis.

[0038] The beneficial effects of the present invention are:

[0039] 1. This invention adopts an adaptive matrix block weighted denoising method, which can adaptively adjust the size of the matrix blocks involved in the weighted calculation and the weight of each data within the matrix block according to the characteristics of the local area of ​​the image. Compared with traditional denoising methods, it can better remove noise while effectively preserving signal details and avoiding over-smoothing. It is particularly suitable for preserving details in large-scale distributed strain sensing measurement signals.

[0040] 2. Compared with traditional denoising methods, the algorithm of this invention relies only on local matrix block information, resulting in lower computational complexity and particularly suitable for denoising large amounts of data signals. Its high computational efficiency makes it suitable for real-time or high-efficiency processing scenarios.

[0041] 3. By introducing an adaptive weight calculation mechanism, the algorithm can dynamically adjust the denoising strategy based on local changes in the image, thereby adapting to different types of noise, including Gaussian and non-Gaussian noise. This makes the method highly adaptable and flexible in practical applications.

[0042] 4. This invention can suppress noise interference in long-distance, high-spatial-resolution, and large-scale distributed strain sensing measurements, improving signal accuracy and reliability. It also has broad application potential in areas such as intelligent fiber optic sensing, structural health monitoring, and temperature sensing, enhancing measurement accuracy and data quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Schematic diagram of the adaptive matrix block weighted distributed optical fiber strain measurement system used in the present invention.

[0044] Figure 2 It is a processing flow chart of the present invention.

[0045] Figure 3 This is the result diagram of the sensing fiber spectrum offset obtained using the existing measurement method. The sensing fiber is subjected to strain information at 54.4-55.4m, with a spatial resolution of 2mm, a strain range of 500με-5500με, and a strain interval range of 500με.

[0046] Figure 4 This is the result diagram of the sensing fiber spectrum shift obtained by the measurement method of the present invention. The sensing fiber is subjected to strain information at 54.4-55.4m, with a spatial resolution of 2mm, a strain range of 500με-5500με, and a strain interval range of 500με. DETAILED DESCRIPTION

[0047] The following will be combined with the accompanying drawings to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, not all of them. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in a variety of different configurations.

[0048] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but is merely intended to represent selected embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative work are within the scope of protection of the present invention.

[0049] refer to Figure 2The adaptive matrix block weighted distributed optical fiber strain measurement method of this embodiment dynamically adjusts the weighting coefficients of neighboring blocks, solving the common problems of signal detail loss and insufficient peak removal in traditional processing methods. This method preserves more signal details during processing, solves the multi-peak and false peak problems of traditional OFDR systems, and significantly improves the accuracy and signal quality of long-distance, high-spatial-resolution, and large-scale distributed strain sensing measurements. The specific steps are as follows:

[0050] 1. Collect signals twice, one without strain information, which is the reference signal; the other with strain information, which is the test signal.

[0051] 2. Map the reference and test signals to the range domain using a fast Fourier transform (FFT) algorithm. Divide the signal into N equal parts using a window size C, where the window size C determines the spatial resolution of the system.

[0052] 3. Perform a fast inverse Fourier transform on the local distance domain information of the first reference signal and the measurement signal;

[0053] 4. The reference signal and the measurement signal after the fast inverse Fourier transform are cross-correlated to obtain a one-dimensional cross-correlation result;

[0054] 5. Repeat 3-4 to obtain the cross-correlation results at each corresponding position of the optical fiber, and reconstruct all the obtained one-dimensional cross-correlation results into a two-dimensional image signal at the optical fiber distance.

[0055] 6. Calculate the local variance using a 3×3 matrix block for each two-dimensional data: The noise intensity of the local matrix block is determined according to the variance. This operation is used to preliminarily estimate the intensity of the change in the area around each pixel, so as to adaptively determine the size of the matrix block involved in the weighted calculation.

[0056] in, is the pixel value of the k-th pixel in the window, is the average value of all pixels in the window, n is the total number of pixels in the window, is the variance of the current window, indicating the intensity of local changes.

[0057] 7. Calculate the average value of all matrix block variances for the entire two-dimensional image: .in, is the average value of all matrix block variances, M is the total number of image rows, and N is the total number of image columns.

[0058] Since the variance can reflect the strength of the noise, an exponential mapping function is introduced to convert the local variance of each pixel into the matrix block size of the weighted calculation, so as to facilitate the adaptive adjustment of the weighted matrix block size: ;

[0059] in, Represents the radius of the matrix block involved in weight calculation, 、 is the set minimum and maximum matrix block radius values, k is the slope coefficient of the mapping function, the larger the value, the more sensitive the matrix block size is to the variance response. is the variance of the local matrix block, is the average of all matrix block variances.

[0060] The adaptively determined matrix block size, the pixel value of the matrix block is: . Calculate the similarity weight between the current pixel and each pixel in the neighborhood using an adaptive weighting function: . Normalize the obtained weights: .

[0061] Where r is the radius of the matrix block, i and j are the row and column coordinates of the pixel in the image, is the pixel value at the center of the matrix block, is the pixel value of other points in the matrix block, Is the value that controls the degree of weight decay, weight Adaptive dynamic adjustment based on the similarity between pixels, is the normalized weight value.

[0062] 8. Perform weighted averaging on the pixels in the matrix block using the calculated similarity weights to obtain the denoised pixel values. The weighted averaging formula is:

[0063] in, is the normalized weight value, is the pixel value of other points in the matrix block, It is the pixel value after weighted average processing, which is the result obtained by weighting the pixel values ​​in the neighborhood. This new pixel value can effectively remove noise while retaining signal details.

[0064] 9. Analyze the strain information along the sensing fiber by determining the appropriate spectral shift between the reference signal and the measured signal through cross-correlation analysis.

[0065] Figure 1The adaptive matrix block weighted distributed optical fiber strain measurement system shown in the figure includes: the continuous laser output of the tunable laser source is divided into two parts by coupler 1 (10 / 90 optical coupler), 10% of which is incident on an unbalanced Mach-Zehnder triggered interferometer to provide a trigger signal for the acquisition card, and the remaining part of the light enters coupler 2; then coupler 2 (1 / 99 optical coupler) is divided into two parts, 1% of which is adjusted by polarization controller 1 so that the "p" and "s" light components have the same power, and 99% enters the sensing fiber detection through a circulator and polarization controller 2, and the sensing fiber is a standard single-mode fiber; then the interference signal obtained by combining the Rayleigh scattering signal with the 1% laser output from coupler 3 (50 / 50 optical coupler) is decomposed into "p" and "s" components by a polarization beam splitter; finally, the "p" and "s" lights are collected by the acquisition card.

[0066] The same sensing fiber is measured using the existing measurement method and the measurement method of the present invention. The strain information at 54.4-55.4m of the sensing fiber is measured, the spatial resolution is 2mm, the strain range is 500με-5500με, and the strain interval range is 500με. Figure 3 、 Figure 4 Result diagram of the sensing fiber spectrum shift. Figure 3 The original measurement result image is generated by the traditional measurement method directly based on the collected cross-correlation result image. Signals of different colors represent different strain signals. It can be seen that without using any image processing means, the specific measurement results of the signal cannot be observed. Figure 4 The strain distribution result diagram along the optical fiber obtained by the measurement method of the present invention is shown in FIG. Signals of different colors represent different strain signals. It can be seen that compared with Figure 3 In general, the present invention solves the common problems of signal detail loss and insufficient noise removal in traditional denoising methods during measurement, so that different strain information can be clearly distinguished. Compared with traditional measurement methods, the measurement effect of the present invention is significantly improved.

[0067] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not limiting. Other modifications or equivalent substitutions made to the technical solution of the present invention by ordinary technicians in this field should be included in the scope of the claims of the present invention as long as they do not depart from the spirit and scope of the technical solution of the present invention.

Claims

1. Adaptive matrix block weighted distributed optical fiber strain measurement method, characterized by: The following steps are involved: S1, collects reference signal and test signal respectively; S2, perform frequency domain windowing on the reference signal and the test signal, and perform local cross-correlation analysis on each windowed signal to obtain a one-dimensional cross-correlation result; S3, integrates the one-dimensional cross-correlation results of all sub-windows and reconstructs them into a two-dimensional image signal over the fiber distance; S4, denoising the image by using an adaptive matrix block weighting processing method; S5, output strain measurement results; S4 is specifically: S41, calculate the local variance using a 3×3 matrix block for each two-dimensional data: , the noise intensity of the local matrix block is judged according to the variance. This operation is used to preliminarily estimate the intensity of the change in the area around each pixel, so as to adaptively determine the size of the matrix block involved in the weighted calculation; in, is the pixel value of the k-th pixel in the window, is the average value of all pixels in the window, n is the total number of pixels in the window, is the variance of the current window, indicating the intensity of local changes; S42, calculate the average value of all matrix block variances of the entire two-dimensional image: ; in, is the average value of all matrix block variances, M is the total number of image rows, and N is the total number of image columns; S43, introduces an exponential mapping function to convert the local variance of each pixel into the matrix block size of the weighted calculation, so as to facilitate the adaptive adjustment of the weighted matrix block size: ; in, Represents the radius of the matrix block involved in weight calculation, 、 is the set minimum and maximum matrix block radius values, k is the slope coefficient of the mapping function, the larger the value, the more sensitive the matrix block size is to the variance response. is the variance of the local matrix block, is the average value of all matrix block variances; S44, calculate the similarity weight between the current pixel and each pixel in the neighborhood, using an adaptive weighting function: ; Normalize the obtained weights: ； in, is the pixel value at the center of the matrix block, is the pixel value of other points in the matrix block, Is the value that controls the degree of weight decay, weight Adaptive dynamic adjustment based on the similarity between pixels, is the normalized weight value; S45, performing weighted averaging processing on the pixels in the matrix block using the calculated similarity weights to obtain denoised pixel values. The weighted averaging formula is: ; in, is the normalized weight value, is the pixel value of other points in the matrix block, It is the new pixel value after weighted averaging.

2. The adaptive matrix block weighted distributed optical fiber strain measurement method according to claim 1, characterized in that: S1 is specifically: The reference signal is a signal that does not contain strain information, and the test signal is a signal that contains strain information.

3. The adaptive matrix block weighted distributed optical fiber strain measurement method according to claim 1, characterized in that: S2 is specifically, S21, the reference signal and the test signal are mapped to the range domain through fast Fourier transform, and the signal is divided into N equal parts using a window size C, where the window size C determines the spatial resolution of the system; S22, performing a fast inverse Fourier transform on the local distance domain information of the first reference signal and the measurement signal; S23, performing a cross-correlation calculation on the reference signal and the measurement signal after the fast inverse Fourier transform to obtain a one-dimensional cross-correlation result.

4. The adaptive matrix block weighted distributed optical fiber strain measurement method according to claim 1, characterized in that: S3 is specifically: Repeat S22-S23 to obtain the cross-correlation result of each corresponding position of the optical fiber, and reconstruct all the obtained one-dimensional cross-correlation results into a two-dimensional image signal over the optical fiber distance.

5. The adaptive matrix block weighted distributed optical fiber strain measurement method according to claim 1, characterized in that: S4 is specifically: In S43, the matrix block size is adaptively determined, and the pixel value of the matrix block is: ； Where r is the radius of the matrix block, i and j are the row and column coordinates of the pixel in the image, The pixel values ​​of other points in the matrix block.

6. The adaptive matrix block weighted distributed optical fiber strain measurement method according to claim 1, characterized in that: S5 is specifically: The strain information along the sensing fiber is analyzed by determining the appropriate spectral shift between the reference signal and the measured signal through cross-correlation analysis.

Images