A method for measuring mode-dependent loss of a randomly coupled multicore optical fiber

Abstract

The application discloses a kind of random coupling multicore optical fiber mode correlation loss measurement method, comprising the following steps: S1, based on the actual measurement MDL of random coupling multicore optical fiber based on disturbance method;S2, based on the MDL of random coupling multicore optical fiber measured based on optical transfer matrix method;Original output-input data set for inversion transmission matrix is constructed;S3, the original light spot image of CCD acquisition is executed pre-processing, and output-input matrix is constructed;S4, the transmission matrix of random coupling multicore optical fiber is calculated based on least square method using output-input matrix, and the MDL of transmission matrix is calculated by singular value decomposition to obtain;S5, the MDL of step S1 is compared with the MDL of step S4, and the reliability of transmission matrix method is verified, while the loss coefficient setting in MDL simulation model is corrected, and the measurement of MDL is completed.The method can realize the evaluation of the transmission characteristics of random coupling multicore optical fiber under the lower hardware configuration requirement.

Description

Technical Field

[0001] This invention belongs to the field of optical fiber communication technology, specifically relating to a method for measuring mode-correlated loss in randomly coupled multi-core optical fibers. Background Technology

[0002] Randomly coupled multi-core fiber (RC-MCF) is considered an ideal medium for long-distance, high-capacity optical transmission systems due to its high spatial density, low transmission loss, and unique advantages in suppressing group delay spread (GDS) and nonlinear effects through strong random mode coupling.

[0003] A key linear impairment faced by RC-MCF in practical applications is mode-dependent loss (MDL). MDL refers to the difference in attenuation experienced by different spatial modes or supermodes when an optical signal propagates in an optical fiber and its components. This difference originates from minute asymmetries in the fiber manufacturing process, mode leakage caused by macro-bending and micro-bending, and alignment errors at fusion splices or connectors. In high-order modulation formats and multiple-input-output (MIMO) digital signal processing systems, MDL disrupts channel orthogonality, leading to reduced channel capacity and significantly increasing the probability of system outages. Therefore, accurate measurement of RC-MCF MDL is crucial for fiber design optimization, manufacturing quality control, and link budget assessment of transmission systems.

[0004] Currently, there are reports on MDL measurement techniques for space-division multiplexing fibers. For example, in 2013, NK. Fontaine et al. proposed the SWI-OVNA technique based on swept wavelength interferometry. However, this technique has a relatively complex hardware architecture and must include a reference arm that is precisely matched to the length of the fiber under test to maintain coherence. In 2023, M. Uyama et al. proposed Linear Optical Sampling (LOS) or dual-comb interferometry. Although LOS technology overcomes the length limitation of the physical reference arm and is suitable for long-distance measurements, its system cost is high (two high-stability pulse sources), and the measurement accuracy is extremely sensitive to the time jitter between the two lasers. It also requires the introduction of dispersive fiber pulse broadening to avoid nonlinear effects, making the operation process cumbersome. In 2021, T. Hasegawa et al. proposed the Scrambing Method, which has high sensitivity and a simple device, making it suitable for long-distance fibers. Summary of the Invention

[0005] The main objective of this invention is to overcome the shortcomings and deficiencies of the prior art and to propose a method for measuring mode correlation loss in randomly coupled multi-core optical fibers.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A method for measuring mode correlation loss in randomly coupled multi-core optical fibers includes the following steps:

[0008] S1. The MDL value of randomly coupled multi-core optical fibers was actually measured using an MDL measurement system based on the perturbation method.

[0009] S2. The MDL of randomly coupled multi-core optical fiber is actually measured using an MDL measurement system based on the optical transfer matrix method. An orthogonal input mode field sequence is dynamically generated using a spatial light modulator and coupled into the fiber under test. The corresponding output response image and input end reference image are acquired simultaneously to construct the original output-input data set for inverting the transfer matrix.

[0010] S3. Perform preprocessing on the raw spot images acquired by the CCD, calibrate the inherent error of the MDL measurement system and perform spot localization and region segmentation operations, extract the feature parameters of each fiber core spot, provide a standardized data foundation for the subsequent construction of the output-input matrix, and construct the output-input matrix;

[0011] S4. Calculate the transmission matrix of the randomly coupled multi-core optical fiber using the least squares method based on the output-input matrix, and then perform singular value decomposition on the transmission matrix to obtain its MDL, where MDL is the result of multiple averaging.

[0012] S5. Compare the measured MDL mean value from step S1 with the MDL calculated in step S4 to verify the reliability of the transfer matrix method. At the same time, correct the loss coefficient setting in the MDL simulation model to complete the MDL measurement.

[0013] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0014] The method of this invention first obtains the reference mean of mode-dependent loss (MDL) of randomly coupled multi-core optical fibers using a perturbation method; simultaneously, it acquires the raw input-output matrix data of the fiber using the optical transfer matrix method, and performs noise reduction and system error calibration; subsequently, it obtains the transfer matrix based on the least squares method and performs singular value decomposition on it to calculate the MDL measurement value; finally, it verifies the reliability of the system by comparing the measured value obtained by the perturbation method with the calculated value obtained by the transfer matrix method, and corrects the loss coefficient of the simulation model accordingly. The method of this invention can achieve good characterization of the transmission characteristics of randomly coupled multi-core optical fibers with low hardware configuration requirements and good consistency, providing a solution that balances cost and accuracy for MDL measurement of randomly coupled multi-core optical fibers. Attached Figure Description

[0015] Figure 1This is a flowchart of the method of the present invention.

[0016] Figure 2 This is a refractive index profile of the randomly coupled four-core optical fiber in Example 1.

[0017] Figure 3 This is a schematic diagram of the MDL measurement system based on the perturbation method in Example 1.

[0018] Figure 4 It is the total output power of the 4C-CF measured by the optical power meter in Example 1.

[0019] Figure 5a It is the MDL output measurement system based on the optical transfer matrix method in Example 1.

[0020] Figure 5b It is the MDL input measurement system based on the optical transfer matrix method in Example 1.

[0021] Figure 6a This is the image preprocessing flow for measuring MDL based on the optical transfer matrix method in Example 1.

[0022] Figure 6b This is a flowchart of the light spot amplitude calculation in Example 1.

[0023] Figure 7 It is in Example 1 With MDL Scraming The data measurement values.

[0024] Figure 8 yes Figure 7 The relative deviation diagram after data fitting using the optical transfer matrix method and the perturbation method.

[0025] Figure 9 This is a flowchart of the spot localization method based on connected component analysis in Example 2. Detailed Implementation

[0026] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.

[0027] Example 1; This example uses a long-stretched, randomly coupled four-core fiber (4C-CF) as an example. Figure 2 As shown, the core radius Core spacing The relative refractive index difference between the core and the cladding A total of four bundles of optical fibers were used for the measurement. Each bundle consisted of a single 25km long fiber with a core pitch of 18µm, cut into lengths of 0.57km, 3.04km, 7.08km, and 14.03km respectively. To maintain the fiber's twist, the cut fibers were rewound onto a fiber optic disc with a disc radius R. b =80 mm.

[0028] like Figure 1 As shown, a method for measuring mode correlation loss in randomly coupled multi-core optical fibers includes the following steps:

[0029] S1. The MDL value of randomly coupled multi-core optical fibers is actually measured using an MDL measurement system based on the perturbation method. For a single bundle of optical fibers, the measurement is repeated multiple times and then the average is taken.

[0030] like Figure 3 As shown, the MDL measurement system based on the perturbation method includes a light source module, a coupling module of randomly coupled multi-core optical fibers, and a data detection and acquisition module at the receiving end.

[0031] The light source module uses a tunable laser source (TLS) to emit measurement light. In this embodiment, the light source module uses a tunable laser Santec TSL-570 and its matching wavelength scanning control system. Its frequency is continuously scanned at a rate of 1 nm / s within a range of 1550 ± 1 nm to randomly change the modes of the randomly coupled multi-core optical fiber. In order to achieve sufficient scrambling, the wavelength scanning range should be sufficiently wide than the spectral correlation width of the transmission matrix.

[0032] The coupling module for randomly coupled multi-core optical fibers is used for coupling the fiber under test (DUT) to a single-mode patch cord. It utilizes a fusion splicer to perform a core-to-core alignment operation between the single-mode patch cord and one core of the randomly coupled multi-core DUT. By adjusting the fusion splicer parameters, the coupling effect is maximized, achieving precise core alignment. In this embodiment, a Fujikura fusion splicer is used, where the single-mode patch cord is aligned with one core of a 4C-CF fiber. The single-mode patch cord is placed on the left side of the fusion splicer, and the DUT is placed on the right side. Editing the splicing mode involves first changing the parameters of the left and right fibers: setting the coating diameter of the right fiber to 250µm, the cladding diameter to 125µm, and the core diameter to 9µm; secondly, setting the alignment position of the right fiber to PAS, the XY alignment to PAS, and the alignment mode to cladding; subsequently, setting the main discharge power to standard +20 bits, and finally, turning on the motor for precise control. Adjust the xy position of the left fiber in the coordinate axis. It is known that the core spacing of the fiber under test is about 18 μm. By coarse and fine adjustment, the coupling efficiency is made to reach more than 90%, thus achieving the purpose of core coupling.

[0033] The data detection and acquisition module uses an optical power meter and a corresponding data acquisition cable for accurate data acquisition. In this embodiment, a Santec MPM-210H (optical power meter) and a GPIB-USB-HS data acquisition cable are used. The Santec MPM-210H has a built-in PD module and DAQ module, which is the data acquisition system, and can achieve accurate data acquisition.

[0034] The measurement system obtains the MDL value based on the following formula:

[0035] (1);

[0036] (2);

[0037] Among them, MDL rms The normalized RMS deviation of the total transmittance T of all cores of the optical fiber under test from its average value. This represents the total output power measured by the optical power meter, which means that the cores of all the optical fibers under test are simultaneously coupled to the optical power meter. This represents the input power measured by the optical power meter; in this embodiment, 1% of the TLS output power is extracted to compensate for power variations.

[0038] In this embodiment, considering the random coupling during the propagation of the 4C-CF supermode and noise interference in the environment, multiple repeated measurements are required before averaging. Furthermore, to enrich the data, the fiber placement is randomly adjusted after each measurement, and measurements are performed again only after the fibers have stabilized. In practice, 30 measurements are performed on each bundle of fibers. Figure 4 As shown, the total output power was measured 30 times by an optical power meter. The five largest and five smallest results were removed from the calculation results, and the final result was obtained by averaging the remaining 20 results.

[0039] S2. The MDL of randomly coupled multi-core optical fiber is actually measured using an MDL measurement system based on the optical transfer matrix method. An orthogonal input mode field sequence is dynamically generated using a spatial light modulator and coupled into the fiber under test. The corresponding output response image and input end reference image are acquired simultaneously to construct the original output-input data set for inverting the transfer matrix.

[0040] The MDL measurement system based on the optical transfer matrix method includes an optical signal generation and modulation module, a coupling module for randomly coupled multi-core optical fibers, and an optical signal imaging module.

[0041] In this embodiment, the MDL measurement system based on the optical transfer matrix method is as follows: Figure 5a and Figure 5b As shown.

[0042] The optical signal generation and modulation module includes a light source (1550 nm Laser), a spatial light modulator (SLM), a beam splitter (BS), a λ / 2 waveplate, a polarizer, and a collimator. These components are used to generate the incident light signal and to modulate and collimate its polarization, spatial mode, and other parameters. The module generates the probe beam by digitally simulating the optical function of a spiral phase plate (SPP) on the SLM. The control process involves displaying a phase modulation pattern on the SLM's liquid crystal pixel array that is equivalent to the thickness distribution of the SPP. This pattern exhibits a spiral phase transition around the optical axis center, and its phase transformation satisfies… , where l corresponds to the topological load parameter set in the software interface; by adjusting the value of the topological load l, a set of orthogonal input mode fields are dynamically generated without changing the mechanical structure, which are used to construct the input vector basis for the transfer matrix measurement.

[0043] The coupling module for randomly coupled multi-core optical fibers is used to couple the incident light emitted from the optical signal generation and control module into the optical fiber under test. In this embodiment, the coupling module for randomly coupled multi-core optical fibers uses a 20× objective lens to couple the incident light emitted from the optical signal generation and control module into the optical fiber under test. It should be noted that after the coupling is debugged, the coupling end will not be debugged in subsequent experimental processes, that is, the input end of the corresponding optical fiber under test will remain unchanged.

[0044] The optical signal imaging module is used to focus and receive the transmitted optical signal, and to complete optical signal imaging and data acquisition through a CCD, which is then used for data post-processing. In this embodiment, the optical signal imaging module completes optical signal imaging and data acquisition through a CCD. Before the experiment begins, the exposure of the CCD is adjusted by testing a 4C-CF with a length similar to RC-MCF1 (approximately 60-70cm), to ensure that the output-input light spot pairs of RC-MCF2, i.e., the actual fiber under test, are not overexposed, thus avoiding affecting the reliability of data post-processing. Specifically, after each image is acquired, data processing is performed to obtain a grayscale matrix, and the data quality is judged by whether it exhibits a Gaussian-like shape.

[0045] The optical signal generation and modulation module includes the following steps:

[0046] A spatial light modulator is used to simulate the optical function of a spiral phase plate to generate a probe beam;

[0047] By loading a phase modulation pattern corresponding to a preset topological charge parameter l onto a spatial light modulator, the probe beam is made to generate a spiral phase change around the center of the optical axis; by adjusting the value of the topological charge parameter l, a set of orthogonal input mode fields is dynamically generated to construct the input vector basis for transmission matrix measurement.

[0048] Specifically, optical signal imaging and data acquisition include:

[0049] CCD is used to acquire light spot images; before measuring the randomly coupled multi-core optical fiber under test, the exposure parameters of the system under test are calibrated using a reference optical fiber to ensure that the acquired output-input light spot image sequence is in a non-saturated state; the acquired images are grayscaled to obtain a grayscale matrix, and a square root is performed based on the sum of pixels within the calculation area, i.e., by... Acquire the light spot intensity; determine whether the collected data meets the quality requirements of subsequent processing based on whether the spatial distribution of the light spot intensity conforms to the preset Gaussian distribution characteristics.

[0050] S3. Perform preprocessing on the raw spot images acquired by the CCD, calibrate the inherent error of the MDL measurement system, perform spot localization and region segmentation operations, extract the feature parameters of each fiber core spot, and provide a standardized data foundation for the subsequent construction of the output-input matrix;

[0051] In step S3, the preprocessing includes:

[0052] Median filtering is applied to the raw spot images acquired by CCD to suppress specific noise while preserving the edge and morphological details of the spot to the maximum extent, providing high-quality image data for subsequent spot localization and region segmentation;

[0053] For spot localization and region segmentation, multiple algorithms can be selected for comprehensive localization. By optimizing edge details and preserving features, high-precision raw data can be obtained.

[0054] In this embodiment, the preprocessing process is as follows: Figure 6a As shown, by performing median filtering (using a 3×3 pixel window) on the raw light spot acquired by the CCD, the discrete noise of the detector and environmental background interference are effectively removed while preserving the edge morphology and pixel distribution details of the light spot to the maximum extent. Subsequently, light spot localization is performed by generating a corresponding circular ROI logical mask for each light spot and overlaying it with the filtered image, as shown. Figure 6b As shown, using the formula Sequentially calculate and extract the beam amplitude A of each core. k .

[0055] In this embodiment, 12 different input optical field vectors are generated by adjusting the topological charge l of the SLM. After passing through the fiber under test (RC-MCF2), the optical signal imaging module (with exposure parameters pre-set) collects the data to obtain the corresponding output optical field matrix. ,here In , ..., a xi With a yi The amplitudes are calculated for the same fiber core i under x-polarization and y-polarization, respectively. In summary, the output optical field amplitude matrix Y is... FUT As in formula (3):

[0056] (3);

[0057] For image acquisition at the input end of the fiber under test, a "cut off" point (the input end face of the RC-MCF2) was set in the experimental setup to cut the fiber and remove it from the subsequent optical path. The objective lens and CCD were then directly aligned with the cut plane, and the SLM was again used to traverse all topological charges l, recording the light spot images at this point, i.e., 12 input light spot images. (It should be noted that to ensure the coupling end of the fiber under test remains unchanged during the cut, the fiber frame in the coupling end module was reinforced with tape beforehand.) These images constitute the input matrix X of the fiber under test. FUT As in formula (4):

[0058] (4);

[0059] S4. Using the output-input matrix preprocessed in step S3, calculate the transmission matrix of the randomly coupled multi-core optical fiber based on the least squares method, and then perform singular value decomposition on the transmission matrix to obtain its MDL, where the MDL is the result of multiple averaging; in this embodiment, specifically:

[0060] The output-input light field amplitude matrix obtained in step S3 and In practical sampling, k > 2n (here k = 12, n = 4) is generally used to improve noise immunity. Here, the Moore-Penrose Pseudoinverse is used:

[0061] (5);

[0062] Since the input states are linearly independent and strictly orthogonal and uniform, that is... The solution then simplifies to At angular frequency At this point, for the overall transfer matrix Perform singular value decomposition:

[0063] (6);

[0064] in It is a diagonal matrix containing singular values, denoted as ,but , are real numbers and Here, we consider using the root mean square (rms) of the overall MDL. rms Composed of 2×4 mode-related loss factors The calculations provide a relatively complete loss distribution:

[0065] (7);

[0066] For h in the output-input light field of k=12 groups ( The output-input amplitude matrix is ​​obtained by arranging and combining groups of different columns. It has been verified that when h is 10, the mean value of the calculated MDL shows a convergent trend. The corresponding transfer matrix T can be obtained by least squares calculation using equation (5), and then the MDL under each combination can be obtained by formulas (6) and (7). rms It should be noted that, to suppress data distortion caused by experimental environmental noise and non-ideal factors, the largest singular value with significantly abnormal values ​​was removed from each set of singular values, and the root mean square operation was performed on the remaining singular value sequence. Considering the comparability of results, 30 calculations were performed on each bundle of optical fibers, and the five largest and five smallest results were removed from each bundle. The final result was obtained by averaging the remaining 20 calculations.

[0067] S5. Compare the measured MDL mean from step S1 with the MDL calculated in step S4 to verify the reliability of the transfer matrix method. Simultaneously, correct the loss coefficient settings in the MDL simulation model to complete the MDL measurement; specifically:

[0068] Based on the calculation in step S4 The MDL obtained from the actual measurement in step S1 Scraming The data from both methods are compared to calculate the relative deviation between them. The loss coefficient in the MDL simulation model is then corrected based on the actual measured values ​​to further improve the MDL simulation model.

[0069] In this embodiment, the reliability of the system is verified by comparing the measurement data of the scrambling method and the optical transfer matrix method under different fiber sample lengths (0.57km-14.03km). Figure 7 and Figure 8As shown. Experimental results show that the MDL of the optical fiber measured by the scrambing method for the four bundle lengths are as follows: 0.0484±0.0340dB for a sample length of 0.57km, 0.1134±0.0405dB for a sample length of 3.04km, 0.1305±0.0506dB for a sample length of 7.08km, and 0.1805±0.0396dB for a sample length of 14.03km; while the method of this invention (Optical Transfer Matrix) The mode-dependent loss growth coefficients (MDLs) of the optical fiber measured by the two methods were as follows: 0.0615 ± 0.0179 dB for a sample length of 0.57 km, 0.1378 ± 0.0564 dB for a sample length of 3.04 km, 0.1686 ± 0.0469 dB for a sample length of 7.08 km, and 0.2219 ± 0.0586 dB for a sample length of 14.03 km. The square root fit of the data from the two methods yielded the mode-dependent loss growth coefficients of the optical fiber as 0.0509 ± 0.00346 (Scrambling Method) and 0.0654 ± 0.00417 (Optical Transfer Matrix Method), i.e., MDLs. Scraming and The relative deviation between the two is approximately 28.4%. Based on this measurement result, the measured MDL growth coefficient is fed back into the MDL simulation model to calibrate the loss factor of the simulation model, thereby improving the simulation model.

[0070] Example 2; The difference between this example and Example 1 is that in step S3, the method used is as follows: Figure 9 The spot localization method shown is based on connected component analysis instead of circular Hough transform. It is based on geometric shape detection and uses edge gradients to vote for regions that conform to the circular analytical expression.

[0071] like Figure 9As shown, image binarization first binarizes the input filtered grayscale image through a set threshold. This step transforms the image into a logically clear black-and-white binary mask, achieving initial separation between the light spot region (logic 1) and the background region (logic 0). Next, morphological cleaning refines the binarized image to improve extraction accuracy. The `imfill` operation fills any holes within the light spot, and `bwareaopen` removes residual minute noise from the background, ensuring the integrity and purity of the light spot outline. Feature extraction and filtering then uses the `regionprops` function to perform a depth scan on the cleaned connected regions. The algorithm calculates the centroid, area, and equivalent diameter of each independent region. Subsequently, area thresholding automatically eliminates interference regions that do not conform to the fiber core specifications. Finally, the system summarizes the filtered data from the four fiber cores, outputting a high-precision center coordinate matrix and equivalent radius sequence, thus completing the transformation from pixel layers to physical parameters. The centroid obtained by this method is compared with the centroid obtained in Example 1, and the difference is the mean value. ; Calculated relative amplitude difference The average relative deviation between the final MDL calculation result and Example 1 is approximately 5%. In image processing, this invention can not only use one algorithm, but also optimize image processing for randomly coupled multi-core optical fibers with a larger number of cores and different arrangements, thereby obtaining high-precision spot positioning and region segmentation.

[0072] It should also be noted that, in this specification, terms such as "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0073] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for measuring mode correlation loss in randomly coupled multi-core optical fibers, characterized in that, Includes the following steps: S1. The MDL value of randomly coupled multi-core optical fibers was actually measured using an MDL measurement system based on the perturbation method. S2. The MDL of randomly coupled multi-core optical fibers was actually measured using an MDL measurement system based on the optical transfer matrix method. A spatial light modulator is used to dynamically generate an orthogonal input mode field sequence and couple it into the fiber under test. The corresponding output response image and input reference image are acquired simultaneously to construct the original output-input data set for inverting the transmission matrix. S3. Perform preprocessing on the raw spot images acquired by the CCD, calibrate the inherent error of the MDL measurement system and perform spot localization and region segmentation operations, extract the feature parameters of each fiber core spot, provide a standardized data foundation for the subsequent construction of the output-input matrix, and construct the output-input matrix; S4. Calculate the transmission matrix of the randomly coupled multi-core optical fiber using the least squares method based on the output-input matrix, and then perform singular value decomposition on the transmission matrix to obtain its MDL, where MDL is the result of multiple averaging. S5. Compare the measured MDL mean value from step S1 with the MDL calculated in step S4 to verify the reliability of the transfer matrix method. At the same time, correct the loss coefficient setting in the MDL simulation model to complete the MDL measurement.

2. The method for measuring mode correlation loss in a randomly coupled multi-core optical fiber according to claim 1, characterized in that, The MDL measurement system based on the perturbation method includes a light source module, a coupling module of randomly coupled multi-core optical fibers, and a data detection and acquisition module at the receiving end. The light source module uses a tunable laser light source to emit measurement light; The coupling module for randomly coupled multi-core optical fiber is used to couple the fiber under test to a single-mode patch cord. It uses a fusion splicer to perform a core-to-core operation between the single-mode patch cord and one of the cores of the randomly coupled multi-core optical fiber under test. By adjusting the parameters of the fusion splicer, the coupling effect can be maximized, and precise core-to-core alignment can be achieved. The data detection and acquisition module uses an optical power meter and a corresponding data acquisition line for accurate data acquisition. The perturbation-based MDL measurement system obtains the MDL value based on the following formula: （1）； （2）； Among them, MDL rms The normalized RMS deviation of the total transmittance T of all cores of the optical fiber under test from its average value. This represents the total output power measured by the optical power meter, which means that the cores of all the optical fibers under test are simultaneously coupled to the optical power meter. This represents the input power measured by the optical power meter.

3. The method for measuring mode correlation loss in a randomly coupled multi-core optical fiber according to claim 1, characterized in that, The MDL measurement system based on the optical transfer matrix method includes an optical signal generation and modulation module, a coupling module for randomly coupled multi-core optical fibers, and an optical signal imaging module. The optical signal generation and control module includes a light source, a spatial light modulator, a beam splitter, a λ / 2 waveplate, a polarizer, and a collimator, which are used to generate incident light signals and control and collimate their polarization and spatial mode parameters. A coupling module for randomly coupled multi-core optical fibers is used to couple incident light emitted from the optical signal generation and modulation module into the optical fiber under test. The optical signal imaging module is used to focus and receive the transmitted optical signal, and to complete the optical signal imaging and data acquisition through CCD, and to use it for data post-processing.

4. The method for measuring mode correlation loss in a randomly coupled multi-core optical fiber according to claim 3, characterized in that, The optical signal generation and modulation module includes the following steps: A spatial light modulator is used to simulate the optical function of a spiral phase plate to generate a probe beam; By loading a phase modulation pattern corresponding to a preset topological charge parameter l onto a spatial light modulator, the probe beam is made to generate a spiral phase change around the center of the optical axis; by adjusting the value of the topological charge parameter l, a set of orthogonal input mode fields is dynamically generated to construct the input vector basis for transmission matrix measurement.

5. The method for measuring mode correlation loss in a randomly coupled multi-core optical fiber according to claim 1, characterized in that, Optical signal imaging and data acquisition specifically include: CCD is used to acquire light spot images; before measuring the randomly coupled multi-core optical fiber under test, the exposure parameters of the system under test are calibrated using a reference optical fiber to ensure that the acquired output-input light spot image sequence is in a non-saturated state; the acquired images are grayscaled to obtain a grayscale matrix, and a square root is performed based on the sum of pixels within the calculation area, i.e., by... Acquire the light spot intensity; determine whether the collected data meets the quality requirements of subsequent processing based on whether the spatial distribution of the light spot intensity conforms to the preset Gaussian distribution characteristics.

6. The method for measuring mode correlation loss in a randomly coupled multi-core optical fiber according to claim 1, characterized in that, In step S3, the preprocessing includes: Median filtering is applied to the raw spot images acquired by CCD to suppress specific noise while preserving the edge and morphological details of the spot to the maximum extent, providing high-quality image data for subsequent spot localization and region segmentation. For image processing algorithms for spot localization and region segmentation, more than one algorithm can be selected for comprehensive localization. By optimizing edge details and preserving features, high-precision raw data can be obtained.

7. The method for measuring mode correlation loss in a randomly coupled multi-core optical fiber according to claim 1, characterized in that, The construction of the output-input matrix includes the following steps: k different input light field vectors are generated by adjusting the topological charge l of the spatial light modulator. And couple it into the fiber under test, where k ≥ 2n, and n is the number of fiber cores; A sequence of output light field images is acquired using a CCD with preset exposure parameters, and the amplitude information of each fiber core under different polarization states is extracted to construct the output light field amplitude matrix. As shown in equation (3): （3）； For image acquisition at the input end of the fiber under test, on the reference plane at the input end of the randomly coupled multi-core fiber under test, the truncation method is used to obtain the input spot image sequence corresponding to each topological charge l, and the corresponding input amplitude information is extracted to construct the input optical field amplitude matrix X. FUT As shown in formula (4): （4）。 8. The method for measuring mode correlation loss in a randomly coupled multi-core optical fiber according to claim 7, characterized in that, Step S4 is as follows: Based on the output-input matrix and Using the Moore-Penrose generalized inverse: (5)； Since the input states are linearly independent and strictly orthogonal and uniform, that is... The solution then simplifies to At angular frequency At this point, for the overall transfer matrix Perform singular value decomposition: （6）； in, It is a diagonal matrix containing singular values, denoted as ,but , are real numbers and ; Consider using the root mean square (RMS) of the overall MDL. rms Composed of 2n mode-related loss factors The calculation yielded: （7）； The h groups in the k groups of output-input optical fields are arranged and combined to obtain the output-input amplitude matrix with different columns. The corresponding transmission matrix T is obtained by least squares calculation using formula (5). Then, the MDL under each combination is obtained by formulas (6) and (7). rms , where 2n ≤ h ≤ k.

9. The method for measuring mode correlation loss in a randomly coupled multi-core optical fiber according to claim 1, characterized in that, Step S5 is as follows: Based on the MDL calculated in step S4 Optical TM The MDL obtained from the actual measurement in step S1 Scraming The data from both methods are compared to calculate the relative deviation between them. The loss factor in the MDL simulation model is then corrected based on the actual measured values ​​to further improve the MDL simulation model.

Images