Multi-core optical fiber torsion identification compensation method, system, equipment and medium

Abstract

The invention belongs to the technical field of optical fiber shape sensing, and discloses a multi-core optical fiber torsion identification compensation method, system, device and medium, and the method comprises the steps: reading and preprocessing the strain monitoring data of a multi-core optical fiber shape sensor; inputting the strain input sequence into a strain curvature mapping model; calculating a first-order derivative of a distribution curve of the first curvature component and the second curvature component along the length direction, and detecting an abnormal point of the first-order derivative through a local abnormal factor algorithm so as to identify the torsion position of the optical fiber; setting an optical fiber torsion angle range, and performing global optimization on a torsion angle of an optical fiber torsion position by taking curvature component second-order derivability as a target; and based on the optimal torsion angle set, calculating a compensated first curvature component and a compensated second curvature component of each monitoring point after torsion compensation, and realizing spatial position reconstruction of each monitoring point through a shape reconstruction algorithm. According to the invention, the precision of the optical fiber shape sensing technology in earth and rockfill dam deformation monitoring is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of optical fiber shape sensing, and particularly relates to a multi-core optical fiber twist identification and compensation method, system, device and medium. BACKGROUND

[0002] With the advantages of low cost, convenient construction, and strong terrain applicability, earth-rock dams are widely used in global water conservancy projects. In recent years, the construction height of earth-rock dams has been increasing. Under the high stress environment inside the high dam, the nonlinear mechanical properties of earth-rock fillers are significantly enhanced, making the dam deformation exhibit stronger nonlinearity and time variability. To ensure the safe operation of earth-rock dams throughout their life cycle, it is crucial to carry out long-term and comprehensive internal deformation monitoring. However, traditional internal deformation monitoring mainly relies on instruments such as extensometer displacement meters and water pipe settlement meters, which have problems such as being easily disturbed by the environment, complex installation and maintenance, low automation level, limited measurement range and spatial coverage, and are difficult to meet the needs of long-distance internal deformation monitoring of high earth-rock dams.

[0003] Optical fiber shape sensing is a newly emerging deformation sensing technology in recent years. It mainly relies on the differential strain response generated by the deformation of multi-core optical fibers or fiber bundles with specific spatial distribution, and combines shape reconstruction algorithms to realize the reconstruction of the shape and spatial position of the measured object. Compared with electromagnetic tracking systems, inertial sensor networks, and other deformation sensing technologies, optical fiber shape sensing technology has advantages such as resistance to electromagnetic interference, easy integration and installation, compact and flexible structure, strong spatial continuity, and adaptability to harsh environments. In particular, distributed optical fiber shape sensing technology based on fiber scattering effect can achieve high-precision measurement under long-distance conditions and has shown good application potential in the field of earth-rock dam deformation monitoring.

[0004] However, in the process of long-distance monitoring, optical fiber twisting becomes the main error source affecting the accuracy of shape reconstruction. Twisting can be introduced during fiber packaging or bending, which causes local circumferential angle deviation of off-axis fibers, and further leads to curvature information and spatial position calculation deviation. Since conventional shape reconstruction algorithms such as Frenet-Serret method and parallel transmission method mainly rely on recursive operations of the moving frame, the twisting error will continuously accumulate during the calculation process, causing the shape sensing to lose stability and accuracy. Currently, the main approaches to solving the problem of multi-core optical fiber twisting include using spiral multi-core optical fibers, using multi-core optical fibers with special cross-sections, and using high torsional stiffness substrates to make sensors. Among them, spiral multi-core optical fibers are subjected to periodic axial twisting during the manufacturing process, and the common mode response generated after bending is used to achieve strain decoupling; multi-core optical fibers with special cross-sections have strong bending bias by changing the geometric shape of the sensor cross-section, such as asymmetric D-shaped cross-section, thereby suppressing twisting; high torsional stiffness substrates reduce the impact of twisting by improving the anti-twisting ability of the sensor.

[0005] The above existing solutions can alleviate the influence of fiber twisting to a certain extent, but each has its own limitations. Among them, the spiral multi-core optical fiber has a complex manufacturing process and high cost, and long-distance axial twisting can significantly increase the fiber transmission loss; the special-shaped multi-core optical fiber has anisotropic characteristics in bending sensitivity, limiting the comprehensiveness of deformation monitoring; the high-rigidity sensor has large volume and weight, low flexibility, is not conducive to large-scale layout in engineering, and may reduce the deformation coordination with the object to be measured.

[0006] Therefore, how to provide a solution capable of accurately identifying and compensating the twisting of parallelly arranged conventional multi-core optical fibers and improving the accuracy of optical fiber shape sensing technology in earth-rock dam deformation monitoring has become a problem to be solved at present. SUMMARY

[0007] The embodiments of the present application provide a multi-core optical fiber twisting identification and compensation method, system, device and medium, which can identify and compensate the twisting of parallelly arranged conventional multi-core optical fibers, thereby improving the accuracy of optical fiber shape sensing technology in earth-rock dam deformation monitoring, and has the advantages of not depending on specific optical fiber structure and type, low cost, easy implementation, high sensitivity, high compensation accuracy, strong compatibility and the like, so as to solve the problems of low shape reconstruction accuracy and cumulative influence of twisting error on monitoring reliability caused by multi-core optical fiber twisting in the prior art.

[0008] According to a first aspect of the embodiments of the present application, a multi-core optical fiber twisting identification and compensation method is provided.

[0009] In one embodiment, the multi-core optical fiber twisting identification and compensation method comprises: reading strain monitoring data of a multi-core optical fiber shape sensor and preprocessing to construct a strain input sequence; inputting the strain input sequence into a strain-curvature mapping model to calculate a first curvature component and a second curvature component of each monitoring point; calculating the first-order derivative of the distribution curve of the first curvature component and the second curvature component along the length direction, detecting the abnormal points of the first-order derivative by a local anomaly factor algorithm, and identifying the monitoring points corresponding to the abnormal points as the fiber twisting positions; setting a fiber twisting angle range, and taking the second-order derivable of the curvature component as the target, globally optimizing the twisting angle of the identified fiber twisting positions to obtain an optimal twisting angle set satisfying the curvature component smoothness condition; based on the optimal twisting angle set, calculating the compensation first curvature component and the compensation second curvature component of each monitoring point after twisting compensation, and realizing the spatial position reconstruction of each monitoring point through the recursive operation of the shape reconstruction algorithm.

[0010] In one embodiment, the strain-curvature mapping model is based on a symmetric parallel multi-core optical fiber shape sensor, and the first curvature component and the second curvature component of each monitoring point are calculated according to the mapping relationship between the fiber strain and the curvature information under three-dimensional conditions; the calculation of the first curvature component and the second curvature component of each monitoring point includes: obtaining the total number of optical fibers, the serial numbers of the optical fibers, the strain of each optical fiber at the monitoring point, the angular deviation of each optical fiber core to the positive direction of the first direction, and the distance of each optical fiber core to the central axis; based on the strain of each optical fiber at the monitoring point, the angular deviation of each optical fiber core to the positive direction of the first direction, and the distance of each optical fiber core to the central axis, the sum of all optical fibers is calculated to obtain the first curvature component and the second curvature component.

[0011] In one embodiment, the first derivative is calculated by using a first-order forward difference approximation; and the abnormal points of the first derivative are detected by using a local outlier factor algorithm, which includes: calculating the local outlier factor value of the first derivative of each monitoring point to represent the degree of outlying; setting a critical local outlier factor value, and identifying the monitoring points with the local outlier factor value exceeding the critical local outlier factor value as abnormal points, and identifying the monitoring points corresponding to the abnormal points as the fiber twist positions.

[0012] In one embodiment, the calculation of the local outlier factor value of the first derivative of each monitoring point includes: for a sample point in a data set, defining the specified neighbor distance of the sample point as the distance from the sample point to its first specified number of nearest neighbor points, defining the specified neighbor neighborhood of the sample point as a set of data points with a distance less than the specified neighbor distance from the sample point; defining the reachable distance of the sample point and the neighborhood point as the larger value of the specified neighbor distance of the neighborhood point and the Euclidean distance between the sample point and the neighborhood point; calculating the local reachable density of the sample point based on the average reachable distance of the specified neighbor neighborhood points of the sample point to the sample point; and obtaining the local outlier factor value of the sample point based on the average ratio of the local reachable density of all neighborhood points in the specified neighbor neighborhood to the sample point.

[0013] In one embodiment, the global optimization of the twist angle of the identified fiber twist position is performed by taking the second-order derivability of the curvature component as the target, which includes: applying a twist compensation angle to the identified fiber twist position, and calculating the compensated first curvature component and the compensated second curvature component of each monitoring point after the twist compensation angle is applied; calculating the second-order right derivative of the compensated first curvature component and the compensated second curvature component at each fiber twist position by using a second-order forward difference approximation, and calculating the second-order left derivative of the compensated first curvature component and the compensated second curvature component at each fiber twist position by using a second-order backward difference approximation; setting a fitness function of a global optimization algorithm; and optimizing the twist compensation angle by using the global optimization algorithm to obtain an optimal twist angle set that minimizes the fitness function.

[0014] According to a second aspect of the embodiments of the present application, a multi-core optical fiber twist identification and compensation system is provided.

[0015] In one embodiment, the multi-core fiber torsion identification and compensation system includes: a data preprocessing module for reading and preprocessing strain monitoring data from a multi-core fiber shape sensor to construct a strain input sequence; a curvature calculation module for inputting the strain input sequence into a strain curvature mapping model to calculate the first curvature component and the second curvature component at each monitoring point; a torsion identification module for calculating the first derivative of the distribution curves of the first and second curvature components along the length direction, detecting anomalous points of the first derivative using a local anomaly factor algorithm, and identifying the monitoring point corresponding to the anomalous point as the fiber torsion location; an angle optimization module for setting the fiber torsion angle range, and globally optimizing the torsion angle of the identified fiber torsion location with the goal of second-order differentiability of the curvature component, to obtain the optimal set of torsion angles that satisfy the curvature component smoothness condition; and a shape reconstruction module for calculating the compensated first curvature component and compensated second curvature component at each monitoring point after torsion compensation based on the optimal set of torsion angles, and reconstructing the spatial position of each monitoring point through recursive operation of the shape reconstruction algorithm.

[0016] According to a third aspect of the present invention, a computer device is provided.

[0017] In one embodiment, the computer device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method described above.

[0018] According to a fourth aspect of the present invention, a computer-readable storage medium is provided.

[0019] In one embodiment, a computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of the method described above.

[0020] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0021] (1) The multi-core optical fiber torsion identification and compensation method proposed in this invention is based on the curve properties of curvature components and their derivatives. It integrates the local anomaly factor algorithm and the intelligent optimization algorithm. The first curvature component and the second curvature component of each monitoring point are calculated through the strain curvature mapping model. Based on the first derivative of the curvature component distribution curve along the length direction, the local anomaly factor algorithm is used to detect anomaly points and identify the torsion position of the optical fiber. Then, with the second-order differentiability of the curvature component as the target, the optimal torsion angle set is calculated through the global optimization algorithm for compensation. Finally, based on the compensated curvature component, the spatial position of each monitoring point is reconstructed through the shape reconstruction algorithm. This method can achieve high-precision identification and compensation of conventional multi-core optical fiber torsion, thereby significantly improving the accuracy and reliability of optical fiber shape sensing technology in the deformation monitoring inside earth-rock dams.

[0022] (2) Compared with existing multi-core fiber torsion solutions, the present invention has high sensitivity in torsion detection and high compensation accuracy. It does not depend on specific fiber structures and types, and does not require the use of spiral multi-core fibers, irregular cross-section multi-core fibers or high torsional stiffness substrates. This greatly reduces manufacturing difficulty and production costs. Furthermore, it can maintain the original slender characteristics and compact structure of the sensor, making it easy to deploy in the limited space and complex environment inside earth-rock dams. At the same time, it has strong compatibility and is easy to implement, and can be directly embedded in existing multi-core fiber sensing systems.

[0023] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0024] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0025] Figure 1 This is a flowchart illustrating a multi-core optical fiber torsion identification and compensation method according to an exemplary embodiment;

[0026] Figure 2 This is a detailed implementation diagram illustrating a multi-core optical fiber torsion identification and compensation method according to an exemplary embodiment;

[0027] Figure 3 This is a schematic cross-sectional view of the multi-core fiber shape sensor before and after twisting in a multi-core fiber twisting identification and compensation method according to an exemplary embodiment;

[0028] Figure 4 This is a schematic diagram of the curvature component distribution curve after the midpoint of the sensor is twisted, according to an exemplary embodiment of a multi-core optical fiber torsion identification and compensation method;

[0029] Figure 5 This is a spatial position diagram of a simulation model of a symmetrical four-core fiber shape sensor built in a multi-core fiber torsion identification and compensation method according to an exemplary embodiment.

[0030] Figure 6 This is a strain distribution diagram of each off-axis fiber in a multi-core fiber simulation model after torsion, according to an exemplary embodiment of a multi-core fiber torsion identification and compensation method.

[0031] Figure 7 This is a distribution diagram of the LOF values ​​of the first-order derivative of the first curvature component of a multi-core optical fiber simulation model after torsion, in a multi-core optical fiber torsion identification and compensation method according to an exemplary embodiment.

[0032] Figure 8This is a distribution diagram of the LOF values ​​of the first derivative of the second curvature component of a multi-core optical fiber simulation model after torsion, in a multi-core optical fiber torsion identification and compensation method according to an exemplary embodiment.

[0033] Figure 9 This is a schematic diagram illustrating the shape restoration results of a multi-core optical fiber simulation model before and after torsion compensation in a multi-core optical fiber torsion identification and compensation method according to an exemplary embodiment;

[0034] Figure 10 This is a schematic diagram illustrating the shape restoration error of a multi-core optical fiber simulation model before and after torsion compensation in a multi-core optical fiber torsion identification and compensation method according to an exemplary embodiment;

[0035] Figure 11 This is a schematic block diagram illustrating a multi-core fiber optic torsion identification and compensation system according to an exemplary embodiment;

[0036] Figure 12 This is a schematic diagram of the structure of a computer device according to an exemplary embodiment. Detailed Implementation

[0037] The following description and accompanying drawings fully illustrate specific embodiments described herein to enable those skilled in the art to practice them. Some portions and features of certain embodiments may be included in or replace portions and features of other embodiments. The scope of the embodiments herein includes the entire scope of the claims and all available equivalents thereof. The various embodiments described herein are presented in a progressive manner, with each embodiment focusing on its differences from other embodiments; similar or identical parts between embodiments can be referred to interchangeably.

[0038] The modules in the apparatus or system of this application can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0039] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0040] Figures 1-2 An embodiment of a multi-core optical fiber torsion identification and compensation method of the present invention is shown.

[0041] In this optional embodiment, the multi-core fiber torsion identification and compensation method includes:

[0042] Step S101: Read and preprocess the strain monitoring data of the multi-core fiber shape sensor to construct a strain input sequence;

[0043] Step S102: Input the strain input sequence into the strain curvature mapping model to calculate the first curvature component and the second curvature component of each monitoring point; wherein, the first curvature component is the curvature component in the first direction of the local coordinate system, and the second curvature component is the curvature component in the second direction of the local coordinate system.

[0044] Step S103: Calculate the first derivative of the distribution curves of the first curvature component and the second curvature component along the length direction, detect the abnormal points of the first derivative using the local anomaly factor algorithm, and identify the monitoring point corresponding to the abnormal point as the fiber torsion position.

[0045] Step S104: Set the range of fiber twist angles, and with the goal of second-order differentiability of curvature components, perform global optimization on the twist angles of the identified fiber twist positions to obtain the optimal set of twist angles that satisfy the curvature component smoothness condition.

[0046] Step S105: Based on the optimal set of torsion angles, calculate the first curvature component and the second curvature component after torsion compensation for each monitoring point, and reconstruct the spatial position of each monitoring point through the recursive operation of the shape reconstruction algorithm.

[0047] In this optional embodiment, the strain-curvature mapping model is based on a symmetrically arranged multi-core fiber shape sensor. According to the mapping relationship between fiber strain and curvature information under three-dimensional conditions, the first curvature component and the second curvature component of each monitoring point are calculated. The calculation of the first curvature component and the second curvature component of each monitoring point includes: obtaining the total number of fibers, the serial number of each fiber, the strain of each fiber at the monitoring point, the angular offset of each fiber core from the positive direction of the first direction, and the distance of each fiber core from the central axis. Based on the strain of each fiber at the monitoring point, the angular offset of each fiber core from the positive direction of the first direction, and the distance of each fiber core from the central axis, the summation is performed on all fibers to obtain the first curvature component and the second curvature component.

[0048] In this optional embodiment, the expression for calculating the first curvature component at each monitoring point is:

[0049]

[0050] The expression for calculating the second curvature component at each monitoring point is as follows:

[0051]

[0052] In the formula, κ1 and κ2 are the curvature components in the y and z directions of the local coordinate system, respectively; n is the total number of optical fibers; i is the fiber index, ranging from 1 to n; ε i Let θ be the strain of fiber i at the monitoring point; i r is the angular offset of the i-th fiber core from the positive y-axis;i Let be the distance from the core of the i-th fiber to the central axis.

[0053] In this optional embodiment, the first derivative is calculated using the first-order forward difference approximation; outliers of the first derivative are detected by the local anomaly factor algorithm, including: calculating the local anomaly factor value of the first derivative of each monitoring point to characterize the degree of outlier; setting a critical local anomaly factor value, identifying monitoring points whose local anomaly factor value exceeds the critical local anomaly factor value as outliers, and identifying the monitoring points corresponding to the outliers as fiber torsion positions.

[0054] In this optional embodiment, calculating the local anomaly factor value of the first derivative of each monitoring point includes: for sample points in the dataset, defining the specified nearest neighbor distance of a sample point as the distance from the sample point to its specified number of nearest neighbors, defining the specified nearest neighbor neighborhood of a sample point as the set of data points whose distance from the sample point is less than the specified nearest neighbor distance; defining the reachable distance between a sample point and its neighboring points as the larger value between the specified nearest neighbor distance of the neighboring points and the Euclidean distance between the sample point and its neighboring points; calculating the local reachability density of the sample point based on the average reachable distance from the specified nearest neighbor points to the sample point; and obtaining the local anomaly factor value of the sample point based on the average ratio of the local reachability density of all neighboring points in the specified nearest neighbor neighborhood to that of the sample point.

[0055] In this optional embodiment, the expression for the reachability distance between a sample point and its neighboring points is:

[0056] reach-dist k (x,y)=max{k-dist(y),d(x,y)};

[0057] In the formula, d(x,y) is the Euclidean distance between sample points x and y; k is the distance from the k-th nearest neighbor to sample point x;

[0058] The expression for the local reachability density of a sample point is:

[0059]

[0060] In the formula, lrd k (x) represents the local reachability density of sample point x; |N k (x)| represents the number of elements in the k-distance neighborhood of sample point x;

[0061] The expression for the local anomaly factor value of the sample point is:

[0062]

[0063] In the formula, LOF k (x) represents the local anomaly factor value of sample point x.

[0064] In this optional embodiment, with the goal of achieving second-order differentiability of the curvature components, the global optimization of the torsion angle at the identified fiber torsion positions includes: applying a torsion compensation angle to the identified fiber torsion positions, and calculating the compensated first curvature component and the compensated second curvature component at each monitoring point after applying the torsion compensation angle; approximating the second-order right derivative of the compensated first curvature component and the compensated second curvature component at each fiber torsion position using second-order forward difference, and approximating the second-order left derivative of the compensated first curvature component and the compensated second curvature component at each fiber torsion position using second-order backward difference; setting the fitness function of the global optimization algorithm; wherein, the fitness function is the sum of the squares of the differences between the second-order left derivative and the second-order right derivative of the compensated first curvature component and the compensated second curvature component at each fiber torsion position; and optimizing the torsion compensation angle using the global optimization algorithm to obtain the optimal set of torsion angles that minimizes the fitness function.

[0065] In this optional embodiment, the expression for the fitness function is:

[0066]

[0067] In the formula, y is the fitness function; Δs is the strain sampling interval; M is the total number of fiber torsion positions identified by the local anomaly factor algorithm; t i The monitoring point number corresponding to the i-th torsion point; To compensate for the first curvature component; This is for the second curvature component after compensation.

[0068] In this optional embodiment, the shape reconstruction algorithm employs a parallel transmission method, using a fourth-order Runge-Kutta method for recursive calculation. Specifically, it includes: constructing a parallel transmission frame containing a unit tangent vector, a first normal vector, and a second normal vector; establishing the variation relationship of the parallel transmission frame's arc-length parameters based on the compensated first curvature component and the compensated second curvature component; recursively solving this variation relationship using the fourth-order Runge-Kutta method to obtain the parallel transmission frame for each monitoring point; and integrating the unit tangent vector of each monitoring point to reconstruct the spatial position of each monitoring point. The expression for the variation relationship of the parallel transmission frame's arc-length parameters is as follows:

[0069]

[0070] In the formula, l is the arc length parameter, and T(l), N1(l), and N2(l) are the unit tangent vector, the first normal vector, and the second normal vector of the parallel transmission frame, respectively.

[0071] In this optional embodiment, the preprocessing includes gross error removal, multiple sampling averaging, and initial strain correction; multiple sampling averaging includes: continuously collecting at least three strain data during monitoring and calculating the average value to reduce random errors; initial strain correction includes: subtracting the strain reference value under the initial straight-line state before performing curvature calculation.

[0072] It should be noted that, as Figure 1 and Figure 2 As shown, the multi-core optical fiber torsion identification and compensation method proposed in this invention includes the following steps:

[0073] Step 1: Read the strain monitoring data from the multi-core fiber shape sensor, and perform data preprocessing such as gross error removal, multiple sampling averaging and initial strain correction to construct the strain input sequence.

[0074] Among them, multiple sampling average refers to collecting strain data three or more times consecutively during monitoring and reducing random errors by calculating the average value; initial strain correction is to subtract the strain reference value under the initial straight line state before performing curvature calculation.

[0075] Step ②: Input the strain data preprocessed in Step ① into the mapping model of fiber strain and curvature information, and calculate the curvature components κ1 and κ2 of each monitoring point after being affected by fiber torsion.

[0076] For a symmetrically arranged multi-core fiber shape sensor, based on the mapping relationship between fiber strain and curvature information under three-dimensional conditions, the curvature components (first curvature component and second curvature component) at each monitoring point can be calculated by the following formula:

[0077]

[0078] In the formula: κ1 and κ2 are the curvature components in the y and z directions of the local coordinate system, respectively; n is the total number of optical fibers; i is the fiber index, ranging from 1 to n; ε i Let θ be the strain of fiber i at the monitoring point; i r is the angular offset of the i-th fiber core from the positive y-axis; i Let be the distance from the core of the i-th fiber to the central axis.

[0079] Specifically, a symmetrical four-core fiber is selected as an example of a multi-core fiber, such as... Figure 3 As shown, when the sensor is bent but not twisted, the strain of fibers 1, 2, 3, and 4 is linearly related to their distance from the neutral axis. Fibers 1 and 4, located on the same side of the bending direction, experience compressive strain, while those on the opposite side experience tensile strain. Therefore, the curvature components can be directly calculated using equations (1) to (2). If the sensor bends at an angle of... The twisting, that is, the simultaneous occurrence of an angle of 1 to 4 in optical fibers. A local circumferential angular shift will change the position of the optical fiber in the local coordinate system to 1′~4′, accompanied by a sudden change in fiber strain. If the initial angular shift θ is still maintained at this point... i Performing curvature calculations using (i = 1~4) will cause misalignment of the curvature component distribution curves, such as... Figure 4 As shown. From the perspective of curve properties, fiber twisting causes the distribution curve of the curvature component along the length direction to lose smoothness at the twist point, that is, the first derivative of the curvature component is discontinuous at the twist point. Therefore, the anomaly points in the distribution of the first derivative of the curvature component can be used as the criterion for twisting in multi-core fiber shape sensors.

[0080] Step 3: Calculate the first derivative of the curvature component distribution curve, detect and identify the abnormal points of the curvature component through the Local Anomaly Factor (LOF) algorithm, treat them as fiber torsion positions and output them to the intelligent optimization algorithm.

[0081] In this study, considering that the strain sampling sequence is discrete data based on the sampling interval, the first derivative of the curvature component is approximated using the first-order forward difference. The specific steps for identifying fiber optic torsion using the local anomaly factor algorithm are as follows: Based on obtaining the first derivative of the curvature component at each monitoring point, the LOF value of the first derivative of the curvature component at each point is calculated to characterize its outlier degree; by setting a critical LOF value, outlier points with a large outlier degree are screened, that is, monitoring points whose LOF value exceeds the set range are considered to be fiber torsion locations (i.e., if the LOF value of a certain point exceeds the set range, it is considered that the fiber has torsion at that point).

[0082] Specifically, the Local Outlier Factor (LOF) algorithm is an anomaly detection algorithm based on data point density. Unlike traditional anomaly detection methods, this algorithm does not rely on the global distribution of the data or prior training. Instead, it determines whether a sample deviates from dense data by calculating the local outlier factor of the data sample, thus identifying anomalies. For a dataset D = {x1, x2, ..., x...} n In the context of a sample point x, its k-distance (specified nearest neighbor distance) k-dist(x) is defined as the distance from the k-th nearest neighbor (the specified number of nearest neighbors) to sample point x. Data points in D whose distance to sample point x is less than k-dist(x) are extracted into a set (excluding sample point x), and this set is defined as the k-distance neighborhood (specified nearest neighbor neighborhood) of sample point x. k (x).

[0083] Specifically, the reach-dist between sample point x and sample point y is defined. k (x,y) is:

[0084] reach-dist k(x,y)=max{k-dist(y),d(x,y)}(3)

[0085] In the formula, d(x,y) is the Euclidean distance between sample points x and y.

[0086] Specifically, the local density of x is estimated by the average reachability distance from k-distance neighboring points to x:

[0087]

[0088] In the formula, lrd k (x) represents the local reachability density of sample point x; |N k (x)| represents the number of elements in the k-distance neighborhood of the sample point x.

[0089] Specifically, the Local Outlier Factor (LOF) of sample point x can be derived based on the local reachability density. k (x), which is defined as N k The average ratio of the local reachability density of all data points in (x) to that of the sample point x can be expressed as:

[0090]

[0091] Specifically, the LOF value characterizes the outlier degree of a sample point. If the LOF value of a sample point is significantly greater than 1, it indicates that the local density of that point differs greatly from the local density of its neighbors, and it is considered an outlier. If the LOF value of a sample point is close to 1, it indicates that the sample point and its neighbors have roughly the same density, belonging to the same cluster. The method of this invention identifies fiber torsion by detecting whether the LOF value of the first derivative of the curvature component at each monitoring point exceeds a critical value. The LOF critical value can be selected according to the actual situation; in this embodiment, it is set to 2.

[0092] Step 4: Set the range of fiber optic twist angles. With the goal of second-order differentiability of the curvature component, use an intelligent optimization algorithm (global optimization algorithm) to globally optimize the twist angles of each twist point identified in Step 3, and obtain the optimal set of twist angles that satisfy the curvature component smoothness condition.

[0093] The initial optimization range for the fiber torsion angle is usually set to [-0.5π, 0.5π], and can be adjusted according to the quality of the optimization results. If it is difficult to obtain the optimal torsion angle of each torsion point through multiple optimizations, the optimization range can be gradually expanded. The intelligent optimization algorithm can adopt a metaheuristic algorithm or a hybrid intelligent algorithm with superior global optimization capabilities.

[0094] Specifically, assuming step ③ identifies a fiber optic twist at point M, then for the twist point t... m (m=1,2,...,M) The applied angle is Torsional compensation (torsional compensation angle) results in the curvature components at each monitoring point after compensation. (The first curvature component and the second curvature component after compensation) can be expressed as:

[0095]

[0096] In the formula, s is the monitoring point number, s = 0, 1, ..., L, and s = 0 and s = L correspond to the sensor start and end points, respectively; t m , These represent the monitoring point number and torsional compensation angle corresponding to the m-th torsion point, respectively, where m = 1, 2, ..., M, and t. M+1 =L.

[0097] Specifically, to ensure the curvature component distribution curve satisfies the smoothness condition after torsional compensation, i.e., the first derivative of the curvature component is continuous at the torsion point, the second-order differentiability of the curvature component is used as the optimization objective of the intelligent optimization algorithm. Considering the discrete characteristics of the strain sampling sequence, when calculating the second derivative of the curvature component, the second-order forward difference is used to approximate the second-order right derivative, and the second-order backward difference is used to approximate the second-order left derivative. The minimum difference between the second-order left and right derivatives of the curvature component is used to approximate second-order differentiability. Therefore, the fitness function y of the intelligent optimization algorithm can be set as the sum of the squares of the differences between the second-order left and right derivatives of the curvature component at the torsion point, which can be expressed as:

[0098]

[0099] In the formula, y is the fitness function; Δs is the strain sampling interval; M is the total number of fiber torsion positions identified by the LOF algorithm; t i The monitoring point number corresponding to the i-th torsion point; To compensate for the first curvature component; This is for the second curvature component after compensation.

[0100] Step 5: Based on the optimization results of the torsion angle at each torsion point, calculate the curvature components of each monitoring point after torsion compensation using equations (6) to (7). The spatial location of each monitoring point is reconstructed through recursive operations of the shape reconstruction algorithm.

[0101] Among them, the spatial position of each monitoring point is reconstructed based on the curvature components after torsional compensation. The shape reconstruction algorithm adopts the parallel transmission method, which can be recursively calculated using numerical methods such as the fourth-order Runge-Kutta method. The core formula of the parallel transmission method is:

[0102]

[0103] In the formula, l is the arc length parameter; T(l), N1(l), and N2(l) are the unit tangent vector, first normal vector, and second normal vector of the Bishop frame (parallel transmission frame), respectively; T′(l), N1′(l), and N2′(l) are the first derivatives of the unit tangent vector, first normal vector, and second normal vector of the Bishop frame, respectively.

[0104] Specifically, taking the fourth-order Runge-Kutta method as an example to solve equation (9), the Bishop frame at the (n+1)th monitoring point can be solved by the following equation:

[0105]

[0106] In the formula, T((n+1)Δs) is the unit tangent vector at the (n+1)th monitoring point; N1((n+1)Δs) is the first normal vector at the (n+1)th monitoring point; and N2((n+1)Δs) is the second normal vector at the (n+1)th monitoring point.

[0107] Where, parameter a i b i c i The expressions for (i = 1 to 4) are as follows:

[0108]

[0109] In the formula, a i b i c i To solve the intermediate parameters of the Bishop frame using the fourth-order Runge-Kutta method.

[0110] Specifically, after solving for the parallel transmission frame along the sensor line, the spatial position can be reconstructed by integrating the unit tangent vector. In the discrete form, numerical methods such as the Euler method and the trapezoidal rule are commonly used for calculation. Taking the trapezoidal rule as an example, the spatial position of the (n+1)th monitoring point can be expressed as:

[0111]

[0112] In the formula, R(s) is the position vector at the sensor monitoring point s.

[0113] To facilitate understanding of the above technical solution of the present invention, a specific explanation is given below using a four-core optical fiber simulation model as an example:

[0114] To verify the performance of the proposed method, in one embodiment, the multi-core fiber torsion identification and compensation method of the present invention is used to conduct a torsion compensation experiment on a multi-core fiber simulation model, specifically including the following steps:

[0115] Step S101, with Figure 3Taking the symmetrical four-core fiber shape sensor shown as an example, a simulation model of the deformation of multi-core optical fibers under stress is constructed. It should be understood that the four-core fiber structure used is exemplary; this invention is applicable to all parallel-arranged multi-core optical fibers or fiber bundles, but is not limited to them. Assuming the total length of the sensor is 10m, and the shape of the central axis after deformation is a helix, its parametric equation is:

[0116]

[0117] The spatial position of the multi-core fiber simulation model is as follows: Figure 5 As shown (for ease of viewing, the distance from the fiber core to the central axis has been appropriately enlarged; the actual distance is taken as 636.40 μm), with the strain sampling interval set to 20 mm, the number of monitoring points for the sensor is 501. To verify the torsional compensation performance of the method proposed in this invention, a torsion angle of 10° was added at the 150th, 300th, and 450th monitoring points, respectively. The strain distribution of each off-axis fiber after torsion is shown below. Figure 6 As shown.

[0118] Step S102: Input the strain data sequence after fiber torsion into the mapping model of fiber strain and curvature information, calculate the curvature components κ1 and κ2 of each monitoring point after being affected by fiber torsion using equations (1) to (2), and use the first-order forward difference approximation to calculate its first derivative.

[0119] Step S103: Set the critical LOF value to 2, and use the LOF algorithm to detect outliers in the first derivatives of the curvature components at each point. The LOF value distributions of the first derivatives of curvature components κ1 and κ2 are as follows: Figure 7 and Figure 8 As shown in the figure, there are three obvious abrupt changes in the first derivatives of both κ1 and κ2. The LOF values ​​of the abnormal samples of κ1 are 310.70 (monitoring point 150), 126.00 (monitoring point 300), and 3132.69 (monitoring point 450), respectively. The LOF values ​​of the abnormal samples of κ2 are 423.64 (monitoring point 150), 784.96 (monitoring point 300), and 534.07 (monitoring point 450), respectively.

[0120] The anomaly detection results of the LOF algorithm show that the locations of the abnormal samples identified by the algorithm in the first derivatives of κ1 and κ2 are consistent with the actual twist locations. Furthermore, the LOF values ​​at the abnormal sample points are all greater than the critical value (much greater than 1), indicating that the multi-core optical fiber has undergone significant twisting at these points. This fully verifies the effectiveness and accuracy of the method of the present invention in optical fiber twisting identification.

[0121] Step S104: Input the fiber torsion position information identified in step S3 into the intelligent optimization algorithm, set the fiber torsion angle range to [-0.5π, 0.5π], and the fitness function to be the sum of the squares of the differences between the second left and right derivatives of the curvature components κ1 and κ2 at all torsion points. Search for the optimal torsion angle at each torsion point through the intelligent optimization algorithm.

[0122] In this embodiment, the Sand Cat Swarm Optimization (SCSO) algorithm, known for its superior optimization performance, is employed. Inspired by the sand cat's hunting and catching methods, this algorithm incorporates adaptively adjusted auditory sensitivity, a randomly initialized search space, and random angles, effectively avoiding local optima and solving complex optimization problems. Compared to other metaheuristic algorithms, it has fewer parameters and converges faster. During the search process, the relevant parameters of the SCSO algorithm are set as follows: maximum number of iterations is 300, sand cat population size is 30, and maximum sensitivity range is 2. After the 198th iteration, the SCSO algorithm converges to the optimal result, yielding a sum of squares of the differences in the least second derivatives of 0.168. The corresponding optimal torsional compensation angles are 10.002°, 10.012°, and 10.014°, showing minimal deviation from the actual torsional angle values.

[0123] Step S105: Based on the optimization results of the torsion angle at each torsion point, calculate the curvature components of each monitoring point after torsion compensation using equations (6) to (7). The shape is reconstructed using a parallel transmission method.

[0124] The shape restoration results and restoration errors before and after torsion compensation are as follows: Figure 9 and Figure 10 As shown in the figure, the shape reconstruction results and error distribution diagram show that after using the method proposed in this invention for torsion identification and compensation, the reconstructed curve is highly consistent with the original curve, and the error accumulation is significantly improved. The maximum error decreases from 922.378 mm after torsion to 1.476 mm, indicating that the method proposed in this invention has strong torsion compensation performance, and the fiber shape reconstruction accuracy is significantly improved after torsion compensation.

[0125] Based on the analysis of the influence of fiber torsion on the smoothness of curvature components, the method of this invention detects and identifies the torsion position of the fiber through a local anomaly factor algorithm, and then uses an intelligent optimization algorithm to calculate the optimal torsion compensation angle. Finally, based on the curvature information after torsion compensation, high-precision shape reconstruction is achieved. It has the advantages of low cost, easy implementation, strong compatibility and high reliability, and can effectively improve the accuracy and reliability of fiber optic shape sensing technology in the internal deformation monitoring of earth-rock dams.

[0126] Figure 11An embodiment of a multi-core fiber optic torsion identification and compensation system of the present invention is shown.

[0127] In this optional embodiment, the multi-core fiber optic torsion recognition and compensation system includes: a data preprocessing module 201, used to read and preprocess strain monitoring data from a multi-core fiber optic shape sensor to construct a strain input sequence; a curvature calculation module 202, used to input the strain input sequence into a strain curvature mapping model to calculate the first curvature component and the second curvature component at each monitoring point; wherein the first curvature component is the curvature component in the first direction of the local coordinate system, and the second curvature component is the curvature component in the second direction of the local coordinate system; and a torsion recognition module 203, used to calculate the first derivative of the distribution curves of the first curvature component and the second curvature component along the length direction. The system employs a local anomaly factor algorithm to detect anomalies in the first derivative and identifies the monitoring points corresponding to these anomalies as fiber torsion locations. An angle optimization module 204 sets the range of fiber torsion angles and, with the goal of second-order differentiability of the curvature components, performs global optimization on the torsion angles of the identified fiber torsion locations to obtain the optimal set of torsion angles that satisfy the curvature component smoothness condition. A shape reconstruction module 205, based on the optimal set of torsion angles, calculates the compensated first curvature component and the compensated second curvature component of each monitoring point after torsion compensation and reconstructs the spatial position of each monitoring point through recursive operations of the shape reconstruction algorithm.

[0128] In this optional embodiment, when calculating the first curvature component and the second curvature component at each monitoring point, the curvature calculation module 202 includes: obtaining the total number of optical fibers, the serial number of each optical fiber, the strain of each optical fiber at the monitoring point, the angular offset of each optical fiber core from the positive direction of the first direction, and the distance of each optical fiber core from the central axis; and summing all optical fibers based on the strain of each optical fiber at the monitoring point, the angular offset of each optical fiber core from the positive direction of the first direction, and the distance of each optical fiber core from the central axis to obtain the first curvature component and the second curvature component.

[0129] In this optional embodiment, when the torsion identification module 203 detects outliers of the first derivative using the local anomaly factor algorithm, it includes: calculating the local anomaly factor value of the first derivative of each monitoring point to characterize the degree of outlier; setting a critical local anomaly factor value; identifying monitoring points whose local anomaly factor values ​​exceed the critical local anomaly factor value as outliers; and identifying the monitoring points corresponding to the outliers as fiber torsion positions.

[0130] In this optional embodiment, when the angle optimization module 204 performs global optimization of the torsion angle of the identified fiber torsion position with the goal of achieving second-order differentiability of the curvature component, it includes: applying a torsion compensation angle to the identified fiber torsion position and calculating the compensated first curvature component and the compensated second curvature component at each monitoring point after applying the torsion compensation angle; approximating the second-order right derivative of the compensated first curvature component and the compensated second curvature component at each fiber torsion position using second-order forward difference, and approximating the second-order left derivative of the compensated first curvature component and the compensated second curvature component at each fiber torsion position using second-order backward difference; setting the fitness function of the global optimization algorithm; wherein the fitness function is the sum of squares of the differences between the second-order left derivative and the second-order right derivative of the compensated first curvature component and the compensated second curvature component at each fiber torsion position; and optimizing the torsion compensation angle through the global optimization algorithm to obtain the optimal set of torsion angles that minimizes the fitness function.

[0131] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 12 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores static and dynamic information data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the above method embodiments.

[0132] Those skilled in the art will understand that Figure 12 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0133] In addition, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0134] In addition, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0135] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0136] This invention is not limited to the structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this invention is limited only by the appended claims.

Claims

1. A method for identifying and compensating for torsion in multi-core optical fibers, characterized in that, The method comprises the following steps: reading and preprocessing strain monitoring data of a multi-core optical fiber shape sensor to construct a strain input sequence; inputting the strain input sequence into a strain curvature mapping model to calculate a first curvature component and a second curvature component of each monitoring point; wherein the first curvature component is a curvature component in the first direction of a local coordinate system, and the second curvature component is a curvature component in the second direction of the local coordinate system; calculating the first-order derivative of the distribution curve of the first curvature component and the second curvature component along the length direction, detecting an abnormal point of the first-order derivative by a local anomaly factor algorithm, and identifying the monitoring point corresponding to the abnormal point as a fiber twist position; setting a fiber twist angle range, and performing global optimization on the twist angle of the identified fiber twist position by taking the second-order derivative of the curvature component as the target to obtain an optimal twist angle set that satisfies the smoothness condition of the curvature component; based on the optimal twist angle set, calculating the first curvature component and the second curvature component of each monitoring point after twist compensation, and realizing the spatial position reconstruction of each monitoring point through recursive operation of the shape reconstruction algorithm.

2. The method of claim 1, wherein, The strain curvature mapping model is based on a symmetric parallel multi-core optical fiber shape sensor, and calculates the first curvature component and the second curvature component of each monitoring point according to the mapping relationship between the fiber strain and the curvature information under three-dimensional conditions. The calculation of the first curvature component and the second curvature component of each monitoring point comprises: obtaining the total number of optical fibers, the serial number of each optical fiber, the strain of each optical fiber at the monitoring point, the angular offset of each optical fiber core to the positive direction of the first direction, and the distance of each optical fiber core to the central axis; based on the strain of each optical fiber at the monitoring point, the angular offset of each optical fiber core to the positive direction of the first direction, and the distance of each optical fiber core to the central axis, performing summation calculation on all optical fibers to obtain the first curvature component and the second curvature component.

3. The method of claim 2, wherein the method further comprises: The expression for calculating the first curvature component of each monitoring point is: The expression for calculating the second curvature component of each monitoring point is: wherein κ1 and κ2 are the curvature components in the y and z directions of the local coordinate system, respectively. n is the total number of optical fibers; i is the optical fiber serial number and takes the value from 1 to n; ε i εi is the strain of optical fiber i at the monitoring point; θ i θi is the angular deviation of the i-th optical fiber core to the positive direction of the y-axis; r i ri is the distance of the i-th optical fiber core to the central axis.

4. The method of claim 1, wherein, The first-order derivative is calculated by using first-order forward difference approximation. The detection of the abnormal point of the first-order derivative by the local anomaly factor algorithm comprises: calculating the local anomaly factor value of the first-order derivative of each monitoring point to represent the degree of outlying; setting a critical local anomaly factor value, identifying the monitoring point with a local anomaly factor value exceeding the critical local anomaly factor value as an abnormal point, and identifying the monitoring point corresponding to the abnormal point as a fiber twist position.

5. The method of claim 4, wherein the method further comprises: The calculation of the local anomaly factor value of the first-order derivative of each monitoring point comprises: for a sample point in the data set, defining the specified neighbor distance of the sample point as the distance from the sample point to its first specified number of nearest neighbor points, and defining the specified neighbor neighborhood of the sample point as a set of data points with a distance less than the specified neighbor distance from the sample point; defining the reachable distance between the sample point and the neighborhood point as the larger value between the specified neighbor distance of the neighborhood point and the Euclidean distance between the sample point and the neighborhood point; based on the average reachable distance of the specified neighbor neighborhood points of the sample point to the sample point, calculating the local reachable density of the sample point; The local anomaly factor value of the sample point is obtained based on an average ratio of local reachable density of all neighborhood points in a specified neighborhood neighborhood to the sample point.

6. The method of claim 5, wherein the method further comprises: An expression of the reachable distance of the sample point and the neighborhood point is: reach-dist k (x,y) = max{k-dist(y), d(x,y)}; In the formula, d(x, y) is the Euclidean distance between the sample point x and y; k is the distance from the kth nearest neighbor point to the sample point x; An expression of the local reachable density of the sample point is: where lrd k (x) is the local reachable density of sample point x; |N k (x) is the number of elements in the k-distance neighborhood of sample point x; An expression of the local anomaly factor value of the sample point is: where LOF k (x) is the local outlier factor value for sample point x.

7. The method of claim 1, wherein the method further comprises: The global optimization of the twist angle of the identified fiber twist position includes: A twist compensation angle is applied to the identified fiber twist position, and a compensated first curvature component and a compensated second curvature component of each monitoring point after the twist compensation angle is applied are calculated; The second-order right derivatives of the compensated first curvature component and the compensated second curvature component at each fiber twist position are calculated by a second-order forward difference approximation, and the second-order left derivatives of the compensated first curvature component and the compensated second curvature component at each fiber twist position are calculated by a second-order backward difference approximation; A fitness function of the global optimization algorithm is set; wherein the fitness function is the sum of squares of the difference between the second-order left derivative and the second-order right derivative of the compensated first curvature component and the compensated second curvature component at each fiber twist position; The twist compensation angle is optimized by the global optimization algorithm to obtain an optimal twist angle set that minimizes the fitness function.

8. The method of claim 7, wherein the method further comprises: An expression of the fitness function is: In the formula, y is a fitness function; Δs is a strain sampling interval; M is a total number of fiber twist positions identified by the local anomaly factor algorithm; t i is a serial number of a monitoring point corresponding to the ith twist point; is a first curvature component after compensation; is a second curvature component after compensation.

9. The method of claim 1, wherein, The shape reconstruction algorithm adopts a parallel transmission method and performs recursive calculation by a fourth-order Runge-Kutta method, and specifically includes: A parallel transmission frame including a unit tangent vector, a first normal vector and a second normal vector is constructed; Based on the compensated first curvature component and the compensated second curvature component, a change relationship of the parallel transmission frame along an arc length parameter is established; The change relationship is recursively solved by the fourth-order Runge-Kutta method to obtain the parallel transmission frame of each monitoring point; The unit tangent vector of each monitoring point is integrated to reconstruct the spatial position of each monitoring point; An expression of the change relationship of the parallel transmission frame along the arc length parameter is: In the formula, l is the arc length parameter, T(l), N1(l) and N2(l) are respectively the unit tangent vector, the first normal vector and the second normal vector of the parallel transmission frame.

10. The method of claim 1, wherein, The preprocessing includes outlier rejection, multiple sampling averaging and initial strain correction; The multiple sampling averaging includes: continuously collecting strain data at least three times during monitoring and calculating an average value to reduce random error; the initial strain correction includes: deducting a strain reference value in a straight initial state before curvature calculation.

11. A multi-core fiber twist identification compensation system, characterized by, The multi-core optical fiber twist identification and compensation system includes: A data preprocessing module configured to read strain monitoring data of a multi-core optical fiber shape sensor and pre-process the strain monitoring data to construct a strain input sequence; A curvature calculation module configured to input the strain input sequence into a strain curvature mapping model to calculate a first curvature component and a second curvature component of each monitoring point; wherein the first curvature component is a curvature component in a first direction of a local coordinate system, and the second curvature component is a curvature component in a second direction of the local coordinate system; A curvature calculation module configured to input the strain input sequence into a strain curvature mapping model to calculate a first curvature component and a second curvature component of each monitoring point; wherein the first curvature component is a curvature component in a first direction of a local coordinate system, and the second curvature component is a curvature component in a second direction of the local coordinate system; The torsion identification module is configured to calculate first curvature components and first-order derivatives of length direction distribution curves of the second curvature components, detect abnormal points of the first-order derivatives by using a local outlier factor algorithm, and identify monitoring points corresponding to the abnormal points as fiber torsion positions. The angle optimization module is configured to set a fiber torsion angle range, and globally optimize torsion angles of the identified fiber torsion positions by taking second-order derivatives of the curvature components as a target to obtain an optimal torsion angle set satisfying a curvature component smoothness condition. The shape reconstruction module is configured to calculate compensated first curvature components and compensated second curvature components of the monitoring points after torsion compensation based on the optimal torsion angle set, and realize spatial position reconstruction of the monitoring points by recursive operation of a shape reconstruction algorithm.

12. The multi-core fiber twist identification compensation system according to claim 11, wherein In the curvature calculation module, when calculating the first curvature components and the second curvature components of the monitoring points, the total number of optical fibers, the serial numbers of the optical fibers, the strains of the optical fibers at the monitoring points, the angular deviations of the cores of the optical fibers from a positive direction of a first direction, and the distances of the cores of the optical fibers to a central axis are obtained; the first curvature components and the second curvature components are obtained by summing all the optical fibers based on the strains of the optical fibers at the monitoring points, the angular deviations of the cores of the optical fibers from the positive direction of the first direction, and the distances of the cores of the optical fibers to the central axis.

13. The multi-core fiber twist identification compensation system of claim 11, wherein, In the torsion identification module, when detecting the abnormal points of the first-order derivatives by using the local outlier factor algorithm, the local outlier factor values of the first-order derivatives of the monitoring points are calculated to represent the degrees of outlying; a critical local outlier factor value is set, the monitoring points with the local outlier factor values exceeding the critical local outlier factor value are identified as the abnormal points, and the monitoring points corresponding to the abnormal points are identified as the fiber torsion positions.

14. The multi-core fiber twist identification compensation system of claim 11, wherein, In the angle optimization module, when globally optimizing the torsion angles of the identified fiber torsion positions by taking the second-order derivatives of the curvature components as a target, a torsion compensation angle is applied to the identified fiber torsion positions, and the compensated first curvature components and the compensated second curvature components of the monitoring points after the application of the torsion compensation angle are calculated; the second-order right derivatives of the compensated first curvature components and the compensated second curvature components at the fiber torsion positions are calculated by using a second-order forward difference approximation, and the second-order left derivatives of the compensated first curvature components and the compensated second curvature components at the fiber torsion positions are calculated by using a second-order backward difference approximation; a fitness function of a global optimization algorithm is set; the fitness function is a sum of squares of differences between the second-order left derivatives and the second-order right derivatives of the compensated first curvature components and the compensated second curvature components at the fiber torsion positions; and the torsion compensation angle is optimized by using the global optimization algorithm to obtain the optimal torsion angle set minimizing the fitness function.

15. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to implement the steps of the method in any one of claims 1 to 10.

16. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 10.

Images