An optical communication network operation and maintenance fault intelligent positioning and tracking method and system

Abstract

The application discloses an optical communication network operation and maintenance fault intelligent positioning and tracking method and system, relates to the technical field of optical communication operation and maintenance decision, and comprises the following steps: collecting optical cable stress field data and environment image data in real time, identifying cable trench markers through a pre-trained convolutional neural network, detecting abnormal events and generating an abnormal event coordinate set and an image recognition abnormal sequence; constructing an image-stress fusion positioning model, inputting the optical cable stress field data and the abnormal event coordinate set into the image-stress fusion positioning model, and generating accurate positioning coordinates; performing space-time sequence analysis on the optical cable stress field data, combining the image recognition abnormal sequence, and generating a stress propagation rate and direction; generating a dynamic topology graph according to the accurate positioning coordinates; and performing coloring and grading processing on the dynamic topology graph to form a risk conduction graph. Through the construction of the image-stress fusion positioning model, the application realizes space-time unification and feature optimization of data, and improves the accuracy and robustness of fault positioning.

Description

Technical Field

[0001] This invention relates to the field of optical communication operation and maintenance decision-making technology, and in particular to an intelligent location and tracking method and system for optical communication network operation and maintenance faults. Background Technology

[0002] In recent years, optical communication network operation and maintenance fault location technology has gradually integrated fiber optic sensing and machine vision methods. Distributed fiber optic vibration monitoring systems (DVS) collect stress wave signals from optical cables and combine this with time-frequency analysis (such as short-time Fourier transform) to identify events such as fiber breakage and compression. Simultaneously, deep learning-based image recognition technologies (such as YOLOv5 and Faster R-CNN) are applied to the real-time analysis of cable trench inspection videos, automatically detecting anomalies such as support displacement and foreign object intrusion. At the data fusion level, existing research has used Kalman filtering or weighted averaging methods to integrate multi-source data to improve location accuracy.

[0003] Existing intelligent location and tracking methods for optical communication network operation and maintenance faults have shortcomings. First, the lack of a spatial alignment mechanism in multimodal data collaborative processing makes it impossible to eliminate affine transformation deviations between coordinate systems of different source devices, thus restricting the improvement of multimodal fusion positioning accuracy. In addition, the dynamic risk transmission quantification capability is insufficient, failing to achieve quantitative characterization of stress wave propagation rate and direction vector. The time domain analysis only focuses on the detection of isolated event peaks and fails to construct a gridded spatiotemporal matrix to characterize the continuous propagation process, resulting in operation and maintenance response lagging behind the risk diffusion process. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an intelligent location and tracking method for optical communication network operation and maintenance faults to solve the problems of insufficient data alignment accuracy and weak risk transmission modeling capabilities.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides an intelligent location and tracking method for optical communication network operation and maintenance faults, comprising: real-time acquisition of optical cable stress field data and environmental image data; identification of cable trench markers through a pre-trained convolutional neural network; detection of abnormal events and generation of an abnormal event coordinate set and image recognition anomaly sequence; construction of an image-stress fusion positioning model; inputting the optical cable stress field data and the abnormal event coordinate set into the image-stress fusion positioning model to generate precise positioning coordinates; performing spatiotemporal sequence analysis on the optical cable stress field data; combining the image recognition anomaly sequence to generate stress propagation rate and direction; generating a dynamic topology map based on the precise positioning coordinates; performing coloring and hierarchical processing on the dynamic topology map to form a risk transmission map; converting the colored areas in the risk transmission map into three-dimensional spatial coordinates, and inputting the three-dimensional spatial coordinates into a BIM model to generate an alarm signal coordinate set; wherein the colored areas include red high-risk areas and yellow medium-risk areas; based on the alarm signal coordinate set, triggering optical path switching commands for the red high-risk area coordinates, generating monitoring parameter adjustment commands for the yellow medium-risk area coordinates, and outputting an optical communication operation and maintenance tracking execution command set.

[0008] As a preferred embodiment of the intelligent location and tracking method for optical communication network operation and maintenance faults described in this invention, the specific steps for generating the set of abnormal event coordinates and the image recognition abnormal sequence are as follows:

[0009] Real-time measurement of optical cable stress field data and capture of cable trench environmental image data, recording acquisition timestamps and geographic coordinates;

[0010] The cable trench environment image data is input into a pre-trained convolutional neural network, which outputs image recognition results.

[0011] Based on image recognition results, analyze cable trench markers and their status; analyze optical cable stress field data to obtain stress values;

[0012] If either the cable trench marker is in an abnormal state or the stress value changes abruptly, mark the abnormal event.

[0013] Extract the geographic coordinates of the abnormal events to form a set of abnormal event coordinates; record the abnormal event type, geographic coordinates and timestamp in chronological order to form an image recognition abnormal sequence.

[0014] As a preferred embodiment of the intelligent location and tracking method for optical communication network operation and maintenance faults described in this invention, the specific steps for constructing the image-stress fusion location model are as follows:

[0015] A parameter establishment layer is constructed based on spatiotemporal benchmark alignment, a spatial registration layer is constructed based on spatial mapping, a regression spline layer is constructed based on multimodal feature collaborative analysis, and an optimization output layer is constructed based on numerical convergence.

[0016] Based on the parameters, an image-stress fusion localization model is constructed by establishing a layer, a spatial registration layer, a regression spline layer, and an optimized output layer.

[0017] As a preferred embodiment of the intelligent location and tracking method for optical communication network operation and maintenance faults described in this invention, the specific steps for generating precise location coordinates are as follows:

[0018] The optical cable stress field data and abnormal event coordinate set are input into the image-stress fusion localization model. A unified spatiotemporal reference is established through parameters and a parameter tensor is output. Based on the spatial registration layer, an affine transformation is performed on the parameter tensor to generate spatial alignment coordinates.

[0019] Based on spatially aligned coordinates, multimodal features are fused through a regression spline layer to generate a residual vector. The output layer is then optimized to iteratively converge the residual vector, thereby generating precise positioning coordinates.

[0020] As a preferred embodiment of the intelligent location and tracking method for optical communication network operation and maintenance faults described in this invention, the specific steps for generating the stress propagation rate and direction are as follows:

[0021] The optical cable stress field data and image recognition anomaly sequences are spatiotemporally aligned and fused to construct a spatiotemporal data matrix.

[0022] The stress propagation rate and direction are calculated based on the spatiotemporal data matrix.

[0023] As a preferred embodiment of the intelligent location and tracking method for optical communication network operation and maintenance faults described in this invention, the specific steps for forming the risk transmission map are as follows:

[0024] A dynamic topology graph is constructed using precise positioning coordinates as nodes, optical cable connections as edges, and stress propagation rate as edge weights.

[0025] Based on the magnitude of stress propagation rate and the type of abnormal event, the dynamic topology map is colored and graded, and the colored dynamic topology map is integrated to form a risk transmission map.

[0026] As a preferred embodiment of the intelligent location and tracking method for optical communication network operation and maintenance faults described in this invention, the specific steps for generating the alarm signal coordinate set are as follows:

[0027] Separate the boundary point sequences of red high-risk areas and yellow medium-risk areas from the risk transmission map; map the boundary point sequences to a three-dimensional spatial coordinate system to generate three-dimensional spatial coordinates;

[0028] Call the BIM model and input the three-dimensional spatial coordinates into the BIM model. Generate an alarm signal coordinate set through spatial matching and attribute binding.

[0029] As a preferred embodiment of the intelligent location and tracking method for optical communication network operation and maintenance faults described in this invention, the specific steps for outputting the optical communication operation and maintenance tracking execution instruction set are as follows:

[0030] Separate the red high-risk area coordinates and the yellow medium-risk area coordinates from the alarm signal coordinate set, and extract the device type and location information;

[0031] Based on the coordinates of the red high-risk area, the location of the core power supply equipment is associated with the optical path switching command; based on the coordinates of the yellow medium-risk area, the location of the auxiliary connection equipment is used to generate the monitoring parameter adjustment command.

[0032] Integrate optical path switching commands and monitoring parameter adjustment commands to generate a set of execution commands for optical communication operation and maintenance tracking.

[0033] As a preferred embodiment of the intelligent location and tracking method for optical communication network operation and maintenance faults described in this invention, the monitoring parameter adjustment instructions include focal length adjustment, frame rate increase, and infrared mode activation instructions.

[0034] Secondly, this invention provides an intelligent fault location and tracking system for optical communication network operation and maintenance, comprising an anomaly detection module, a fusion positioning module, a risk transmission module, a coordinate transformation module, and an instruction generation module; the anomaly detection module is used to collect optical cable stress field data and environmental image data in real time, identify cable trench markers through a pre-trained convolutional neural network, detect abnormal events, and generate an abnormal event coordinate set and image recognition anomaly sequence; the fusion positioning module is used to construct an image-stress fusion positioning model, input the optical cable stress field data and the abnormal event coordinate set into the image-stress fusion positioning model, and generate accurate positioning coordinates; the risk transmission module is used to process the optical cable stress field data... The system performs spatiotemporal sequence analysis, combines image recognition of abnormal sequences, and generates stress propagation rate and direction. It generates a dynamic topology map based on precise positioning coordinates. The dynamic topology map is then colored and graded to form a risk transmission map. A coordinate transformation module converts the colored areas in the risk transmission map into three-dimensional spatial coordinates, inputs these coordinates into the BIM model, and generates an alarm signal coordinate set. The colored areas include red high-risk areas and yellow medium-risk areas. An instruction generation module, based on the alarm signal coordinate set, triggers optical path switching instructions for the red high-risk area coordinates and generates monitoring parameter adjustment instructions for the yellow medium-risk area coordinates, outputting a set of optical communication operation and maintenance tracking execution instructions.

[0035] The beneficial effects of this invention are as follows: By constructing an image-stress fusion localization model, the spatiotemporal benchmark unification and feature co-optimization of multimodal data are achieved, significantly improving the accuracy and robustness of fault localization in complex environments, and overcoming the fusion deviation problem caused by the lack of spatial alignment mechanism for heterogeneous data in traditional methods; through quantitative analysis of dynamic topology maps and risk transmission maps, the dynamic characteristics of stress wave propagation are transformed into visualized risk diffusion paths, enabling accurate prediction of fault transmission rate and direction, breaking through the limitation that static topology labeling cannot respond to dynamic risks, and greatly improving the intelligent operation and maintenance level and fault handling efficiency of optical communication networks. Attached Figure Description

[0036] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 A flowchart for an intelligent fault location and tracking method for optical communication network operation and maintenance.

[0038] Figure 2 This is a schematic diagram of an intelligent fault location and tracking system for optical communication network operation and maintenance.

[0039] Figure 3 A flowchart for generating precise positioning coordinates.

[0040] Figure 4 A flowchart for generating a risk transmission map.

[0041] Figure 5 A flowchart for generating the instruction set for optical communication operation and maintenance tracing. Detailed Implementation

[0042] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0043] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0044] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0045] Reference Figures 1-5 As one embodiment of the present invention, this embodiment provides an intelligent location and tracking method for optical communication network operation and maintenance faults, including the following steps:

[0046] S1. Real-time acquisition of optical cable stress field data and environmental image data; identification of cable trench markers through pre-trained convolutional neural network; detection of abnormal events and generation of abnormal event coordinate set and image recognition abnormal sequence.

[0047] S1.1 Use stress sensors to measure optical cable stress field data in real time, use cameras to capture cable trench environment image data in real time, and record timestamps and location information simultaneously;

[0048] It should be noted that stress sensors are installed at optical cable nodes (such as junction boxes and bending points) and the detection sampling rate is set to be greater than the reference sampling rate (defined based on the dynamic characteristics of optical cable stress events and fault detection sensitivity requirements, such as 100Hz). The detection sampling rate is used to measure the micro-strain changes of the optical cable in real time, generating a raw data stream containing stress values, optical cable stress peak frequencies, and sensor identifiers. Explosion-proof high-definition cameras are deployed in the cable trench to capture environmental image data focused on the cable supports, signs, and the condition of the optical cable sheath, and the latitude and longitude coordinates are automatically recorded via BeiDou satellite navigation.

[0049] The sampling timestamp of the stress sensor is synchronized with the timestamp of the camera image frame through the PTP precision clock protocol. The stress sensor installation location coordinates and camera geographic coordinates are input into the Bursa transformation matrix for translation, rotation and scale correction to eliminate coordinate system offset caused by differences in equipment installation location. The output is a unified three-dimensional coordinate (longitude, latitude and elevation) in the WGS-84 coordinate system (i.e., geocentric coordinate system), forming a metadata tag with fully aligned spatiotemporal reference.

[0050] S1.2 Input the cable trench environmental image data into a pre-trained convolutional neural network and output the image recognition results; parse the cable trench markers and their status based on the image recognition results; parse the optical cable stress field data to obtain the stress value; mark the abnormal event when there is either an abnormality in the status of the markers or a sudden change in the stress value.

[0051] It should be noted that a historical image dataset containing various markers in cable trenches (such as supports, fiber optic cable sheaths, and signs) is collected and labeled with normal and abnormal state labels, which are the true state labels of the historical image dataset. A stochastic gradient descent algorithm is used, with cross-entropy as the loss function to drive the learning of the convolutional neural network. The response of the convolutional layer to the local features of the image and the spatial dimensionality reduction features of the pooling layer are calculated through forward propagation. Then, the classification probability is output through the fully connected layer to generate the predicted state label. In the backpropagation, the weight parameters of the convolutional kernel are adjusted layer by layer according to the error between the predicted state label and the true state label. The iteration is repeated until the loss function decreases by no more than the loss change threshold for ten consecutive times (defined based on the convergence stability requirement, such as 0.001) or the maximum number of iterations is reached (defined based on the network architecture complexity and dataset size, such as 5000 times), and the trained convolutional neural network is obtained.

[0052] Real-time acquired images of the cable trench environment are input into a convolutional neural network. Feature maps of cable supports, signs, and optical cable sheaths are extracted from the images through multiple convolutional and pooling layers. After classification by a fully connected layer, an abnormal state is determined when any of the following occurs: the confidence level of an image sign exceeds a confidence threshold (based on the fault misjudgment tolerance definition, such as 95%), the sign features are missing (e.g., the text area of ​​the sign is unrecognizable or the features of the optical cable sheath are obscured), or an abnormal feature combination occurs (e.g., the cable trench cover is damaged and liquid leakage occurs simultaneously). The status of the sign and its confidence level are then output.

[0053] The stress value sequence in the optical cable stress field data is analyzed in real time, and the absolute value of the ratio of the stress value between adjacent sampling points to the sampling time interval is calculated to generate the stress change rate. When the image recognition result shows that the status of the marker is abnormal or the stress change rate is greater than the change rate benchmark (the change rate benchmark is defined based on the critical value of optical cable material fracture and the amplitude of environmental noise, such as 10% / millisecond), the abnormal event type (image abnormality or stress mutation) is immediately marked and associated with the current timestamp and three-dimensional coordinate position.

[0054] S1.3 Extract the geographic coordinates of the abnormal events to form a set of abnormal event coordinates; record the abnormal event type, coordinates and timestamp in chronological order to form an image recognition abnormal sequence.

[0055] It should be noted that the three-dimensional location coordinates are extracted from the marked abnormal event records. The coordinate points associated with the same abnormal source (such as cable bracket displacement or stress mutation point) within the same timestamp are aggregated into a single coordinate record, generating a coordinate set containing the precise location information of all abnormal events, which is the abnormal event coordinate set. The abnormal event records are traversed in the order of timestamps, and the abnormal type, corresponding three-dimensional coordinate position and timestamp are extracted in sequence to form a sequence data with time as the primary key. Each record in the sequence follows the triplet format of "timestamp-abnormal type-three-dimensional coordinate position". The output is an image recognition abnormal sequence arranged in ascending order of time.

[0056] S2. Construct an image-stress fusion localization model. Input the optical cable stress field data and the set of abnormal event coordinates into the image-stress fusion localization model to generate accurate localization coordinates.

[0057] S2.1 Construct a parameter establishment layer based on spatiotemporal benchmark alignment, construct a spatial registration layer based on spatial mapping, construct a regression spline layer based on multimodal feature collaborative analysis, and construct an optimization output layer based on numerical convergence;

[0058] It should be noted that a four-dimensional parameterized structural framework (longitude, latitude, elevation, and timestamp) is defined to incorporate the spatiotemporal attributes of optical cable stress field data and anomalous event coordinate sets into a unified dimension; all input coordinates and timestamps are forced to be converted to the WGS-84 coordinate system and UTC time base to generate normalized parameters; a structured output interface is defined to encapsulate the normalized parameters into fixed-dimensional parameter tensors, thus completing the construction of the parameter establishment layer;

[0059] The input parameter tensor is mapped to a two-dimensional grid. Based on the heterogeneity of the spatial coordinate system of multiple source devices, affine transformation rules are formulated to ensure that spatial data from different sources can be accurately matched, generating aligned spatial coordinate data. The aligned spatial coordinate data is then encapsulated in a fixed-dimensional sequence format (using a two-dimensional matrix format with a constant number of rows and columns, where the row dimension is the number of rows in the two-dimensional grid and the column dimension is fixed at three columns, namely xy coordinates and elevation) to form spatial registration coordinates, thus completing the construction of the spatial registration layer.

[0060] It should also be noted that the two-dimensional grid is a regularized coordinate framework. Based on the spatial boundary of the geographical coverage area of ​​the cable trench, a uniform grid array is divided according to a preset grid surface size (defined based on positioning accuracy requirements, such as 0.1m × 0.1m). Longitude, latitude, and elevation are converted into two-dimensional Cartesian coordinates (X,Y) through local Cartesian transformation, where the X-axis represents the distance in the east-west direction, the Y-axis represents the distance in the north-south direction, and the elevation h is an independent attribute value that does not participate in the grid mapping. A two-dimensional grid covering the entire cable trench area is generated.

[0061] Affine transformation rules refer to coordinate transformation functions constructed using six geometric parameters (rotation angle, x-direction translation, y-direction translation, x-direction scaling factor, y-direction scaling factor, and shear coefficient) to map source space data from different source devices to a unified coordinate system. Mathematically, it is a combination of linear transformations (rotation / scaling / shearing) and translation transformations, expressed as follows:

[0062] ;

[0063] in, This represents the transformed east-west coordinates; This represents the transformed north-south coordinates; Indicates the original east-west coordinates of the input device; Indicates the original north-south coordinates of the input device; express Orientation scaling factor; express Orientation scaling factor; Indicates the rotation angle; Indicates the shear coefficient; express Directional translation; express Directional translation;

[0064] The combined formulas of linear transformation and translation transformation achieve coordinate transformation through matrix multiplication. It calculates the product of the original coordinate vector and the transformation matrix. The transformation matrix contains a combination of trigonometric functions of scaling factors and rotation angles, as well as nonlinear transformation components composed of shear coefficients. After completing geometric transformations such as rotation, scaling, and shearing, it is then summed with the translation vector to obtain the new coordinates after the complete affine transformation.

[0065] Define a feature fusion dimension to establish a correlation mapping between spatial registration coordinates, optical cable stress peak frequency, and image marker confidence level: Based on the two-dimensional grid position in the spatial registration coordinates, create a unique spatial identifier for the center point of each grid surface; simultaneously extract the optical cable stress peak frequency and image marker confidence level corresponding to each two-dimensional grid position to form a physical feature binary; using the spatial identifier as the primary key, associate and store the stress frequency and image confidence level as attribute values ​​to construct a feature-coordinate mapping table; generate feature coupling weight coefficients through a cubic B-spline function, and output the residual vector containing the spatial identifier, optical cable stress peak frequency, image marker confidence level, and feature coupling weight coefficients in the order of two-dimensional grid position to complete the construction of the regression spline layer; the expression for generating the feature coupling weight coefficients is as follows:

[0066] ;

[0067] ;

[0068] in, Representing input features The corresponding feature coupling weight coefficients; This represents the normalized fusion eigenvalues; The total number of control points determines the degrees of freedom of the spline curve (the more control points, the more refined the curve). It is defined based on the complexity of eigenvalue distribution and computational resource constraints, such as 20. This represents the control point index, with a value range of 0 to... ; Indicates the first The coefficients for each control point; Representing input features In the Cubic B-spline basis function values ​​for each control point; This refers to the order of the spline basis functions, which is... This indicates that the basis functions are fixed as cubic B-spline functions; Indicates the stress characteristic weight; Represents image feature weights; Indicates the peak frequency of the optical cable stress wave; This represents the stress frequency reference value, which is 100Hz; This represents the stress frequency normalization scale. Indicates the confidence level of image identifiers; Indicates the confidence threshold; This represents the normalization scale of image confidence.

[0069] The formula for calculating the feature coupling weight coefficient is to substitute the normalized fused feature value into the cubic B-spline basis function. The cubic B-spline basis function generates corresponding basis function values ​​according to the distribution of the normalized fused feature value in different control point intervals. These basis function values ​​are multiplied and summed with the corresponding control point coefficients to generate a smooth and continuous feature coupling weight output.

[0070] The convergence termination condition is set as follows: the position deviation value is less than the position deviation threshold (based on the physical diameter of the optical cable, such as 0.01 meters) or the number of operations exceeds the upper limit of the number of operations (based on the convergence characteristics of the gradient descent algorithm and the edge computing power constraint definition, such as 200 times); a spatial position correction interface is defined, and a mapping rule from the residual vector to coordinate fine-tuning is established (the product of the residual vector and the calibration coefficient of the corresponding direction is calculated to generate the three-dimensional coordinate offset); the precise positioning coordinate output format is encapsulated, and the accuracy standard of the precise positioning coordinate is specified (based on the spatial resolution of the cable trench and the positioning error tolerance definition, such as ±0.05 meters), thus completing the construction of the optimized output layer;

[0071] The calibration coefficients are defined based on the error distribution patterns of historical positioning data of cable trenches. For example, the calibration coefficient for the east-west direction is 0.7, and the calibration coefficient for the north-south direction is 0.6.

[0072] S2.2 Construct an image-stress fusion localization model based on parameter-based layer, spatial registration layer, regression spline layer, and optimized output layer;

[0073] It should be noted that the image-stress fusion localization model is constructed by connecting the parameter establishment layer, spatial registration layer, regression spline layer, and optimized output layer in series.

[0074] Collect historical operation and maintenance data of cable trenches (including historical fault reports), and divide the historical operation and maintenance data of cable trenches into training sets and validation sets; unify the training set to the WGS-84 coordinate system and UTC time base to generate a parameter tensor; input the parameter tensor into the spatial registration layer, generate aligned spatial coordinate data through affine transformation rules, and encapsulate to generate spatial registration coordinates; input the spatial registration coordinates into the regression spline layer, generate feature coupling weight coefficients through cubic B-spline basis functions, and generate residual vectors based on the feature coupling weight coefficients; input the residual vector into the optimization output layer, and generate predicted precise positioning coordinates through the gradient descent algorithm; verify the accuracy of the predicted precise positioning coordinates through the validation set. If the error between the predicted precise positioning coordinates and the actual precise positioning coordinates (derived from historical fault reports) is lower than the error threshold (defined based on the physical spatial constraints and modeling accuracy requirements of the cable trench, such as 0.05) for multiple consecutive rounds, the training ends; otherwise, repeat the optimization process until the error between the predicted precise positioning coordinates and the actual precise positioning coordinates is lower than the error threshold for multiple consecutive rounds, thus completing the training of the image-stress fusion positioning model.

[0075] S2.3 Input the optical cable stress field data and abnormal event coordinate set into the image-stress fusion positioning model, and generate accurate positioning coordinates through spatial registration and regression spline algorithms.

[0076] It should be noted that the optical cable stress field data and the set of anomalous event coordinates are aligned to a spatiotemporal reference through a parameter establishment layer. Timestamps are unified to the UTC time axis, and geographic coordinates are converted to three-dimensional Cartesian coordinates in the WGS-84 coordinate system, outputting a parameter tensor with a unified spatiotemporal reference. This parameter tensor is input to a spatial registration layer, which maps it to a two-dimensional mesh using affine transformation rules, outputting spatially aligned spatial registration coordinates. These spatial registration coordinates, along with the real-time extracted optical cable stress peak frequency and image marker confidence scores, are input to a regression spline layer to generate a residual vector. The residual vector is then input to an optimization output layer, where the coordinate offset is iteratively adjusted using a gradient descent algorithm. Optimization terminates when the position deviation is less than 0.01 meters or the number of iterations exceeds 200, outputting precise positioning coordinates. The expression for calculating the precise positioning coordinates is...

[0077] ;

[0078] ;

[0079] ;

[0080] in, Indicates precise location coordinates; This indicates the east-west coordinates of the precise location. The north-south coordinates represent the precise positioning coordinates; Indicates the elevation of the precisely located coordinates; In The starting value of the control point index is indicated by That is, from the first The calculation begins at each control point; Indicates spatial registration coordinates; Indicates the calibration coefficient; Indicates the feature coupling weight coefficient; The east-west coordinates represent the coordinates of the spatial registration coordinates; The north-south coordinates representing the spatial registration coordinates; The elevation represents the coordinates of the spatial registration.

[0081] S3. Perform spatiotemporal sequence analysis on the optical cable stress field data, combine with image recognition of abnormal sequences, and generate stress propagation rate and direction; generate a dynamic topology map based on precise positioning coordinates; perform coloring and hierarchical processing on the dynamic topology map to form a risk transmission map;

[0082] S3.1. Spatiotemporally align and fuse the optical cable stress field data with the image recognition anomaly sequence to construct a spatiotemporal data matrix;

[0083] It should be noted that, using the timestamp of the precisely positioned coordinates as the reference time axis, the stress values ​​of the optical cable stress field data and the timestamps of the image marker confidence scores are aligned by comparing their numerical values. When the absolute value of the difference between the timestamp of the stress value and the timestamp of the image marker confidence score is less than the alignment threshold (based on the synchronization accuracy definition of the PTP precision clock protocol, such as 1 millisecond), it is considered to be the same time point. The stress value of the current time is paired with the image marker confidence score and marked as the same spatiotemporal data point. The stress value and the image marker confidence score of each time point are integrated according to the time dimension, spatial dimension, and data dimension to construct a spatiotemporal data matrix. The X-axis of the spatiotemporal data matrix represents the time series, and the Y-axis and Z-axis represent the spatial location index of the two-dimensional grid. The matrix elements store the stress value and image marker confidence score of the corresponding spatiotemporal point.

[0084] S3.2 Analyze the stress wave propagation characteristics based on the spatiotemporal data matrix, and calculate the stress propagation rate and stress propagation direction;

[0085] It should be noted that, based on the spatiotemporal data matrix, the stress value sequence at each grid surface index position is extracted along the time dimension. The peak value and occurrence time of the stress value within each sliding time window are detected by the sliding time window, and the arrival time point of the wavefront in the stress propagation process is identified. The grid surface position where the stress peak first appears is taken as the wave source starting point, and the time delay of the wavefront reaching other grid surface positions is calculated. The stress propagation rate of the stress wave between grid surfaces is calculated based on the ratio of the known physical distance between the center points of adjacent grid surfaces to the time delay.

[0086] Using the spatial coordinates of the grid surface as the independent variable and the wavefront arrival time as the dependent variable, a linear function is fitted using the least squares method. The expression for the linear function is:

[0087] ;

[0088] ;

[0089] in, Indicates the wavefront arrival time; Indicates the east-west direction of the grid surface; Indicates the north-south direction of the grid surface; express The slope of the direction; express The slope vector of the direction; This represents the stress propagation direction vector; the opposite direction of the slope vector of the linear function is the stress propagation direction.

[0090] S3.3 Construct a dynamic topology graph using precise positioning coordinates as nodes, optical cable connections as edges, and stress propagation rate as edge weights;

[0091] It should be noted that the sequence of grid surfaces pointed to by the stress propagation direction vector is used as the node connection path, and the precise positioning coordinates of the center point of each grid surface contained in the node connection path are extracted as topology graph nodes; according to the physical laying route map of optical cables in the cable trench, topology graph node pairs with real optical cable connection relationships between adjacent topology graph nodes are connected with undirected edges; the stress propagation rate is used as the edge weight of the corresponding edge; and all topology graph nodes, edges and edge weights are integrated to construct a dynamic topology graph.

[0092] S3.4. Based on the magnitude of stress propagation rate and the type of abnormal event, the dynamic topology map is colored and graded, and the colored dynamic topology map is integrated to form a risk transmission map.

[0093] It should be noted that, taking the dynamic topology graph as input, the speed levels are divided according to the value of the edge weight of each edge in the dynamic topology graph. Edges with a speed greater than the high speed threshold (based on the statistical definition of historical fault data, such as 0.08 m / ms) are marked in red, edges with a speed between the medium speed threshold (based on the statistical definition of historical fault data, such as 0.04 m / ms) and the high speed threshold are marked in yellow, and edges with a speed less than the medium speed threshold are marked in green.

[0094] The node color is set according to the type of abnormal event associated with the node in the topology graph. The topology graph nodes with image abnormality are marked as dark red, the topology graph nodes with stress mutation are marked as orange-red, and the topology graph nodes without associated abnormal events are not colored.

[0095] Topological graph nodes and edges that are spatially adjacent and of the same color are visually aggregated and rendered as continuous color blocks: connected regions composed of dark red or orange-red nodes and the red edges connecting them form red high-risk areas; connected regions composed of dark red or orange-red nodes and the yellow edges connecting them, or yellow edges in uncolored topological graph node regions, form yellow medium-risk areas; connected regions composed of green edges and the connected uncolored topological graph nodes form green safe areas; and all colored areas are integrated to form a risk transmission map that visually represents the risk distribution using continuous color blocks.

[0096] It should also be noted that when rendering continuous color blocks, based on the two-dimensional grid defined by the spatial registration layer, the grid faces that are spatially adjacent and have the same color are visually merged and rendered into a continuous color block.

[0097] S4. Convert the colored areas in the risk transmission map into three-dimensional spatial coordinates, and input the three-dimensional spatial coordinates into the BIM model to generate an alarm signal coordinate set; the colored areas include red high-risk areas and yellow medium-risk areas;

[0098] S4.1 Separate the boundary point sequences of the red high-risk area and the yellow medium-risk area from the risk transmission map; map the boundary point sequences to a three-dimensional spatial coordinate system to generate three-dimensional spatial coordinates;

[0099] It should be noted that, taking the bottom grid surface of the risk transmission map as input, the color indicators of all grid surfaces in the risk transmission map are traversed, and the color indicators of each grid surface are compared with those of adjacent grid surfaces in the four directions of east, south, west, and north. When it is found that the color indicators of adjacent grid surfaces are different and the current grid surface is red or yellow, the spatial position index of the current grid surface is recorded as a boundary point. All recorded boundary points are sorted according to the spatial connection order to form a sequence of boundary points between red high-risk areas and yellow medium-risk areas.

[0100] The precise location coordinates of all abnormal events are aggregated to form a precise location coordinate set; each point in the boundary point sequence is essentially an index coordinate of a two-dimensional grid, and the corresponding three-dimensional spatial coordinates of the boundary point are directly found in the precise location coordinate set based on the index coordinates.

[0101] S4.2 Call the BIM model and input the three-dimensional spatial coordinates into the BIM model. Generate an alarm signal coordinate set through spatial matching and attribute binding.

[0102] It should be noted that the existing BIM model stored in the secure storage area of ​​the power grid data center is called through the application programming interface; the three-dimensional spatial coordinates are input into the BIM model, and the BIM component where each three-dimensional spatial coordinate point is located is determined by ray detection, and the engineering attribute data of the corresponding BIM component (including the component's unique identifier, component type, and spatial location) is queried; the three-dimensional spatial coordinates and BIM component attribute data are bound and encapsulated into an alarm signal coordinate set.

[0103] S5. Based on the alarm signal coordinate set, trigger optical path switching commands for red high-risk areas, generate monitoring parameter adjustment commands for yellow medium-risk areas, and output the optical communication operation and maintenance tracking execution command set.

[0104] S5.1 Separate the red high-risk area and yellow medium-risk area where the alarm signal coordinates are concentrated, and extract the equipment type and location information;

[0105] It should be noted that when a grid surface is colored because the components within its own spatial range are marked as dark red or orange-red, or because the edges it connects to are marked as red or yellow, the unique identifier of the component, the spatial index of the grid surface, and the color are recorded as a record. Multiple records are sorted and merged to form a mapping table that records the mapping relationship between all components and the color status in the risk transmission map.

[0106] Using the alarm signal coordinate set as input, and based on the unique identifier of the component associated with each alarm signal coordinate, the unique identifier of the component is matched with the grid surface through a mapping table to find the grid surface corresponding to the component in the risk transmission map, and the color status of the grid surface is read. If the location belongs to the red high-risk area in the risk transmission map, the alarm signal coordinate is classified as red high-risk area coordinate; if it belongs to the yellow medium-risk area, it is classified as yellow medium-risk area coordinate. The component type and spatial location in the BIM component attributes corresponding to the red high-risk area coordinate and the yellow medium-risk area coordinate are extracted.

[0107] S5.2. Generate optical path switching instructions based on the location of core power supply equipment associated with the red high-risk area; generate monitoring parameter adjustment instructions based on the location of auxiliary connection equipment in the yellow medium-risk area; the monitoring parameter adjustment instructions include focus adjustment, frame rate increase and infrared mode activation instructions;

[0108] It should be noted that during the construction phase of the cable communication network, the backup transmission path of each backbone optical cable is determined according to the topological connection relationship of the physical route of the optical cable, and the priority switching order and triggering conditions are formulated based on historical operation and maintenance records, forming policy entries containing optical cable identifiers, backup port numbers and switching priorities, etc. All policy entries are entered and stored through the operation and maintenance configuration interface to complete the preset of the optical path protection policy library.

[0109] Power supply equipment is selected based on the equipment type recorded in the component attributes associated with the coordinates of the red high-risk area. The optical path identifier and the connection port information of the other end are read. The optical path switching instruction containing the target optical path identifier, switching action and target port is generated by combining the optical path protection strategy library.

[0110] The auxiliary connection device is located based on the component attributes associated with the coordinates of the yellow medium-risk area. The physical installation identifier of the camera is extracted from the spatial location information. Based on the relative orientation and distance between the auxiliary connection device and the yellow medium-risk area, monitoring parameter adjustment instructions are generated, including adjusting the optical focal length to align with the risk point, increasing the video acquisition frame rate, and activating the infrared thermal imaging mode to enhance monitoring capabilities in nighttime or low-light environments. The optical path switching instructions and monitoring parameter adjustment instructions are output in parallel.

[0111] S5.3 Integrate optical path switching commands and monitoring parameter adjustment commands to generate an optical communication operation and maintenance tracking execution command set.

[0112] It should be noted that a unique instruction identifier and timestamp are added to each optical path switching instruction and monitoring parameter adjustment instruction to ensure that the instructions are traceable. The optical path switching instructions are sorted according to their priority order over the monitoring parameter adjustment instructions. The sorted optical path switching instructions and monitoring parameter adjustment instructions are associated and bound with the corresponding risk area coordinates and equipment information. The optical path switching instructions, monitoring parameter adjustment instructions, associated coordinates and equipment information are integrated and packaged to generate an optical communication operation and maintenance tracking execution instruction set.

[0113] This embodiment also provides an intelligent location and tracking system for optical communication network operation and maintenance faults, including: an anomaly detection module, a fusion location module, a risk transmission module, a coordinate transformation module, and an instruction generation module;

[0114] The anomaly detection module is used to collect optical cable stress field data and environmental image data in real time, identify cable trench markers through a pre-trained convolutional neural network, detect abnormal events, and generate a set of abnormal event coordinates and an image recognition anomaly sequence.

[0115] The fusion positioning module is used to construct an image-stress fusion positioning model. It inputs the optical cable stress field data and the set of abnormal event coordinates into the image-stress fusion positioning model to generate accurate positioning coordinates.

[0116] The risk transmission module is used to perform spatiotemporal sequence analysis on optical cable stress field data, combine image recognition of abnormal sequences, generate stress propagation rate and direction; generate dynamic topology map based on precise positioning coordinates; and perform coloring and hierarchical processing on the dynamic topology map to form a risk transmission map.

[0117] The coordinate transformation module is used to convert the colored areas in the risk transmission map into three-dimensional spatial coordinates, and input the three-dimensional spatial coordinates into the BIM model to generate an alarm signal coordinate set; the colored areas include red high-risk areas and yellow medium-risk areas;

[0118] The instruction generation module is used to trigger optical path switching instructions for the coordinates of red high-risk areas and generate monitoring parameter adjustment instructions for the coordinates of yellow medium-risk areas based on the alarm signal coordinate set, and output the optical communication operation and maintenance tracking execution instruction set.

[0119] This embodiment also provides a computer device applicable to the intelligent location and tracking method for optical communication network operation and maintenance faults, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the intelligent location and tracking method for optical communication network operation and maintenance faults as proposed in the above embodiment.

[0120] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0121] This embodiment also provides a storage medium storing a computer program. When executed by a processor, the program implements the intelligent location and tracking method for optical communication network operation and maintenance faults as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0122] In summary, this invention achieves unified spatiotemporal benchmarks and collaborative feature optimization for multimodal data by constructing an image-stress fusion localization model, significantly improving the accuracy and robustness of fault localization in complex environments and overcoming the fusion deviation problem caused by the lack of spatial alignment mechanisms for heterogeneous data in traditional methods. Through quantitative analysis of dynamic topology maps and risk propagation maps, the dynamic characteristics of stress wave propagation are transformed into visualized risk diffusion paths, enabling accurate prediction of fault propagation rates and directions. This overcomes the limitation that static topology annotations cannot respond to dynamic risks, and significantly improves the intelligent operation and maintenance level and fault handling efficiency of optical communication networks.

[0123] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intelligent location and tracking of faults in the operation and maintenance of optical communication networks, characterized in that: The application relates to a method for generating an alarm signal coordinate set based on a BIM model, and belongs to the field of optical communication operation and maintenance. Real-time acquisition of optical cable stress field data and environmental image data, identification of cable trench markers through a pre-trained convolutional neural network, detection of abnormal events and generation of abnormal event coordinate sets and image-identified abnormal sequences; An image-stress fusion positioning model is constructed, and the specific steps are as follows, A parameter establishment layer is constructed based on space-time reference alignment, a space registration layer is constructed based on space mapping, a regression spline layer is constructed based on multi-modal feature collaborative analysis, and an optimization output layer is constructed based on numerical convergence; An image-stress fusion positioning model is constructed according to the parameter establishment layer, the space registration layer, the regression spline layer and the optimization output layer; The optical cable stress field data and the abnormal event coordinate set are input into the image-stress fusion positioning model to generate accurate positioning coordinates, and the specific steps are as follows, The optical cable stress field data and the abnormal event coordinate set are input into the image-stress fusion positioning model, the space-time reference is unified through the parameter establishment layer, and a parameter tensor is output, and the parameter tensor is subjected to affine transformation according to the space registration layer to generate space alignment coordinates; Based on the space alignment coordinates, the regression spline layer is used to fuse multi-modal features and generate a residual vector, and the optimization output layer is used for iterative convergence of the residual vector to generate accurate positioning coordinates; Space-time sequence analysis is performed on the optical cable stress field data, and the image-identified abnormal sequence is combined to generate a stress propagation rate and direction; and a dynamic topology graph is generated according to the accurate positioning coordinates; The dynamic topology graph is subjected to coloring and hierarchical processing to form a risk transmission atlas; The colored areas in the risk transmission atlas are converted into three-dimensional space coordinates, and the three-dimensional space coordinates are input into a BIM model to generate an alarm signal coordinate set; the colored areas include a red high-risk area and a yellow medium-risk area; Based on the alarm signal coordinate set, a light path switching instruction is triggered for the red high-risk area coordinates, a monitoring parameter adjustment instruction is generated for the yellow medium-risk area coordinates, and an optical communication operation and maintenance tracking execution instruction set is output. 2.The method of claim 1, wherein: The specific steps of generating the abnormal event coordinate set and the image-identified abnormal sequence are as follows, Real-time measurement of optical cable stress field data and capture of cable trench environmental image data, recording of collection time stamps and geographical position coordinates; The cable trench environmental image data are input into a pre-trained convolutional neural network to output image recognition results; The image recognition results are analyzed to analyze the cable trench markers and the state of the cable trench markers; The stress value is obtained by analyzing the optical cable stress field data; If any of the state of the cable trench marker and the stress value mutation is abnormal, the abnormal event is marked; The geographical position coordinates of the abnormal event are extracted to form an abnormal event coordinate set; The types of abnormal events, geographical position coordinates and time stamps are recorded in chronological order to form an image-identified abnormal sequence. 3.The method of claim 2, wherein: The specific steps of generating the stress propagation rate and direction are as follows, The optical cable stress field data and the image-identified abnormal sequence are subjected to space-time alignment and fusion to construct a space-time data matrix; The stress propagation rate and direction are calculated based on the space-time data matrix.

4. The intelligent positioning and tracking method for operation and maintenance fault of optical communication network according to claim 3, characterized in that: The specific steps of forming the risk transmission atlas are as follows, A dynamic topology graph is constructed with the accurate positioning coordinates as nodes, optical cable connections as edges and stress propagation rates as edge weights; According to the stress propagation rate and the type of abnormal event, the dynamic topology graph is colored and graded, the colored dynamic topology graph is integrated, and a risk transmission atlas is formed.

5. The intelligent positioning and tracking method for operation and maintenance fault of optical communication network according to claim 4, characterized in that: The specific steps of generating the alarm signal coordinate set are as follows, The boundary point sequence of the red high-risk area and the yellow medium-risk area is separated from the risk transmission atlas, and the boundary point sequence is mapped to a three-dimensional space coordinate system to generate a three-dimensional space coordinate; The BIM model is called, and the three-dimensional space coordinate is input into the BIM model to generate an alarm signal coordinate set through space matching and attribute binding.

6. The intelligent positioning and tracking method for operation and maintenance fault of optical communication network according to claim 5, characterized in that: The specific steps of outputting the optical communication operation and maintenance tracking execution instruction set are as follows, The red high-risk area coordinate and the yellow medium-risk area coordinate in the alarm signal coordinate set are separated, and the device type and position information are extracted; The optical path switching instruction is generated based on the red high-risk area coordinate associated with the position of the core power supply device; The monitoring parameter adjustment instruction is generated based on the yellow medium-risk area coordinate positioning auxiliary connection device; The optical path switching instruction and the monitoring parameter adjustment instruction are integrated to generate an optical communication operation and maintenance tracking execution instruction set.

7. The intelligent positioning and tracking method for operation and maintenance fault of optical communication network according to claim 6, characterized in that: The monitoring parameter adjustment instruction includes focus adjustment, frame rate improvement, and infrared mode activation instruction.

8. An optical communication network operation and maintenance fault intelligent positioning and tracking system based on the optical communication network operation and maintenance fault intelligent positioning and tracking method of any one of claims 1-7. It includes an abnormality detection module, a fusion positioning module, a risk transmission module, a coordinate conversion module, and an instruction generation module; The abnormality detection module is used for real-time acquisition of optical cable stress field data and environment image data, identification of cable trench markers through a pre-trained convolutional neural network, detection of abnormal events, and generation of abnormal event coordinate sets and image recognition abnormal sequences; The fusion positioning module is used for constructing an image-stress fusion positioning model, inputting the optical cable stress field data and the abnormal event coordinate set into the image-stress fusion positioning model, and generating accurate positioning coordinates; The risk transmission module is used for performing time-space sequence analysis on the optical cable stress field data, combining the image recognition abnormal sequence, generating stress propagation rate and direction, and generating a dynamic topology graph according to the accurate positioning coordinates; The dynamic topology graph is colored and graded to form a risk transmission atlas; The coordinate conversion module is used for converting the colored area in the risk transmission atlas into a three-dimensional space coordinate, inputting the three-dimensional space coordinate into a BIM model, and generating an alarm signal coordinate set; the colored area includes a red high-risk area and a yellow medium-risk area; The instruction generation module is used for generating an optical communication operation and maintenance tracking execution instruction set based on the alarm signal coordinate set, triggering an optical path switching instruction for the red high-risk area coordinate, and generating a monitoring parameter adjustment instruction for the yellow medium-risk area coordinate.

Images