Optical fiber loss event classification detection method, system, equipment and medium

Abstract

The embodiment of the invention provides an optical fiber loss event classification detection method, system and device and a medium, and belongs to the technical field of optical fiber line detection. The method comprises the following steps: acquiring multi-band echo data of a to-be-measured optical fiber under a preset wavelength set; constructing a cross-band alignment relationship based on the multi-band echo data and executing amplitude comparable processing to generate multi-band comparable echo data; detecting event candidate segments based on the multiband comparable echo data and extracting a corresponding multiband feature set; and outputting an event position and an event category result based on the multi-band feature set. According to the scheme of the invention, through cross-band distance alignment and amplitude comparable processing, spatial consistent expression of multi-band echoes at the same physical event is realized, so that the optical fiber event positioning precision and the event type judgment reliability are improved.

Description

Technical Field

[0001] This invention relates to the field of optical fiber line testing technology, and specifically to a method, system, equipment, and medium for classifying and detecting optical fiber loss events. Background Technology

[0002] During long-term operation in communication and monitoring applications, fiber optic lines are prone to loss events such as fiber breakage, bending, and connection abnormalities due to construction, environmental changes, or external forces. Therefore, it is necessary to detect and analyze the fiber loss distribution and the location of these events. Existing fiber optic testing methods are mostly based on the principle of optical time-domain reflectometry, using the analysis of backscattered and reflected signals generated along the fiber to measure fiber loss and locate faults.

[0003] In practical applications, single-wavelength fiber optic detection methods have limitations in distinguishing event types. Different types of events may exhibit similar echo characteristics, making stable differentiation difficult. To obtain more detection information, some technical solutions introduce multi-band detection methods. However, there are common problems of inconsistencies in distance axes and incomparable amplitude scales between echo data from different bands. This makes it difficult to directly correlate events at the same physical location across different bands, affecting the reliability of event detection and classification.

[0004] Therefore, in multi-band fiber optic detection scenarios, how to effectively align and process echo data from different bands, and on this basis, realize event candidate segment detection and event category determination, still requires further research and improvement. Summary of the Invention

[0005] The purpose of this invention is to provide a method, system, device, and medium for classifying and detecting fiber optic loss events, so as to at least solve the problems of alignment and comparability of multi-band echo data.

[0006] To achieve the above objectives, the first aspect of the present invention provides a method for classifying and detecting optical fiber loss events. The method includes: acquiring multi-band echo data of the optical fiber under test under a preset wavelength set; constructing cross-band alignment relationships based on the multi-band echo data and performing amplitude comparability processing to generate multi-band comparable echo data; detecting event candidate segments based on the multi-band comparable echo data and extracting corresponding multi-band feature sets; and outputting event location and event category results based on the multi-band feature sets.

[0007] Optionally, acquiring multi-band echo data of the optical fiber under test under a preset wavelength set includes: based on the preset wavelength set, controlling a multi-band probe light source to inject probe light of corresponding wavelengths into the optical fiber under test sequentially or according to a preset time sequence to form a multi-band probe process under the same measurement link conditions; during the injection of probe light in each band, collecting backscattered signals and reflected signals generated along the optical fiber under test, and marking the collected backscattered signals and reflected signals according to the corresponding band to form independent echo sequence data for each band; uniformly mapping the echo sequence data of each band to the distance dimension to generate multi-band echo data with optical fiber distance as the independent variable and echo amplitude as the dependent variable.

[0008] Optionally, constructing a cross-band alignment relationship based on the multi-band echo data and performing amplitude comparability processing to generate multi-band comparable echo data includes: selecting a preset reference band in the multi-band echo data, and identifying the echo position corresponding to a preset reference event in each band echo data to characterize the distance correspondence of the same physical location in different bands; constructing a cross-band distance alignment relationship based on the correspondence between the echo position of the reference band and the corresponding echo position in other bands, and performing distance axis correction on the echo data of non-reference bands according to the cross-band distance alignment relationship to obtain distance-aligned multi-band echo data; and performing amplitude normalization processing on the distance-aligned multi-band echo data based on the reference amplitude information corresponding to each band to generate multi-band comparable echo data.

[0009] Optionally, based on the reference amplitude information corresponding to each band, amplitude normalization processing is performed on the distance-aligned multi-band echo data to generate comparable multi-band echo data. This includes: for each band, obtaining reference amplitude information corresponding to the echo data of that band, the reference amplitude information being used to characterize the reference amplitude level of that band under the current measurement conditions; based on the reference amplitude information, determining the amplitude normalization parameter for the corresponding band, the amplitude normalization parameter being used to eliminate amplitude offsets caused by differences in transmit power or link response between different bands; and performing amplitude normalization processing on the distance-aligned multi-band echo data respectively according to the amplitude normalization parameter, so that the echo data of different bands are comparable under a unified amplitude reference, thereby generating comparable multi-band echo data.

[0010] Optionally, detecting event candidate segments and extracting corresponding multi-band feature sets based on the multi-band comparable echo data includes: selecting a preset detection band as the event detection benchmark in the multi-band comparable echo data; performing change analysis on the multi-band comparable echo data along the distance dimension to identify echo change intervals that satisfy preset change criteria or preset reflection criteria, and generating corresponding event candidate segments; mapping the event candidate segments across bands in the multi-band comparable echo data to determine the echo intervals corresponding to each event candidate segment in each band; for each event candidate segment, extracting multi-band feature parameters to characterize the event within the corresponding echo intervals in each band, and aggregating the multi-band feature parameters to form a corresponding multi-band feature set.

[0011] Optionally, the preset change criterion is as follows: Compare adjacent sampling points of the echo data along the distance dimension to obtain adjacent difference sequences, wherein each difference value of the adjacent difference sequence represents the difference in echo amplitude at adjacent distance positions; when the absolute value of the difference values ​​of the adjacent difference sequences within a preset number of consecutive sampling points is greater than a preset change threshold, the corresponding distance range is determined as the echo change interval; the preset reflection criterion is as follows: Identify local peaks in the echo data and set preset background windows on both sides of the peak position of the local peak to obtain a background amplitude reference; wherein the background amplitude reference is statistically obtained from the echo amplitude within the preset background window; when the difference between the peak amplitude of the local peak and the background amplitude reference is greater than a preset reflection threshold, and the distance span of the local peak corresponding to a preset amplitude ratio is less than a preset peak width threshold, the distance range corresponding to the local peak is determined as the echo change interval.

[0012] Optionally, outputting event location and event category results based on the multi-band feature set includes: for each event candidate segment, reading its corresponding multi-band feature set, and constructing a discriminative feature vector to characterize the event characteristics based on the multi-band feature set, wherein the discriminative feature vector is formed by combining multiple feature parameters in the multi-band feature set; comparing the discriminative feature vector with preset event category feature descriptions one by one, and determining the event category corresponding to the event candidate segment based on the degree of matching between the discriminative feature vector and each event category feature description; determining the event location corresponding to the event candidate segment based on the location range of the event candidate segment in the distance-aligned multi-band comparable echo data, and associating the event location with the event category to output the corresponding event location and event category results.

[0013] A second aspect of the present invention provides an optical fiber loss event classification and detection system, the system comprising: an acquisition unit for acquiring multi-band echo data of the optical fiber under test under a preset wavelength set; a processing unit for constructing cross-band alignment relationships based on the multi-band echo data and performing amplitude comparability processing to generate multi-band comparable echo data; a feature extraction unit for detecting event candidate segments based on the multi-band comparable echo data and extracting corresponding multi-band feature sets; and an output unit for outputting event location and event category results based on the multi-band feature sets.

[0014] A third aspect of the present invention provides an electronic device, comprising: one or more processors; and a storage device having stored one or more programs thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the above-described fiber optic loss event classification and detection method.

[0015] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described fiber optic loss event classification and detection method.

[0016] Through the above technical solution, this invention acquires echo data of the optical fiber under test at multiple preset wavelengths and performs unified distance alignment and amplitude comparability processing on the echo data of different bands. This allows multi-band echo responses at the same physical location to be compared and analyzed under consistent distance coordinates and amplitude benchmarks, thereby avoiding position offset and amplitude mismatch problems caused by differences in propagation characteristics between different band echoes. Based on this, by detecting event candidate segments and extracting corresponding multi-band feature sets from the comparable multi-band echo data, potential loss events in the optical fiber can be stably identified under multi-band information constraints, forming feature descriptions that can be used to distinguish different event types. Furthermore, based on the multi-band feature sets, the event location and event category results are output, establishing event localization and event type determination on the basis of multi-band consistency analysis. This improves the stability and distinguishability of event detection results and enhances the reliability of identifying and classifying different loss events in multi-band optical fiber testing scenarios.

[0017] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings: Figure 1This is a flowchart of the steps of a fiber optic loss event classification and detection method provided in one embodiment of the present invention; Figure 2 This is a comparative schematic diagram of the multi-band echo before and after alignment processing within the event interval, provided by one embodiment of the present invention. Figure 3 This is a system structure diagram of an optical fiber loss event classification and detection system provided in one embodiment of the present invention; Figure 4 This is an internal structural diagram of a computer device provided in one embodiment of the present invention. Detailed Implementation

[0019] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0020] like Figure 1 As shown, embodiments of the present invention provide a method for classifying and detecting fiber optic loss events, the method comprising: Step S10: Obtain multi-band echo data of the optical fiber under test under a preset wavelength set.

[0021] Specifically, based on a preset wavelength set, multi-band probe light sources are controlled to inject probe light of corresponding wavelengths into the fiber under test sequentially or according to a preset time sequence, so as to form a multi-band probe process under the same measurement link conditions. During the injection of probe light in each band, backscattered signals and reflected signals generated along the fiber under test are collected respectively, and the collected backscattered signals and reflected signals are marked according to the corresponding band to form independent echo sequence data for each band. The echo sequence data of each band are uniformly mapped to the distance dimension to generate multi-band echo data with fiber distance as independent variable and echo amplitude as dependent variable.

[0022] In this embodiment of the invention, multi-band echo data of the optical fiber under test under a preset wavelength set is acquired to provide basic data support for subsequent cross-band alignment, amplitude comparability processing, and event candidate segment detection. The process of acquiring multi-band echo data includes multi-band probe light injection, echo signal acquisition, and range domain mapping of the echo data.

[0023] Specifically, based on a preset wavelength set, multi-band probe light sources are controlled to inject probe light of corresponding wavelengths into the fiber under test sequentially or according to a preset timing sequence, so that the probe process at different wavelengths is completed under the same measurement link conditions. The same measurement link conditions here are used to ensure that the access structure, connector status, and measurement path of the fiber under test remain unchanged during multi-band measurement, thereby avoiding additional loss differences introduced by link changes.

[0024] During the probe light injection process at each wavelength, the probe light propagates along the fiber under test. Due to Rayleigh scattering within the fiber material and reflections caused by discontinuities in the fiber, an echo signal is generated at the incident end. The receiving end continuously samples the echo signal. The acquired echo signal simultaneously contains backscattered components continuously distributed along the fiber and reflection components generated by fiber breakage, connector end faces, or local bending discontinuities. To facilitate subsequent unified processing of multi-band data, the echo signals acquired at each wavelength are labeled according to their corresponding wavelengths, forming independent echo sequence data for each wavelength.

[0025] After obtaining the echo sequence data for each band, the echo signals are mapped from the time dimension to the distance dimension. The mapping relationship between echo time and fiber distance is established based on the round-trip propagation characteristics of the probe light in the fiber, and its calculation method is as follows:

[0026] in, Indicates the fiber optic distance location corresponding to the echo signal; This represents the speed of light in a vacuum. This represents the time interval from the start of probe light emission to the reception of the echo, and the time interval is the round-trip propagation time of the probe light in the optical fiber; The effective refractive index of the optical fiber under test is indicated by the corresponding wavelength; the coefficient 2 in the denominator is used to characterize the path relationship of the probe light's round-trip propagation to form the echo signal.

[0027] Based on the above mapping relationship, a time-to-distance conversion is performed on each sampling point in the echo sequence of each band, uniformly representing the echo sequence as distance-domain echo data with fiber distance as the independent variable and echo amplitude as the dependent variable. In specific implementation, the effective refractive index... The value can be obtained from a preset fiber parameter table or equipment configuration parameters, and should remain consistent during the echo data mapping process for the same wavelength; when there are differences in the effective refractive index for different wavelengths, each band should use its corresponding value. Complete the distance mapping and retain the wavelength identifier in the data structure to facilitate the subsequent construction of cross-band distance alignment relationships.

[0028] Through the above processing, multi-band echo data organized under a unified fiber distance coordinate system is obtained, enabling echo responses under different wavelength conditions to correspond and compare in spatial location, providing a reproducible data basis for subsequent cross-band alignment, amplitude comparability processing, and event candidate segment detection.

[0029] Step S20: Based on the multi-band echo data, construct cross-band alignment relationships and perform amplitude comparability processing to generate multi-band comparable echo data.

[0030] Specifically, in the multi-band echo data, a preset reference band is selected, and the echo position corresponding to the preset reference event is identified in the echo data of each band to characterize the distance correspondence of the same physical location in different bands; based on the correspondence between the echo position of the reference band and the corresponding echo position in other bands, a cross-band distance alignment relationship is constructed, and the echo data of non-reference bands is corrected for distance axis according to the cross-band distance alignment relationship to obtain distance-aligned multi-band echo data; based on the reference amplitude information corresponding to each band, amplitude normalization processing is performed on the distance-aligned multi-band echo data to generate multi-band comparable echo data.

[0031] Furthermore, based on the reference amplitude information corresponding to each band, amplitude normalization processing is performed on the distance-aligned multi-band echo data to generate comparable multi-band echo data. This includes: for each band, obtaining reference amplitude information corresponding to the echo data of that band, the reference amplitude information being used to characterize the reference amplitude level of that band under the current measurement conditions; based on the reference amplitude information, determining the amplitude normalization parameter for the corresponding band, the amplitude normalization parameter being used to eliminate amplitude offsets caused by differences in transmit power or link response between different bands; and performing amplitude normalization processing on the distance-aligned multi-band echo data respectively according to the amplitude normalization parameter, so that the echo data of different bands are comparable under a unified amplitude reference, thereby generating comparable multi-band echo data.

[0032] In this embodiment of the invention, cross-band alignment relationships are constructed based on the multi-band echo data and amplitude comparability processing is performed to eliminate the differences between echoes of different bands on the distance axis and amplitude scale, so that multi-band echo responses at the same physical location can be compared and analyzed in a unified data space, thereby generating multi-band comparable echo data.

[0033] Specifically, in the multi-band echo data, a preset reference band is first selected. This reference band serves as a reference coordinate for cross-band distance alignment. Its selection can be based on band stability, signal-to-noise characteristics, or engineering experience, but it remains fixed throughout the same measurement process. Based on the reference band, the echo positions corresponding to preset reference events are identified in the echo data of each band. These reference events are physical events that can be stably observed under multi-band conditions, and their echo characteristics are identifiable in each band, used to characterize the distance correspondence of the same physical location in different bands.

[0034] In a specific implementation, the preset reference event can be a known connector reflection position, a fixed structure reflection position, or other event points with significant characteristics in the echo data within the optical fiber. By determining the reference echo position corresponding to the reference event in the reference band echo data and determining the echo positions corresponding to the reference event in echo data from other bands, a distance correspondence between the reference band and non-reference bands is established. This distance correspondence reflects the offset characteristics between the distance axes in different bands, and its formation process does not rely on additional physical assumptions but is based solely on the echo data itself.

[0035] After obtaining the distance correspondence between the reference band and each non-reference band, a cross-band distance alignment relationship is constructed. Based on this relationship, distance axis correction is performed on the echo data of the non-reference bands. This distance axis correction maps the distance coordinates of the non-reference band echo data to a distance coordinate system consistent with the reference band, thereby aligning the echo data of different bands in spatial location. Through this process, each sampling point in the non-reference band echo data can correspond one-to-one with its corresponding physical location in the reference band, resulting in distance-aligned multi-band echo data.

[0036] After cross-band distance alignment is completed, echo data from different bands have a consistent physical meaning on the distance coordinate system. However, due to differences in transmit power, fiber scattering characteristics, and link response across different bands, inconsistencies in echo amplitude still exist between different bands. Therefore, amplitude normalization processing needs to be performed on the distance-aligned multi-band echo data based on reference amplitude information corresponding to each band to eliminate amplitude offsets between different bands and generate comparable multi-band echo data.

[0037] Specifically, for each band, reference amplitude information corresponding to the echo data of that band is acquired. This reference amplitude information characterizes the reference amplitude level of that band under the current measurement conditions. It can be acquired based on statistical results of echo amplitude within a preset reference interval, or determined based on the echo amplitude corresponding to a preset reference event. The reference amplitude information remains consistent within the same band and serves as a benchmark for amplitude adjustment within that band.

[0038] Based on the reference amplitude information, amplitude normalization parameters for the corresponding band are determined. These parameters are used to perform amplitude scaling adjustments on the echo data of that band to eliminate systematic amplitude shifts caused by differences in transmit power or link response between different bands. In this embodiment, the calculation method for performing amplitude normalization processing on the range-aligned echo data is as follows:

[0039] in, Indicates the wavelength as the first At each band, at the distance position The echo amplitude after amplitude normalization; This indicates that after distance alignment is completed, the wavelength is the [missing value]. At distance position in each band The original echo amplitude at that location; Indicates the relationship with the first The reference amplitude information corresponding to each band is used to characterize the reference amplitude level of that band under the current measurement conditions.

[0040] Through the amplitude normalization process described above, echo data from different bands are represented under a unified amplitude benchmark, thus avoiding inconsistencies in amplitude scale caused by band differences. After normalization, the multi-band echo data is aligned on the distance axis and comparable in amplitude scale, making it directly usable for subsequent event candidate segment detection, multi-band feature extraction, and event location and category determination. Through the aforementioned cross-band distance alignment and amplitude comparability processing, the spatial location and amplitude scale of multi-band echo data are unified, providing a stable and reproducible data foundation for subsequent multi-band information fusion and event analysis.

[0041] In another possible implementation, the construction of cross-band alignment does not rely on a single preset reference event, but is determined based on the joint constraints of multiple reference events. Specifically, multiple spatially distributed reference event locations are simultaneously identified in the echo data of each band, and the corresponding echo locations of each reference event in the reference band and non-reference bands are used as a set of distance correspondence points. Based on the set of distance correspondence points, the range axis of the non-reference band is corrected as a whole, so that the corrected non-reference band echo data is simultaneously aligned with the reference band at multiple reference event locations, thereby reducing the impact of single reference event identification errors on the cross-band alignment results.

[0042] In this embodiment, the reference amplitude information can also be obtained in a segmented manner. Specifically, the distance-aligned echo data is divided into multiple preset distance segments, and the reference amplitude information of the corresponding band is extracted in each segment, thereby determining the segment-level amplitude normalization parameters. By performing amplitude normalization processing on different distance segments separately, the amplitude comparability of multi-band echo data is maintained across the entire distance range, avoiding local amplitude mismatch caused by changes in fiber attenuation along the fiber path. Through the above method, in complex link conditions or long-distance fiber scenarios, the stability of cross-band alignment and amplitude comparability processing can be further improved, providing a more reliable data foundation for subsequent event detection and event classification.

[0043] Step S30: Detect event candidate segments based on the multi-band comparable echo data and extract the corresponding multi-band feature set.

[0044] Specifically, in the multi-band comparable echo data, a preset detection band is selected as the event detection benchmark. Change analysis is performed on the multi-band comparable echo data along the distance dimension to identify echo change intervals that meet preset change criteria or preset reflection criteria, and corresponding event candidate segments are generated. The event candidate segments are mapped across the multi-band comparable echo data to determine the echo intervals corresponding to each event candidate segment in each band. For each event candidate segment, multi-band feature parameters for characterizing the event are extracted in the corresponding echo intervals of each band, and the multi-band feature parameters are aggregated to form a corresponding multi-band feature set.

[0045] Specifically, the preset change criterion is as follows: Adjacent sampling points of the echo data are compared along the distance dimension to obtain adjacent difference sequences, where each difference value of the adjacent difference sequence represents the difference in echo amplitude at adjacent distance positions; when the absolute value of the difference values ​​of the adjacent difference sequences within a preset number of consecutive sampling points is greater than a preset change threshold, the corresponding distance range is determined as the echo change interval; the preset reflection criterion is as follows: Local peaks are identified in the echo data, and preset background windows are set on both sides of the peak position of the local peak to obtain a background amplitude reference; wherein the background amplitude reference is statistically obtained from the echo amplitude within the preset background window; when the difference between the peak amplitude of the local peak and the background amplitude reference is greater than a preset reflection threshold, and the distance span of the local peak corresponding to a preset amplitude ratio is less than a preset peak width threshold, the distance range corresponding to the local peak is determined as the echo change interval.

[0046] In this embodiment of the invention, event candidate segments are detected based on the multi-band comparable echo data, and corresponding multi-band feature sets are extracted. This is used to stably identify potential loss event intervals in the optical fiber under a unified distance coordinate and amplitude reference, and to provide structured feature input for subsequent event location determination and event category classification. This step is based on distance-aligned and amplitude-comparable multi-band echo data, and generates event candidate segments by combining variation criteria and reflection criteria. Multi-band feature parameters are then extracted based on these parameters.

[0047] Specifically, in the multi-band comparable echo data, a preset detection band is first selected as the event detection benchmark. This preset detection band is used for the initial event localization task; its selection can be based on band stability, signal-to-noise ratio, or engineering experience, but remains fixed throughout the same measurement process. Based on the preset detection band, variation analysis is performed on the multi-band comparable echo data along the distance dimension to identify echo variation ranges that may correspond to fiber loss events.

[0048] In the variation analysis process, adjacent sampling points of the echo data are first compared along the distance dimension to obtain the adjacent difference sequence. The adjacent difference sequence is used to characterize the local variation characteristics of the echo amplitude in space, and its calculation method is as follows:

[0049] in, Indicates distance position Adjacent difference values ​​at; Indicates distance position Echo amplitude at the location; Indicates distance position Echo amplitude at the location; and This indicates the distance between adjacent sampling points.

[0050] Based on the adjacent difference sequence, a change determination is further performed within a preset number of consecutive sampling points. When the absolute value of the adjacent difference value is greater than a preset change threshold at multiple consecutive distance sampling points, it is considered that there is a significant echo change within the corresponding distance range, and this distance range is determined as the echo change interval. By introducing a constraint on the number of consecutive sampling points, misjudgments caused by single-point noise or occasional disturbances can be avoided, improving the stability of change interval identification.

[0051] In addition to the change criterion, this embodiment also introduces a preset reflection criterion to identify event intervals with obvious reflection characteristics. Specifically, local peaks are identified in the echo data, and preset background windows are set on both sides of the peak position, centered on the local peak. The background windows are used to obtain the background amplitude level around the local peak, and the background amplitude reference is obtained by performing statistics on the echo amplitude within the background window.

[0052] In the process of determining the reflection criterion, the difference between the peak amplitude of the local peak and the background amplitude reference is first calculated, and the spatial width of the local peak is constrained. The determination relationship of the reflection characteristics can be expressed as follows:

[0053] in, Indicates the significance index of reflection; Indicates the peak echo amplitude corresponding to the local peak. This represents the average echo amplitude within the background window; This represents the standard deviation of the echo amplitude within the background window.

[0054] When the reflection significance index When the distance span corresponding to the local peak is greater than a preset reflection threshold, and the distance span corresponding to the local peak under a preset amplitude ratio is less than a preset peak width threshold, the distance range corresponding to the local peak is determined as the echo variation range. By simultaneously introducing amplitude difference constraints and spatial width constraints, it is possible to effectively distinguish between real reflection events and pseudo-peaks formed by background undulations or noise superposition.

[0055] After generating event candidate segments using change or reflection criteria, these segments are mapped across bands in the multi-band comparable echo data. Since the multi-band echo data has already undergone distance alignment processing, the echo intervals corresponding to each event candidate segment in each band can be directly determined using unified distance coordinates, thereby obtaining the corresponding echo intervals of each event candidate segment in different bands.

[0056] For each candidate event segment, multi-band feature parameters are extracted within the corresponding echo intervals of each band to characterize the event. These multi-band feature parameters include at least echo amplitude variation characteristics, echo morphology characteristics, and relative differences between the multiple bands. In specific implementations, statistical analysis can be performed on the echo amplitudes within each band echo interval to obtain feature parameters such as peak amplitude, average amplitude, and amplitude change rate, and further, a feature description reflecting the differences in multi-band responses can be constructed.

[0057] After extracting the multi-band feature parameters, these parameters are aggregated to form a corresponding multi-band feature set. This multi-band feature set is used to comprehensively characterize the echo response characteristics of the same event candidate segment under different band conditions, enabling a unified expression of information about the event across multiple dimensions, such as spatial location, amplitude variation, and band differences. This provides stable feature input for subsequent event location determination and event category classification.

[0058] Through the above-mentioned event candidate segment detection and multi-band feature set extraction process, stable localization and characteristic description of potential fiber loss events are achieved based on multi-band comparable echo data, providing key support for the overall execution of multi-band fiber loss and event classification detection methods.

[0059] In another possible implementation, the detection of event candidate segments no longer relies solely on a single preset detection band, but instead introduces multi-band consistency constraints to jointly determine candidate segments. Specifically, in multi-band comparable echo data, change analysis and reflection analysis are performed on each band along the distance dimension to generate an initial set of change intervals for the corresponding band. Subsequently, overlap analysis is performed on the initial change intervals of each band along the distance dimension. Only when the same distance range is simultaneously identified as a change interval in no fewer than a preset number of bands is that distance range confirmed as an event candidate segment.

[0060] In this implementation, the extraction of multi-band feature parameters further incorporates cross-band contrast features. Specifically, within each event candidate segment, in addition to extracting the amplitude variation features and reflection features of each band, a consistency index of the echo amplitude variation trend between different bands is calculated to characterize the response coordination of the event under multi-band conditions. By incorporating the consistency index along with the independent features of each band into the multi-band feature set, the influence of noise-type anomalies that only appear in a single band on the event detection results can be effectively suppressed.

[0061] By employing the above methods, the robustness of event candidate segment detection can be further improved in fiber optic link scenarios with high background noise levels or frequent local disturbances, making the generated multi-band feature set more reliable and providing more discriminative input features for subsequent event location determination and event category classification.

[0062] Step S40: Output the event location and event category results based on the multi-band feature set.

[0063] Specifically, for each event candidate segment, its multi-band feature set is read, and a discriminative feature vector is constructed based on the multi-band feature set to characterize the event characteristics. The discriminative feature vector is formed by combining multiple feature parameters from the multi-band feature set. The discriminative feature vector is compared one by one with preset event category feature descriptions. Based on the degree of matching between the discriminative feature vector and each event category feature description, the event category corresponding to the event candidate segment is determined. Based on the position range of the event candidate segment in the distance-aligned multi-band comparable echo data, the event position corresponding to the event candidate segment is determined, and the event position is associated with the event category, outputting the corresponding event position and event category results.

[0064] In this embodiment of the invention, the event location and event category results are output based on the multi-band feature set, which is used to transform the event candidate segments and their multi-band feature information formed in the preceding steps into detection results with clear spatial locations and clear event types. This step, as the result generation stage of the multi-band fiber loss and event classification detection method, takes the event candidate segments and multi-band feature set as inputs and the event location and event category as outputs, completing the mapping from feature description to detection conclusions.

[0065] Specifically, for each event candidate segment, the corresponding multi-band feature set associated with that event candidate segment is first read. The multi-band feature set consists of multiple feature parameters extracted from different bands, used to characterize the echo response characteristics of the same event candidate segment from multiple dimensions, such as amplitude variation, reflection characteristics, and cross-band response differences. In this embodiment, instead of directly judging a single feature parameter, the multi-band feature set is uniformly organized to form a discriminative feature vector characterizing the overall characteristics of the event.

[0066] The discriminative feature vector is obtained by combining multiple feature parameters from the multi-band feature set. Each component corresponds to a feature parameter or a set of feature parameters in the multi-band feature set. By converting the multi-band feature set into a discriminative feature vector, the feature representations of different event candidate segments can have a unified data structure while maintaining the integrity of multi-band information, which facilitates the subsequent event category determination process.

[0067] After obtaining the discriminative feature vector, it is compared one by one with the preset event category feature descriptions. The event category feature descriptions are used to characterize the typical features of different types of fiber optic events under multi-band conditions, and can be composed of historical measurement experience, engineering calibration results, or preset feature ranges. Each event category feature description corresponds to a specific event type and is used as a reference for determining whether the event belongs to that type.

[0068] During the comparison process, the event category corresponding to the event candidate segment is determined based on the degree of matching between the discriminative feature vector and the feature description of each event category. The degree of matching reflects the consistency between the discriminative feature vector and the event category feature description across multiple feature dimensions. The determination process is based on the magnitude of differences, consistency of change trends, or conformity to a preset range among multi-band feature parameters. By comparing the feature descriptions corresponding to different event categories one by one, a unique or highest-priority event category is ultimately determined for each event candidate segment.

[0069] After determining the event category, the event location corresponding to the event candidate segment is further determined. The event location is determined based on the position range of the event candidate segment in the distance-aligned multi-band comparable echo data. Since the distance alignment processing of the multi-band echo data has been completed in the previous steps, the echo intervals corresponding to the event candidate segment in each band have consistent distance coordinates. Therefore, the spatial location of the event in the optical fiber can be directly determined based on the start and end positions of the event candidate segment in the distance dimension.

[0070] The event location is associated with the event category to form a corresponding event detection result, and the event location and event category results are output. This method achieves a complete mapping from a multi-band feature set to event location and event category, ensuring that the detection result not only includes the spatial location information of the event but also clearly indicates the corresponding event type, providing direct evidence for fiber optic operational status assessment, maintenance decisions, or subsequent analysis.

[0071] This implementation method enables the stable output of event detection results with clear physical meaning based on multi-band comparable echo data and multi-band feature sets, thus giving the multi-band fiber loss and event classification detection method good interpretability in engineering applications.

[0072] In one specific implementation, the multi-band fiber loss and event classification detection method proposed in this invention is applied to an online testing scenario of a real communication fiber optic link. The fiber under test is approximately 12km long, with multiple conventional connection nodes along its length. During long-term operation, it may be affected by changes in ambient temperature or external forces, resulting in localized loss or reflection events. To accurately locate and determine the type of these events, the detection system is equipped with a multi-band detection light source, selecting 1310nm, 1550nm, and 1625nm as a preset wavelength set to perform multi-band detection on the fiber under test.

[0073] During the testing process, a multi-band probe light source is controlled to sequentially inject probe light of corresponding wavelengths into the optical fiber under test according to a preset timing sequence. Echo signals generated by each band propagating along the optical fiber are collected under the same measurement link conditions. The receiving end maps the echo signals of each band to a distance dimension based on the round-trip propagation time of the probe light, forming multi-band echo data. Figure 2 The left side shows the distribution of echoes in the distance dimension for each band without cross-band alignment processing.

[0074] like Figure 2 As shown, there is a distinct echo anomaly zone within a distance range of approximately 6km to 8km, which is designated as the event zone. It can be observed that before alignment, the echo peaks corresponding to different bands exhibit significant positional shifts within this event zone. That is, the peak positions of the same physical event do not completely overlap in different bands, which can easily lead to positioning errors when directly determining events based on single-band or unaligned multi-band data.

[0075] To address the aforementioned issues, this implementation selects the 1550nm band as the reference band and uses significant echo peaks within the event interval as reference events. Range axis correction is performed on the echo data of the 1310nm and 1625nm bands to establish a cross-band range alignment relationship. After range alignment, amplitude normalization is further performed on the range-aligned echo data based on the reference amplitude information corresponding to each band, generating comparable multi-band echo data. Figure 2 The right side shows the multi-band echo distribution after cross-band alignment and amplitude comparability processing.

[0076] As can be seen, after processing, the echo peaks of different bands within the event interval achieve good overlap in distance position, indicating that the cross-band alignment relationship is effectively constructed. Based on the aforementioned multi-band comparable echo data, candidate segments of the event are further detected and the corresponding multi-band feature sets are extracted. Finally, it is determined that the event is located at approximately 7km, and based on the multi-band feature response characteristics, it is classified as a loss-type event.

[0077] Through the above implementation methods, accurate location and reliable classification of fiber optic events can be achieved under multi-band conditions. Figure 2 The changes in echo consistency before and after cross-band alignment processing are visually demonstrated, thus showcasing the technical effectiveness of the present invention in multi-band fiber optic testing scenarios.

[0078] like Figure 3 As shown, the present invention also provides an optical fiber loss event classification and detection system 300, which includes: an acquisition unit 301 for acquiring multi-band echo data of the optical fiber under test under a preset wavelength set; a processing unit 302 for constructing cross-band alignment relationships based on the multi-band echo data and performing amplitude comparability processing to generate multi-band comparable echo data; a feature extraction unit 303 for detecting event candidate segments based on the multi-band comparable echo data and extracting corresponding multi-band feature sets; and an output unit 304 for outputting event location and event category results based on the multi-band feature sets.

[0079] The fiber optic loss event classification and detection system 300 provided by this invention can achieve Figure 1 The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0080] The present invention also provides an electronic device, comprising: one or more processors; and a storage device storing one or more programs thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the various steps of the above-described fiber optic loss event classification and detection method embodiments and achieve the same technical effect. To avoid repetition, further details are omitted here.

[0081] The present invention also provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the various steps of the above-described fiber optic loss event classification and detection method embodiment, and achieve the same technical effect. To avoid repetition, these instructions will not be repeated here.

[0082] In one embodiment, the present invention also provides a computer device, which may be a server, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor A01, a network interface A02, a memory (not shown), a database (not shown), a display screen A04, and an input device A05, all connected via a system bus. The processor A01 provides computing and control capabilities. The memory includes internal memory A03 and a non-volatile storage medium A06. The non-volatile storage medium A06 stores an operating system B01, a computer program B02, and a database (not shown). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 stored in the non-volatile storage medium A06. The network interface A02 is used for communication with external terminals via a network connection. When the computer program B02 is executed by the processor A01, it implements various steps of an embodiment of a fiber optic loss event classification and detection method, achieving the same technical effects; therefore, to avoid repetition, it will not be described again here.

[0083] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a microcontroller, chip, or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0084] The optional embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details described above. Within the scope of the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the protection scope of the embodiments of the present invention. It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not further describe the various possible combinations.

[0085] Furthermore, various different embodiments of the present invention can be combined in any way, as long as they do not violate the spirit of the embodiments of the present invention, they should also be regarded as the content disclosed by the embodiments of the present invention.

Claims

1. A method for classifying and detecting optical fiber loss events, characterized in that, The method includes: Acquire multi-band echo data of the optical fiber under test under a preset wavelength set; Based on the multi-band echo data, cross-band alignment relationships are constructed and amplitude comparability processing is performed to generate multi-band comparable echo data. Based on the multi-band comparable echo data, candidate segments of events are detected and corresponding multi-band feature sets are extracted; The event location and event category results are output based on the multi-band feature set.

2. The fiber optic loss event classification and detection method according to claim 1, characterized in that, Acquire multi-band echo data of the optical fiber under test within a preset wavelength set, including: Based on a preset wavelength set, the multi-band detection light source is controlled to inject detection light of the corresponding wavelength into the optical fiber under test sequentially or in a preset time sequence, so as to form a multi-band detection process under the same measurement link conditions. During the probe light injection process in each band, backscattered and reflected signals generated along the optical fiber under test are collected respectively, and the collected backscattered and reflected signals are marked according to the corresponding band to form independent echo sequence data for each band. The echo sequence data of each band are uniformly mapped to the distance dimension to generate multi-band echo data with fiber distance as the independent variable and echo amplitude as the dependent variable.

3. The fiber optic loss event classification and detection method according to claim 1, characterized in that, Based on the multi-band echo data, cross-band alignment relationships are constructed and amplitude comparability processing is performed to generate multi-band comparable echo data, including: In the multi-band echo data, a preset reference band is selected, and the echo position corresponding to the preset reference event is identified in the echo data of each band, which is used to characterize the distance correspondence of the same physical location in different bands. Based on the correspondence between the echo positions of the reference band and the corresponding echo positions in other bands, a cross-band distance alignment relationship is constructed, and the range axis correction is performed on the echo data of non-reference bands according to the cross-band distance alignment relationship to obtain multi-band echo data after distance alignment. Based on the reference amplitude information corresponding to each band, amplitude normalization processing is performed on the distance-aligned multi-band echo data to generate multi-band comparable echo data.

4. The fiber optic loss event classification and detection method according to claim 3, characterized in that, Based on the reference amplitude information corresponding to each band, amplitude normalization processing is performed on the distance-aligned multi-band echo data to generate comparable multi-band echo data, including: For each band, obtain reference amplitude information corresponding to the echo data of that band. The reference amplitude information is used to characterize the reference amplitude level of that band under the current measurement conditions. Based on the reference amplitude information, the amplitude normalization parameter for the corresponding band is determined. The amplitude normalization parameter is used to eliminate amplitude offsets caused by differences in transmit power or link response between different bands. Based on the amplitude normalization parameters, amplitude normalization processing is performed on the distance-aligned multi-band echo data to make the echo data of different bands comparable under a unified amplitude reference, thereby generating multi-band comparable echo data.

5. The fiber optic loss event classification and detection method according to claim 1, characterized in that, Based on the multi-band comparable echo data, candidate segments of events are detected and corresponding multi-band feature sets are extracted, including: In the multi-band comparable echo data, a preset detection band is selected as the event detection benchmark. Change analysis is performed on the multi-band comparable echo data along the distance dimension to identify echo change intervals that meet preset change criteria or preset reflection criteria, and corresponding event candidate segments are generated. The event candidate segments are mapped across bands in the multi-band comparable echo data to determine the echo interval corresponding to each event candidate segment in each band. For each candidate segment of an event, multi-band feature parameters for characterizing the event are extracted in the corresponding echo intervals of each band, and the multi-band feature parameters are aggregated to form a corresponding multi-band feature set.

6. The fiber optic loss event classification and detection method according to claim 5, characterized in that, The preset change criterion is: The echo data is compared along the distance dimension to obtain an adjacent difference sequence, wherein each difference value of the adjacent difference sequence represents the difference in echo amplitude at adjacent distance positions; When the absolute value of the difference between adjacent difference sequences within a preset number of consecutive sampling points is greater than a preset change threshold, the corresponding distance range is determined as the echo change interval. The preset reflection criterion is: Local peaks are identified in the echo data, and preset background windows are set on both sides of the peak position of the local peaks to obtain a background amplitude reference; wherein, the background amplitude reference is obtained by statistical analysis of the echo amplitude within the preset background windows; When the difference between the peak amplitude of the local peak and the background amplitude reference is greater than a preset reflection threshold, and the distance span of the local peak corresponding to the preset amplitude ratio is less than a preset peak width threshold, the distance range corresponding to the local peak is determined as the echo variation range.

7. The fiber optic loss event classification and detection method according to claim 1, characterized in that, Based on the multi-band feature set, the output results of event location and event category include: For each candidate segment of an event, its multi-band feature set is read, and a discriminative feature vector is constructed based on the multi-band feature set to characterize the event characteristics. The discriminative feature vector is formed by combining multiple feature parameters in the multi-band feature set. The discriminative feature vector is compared one by one with the preset event category feature descriptions. Based on the degree of matching between the discriminative feature vector and each event category feature description, the event category corresponding to the event candidate segment is determined. Based on the location range of the event candidate segment in the distance-aligned multi-band comparable echo data, the event location corresponding to the event candidate segment is determined, and the event location is associated with the event category, and the corresponding event location and event category results are output.

8. A fiber optic loss event classification and detection system, characterized in that, The system includes: The acquisition unit is used to acquire multi-band echo data of the optical fiber under test under a preset wavelength set; The processing unit is used to construct cross-band alignment relationships based on the multi-band echo data and perform amplitude comparability processing to generate multi-band comparable echo data; The feature extraction unit is used to detect event candidate segments based on the multi-band comparable echo data and extract the corresponding multi-band feature set; The output unit is used to output the event location and event category results based on the multi-band feature set.

9. An electronic device, characterized in that, include: One or more processors; A storage device having stored one or more programs thereon, which, when executed by the one or more processors, cause the one or more processors to implement the fiber optic loss event classification and detection method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the fiber optic loss event classification and detection method as described in any one of claims 1-7.

Images