Power distribution optical cable anomaly detection and management method, system, equipment and medium

Abstract

The invention discloses a power distribution optical cable anomaly detection and management method, system, device and medium, and belongs to the field of power systems.The method comprises the steps that an initial three-dimensional point cloud model is obtained, a first abnormal area of the model is determined according to point cloud data, and feature parameters of the first abnormal area are determined; adjusting a sensitivity threshold value of the model according to the characteristic parameters, further adjusting a characteristic threshold value corresponding to the second characteristic, and determining a second abnormal region and a damage degree; determining the position of the target abnormal area through the positioning mark, and performing feature analysis on the target abnormal area to determine an influence range; adjusting the model parameters according to the influence range, and refitting the point cloud data to obtain a target three-dimensional point cloud model; and determining an anomaly detection result according to the target three-dimensional point cloud model, and performing power distribution optical cable management according to the anomaly detection result, so that the precision and accuracy of power distribution optical cable anomaly detection can be improved, and the management efficiency is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power systems, in particular to a power distribution optical cable anomaly detection and management method, system, device and medium. BACKGROUND

[0002] In modern power distribution optical cable inventory management, intelligent warehouse technology based on machine vision improves inventory management efficiency and resource utilization through efficient and accurate automation, thereby realizing stable operation and cost control of the power system.

[0003] In the power system, as the demand for power distribution optical cables increases, traditional inventory management methods rely on manual inspection and static records, which are not only inefficient, but also prone to inventory state evaluation errors due to human negligence, making it difficult to adapt to large-scale and complex inventory management needs in the current environment. Moreover, in the case of long-term storage of optical cable reels, existing solutions use fixed feature extraction models for inventory management, which are difficult to adapt to the diversified abnormalities caused by deformation or damage of the optical cable reels, and lack robustness to environmental noise and illumination changes, resulting in low model anomaly recognition accuracy. Especially in the case of complex and variable storage conditions of optical cable reels, misjudgment and missed judgment of anomaly recognition occur frequently, further reducing inventory management efficiency. SUMMARY

[0004] The present application provides a power distribution optical cable anomaly detection and management method, system, device and medium, which can solve the problem of low anomaly detection precision and accuracy in power distribution optical cable inventory management.

[0005] The present application provides a power distribution optical cable anomaly detection and management method, comprising: Obtain point cloud data when detecting each power distribution optical cable, and construct a corresponding initial three-dimensional point cloud model based on each point cloud data; In each initial three-dimensional point cloud model, calculate point cloud density data corresponding to a first feature based on the point cloud data, determine a first abnormal area based on the point cloud density data; determine the feature parameters of the first abnormal area, and adjust the sensitivity threshold of the initial three-dimensional point cloud model in the anomaly detection process according to the feature parameters, and then adjust the feature threshold corresponding to a second feature, determine a second abnormal area and a damage degree; Locate the damage degree of the first abnormal area and the second abnormal area to determine the location of the target abnormal area, and perform feature analysis on the target abnormal area to determine the influence range; adjust the model parameters of the initial three-dimensional point cloud model according to the influence range, and refit each point cloud data according to the adjusted model parameters to obtain a target three-dimensional point cloud model; Determine the anomaly detection result according to the target three-dimensional point cloud model, and manage the power distribution optical cable according to the anomaly detection result.

[0006] The embodiment constructs an initial three-dimensional point cloud model by acquiring point cloud data, determines a first abnormal area based on point cloud density data, adjusts a sensitivity threshold according to a feature parameter, and further adjusts a feature threshold of a second feature to determine a second abnormal area and a damage degree. The abnormal area is marked and analyzed to determine an influence range, and finally the model parameters are adjusted according to the influence range to obtain a target three-dimensional point cloud model. Through multi-dimensional detection and model adjustment, the abnormal area is accurately positioned, and the abnormal detection accuracy and inventory management efficiency are improved.

[0007] Further, the abnormal detection result is determined according to the target three-dimensional point cloud model, and the power distribution optical cable management is performed according to the abnormal detection result, further comprising: determining a target multi-dimensional visual feature vector of the target three-dimensional point cloud model, analyzing the change trend between the multi-dimensional visual feature vector and the target multi-dimensional visual feature vector to obtain index data; Based on the index data and the environment data of the power distribution optical cable, a deformation cumulative value is calculated; if the deformation cumulative value is not in a preset range, the abnormal detection parameters of the target three-dimensional point cloud model are adjusted through a random forest algorithm to obtain the abnormal detection result and the inventory state data corresponding to the environment data; The support vector machine algorithm is used to determine the real-time turnover efficiency by combining the inventory state data with the target three-dimensional point cloud model, and the efficiency evaluation result is determined according to the real-time turnover efficiency; According to the efficiency evaluation result and the index data, a power distribution optical cable management scheme is determined. In this way, by performing feature extraction on the running data, information from different representation subspaces in different positions can be focused on, and data from a long distance can be combined to comprehensively capture multi-scale features.

[0008] In this way, multi-dimensional visual features and environment data can be combined to optimize abnormal detection parameters and improve inventory management efficiency.

[0009] Further, the point cloud density data corresponding to the first feature is calculated based on the point cloud data, and the first abnormal area is determined based on the point cloud density data, specifically: The initial three-dimensional point cloud model is statistically analyzed to obtain the point cloud density data, and the area whose point cloud density data does not meet the preset range condition is determined as the first abnormal area; The sensitivity threshold is adjusted according to the feature parameter, and an adjustment coefficient is calculated according to the adjusted sensitivity threshold. The initial three-dimensional point cloud model is updated through the adjustment coefficient to obtain updated point cloud density data; If the preset range condition is still not met based on the updated point cloud density data, iterative updating is performed until the preset range condition is met, and the first abnormal area is obtained.

[0010] In this way, the point cloud density distribution is gradually optimized in a data-driven manner, which not only improves the consistency of the model with the actual object, but also makes the positioning of the abnormal area more accurate, and the sensitivity adjustment provides flexibility for optimization.

[0011] Further, the sensitivity threshold of the initial three-dimensional point cloud model in the abnormality detection process is adjusted according to the feature parameter, and then the feature threshold corresponding to the second feature is adjusted to determine the second abnormal area and the damage degree, specifically: The threshold value corresponding to the second feature is updated based on the adjustment coefficient to obtain the feature threshold; Obtain the reflection intensity data corresponding to the second feature, and perform time series analysis on the reflection intensity data to obtain a feature sequence; According to the feature threshold and the feature sequence, a plurality of second abnormal areas are determined, and statistical analysis is performed on each second abnormal area to obtain a damage degree.

[0012] In this way, the matching result between the point cloud model and the actual object is optimized by dynamically adjusting the feature matching strategy, and the abnormality detection accuracy is improved.

[0013] Further, the point cloud data of the power distribution cable resources in the resource library is obtained, and the corresponding initial three-dimensional point cloud model is constructed based on each point cloud data, specifically: The point cloud data is preprocessed, and the point cloud data after data preprocessing is processed by using a statistical filtering algorithm to generate a three-dimensional point cloud model; Extract the edge contour features of the three-dimensional point cloud model, determine the first feature and the second feature based on the edge contour features and the three-dimensional point cloud data, and fuse the first feature and the second feature to generate a multi-dimensional visual feature vector; According to the geometric deviation between the three-dimensional point cloud model of the current iteration and the standard point cloud model, and combining the multi-dimensional visual feature vector, the matching parameters between the three-dimensional point cloud model and the standard point cloud model are updated until the geometric deviation is greater than or equal to a first preset threshold, the iteration is stopped, and the initial three-dimensional point cloud model is output.

[0014] In this way, through data filtering, the initial three-dimensional point cloud model is smoother, which can clearly show the cylindrical structure of the cable drum, and effectively improve the accuracy of subsequent feature extraction. By comparing with the standard cable drum features, it can quickly judge whether the quality meets the standard.

[0015] Further, the first feature and the second feature are determined based on the edge contour feature and the three-dimensional point cloud data, the first feature and the second feature are fused to generate a multi-dimensional visual feature vector, and specifically: A point distance distribution value is calculated from the edge contour feature, and tightness is determined according to the point distance distribution value. The surface protection layer of the power distribution optical cable resource is extracted, if the point cloud density value of the surface protection layer is lower than the second preset threshold value, it is determined that the surface protection layer is abnormal, and a protection integrity feature is obtained; The multi-dimensional visual feature vector is fused based on the tightness feature and the protection integrity feature.

[0016] In this way, by updating the sensitivity threshold, the protection layer damage is accurately identified and classified, and the abnormal detection accuracy is improved.

[0017] Further, the feature analysis is performed on the target abnormal area to determine the influence range, and specifically: The shape feature and the distribution feature of the target abnormal area are extracted, the influence range is determined based on the shape feature and the distribution feature, and if the influence range is greater than or equal to the third preset threshold value, the model parameters are adjusted.

[0018] In this way, by positioning the mark and adjusting the model parameters, the damage abnormal area is accurately determined, and the abnormal detection accuracy is improved.

[0019] Another embodiment of the application also provides a power distribution optical cable abnormality detection and management system, comprising: an acquisition module, an abnormal area determination module, a model adjustment module and a management module; The acquisition module is used to acquire point cloud data when detecting each power distribution optical cable, and construct a corresponding initial three-dimensional point cloud model based on each point cloud data; The abnormal area determination module is used to calculate point cloud density data corresponding to a first feature based on the point cloud data in each initial three-dimensional point cloud model, determine a first abnormal area based on the point cloud density data, determine feature parameters of the first abnormal area, and adjust a sensitivity threshold of the initial three-dimensional point cloud model in the abnormality detection process according to the feature parameters, and then adjust a feature threshold corresponding to a second feature, determine a second abnormal area and a damage degree; The model adjustment module is used to locate and mark the damage degree of the first abnormal area and the second abnormal area to determine the position of a target abnormal area, perform feature analysis on the target abnormal area to determine an influence range, adjust model parameters of the initial three-dimensional point cloud model according to the influence range, and refit each point cloud data according to the adjusted model parameters to obtain a target three-dimensional point cloud model; The management module is configured to determine an anomaly detection result according to the target three-dimensional point cloud model, and perform power distribution optical cable management according to the anomaly detection result.

[0020] The embodiment constructs an initial three-dimensional point cloud model by acquiring point cloud data, determines a first abnormal area based on point cloud density data, adjusts a sensitivity threshold according to a feature parameter, and further adjusts a feature threshold of a second feature to determine a second abnormal area and a damage degree. The abnormal area is marked and analyzed to determine an influence range, and finally, a target three-dimensional point cloud model is obtained by adjusting model parameters and re-fitting according to the influence range. Through multi-dimensional detection and model adjustment, the abnormal area is accurately positioned, and the abnormal detection accuracy and inventory management efficiency are improved.

[0021] Another embodiment of the present application also provides a terminal device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, the steps of the power distribution optical cable anomaly detection and management method are implemented.

[0022] Another embodiment of the present application also provides a computer readable storage medium item, comprising a stored computer program, and when the computer program runs, the device where the computer readable storage medium is located executes the steps of the power distribution optical cable anomaly detection and management method. BRIEF DESCRIPTION OF DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0024] Figure 1 is a flowchart of an embodiment of the power distribution optical cable anomaly detection and management method; Figure 2 is a structural schematic diagram of an embodiment of the power distribution optical cable anomaly detection and management system. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical scheme and advantages of the present application clearer, the technical scheme in the present application will be described clearly and completely in the following combined with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0026] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of this application; the use of the terms "including," "comprising," or "having" and variations thereof herein is intended to be broad and encompass the terms "consisting of" and "consisting essentially of" and variations thereof. Unless otherwise noted, the terms "including" and "comprising" are open-ended and do not exclude the presence of unrecited elements or limitations.

[0027] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "a plurality of" is two or more, unless otherwise explicitly specified.

[0028] Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearance of the phrase in various places in the specification does not necessarily all refer to the same embodiment, nor is it necessarily independent or alternative embodiments to each other. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0029] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which means that there are three cases of A alone, A and B together, and B alone. In addition, the character " / " in this paper generally represents that the front and rear associated objects are a "or" relationship.

[0030] In the description of the embodiments of the present application, the term "a plurality of" refers to two or more (including two), and similarly, "a plurality of groups" refers to two or more groups (including two groups), and "a plurality of pieces" refers to two or more pieces (including two pieces).

[0031] In the description of the embodiments of the present application, unless otherwise explicitly specified and limited, the technical terms "mounting", "connecting", "connecting", "fixing" and the like should be understood in a broad sense, for example, it can be fixedly connected, or it can be detachably connected, or it can be integrated; it can be mechanical connection, or it can be electrical connection; it can be directly connected, or it can be indirectly connected through an intermediate medium; it can be the internal communication of two elements or the interaction relationship between two elements. For those skilled in the art, the specific meaning of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0032] Reference Figure 1 To solve the problem of low precision and accuracy of power distribution optical cable abnormality detection in the prior art, an embodiment of the present application provides a power distribution optical cable abnormality detection and management method, which comprises the steps ofFigure 1 As shown, the method comprises steps S101-S103, comprising: Step 101, acquiring point cloud data when detecting each power distribution optical cable, and constructing a corresponding initial three-dimensional point cloud model based on each point cloud data.

[0033] In this embodiment, the optical cable drum in the warehouse is scanned by a laser scanning device, three-dimensional point cloud data of the optical cable drum is collected, including geometric shape, winding tightness and surface protection information, and point cloud data is obtained. For example, a high-precision laser scanner is used to scan the optical cable drum to generate a point cloud data set containing millions of points, which not only reflects the outer contour of the optical cable drum, but also includes the winding layer spacing and surface details. It should be noted that the advantage of laser scanning is non-contact measurement, which can obtain high-resolution data without disturbing the state of the optical cable drum, laying a foundation for subsequent analysis. Based on the point cloud data, a three-dimensional point cloud model is obtained by three-dimensional modeling, which can be matched with a standard model to obtain an aligned initial three-dimensional point cloud model.

[0034] As an example of an embodiment of the present application, the point cloud data of the power distribution optical cable resource in the resource library is acquired, and a corresponding initial three-dimensional point cloud model is constructed based on each point cloud data. Specifically, the point cloud data is preprocessed, and the preprocessed point cloud data is processed by a statistical filtering algorithm to generate a three-dimensional point cloud model; the edge contour features of the three-dimensional point cloud model are extracted, the first feature and the second feature are determined based on the edge contour features and the three-dimensional point cloud data, and the first feature and the second feature are fused to generate a multi-dimensional visual feature vector; according to the geometric deviation between the three-dimensional point cloud model of the current iteration and the standard point cloud model, the matching parameters between the three-dimensional point cloud model and the standard point cloud model are updated in combination with the multi-dimensional visual feature vector, until the geometric deviation is greater than or equal to a first preset threshold, the iteration is stopped, and the initial three-dimensional point cloud model is output.

[0035] In this embodiment, the point cloud data is filtered by a statistical filtering algorithm to remove noise in the point cloud data. For example, assuming that the collected point cloud data is mixed with outliers caused by environmental light interference, the statistical filtering calculates the distance distribution of each point and its neighborhood points, and if a point deviates too far from the average distance, such as more than 2 standard deviations, it is considered as noise and is removed. After filtering, the initial three-dimensional point cloud model is smoother, which can clearly show the cylindrical structure of the cable reel and effectively improve the accuracy of subsequent feature extraction. The three-dimensional point cloud model after noise removal is processed using a statistical filtering algorithm to generate a three-dimensional point cloud model. For the three-dimensional point cloud model, the edge contour features corresponding to the geometric shape are extracted, and the Euclidean clustering algorithm is used to determine the contour boundary, wherein the Euclidean clustering algorithm is based on the spatial distance between points for clustering, for example, setting the distance threshold to 5 millimeters, the point cloud of the edge of the cable reel is clustered into a class, and the contour and background are separated. The multi-dimensional visual feature vector can be generated based on the elasticity feature and the protection integrity feature, and the initial three-dimensional point cloud model data is calculated through the multi-dimensional visual feature vector, and the calculation process relies on the multi-dimensional information of the point cloud data. For example, in the detection scene of the cable reel, the collected point cloud data may contain spatial coordinates and reflection intensity information, and the feature vector can be generated by combining these data, for example, assuming that the diameter of the cable reel is 1 meter, and the number of collected point cloud data points is 100,000, the edge features can be extracted by clustering method to generate a feature vector containing position and density. The geometric deviation value can be obtained by comparing the three-dimensional point cloud with the standard model, and the point-to-point distance calculation method can be used, specifically, the standard model may be a preset ideal geometric shape of the cable reel, such as a cylinder. Assuming that the standard model has a diameter of 1 meter, and the initial three-dimensional point cloud model has a diameter measurement of 1.02 meters, the deviation value is 0.02 meters. If the preset threshold is 0.01 meters, the deviation exceeds the threshold. This comparison can intuitively reflect the deformation of the model and provide a basis for subsequent adjustment. If the geometric deviation exceeds the preset threshold, the feature matching strategy is updated using a dynamic adjustment method. For example, the weight parameters of point cloud registration can be adjusted to preferentially match the key areas such as the edge part, and the initial matching may only focus on the global point cloud distribution, while the weight proportion of the edge point can be increased from 30% to 50% after adjustment, thereby improving the matching accuracy. This strategy optimization can effectively reduce the deviation. New matching parameters are obtained according to the adjusted matching strategy to generate optimized feature matching data, and this process can introduce local feature descriptors, for example, for the winding area of the cable reel, the curvature feature of the point cloud is extracted to generate new matching parameters, assuming that the curvature value is adjusted from 0.1 to 0.05, the matching data is closer to the actual shape. This method can improve the relevance of the data. The updated geometric deviation is calculated through the optimized feature matching data to determine whether it meets the requirements. For example, the updated deviation is reduced from 0.02 meters to 0.008 meters, which is less than the threshold of 0.01 meters, and it meets the requirements. This verification method ensures the reliability of the results.If the updated geometric deviation still exceeds the preset threshold, the matching parameters are adjusted through iterative processing to obtain an optimization result. For example, if the initial parameters are iterated only 5 times and the deviation is still 0.015 meters, the number of iterations can be increased to 10, and the registration angle is adjusted each time to reduce the deviation to 0.009 meters. According to the optimization result, the matching accuracy of the three-dimensional point cloud model and the standard model is determined. Specifically, the matching accuracy can be evaluated by counting the coincidence rate of the matching points. For example, if the coincidence rate increases from 85% to 95%, it indicates that the accuracy has been significantly improved. This high-precision matching helps subsequent quality detection of the optical cable reel and ensures that the product meets the standards. In actual applications, this method can also reduce false positives caused by deviations and improve detection efficiency. For example, through dynamic adjustment and iterative optimization of multi-dimensional features, the entire process can adapt to the detection needs of different optical cable reels, and has strong practicality.

[0036] As an example of an embodiment of the present application, the first feature and the second feature are determined based on the edge contour feature and the three-dimensional point cloud data, and a multi-dimensional visual feature vector is generated by fusing the first feature and the second feature. Specifically, a point distance distribution value is calculated from the edge contour feature, and a tightness feature is determined based on the point distance distribution value. The surface protective layer of the power distribution optical cable resource is extracted, and if the point cloud density value of the surface protective layer is lower than a second preset threshold, it is determined that the surface protective layer is abnormal, and a protection integrity feature is obtained. The multi-dimensional visual feature vector is generated by fusing the tightness feature and the protection integrity feature.

[0037] In this embodiment, the first feature is the tightness feature determined when analyzing the tightness of the cable drum winding, and the second feature is the protection integrity feature determined when analyzing the integrity of the surface protection layer of the cable drum. Specifically, when analyzing the tightness of the winding, the point spacing distribution is calculated from the edge contour feature. For example, in the point cloud of the side surface of the cable drum, if the average value of the point spacing of a certain winding area is 3 mm and the standard deviation is less than 0.5 mm, it can be determined that the winding is uniform; if the standard deviation rises to 1.2 mm, it indicates that the tightness is uneven. This analysis can intuitively reflect the quality stability of the cable drum and help optimize the production process. When obtaining the integrity information of the surface protection layer, the point cloud density is a key indicator. For example, if the density threshold is set to 100 points per square centimeter, and the density of a certain area drops to 70 points, it may be due to the peeling of the protection layer, resulting in sparse point cloud, and it is determined that there is a defect. This method can quickly locate the problem area and improve the detection efficiency. After fusing the tightness analysis result and the protection integrity feature to generate multi-dimensional visual features, the principal component analysis algorithm can be used for dimension reduction processing. For example, the feature set containing tightness uniformity and density distribution is compressed to 3 main dimensions, retaining 95% of the information variance. This dimension reduction can reduce the computational complexity while maintaining the effectiveness of the features, providing convenience for subsequent vectorization. When the optimized visual feature set is vectorized, the final feature representation can be generated through matrix operations. Specifically, the reduced feature set is mapped to a 1x3 vector, such as [0.85, 0.92, 0.78], which represents the comprehensive scores of geometry, winding tightness, and surface protection, respectively. This representation method is convenient for storage and comparison, such as comparing with the standard cable drum features to quickly determine whether the quality meets the standards. In one embodiment, this feature vector can also be input into a machine learning model to further predict the service life of the cable drum, significantly improving the intelligent level of quality inspection.

[0038] In step 102, based on the point cloud data, the point cloud density data corresponding to the first feature is calculated in each of the initial three-dimensional point cloud models, and the first abnormal area is determined based on the point cloud density data; the feature parameters of the first abnormal area are determined, and the sensitivity threshold of the initial three-dimensional point cloud model in the abnormal detection process is adjusted according to the feature parameters, and then the feature threshold corresponding to the second feature is adjusted to determine the second abnormal area and the damage degree.

[0039] In this embodiment, the point cloud data after matching is processed by statistical analysis to calculate the point cloud density distribution corresponding to the tightness, and the point cloud density data is obtained. This process aims to extract regularity information from massive point cloud data. For example, assuming that the point cloud data is derived from the surface scanning of a mechanical part, the statistical analysis can be based on the spacing of each point in the point cloud to calculate the density value, such as the number of points per cubic millimeter, to obtain a density distribution curve. According to the density distribution curve, the preset density interval is determined to be 50-70 points per cubic millimeter. The point cloud density data is compared with the preset density interval, and if it deviates, the deviation state is determined. Specifically, if it is detected that the density of a certain area is only 30 points per cubic millimeter, which is obviously lower than the normal range, it can be determined as a "loose" state. It should be noted that this deviation may be caused by insufficient scanning device accuracy or irregular part surface. Preferably, the deviated area is marked in red by a visualization tool for subsequent processing. For the deviation state, multiple loose tightness abnormal areas are identified using cluster analysis, for example, the K-means clustering method can be used to classify points with a density lower than 40 points per cubic millimeter as an abnormal cluster, and finally multiple loose tightness abnormal areas are obtained. For example, the edge area of a certain part may be sparse due to wear, forming an independent abnormal area. This method helps to quickly locate the problem area and improve analysis efficiency. The sensitivity threshold in the abnormal detection process of the initial three-dimensional point cloud model is adjusted according to the characteristic parameters of the loose tightness abnormal area, and the adjustment coefficient is calculated according to the adjusted sensitivity threshold. The initial three-dimensional point cloud model is updated by the adjustment coefficient to obtain updated point cloud density data. If the updated point cloud density data still does not meet the preset range condition, iterative updating is performed until the preset range condition is met to obtain the first abnormal area. The characteristic parameters of the first abnormal area are determined, and the sensitivity threshold in the abnormal detection process of the initial three-dimensional point cloud model is adjusted according to the characteristic parameters, and then the characteristic threshold corresponding to the second characteristic is adjusted to determine the second abnormal area and the damage degree.

[0040] As an example of an embodiment of the present application, the point cloud density data corresponding to the first characteristic is calculated based on the point cloud data, and the first abnormal area is determined based on the point cloud density data, specifically: statistical analysis of the initial three-dimensional point cloud model is performed to obtain the point cloud density data, and the area where the point cloud density data does not meet the preset range condition is determined as the first abnormal area; the sensitivity threshold is adjusted according to the characteristic parameters, and the adjustment coefficient is calculated according to the adjusted sensitivity threshold. The initial three-dimensional point cloud model is updated by the adjustment coefficient to obtain updated point cloud density data. If the updated point cloud density data still does not meet the preset range condition, iterative updating is performed until the preset range condition is met to obtain the first abnormal area.

[0041] In this embodiment, the features are extracted from the first abnormal region, and the feature parameters of the abnormal region are determined, wherein the feature parameters include the average density, the boundary length or the shape complexity of the region. For example, an abnormal region has an average density of 35 points per cubic millimeter, a boundary length of 10 millimeters, and an irregular polygon shape. These parameters provide data support for subsequent adjustment. Specifically, if the low density indicates that the abnormal detection is too strict and the condition needs to be relaxed, the sensitivity is adjusted according to the feature parameters to calculate an adjustment coefficient. For example, if the density of an abnormal region is low, the sensitivity can be reduced by 20%, and an adjustment coefficient such as 0.8 is generated. After adjustment, the density distribution is recalculated to observe whether it tends to be normal. This adjustment can effectively reduce false positives and improve the robustness of matching. Through the adjustment coefficient, the feature matching result in the abnormal detection is optimized to obtain an optimized point cloud density distribution. For example, the density distribution value of a certain region is 30 points per cubic millimeter, and through the adjustment coefficient, the density distribution value of the region is recalculated to obtain 45 points per cubic millimeter, which is closer to the normal range. Preferably, this optimization can make the point cloud model more realistically reflect the surface characteristics of the part. If the optimized distribution still deviates from the normal range, repeat the clustering analysis and feature extraction to obtain a new adjustment coefficient. For example, the sensitivity can be further reduced to 0.6 in the second iteration until the requirements are met. This iterative approach ensures gradual approximation of the result, improving matching accuracy. It should be noted that the above process gradually optimizes the point cloud density distribution through data-driven methods, which not only improves the consistency of the model with the actual object, but also provides a reliable basis for subsequent quality detection. For example, in part manufacturing, this method can be used to identify processing defects to ensure product quality. It can be understood that the combination of clustering and feature extraction makes the positioning of the abnormal region more accurate, and the sensitivity adjustment provides flexibility for optimization.

[0042] As an example of an embodiment of the application, adjusting the sensitivity threshold of the initial three-dimensional point cloud model in the abnormal detection process according to the feature parameters, and then adjusting the feature threshold corresponding to the second feature to determine the second abnormal region and the damage degree, specifically: updating the threshold corresponding to the second feature based on the adjustment coefficient to obtain the feature threshold; obtaining the reflection intensity data corresponding to the second feature, and performing time series analysis on the reflection intensity data to obtain a feature sequence; determining a plurality of second abnormal regions according to the feature threshold and the feature sequence, and performing statistical analysis on each of the second abnormal regions to obtain a damage degree.

[0043] In this embodiment, the initial intensity threshold is updated to obtain a target intensity threshold through a preset sensitivity threshold adjustment formula and an adjustment coefficient, and the sensitivity threshold adjustment formula is as follows: S=S0(1+kC); wherein S is the target intensity threshold, S0 is the initial intensity threshold, k is the adjustment coefficient, and C is the sensitivity threshold.

[0044] For example, in the surface protection layer detection, the initial intensity threshold S0 is set to 50, representing the reference value of normal reflection intensity, the adjustment factor k is usually taken as 0.2, for indicating the adjustment amplitude, and the sensitivity threshold C is set to 0.5 according to actual needs, obtaining the target intensity threshold S as 55. The core of this way is to adjust flexibly according to the environmental or material characteristics, avoiding misjudgment caused by fixed threshold. It can be understood that this adjustment can more accurately adapt to the aging degree or external interference of different protection layers.

[0045] For example, the reflection intensity data of the surface protection layer is collected by the optical sensor every 1 hour for 24 hours, obtaining the intensity change sequence. For example, the initial intensity is 48, which rises to 56 after 10 hours, and then drops to 52 after 5 hours. Combined with the target intensity threshold 55, if a certain point in the intensity sequence exceeds 55, it is marked as an abnormal point. The advantage of this method is that it can monitor the state of the protection layer in real time and discover potential problems in time. Specifically, the distribution characteristics of the abnormal points are counted, and the concentration degree of the abnormal points on the time axis is analyzed. For example, if there are 5 abnormal points within 24 hours, and they are all concentrated in the 8th to 12th hour, it may indicate that the protection layer is impacted by external factors during this period. Based on this, the damage degree can be divided into three levels of slight, medium and severe, corresponding to the number of abnormal points 1-3, 4-6 and 7 or more. This classification helps to quickly judge whether the protection layer needs maintenance. Preferably, when grouping the damage category set, the K-means method can be used to group the abnormal points according to the intensity value and time interval into 3 groups. For example, the first group has intensity concentrated in 55-60, and the time distribution is uniform; the second group has intensity exceeding 65, and the time is concentrated in a certain 2 hours. Such grouping can clearly reflect the regularity of damage, facilitating subsequent analysis. It should be noted that the clustering result can also reveal whether the anomaly is caused by a single factor. In an embodiment, when extracting key parameters from the protection layer characteristics, attention can be paid to the intensity peak value and the decay rate. For example, the peak value of a certain group of abnormal points is 68, and the decay rate is 2 units per hour, indicating a high possibility of local peeling of the protection layer. Another group has a peak value of 60 and a decay rate of 0.5, which may be a slight wear. The benefit of such parameter extraction is that it can intuitively judge whether the overall state of the protection layer is stable. For example, when judging the overall state of the surface protection layer, if the key parameters show that the peak values of multiple groups of abnormal points all exceed 60 and the decay rate is fast, it may indicate that the protection layer is close to failure and needs to be replaced. If only one group of abnormal points and the decay rate is slow, it may be a local problem and can be repaired.

[0046] In step 103, the damage degree of the first abnormal area and the second abnormal area is marked to determine the location of the target abnormal area, the target abnormal area is analyzed to determine the influence range, the model parameters of the initial three-dimensional point cloud model are adjusted according to the influence range, and each point cloud data is refitted according to the adjusted model parameters to obtain a target three-dimensional point cloud model.

[0047] In this embodiment, the target abnormal area is identified by the damage degree and the tightness abnormal area, and the state of the surface protective layer is accurately judged. For example, when detecting a certain metal protective coating, if the tightness abnormality is manifested as the expansion or contraction of a local area, the stress distribution data can be collected by the sensor to identify the abnormal area. For example, if the stress value of a certain area exceeds the normal range by 20%, combined with the damage category such as "slight peeling" or "deep crack", the abnormality can be preliminarily determined. The feature analysis of the target abnormal area can extract the feature analysis results corresponding to the shape feature and the distribution feature, and the influence range of the abnormal area is determined according to the feature analysis results. If the influence range threshold is set to 10 cm, when the influence range of a certain abnormal area reaches 12 cm, the model parameters are dynamically adjusted, for example, the weight of the fitting algorithm is adjusted from 0.5 to 0.7 to enhance the attention to the abnormal area. This adjustment can optimize the fitting effect of the subsequent geometric shape. In one embodiment, after the geometric shape is refitted using the adjusted model parameters, the preliminary updated three-dimensional point cloud model shows that the edge of the abnormal area is smoother. For example, the jagged point cloud of the original crack edge is corrected to a continuous curve. This fitting process can reduce data noise and improve the reliability of the model. By analyzing the feature changes of the abnormal area through the preliminary updated three-dimensional point cloud model, specifically, the crack width can be observed to decrease from 2 mm to 1.5 mm, indicating that the fitting accuracy has been improved. If the accuracy is not up to standard, for example, the error is still more than 0.5 mm, the geometric shape is adjusted through an iterative optimization algorithm. After 3 iterations, the error is reduced to 0.2 mm, and a target three-dimensional point cloud model closer to the actual state is obtained.

[0048] As an example of an embodiment of the present application, the feature analysis of the target abnormal area to determine the influence range specifically includes: using a positioning and marking technology to determine the area position data of the target abnormal area, extracting the shape feature and the distribution feature of the target abnormal area; determining the influence range based on the shape feature and the distribution feature, and adjusting the model parameters if the influence range is greater than or equal to a third preset threshold.

[0049] In this embodiment, the position of the abnormal area in the three-dimensional point cloud is determined by using the positioning marker technology, and the three-dimensional point cloud data is generated by using the laser scanner. Specifically, the coordinates of the abnormal area are marked as x=50, y=30, and z=10. By analyzing the point cloud density and height change, it is confirmed that the area deviates from the normal plane by about 5 mm. This positioning method can intuitively reflect the spatial distribution of the abnormal area and is helpful for subsequent analysis. In combination with image processing technology, the shape feature and distribution feature of the abnormal area can be extracted, and the connectivity of the abnormal points in the point cloud can be analyzed, thereby improving the accuracy of judgment. For example, the abnormal area may present a long strip-shaped crack with a length of about 15 cm and a width of 2 mm, and the distribution is concentrated at the edge of the protective layer. This shape feature indicates that the damage may be caused by external stress concentration. According to the feature analysis result, the influence range of the abnormal area is judged, and if the influence range is greater than or equal to a third preset threshold, the model parameters are adjusted. Further, the positioning marker of the abnormal area can be verified for the target three-dimensional point cloud model, for example, it can be confirmed that the deviation of the marker point x=50, y=30, and z=10 is only 0.1 mm. At the same time, the damage category is repaired from “deep crack” to “shallow scratch”, and the tightness abnormality is also relieved. This verification method can effectively evaluate the repair degree and provide a basis for the maintenance of the protective layer.

[0050] In step 104, the abnormal detection result is determined according to the target three-dimensional point cloud model, and the power distribution cable management is performed according to the abnormal detection result.

[0051] In this embodiment, based on the target three-dimensional point cloud model, the multi-dimensional visual feature change trend and the influence of the storage environment are analyzed to obtain a quantitative index of deformation accumulation. If the quantitative index exceeds a preset range, the abnormal detection parameter is adjusted to obtain the current inventory state data adapted to the current environment. Through the data combined with the three-dimensional point cloud model and the abnormal detection result, the real-time inventory turnover efficiency of the cable drum is judged, the optimization management parameter is output, and the cable drum is managed according to the optimization management parameter.

[0052] As an example of an embodiment of the present application, the abnormality detection result is determined according to the target three-dimensional point cloud model, and power distribution optical cable management is performed according to the abnormality detection result. Specifically, a target multi-dimensional visual feature vector of the target three-dimensional point cloud model is determined, a change trend between the multi-dimensional visual feature vector and the target multi-dimensional visual feature vector is analyzed to obtain index data; based on the index data and environment data of the power distribution optical cable, a deformation cumulative value is calculated; if the deformation cumulative value is not in a preset range, an abnormality detection parameter of the target three-dimensional point cloud model is adjusted through a random forest algorithm to obtain the abnormality detection result and inventory state data corresponding to the environment data; the inventory state data and the target three-dimensional point cloud model are combined, a support vector machine algorithm is used to determine a real-time turnover efficiency, an efficiency evaluation result is determined according to the real-time turnover efficiency; and a power distribution optical cable management scheme is determined according to the efficiency evaluation result and the index data.

[0053] In this embodiment, extracting multi-dimensional visual features of the target through a 3D point cloud model can be understood as obtaining information such as shape, texture, and distribution from point cloud data. For example, in the inventory management of optical cable reels, the edge curvature and surface flatness of the reel can be captured through point cloud data, and the changing trends of these features over time can be analyzed. For instance, assuming the edge curvature of an optical cable reel increases from 0.5 to 0.8 within 30 days, it indicates potential deformation. The quantification index can be defined as the rate of change of curvature, with a threshold of 0.2. In one possible implementation, the cumulative deformation value is calculated by combining storage environment data such as temperature and humidity. Specifically, if the ambient temperature rises from 20°C to 35°C and the humidity increases from 50% to 70%, the cumulative deformation value can be inferred from historical data to reach 5mm, exceeding the preset range of 3mm. It should be noted that the calculation of this cumulative value relies on the correlation analysis between the environment and point cloud features, which helps in determining the reel's condition. If the cumulative deformation value exceeds the limit, the anomaly detection parameters are adjusted using a random forest algorithm. For example, the initial detection sensitivity is set to 80%, but adjusted to 90% after environmental changes, thus generating inventory status data adapted to high temperature and humidity. This adjustment can more accurately reflect tray anomalies and improve detection reliability. When judging real-time turnover efficiency by combining inventory status data with the target 3D point cloud model and using the support vector machine algorithm, it can be analyzed from two aspects: tray availability and handling difficulty. In one embodiment, if the point cloud shows that tray deformation leads to a reduction in stacking gaps, the turnover efficiency may drop from 95% to 85%, and the efficiency assessment result directly reflects management problems. Based on the efficiency assessment results and quantitative indicators, the optimization parameters can be adjusted, changing the handling priority from "high" to "medium" and optimizing the stacking height from 2 meters to 1.5 meters to determine the optical cable tray management scheme. This adjustment can reduce secondary damage to deformed trays and improve inventory turnover efficiency. Based on the optimization parameters, the features of the target 3D point cloud model are updated. For example, the edge curvature feature can be adjusted from 0.8 back to 0.6 to generate dynamic inventory data adapted to the current environment. This update can more realistically reflect the tray status and facilitate subsequent management decisions. By calculating the long-term trend of turnover efficiency using dynamic inventory data and anomaly detection results, for example, if the efficiency remains stable above 90% for three consecutive months, it can be inferred that the optimized values ​​of management parameters are approaching a reasonable level. For instance, maintaining a stacking height of 1.5 meters and a detection sensitivity of 90% in the long term indicates more efficient inventory management and a reduced anomaly rate. Specifically, this method can dynamically optimize the optical cable reel management process through multi-dimensional feature analysis and environmental adaptability adjustments.

[0054] like Figure 2 As shown, based on the above method embodiments, corresponding system embodiments are provided; One embodiment of the present invention provides a power distribution optical cable anomaly detection and management system 200, including: an acquisition module 201, an anomaly area determination module 202, a model adjustment module 203, and a management module 204; The acquisition module 201 is configured to acquire point cloud data when each power distribution optical cable is detected, and construct a corresponding initial three-dimensional point cloud model based on each point cloud data. The abnormal area determination module 202 is configured to calculate point cloud density data corresponding to a first feature based on the point cloud data in each initial three-dimensional point cloud model, determine a first abnormal area based on the point cloud density data, determine a feature parameter of the first abnormal area, and adjust a sensitivity threshold of the initial three-dimensional point cloud model in an abnormality detection process according to the feature parameter, thereby adjusting a feature threshold corresponding to a second feature, determining a second abnormal area and a damage degree. The model adjustment module 203 is configured to locate and mark the damage degree of the first abnormal area and the second abnormal area to determine the position of a target abnormal area, perform feature analysis on the target abnormal area to determine an influence range, adjust model parameters of the initial three-dimensional point cloud model according to the influence range, and refit each point cloud data according to the adjusted model parameters to obtain a target three-dimensional point cloud model. The management module 204 is configured to determine an abnormality detection result according to the target three-dimensional point cloud model, and perform power distribution optical cable management according to the abnormality detection result.

[0055] It can be understood that the above-mentioned device item embodiments correspond to the method item embodiments of the present application, and can realize the power distribution optical cable abnormality detection and management method provided by any one of the above-mentioned method item embodiments.

[0056] It should be noted that the device embodiments described above are only schematic, and some or all of the modules can be selected to achieve the purpose of the present embodiment. In addition, in the device embodiment provided by the present application, the connection relationship between the modules indicates that there is a communication connection between them, which can be realized as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0057] On the basis of the above-mentioned embodiment of the power distribution optical cable abnormality detection and management method, another embodiment of the present application provides a terminal device, which comprises a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, the power distribution optical cable abnormality detection and management method of any one embodiment of the present application is realized.

[0058] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present application. The one or more module elements can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program in the terminal device.

[0059] The terminal device can be a desktop computer, a notebook computer, a palm computer, a cloud server and other computing devices. The terminal device can include, but is not limited to, a processor and a memory.

[0060] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, and connects all parts of the terminal device through various interfaces and lines.

[0061] On the basis of the above-mentioned method embodiment, another embodiment of the present application provides a computer readable storage medium, including a stored computer program, wherein when the computer program runs, the device where the computer readable storage medium is located executes the power distribution optical cable anomaly detection and management method described in any one of the above-mentioned method embodiments of the present application.

[0062] The modules / units integrated in the device / terminal equipment, if realized in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on this understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. When the computer program is executed by a processor, the steps of each method embodiment can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0063] The above is the preferred embodiment of the present application. It should be noted that, for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, which are also considered within the scope of protection of the present application.

Claims

1. A method for detecting and managing anomalies in power distribution optical cables, characterized in that, include: Acquire point cloud data when detecting each power distribution optical cable, and construct the corresponding initial three-dimensional point cloud model based on the point cloud data; In each of the initial three-dimensional point cloud models, point cloud density data corresponding to the first feature is calculated based on the point cloud data, and a first abnormal region is determined based on the point cloud density data; the feature parameters of the first abnormal region are determined, and the sensitivity threshold of the initial three-dimensional point cloud model in the anomaly detection process is adjusted according to the feature parameters, thereby adjusting the feature threshold corresponding to the second feature, and determining the second abnormal region and the degree of damage; The damage levels of the first and second abnormal regions are located and marked to determine the location of the target abnormal region. Feature analysis is performed on the target abnormal region to determine the range of influence. The model parameters of the initial three-dimensional point cloud model are adjusted according to the range of influence, and the point cloud data are refitted according to the adjusted model parameters to obtain the target three-dimensional point cloud model. The anomaly detection results are determined based on the target 3D point cloud model, and the power distribution optical cable is managed based on the anomaly detection results.

2. The method for detecting and managing abnormalities in power distribution optical cables as described in claim 1, characterized in that, The step of determining the anomaly detection result based on the target 3D point cloud model and managing the power distribution optical cable based on the anomaly detection result specifically involves: Determine the target multidimensional visual feature vector of the target three-dimensional point cloud model, analyze the changing trend between the multidimensional visual feature vector and the target multidimensional visual feature vector, and obtain index data; Based on the aforementioned index data and environmental data of the power distribution optical cable, the cumulative deformation value is calculated; If the cumulative deformation value is not within the preset range, the anomaly detection parameters of the target 3D point cloud model are adjusted by the random forest algorithm to obtain the anomaly detection result and the inventory status data corresponding to the environmental data. Combining the inventory status data with the target 3D point cloud model, the support vector machine algorithm is used to determine the real-time turnover efficiency, and the efficiency evaluation result is determined based on the real-time turnover efficiency. Based on the efficiency evaluation results and the indicator data, a management scheme for power distribution optical cables is determined.

3. The method for detecting and managing abnormalities in power distribution optical cables as described in claim 1, characterized in that, The step of calculating the point cloud density data corresponding to the first feature based on the point cloud data, and determining the first abnormal region based on the point cloud density data, specifically involves: Statistical analysis of the initial three-dimensional point cloud model is performed to obtain the point cloud density data, and the areas where the point cloud density data does not meet the preset range conditions are identified as the first abnormal areas. The sensitivity threshold is adjusted according to the feature parameters, and an adjustment coefficient is calculated based on the adjusted sensitivity threshold. The initial three-dimensional point cloud model is then updated using the adjustment coefficient to obtain the updated point cloud density data. If the updated point cloud density data still does not meet the preset range conditions, then iterative updates are performed until the preset range conditions are met, thus obtaining the first abnormal region.

4. The method for detecting and managing abnormalities in power distribution optical cables as described in claim 1, characterized in that, The step of adjusting the sensitivity threshold of the initial 3D point cloud model in the anomaly detection process based on the feature parameters, and then adjusting the feature threshold corresponding to the second feature to determine the second anomaly region and the degree of damage, specifically involves: The threshold corresponding to the second feature is updated based on the adjustment coefficient to obtain the feature threshold; Obtain the reflection intensity data corresponding to the second feature, and perform time series analysis on the reflection intensity data to obtain the feature sequence; Multiple second abnormal regions are determined based on the feature threshold and the feature sequence, and the degree of damage is obtained by statistical analysis of each second abnormal region.

5. The method for detecting and managing abnormalities in power distribution optical cables as described in claim 1, characterized in that, The step of acquiring point cloud data of power distribution optical cable resources in the resource library and constructing corresponding initial three-dimensional point cloud models based on each point cloud data specifically involves: The point cloud data is preprocessed, and a statistical filtering algorithm is used to process the preprocessed point cloud data to generate a three-dimensional point cloud model. Extract the edge contour features of the three-dimensional point cloud model, determine the first feature and the second feature based on the edge contour features and the three-dimensional point cloud data, and fuse the first feature and the second feature to generate a multi-dimensional visual feature vector. Based on the geometric deviation between the current iteration of the 3D point cloud model and the standard point cloud model, and combined with the multidimensional visual feature vector, the matching parameters between the 3D point cloud model and the standard point cloud model are updated until the geometric deviation is greater than or equal to a first preset threshold, at which point the iteration stops and the initial 3D point cloud model is output.

6. The method for detecting and managing abnormalities in power distribution optical cables as described in claim 5, characterized in that, The process of determining a first feature and a second feature based on the edge contour features and the 3D point cloud data, and fusing the first feature and the second feature to generate a multidimensional visual feature vector, specifically involves: Calculate the point spacing distribution value from the edge contour features, and determine the tightness feature based on the point spacing distribution value; Extract the surface protective layer of the power distribution optical cable resource. If the point cloud density value of the surface protective layer is lower than the second preset threshold, it is determined that there is an anomaly in the surface protective layer, and the protective integrity feature is obtained. The multidimensional visual feature vector is generated by fusing the tightness feature and the protective integrity feature.

7. The method for detecting and managing abnormalities in power distribution optical cables as described in claim 1, characterized in that, The step of performing feature analysis on the target anomaly region to determine the scope of influence specifically involves: Location marking technology is used to determine the regional location data of the target anomaly region, and the shape and distribution features of the target anomaly region are extracted. The influence range is determined based on the shape features and the distribution features. If the influence range is greater than or equal to a third preset threshold, the model parameters are adjusted.

8. A power distribution optical cable anomaly detection and management system, characterized in that, include: Acquisition module, abnormal region identification module, model adjustment module, and management module; The acquisition module is used to acquire point cloud data when detecting each power distribution optical cable, and to construct a corresponding initial three-dimensional point cloud model based on each point cloud data. The abnormal region determination module is used to calculate the point cloud density data corresponding to the first feature based on the point cloud data in each of the initial three-dimensional point cloud models, determine the first abnormal region based on the point cloud density data, determine the feature parameters of the first abnormal region, and adjust the sensitivity threshold of the initial three-dimensional point cloud model in the abnormal detection process according to the feature parameters, thereby adjusting the feature threshold corresponding to the second feature, and determining the second abnormal region and the degree of damage. The model adjustment module is used to locate and mark the damage degree of the first abnormal region and the second abnormal region to determine the location of the target abnormal region, perform feature analysis on the target abnormal region to determine the influence range, adjust the model parameters of the initial three-dimensional point cloud model according to the influence range, and refit each point cloud data according to the adjusted model parameters to obtain the target three-dimensional point cloud model. The management module is used to determine the anomaly detection result based on the target 3D point cloud model, and to manage the power distribution optical cable based on the anomaly detection result.

9. A terminal device, characterized in that, The method includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements the distribution optical cable anomaly detection and management method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, include: A stored computer program, wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform the power distribution optical cable anomaly detection and management method as described in any one of claims 1-7.

Images