Routing anomaly detection method, device, equipment and storage medium

Abstract

The application discloses a routing anomaly detection method and device, equipment and a storage medium, and specifically discloses the following: a physical topology routing picture of a transmission network is acquired; the physical topology routing picture comprises routing line segments between a routing starting point, an ending point, a switching point and a connecting point; the physical topology routing comprises at least one of an optical path topology routing and an electrical circuit topology routing; the physical topology routing picture is input into a pre-trained target detection model; wherein the target detection model is obtained by training sample pictures labeled with a routing anomaly type and an abnormal area position; a routing anomaly detection result output by the target detection model is acquired, and the routing anomaly detection result comprises an abnormal type of a detected abnormal routing and an abnormal position in the physical topology routing picture.

Description

Technical Field

[0001] This invention relates to the field of transmission technology, and in particular to a method, apparatus, device, and storage medium for detecting routing anomalies. Background Technology

[0002] Currently, the identification of transmission physical topology routing problems mainly relies on solutions based on dummy resource digitalization and solutions based on predefined rule verification. The dummy resource digitalization solution involves deploying RFID (Radio Frequency Identification) products and technologies or NFC (Near Field Communication)... Near-field communication (NFC) tags digitize physical resources such as optical cables, mapping the coordinates of optical cable nodes to a GIS (Geographic Information System) to form a basic topology view, thereby indirectly supporting route anomaly identification. The predefined rule-based verification scheme, on the other hand, constructs a computational model by encoding business logic rules, such as connection compliance and routing standardization, and analyzes transmission resource data to directly identify anomalies at the routing layer. For example, by checking the connectivity of optical or electrical routes, it identifies missing issues (route breakpoints) or verifies logical errors such as route detours.

[0003] However, both approaches have significant drawbacks: the approach based on dummy resource digitalization relies heavily on external hardware deployment, making it difficult to apply to existing network resource scenarios and unable to cover topological issues at the logical resource layer, such as optical paths and circuits. The approach based on predefined rule verification, on the other hand, is limited by the completeness and complexity of rule design, lacking the ability to identify topological anomalies such as ambiguous business logic, abnormal data synchronization, or rules not covering them. This results in a limited verification scope, low identification efficiency, and numerous detection blind spots. Summary of the Invention

[0004] The main objective of this invention is to provide a routing anomaly detection method, apparatus, device, and storage medium, which aims to solve the problems of existing methods for identifying transmission physical topology routing problems, which rely entirely on the deployment of intelligent hardware for dumb resources and the verification of predefined rules. These methods suffer from limited identification scope, inability to cover logical resource topology problems, low identification efficiency, large number of detection blind spots, poor applicability to existing resource scenarios, and low accuracy in identifying anomalies with undefined rules.

[0005] In a first aspect, embodiments of the present invention provide a routing anomaly detection method, including: Obtain a physical topology routing image of the transmission network; the physical topology routing image includes route start point, end point, transfer point, and routing line segments between connection points; the physical topology routing includes at least one of optical path topology routing and circuit topology routing; The physical topology routing image is input into a pre-trained target detection model; wherein the target detection model is trained using sample images labeled with routing anomaly types and anomaly region locations; Obtain the routing anomaly detection results output by the target detection model. The routing anomaly detection results include the anomaly type of the detected abnormal routes and the abnormal location in the physical topology routing image.

[0006] Secondly, embodiments of the present invention provide a routing anomaly detection device, comprising: The acquisition unit is used to acquire a physical topology routing image of the transmission network; the physical topology routing image includes route start point, end point, transfer point, and routing line segments between connection points; the physical topology routing includes at least one of optical path topology routing and circuit topology routing. The input unit is used to input the physical topology routing image into a pre-trained target detection model; wherein the target detection model is trained using sample images labeled with routing anomaly types and anomaly region locations; The detection unit is used to obtain the routing anomaly detection results output by the target detection model. The routing anomaly detection results include the anomaly type of the detected abnormal routes and the abnormal location in the physical topology routing image.

[0007] Thirdly, embodiments of the present invention provide an electronic device, including: a processor; and a memory configured to store computer-executable instructions, which, when executed, cause the processor to perform the steps of the method described in the first aspect above.

[0008] Fourthly, embodiments of the present invention provide a computer-readable storage medium for storing computer-executable instructions that, when executed by a processor, implement the steps of the method described in the first aspect above.

[0009] Fifthly, embodiments of the present invention provide a computer program product, the computer program product including a computer program, which, when executed by a processor, implements the steps of the method described in the first aspect above.

[0010] The at least one technical solution provided by the embodiments of the present invention can achieve the following technical effects: In this embodiment of the invention, physical topology routing images of the transmission network can be acquired first. These images include routing start points, end points, transfer points, and routing segment information between connection points, covering at least one scenario of optical path topology routing and circuit topology routing, ensuring the comprehensiveness and representativeness of the input data. Then, these physical topology routing images are input into a pre-trained target detection model. This model is trained using sample images labeled with routing anomaly types and anomaly location information, enabling automatic identification of potential problems in the routing. Finally, the routing anomaly detection results output by the target detection model are obtained, including the specific anomaly type and the anomaly location information in the physical topology routing images, thereby achieving accurate location and classification of routing vulnerabilities.

[0011] This invention, through the construction of an intelligent identification process based on target detection, effectively solves the limitations of existing technologies that rely on hardware deployment or predefined rules. This invention does not rely on external hardware devices such as RFID or NFC, avoiding the applicability issues of dumb resource digitalization solutions in existing network scenarios, and can cover topology anomaly detection at the logical resource layer, including optical paths and circuits. Simultaneously, by replacing complex rule design with a target detection model, it overcomes the identification blind spots and inefficiencies caused by ambiguous business logic and incomplete rules in predefined rule verification schemes, thus improving the verification scope and accuracy. Attached Figure Description

[0012] Figure 1 This is one of the flowcharts illustrating a routing anomaly detection method provided in an embodiment of the present invention; Figure 2 This is one of the scenario diagrams illustrating the routing anomaly detection method provided in an embodiment of the present invention; Figure 3 This is a second scenario illustration of a routing anomaly detection method provided in an embodiment of the present invention; Figure 4 This is a third scenario illustration of a routing anomaly detection method provided in an embodiment of the present invention; Figure 5 The fourth scenario diagram illustrates the routing anomaly detection method provided in one embodiment of the present invention; Figure 6 Fifth scenario illustration of a routing anomaly detection method provided in an embodiment of the present invention; Figure 7 A sixth scenario illustration of a routing anomaly detection method provided in an embodiment of the present invention; Figure 8 This is a second flowchart illustrating a routing anomaly detection method provided in one embodiment of the present invention. Figure 9 The third flowchart illustrates a routing anomaly detection method provided in one embodiment of the present invention. Figure 10 This is a schematic diagram of the architecture of a routing anomaly detection method provided in one embodiment of the present invention; Figure 11 A schematic diagram of the module composition of a routing anomaly detection device 1100 provided in an embodiment of the present invention; Figure 12 This is a schematic diagram of the hardware structure of an electronic device provided in one embodiment of the present invention. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0014] The technical solutions provided by the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0015] Please see Figure 1 , Figure 1 This is one of the flowcharts illustrating a routing anomaly detection method provided in an embodiment of the present invention, such as... Figure 1 As shown, the method includes the following steps: Step 102: Obtain a physical topology routing image of the transmission network; the physical topology routing image includes route start point, end point, transfer point, and routing segments between connection points; the physical topology routing includes at least one of optical path topology routing and circuit topology routing.

[0016] Step 104: Input the physical topology routing image into a pre-trained target detection model; wherein the target detection model is trained using sample images labeled with routing anomaly types and anomaly region locations.

[0017] Step 106: Obtain the routing anomaly detection result output by the target detection model. The routing anomaly detection result includes the anomaly type of the detected abnormal route and the abnormal location in the physical topology routing image.

[0018] In this embodiment of the invention, a physical topology routing image of the transmission network can be obtained first. The physical topology routing image can be derived from the optical path topology routing or the circuit topology routing in the transmission network, and can include the route start point, route end point, route transition point, and the routing lines between these points.

[0019] Optical path topology routing can be obtained through an integrated resource management system. This system is a resource management library and process management tool. Centered on a resource model, it meticulously records and manages the static attribute information, connection relationships, and geospatial information of all network entities, from the physical layer (computer rooms, racks, pipelines, optical cables, and equipment ports) to the logical layer (circuits, optical paths, and IP addresses). This provides accurate data for operational activities such as network planning, service activation, fault analysis, and resource optimization. In this embodiment of the invention, the optical cable route coordinates, equipment connection relationships, and other data stored in the integrated resource management system can be exported or accessed via an interface to generate an image of the optical path topology routing, including the route's start point, end point, transfer points, and the lines.

[0020] Circuit topology routing can be obtained through a transmission workbench. The transmission workbench is a specialized interface or platform for circuit resource management and topology visualization in transmission network operation and maintenance. It primarily focuses on physical topology management at the circuit level, providing a graphical display of the start, end, transition points, and connecting segments of circuit routes. Within the transmission workbench, relevant personnel can view, edit, and monitor circuit routing information through an interactive interface, and it supports automatic or manual generation of topology images. This embodiment of the invention utilizes the transmission workbench as a data source, automatically logging in via circuit identifiers and obtaining a topology view to obtain a standardized circuit physical topology image for subsequent target detection-based routing anomaly analysis.

[0021] In this embodiment of the invention, optical path topology routing images and circuit topology routing images can be obtained through manual or automatic screenshotting. Specifically, when automatically taking screenshots, the corresponding network management system page can be logged into based on the optical path identifier or circuit identifier to capture the displayed physical topology routing view and generate a physical topology routing image. In practice, an automated test script can be created using Python to read the optical path or circuit ID, automatically log in to the browser URL, and then automatically capture the browser's optical path physical topology page to generate an image. The image is then automatically converted to a 1920*1080 resolution and output as a physical topology routing image to be tested, named after the optical path or circuit. The automatic screenshotting method in this embodiment of the invention effectively avoids repetitive manual operations, improves data collection efficiency, and is suitable for scenarios with existing network resources.

[0022] In this embodiment of the invention, when acquiring a physical topology routing image, the detection scenario can be determined first. The scenario includes optical transmission path topology routing or electrical transmission circuit topology routing. The physical topology routing image displays routing elements on a GIS, such as start points, end points, transfer points, and the connection methods of routing line segments between connection points, thus presenting the physical topology of transmission resources such as optical paths or electrical circuits. The automatic screenshot method can achieve automatic browser URL login and page capture by creating automated test scripts, avoiding manual intervention and improving data collection efficiency. After the image is generated, it can be persistently saved in a file and the data can be updated periodically to support subsequent processing.

[0023] In this embodiment of the invention, after obtaining the physical topology routing image, the physical topology routing image can be input into a pre-trained target detection model. The target detection model can be trained using sample images labeled with routing anomaly types and anomaly region locations. Before inputting the image, the model needs to be prepared and trained. The model training process includes data acquisition, data preprocessing, data labeling, data augmentation, and model optimization.

[0024] During the data acquisition phase, sample images containing physical topology routes can be collected. These sample images can cover various routing scenarios and anomaly types. In the data preprocessing phase, the collected data can be cleaned to remove junk, invalid, and redundant data. For example, problematic images such as those with incomplete physical topology loading, incomplete display, or numerous duplicates can be removed. The cleaned images can be persistently saved and updated periodically.

[0025] During the data annotation phase, identification rules for routing anomalies can be constructed based on expert experience. These rules can define six typical problems: spurious lines, detours, missing information, reverse AZ segment laying, closures, and clustering. Among these, spurious lines manifest as a long straight line in the physical topology, deviating from the road. This is typically caused by missing latitude and longitude information at the fiber optic cable laying point or data entry errors. For example, it can be seen as follows: Figure 2 As shown; a detour manifests as a loop, usually caused by routing planning problems or data errors. For example, it can be like... Figure 3 As shown; the missing data manifests as an incomplete link path, split into unconnected sub-links, with breakpoints. This is typically caused by missing data related to maintained optical cables or cable segments. For example, it can be seen as... Figure 4 As shown; the reverse laying pattern of section AZ is a semi-circular closed curve, with one side being a straight line and the other a continuous curve. This is usually caused by the reversed input of the start and end coordinates of the laying section. For example, it can be like this: Figure 5 As shown; a closed loop is characterized by a route appearing as a closed curve, which is usually caused by routing planning or allocation problems. For example, it can be like... Figure 6As shown; clustering occurs when data is distributed in clusters along the same path, usually due to numerous data entry errors. For example, it can be seen as follows: Figure 7 As shown, the dataset can be labeled according to the above labeling rules. During labeling, labeling tools such as LabelImg can be used to generate labeled image samples and a txt format detail file, recording the anomaly type and region location.

[0026] In the data augmentation stage, the labeled sample images can be enhanced. Augmentation methods can include translation, rotation, mirroring, flipping, and brightness adjustment. These methods fall under geometric and color transformations, aiming to expand the number of training samples and improve the model's recognition ability under different angles, positions, and lighting conditions. Specifically, translation augmentation can simulate changes in route position, rotation augmentation can handle angular differences, mirroring and flipping can enhance sample diversity, and brightness augmentation can adapt to different lighting conditions. These augmentation methods can improve the model's generalization ability, thereby ensuring stable recognition in various scenarios.

[0027] During the model training phase, the YOLO object detection algorithm can be used. Before training, the enhanced sample images can be divided into training, validation, and test sets in an 8:1:1 ratio. During training, to improve the model's recognition ability and generalization performance, several optimization strategies are employed sequentially: First, random erasure is used to randomly select rectangular regions in the input images and replace pixel values, forcing the model to learn more comprehensive feature representations. Then, a hybrid augmentation method can be applied to proportionally merge two sample images to generate new training samples, increasing data diversity. A learning rate warm-up strategy is used in the early stages of training, starting with a small learning rate and gradually increasing it to a set value to ensure stable model initialization. Then, a cosine annealing phase is initiated, causing the learning rate to smoothly decrease along a cosine curve, promoting better convergence to the optimal solution. The AdamW optimizer can be used throughout the training process to adjust parameter updates. This optimizer introduces a weight decay mechanism based on the Adam algorithm, effectively controlling the risk of overfitting. Meanwhile, the early stopping technique can be used to continuously monitor performance changes on the validation set. When the model does not show significant improvement in multiple consecutive training rounds, the training process is automatically terminated to prevent overfitting of the training data, thereby ensuring that the final model has good generalization ability and robustness.

[0028] After model training, the resulting target detection model can identify route anomaly types, including flying lines, detours, missing sections, reversed laying, closures, and clusters. Anomaly types can be defined based on visual features. For example, flying lines are caused by missing latitude and longitude information or data entry errors at the fiber optic cable laying points; detours are caused by route planning problems or data errors; missing sections are caused by missing data on fiber optic cables or cable segments; reversed laying is caused by reversed coordinates of the start and end points of the laying segment; closures are caused by route planning or allocation problems; and clusters are caused by numerous data entry errors.

[0029] In this embodiment of the invention, the routing anomaly detection results output by the target detection model can be obtained. The routing anomaly detection results may include the anomaly type of the detected abnormal route and the anomaly location in the physical topology routing image. The routing anomaly detection results may also include the optical path or circuit name, optical path or circuit identifier, and confidence level. The result output can be implemented using a streaming inference method to avoid memory overflow during batch processing. The results can be saved as image and text information; the text information may include the problem type, problem point coordinates, confidence level, etc., facilitating subsequent processing or integration with external systems.

[0030] In this embodiment of the invention, after generating the routing anomaly detection results, the problem area can be located based on the routing anomaly detection results. For example, a flying line anomaly is displayed as a long straight line in the image, the model outputs the anomaly type as flying line, and marks the coordinates of the anomaly location; a detour anomaly is displayed as a route loop, and the model provides location information after identification; a missing anomaly is displayed as a breakpoint, and the model marks the breakpoint area; a reverse laying anomaly is displayed as a semi-circular curve, and the model identifies the curve area; a closed anomaly is displayed as a closed loop, and the model marks the loop range; a clustered anomaly is displayed as a scattered path, and the model identifies the scattered points.

[0031] In the results output stage, the optimal model parameters obtained during training can be loaded into the object detection model, and the physical topology route image to be detected can be used as model input to start the inference process. To avoid the risk of memory accumulation and overflow when processing a large number of images in batches, the inference code adopts a streaming processing method, that is, it reads images one by one, releases memory immediately after completing inference, and then processes the next image, thereby effectively controlling memory usage. After inference is completed, the generated structured detection results are encapsulated into a standardized interface data format, which is convenient for transmission back to external systems such as integrated resource management systems or transmission workbenches. The detailed information of the returned results includes the name and unique identifier of the optical path or circuit, the classification of the identified problem type, the precise coordinate position of the problem point in the image, and the confidence index of the model judgment. This complete information provides direct data support for subsequent hazard location and resource rectification.

[0032] In this embodiment of the invention, the identification of physical topology routing problems can be based entirely on the analysis of image visual features. Specifically, when identifying flying line problems, the method can identify them by detecting straight long line segments that do not extend along roads in the topology map, thus overcoming the technical shortcomings of traditional methods that require explicitly defining complex topology rules. When identifying detour problems, the method can rely on detecting obvious loops or ring patterns in the routing path, without needing to pre-encode business constraints such as connection compliance. When identifying missing problems, the method can detect breakpoints or unconnected sub-chain segments in the path, effectively overcoming blind spots caused by untimely data synchronization or missing data. When identifying reverse routing problems, the method can identify them by recognizing the unique semi-circular curve features in the topology map. These curves typically appear as straight lines on one side and continuous arcs on the other, thus eliminating the impact of incorrect coordinate input direction. When identifying closure problems, the method can focus on detecting closed curve features that form complete loops, thus solving the problem that traditional methods struggle to accurately determine the rationality of route allocation. When identifying cluster problems, the method can detect the disordered and scattered distribution patterns of a large number of paths in local areas, effectively handling complex anomalies caused by large-scale data input errors.

[0033] In one example, such as Figure 8 As shown, this illustrates a complete process for intelligent detection of physical topology routing problems.

[0034] like Figure 8 As shown, firstly, physical topology images can be acquired. Optical path topology routing views or circuit topology routing views can be obtained from an integrated resource management system or transmission workbench, and standardized images can be generated through automatic screenshotting. Next, the acquired standardized image data is preprocessed, including data cleaning and resolution standardization to remove invalid data and unify the format, laying the foundation for subsequent analysis. The preprocessed image data can then be used for routing problem labeling. Based on six defined anomaly type rules, tools are used to label abnormal regions in the images, generating label files for training. After routing problem labeling, data augmentation can be performed on the labeled images. Geometric transformations such as translation, rotation, and mirroring, or brightness adjustments, can be used to expand the dataset and improve the model's generalization ability. The augmented data can then be input into the problem detection model training module to train a target detection model capable of recognizing anomalies. Finally, the trained model can be used to perform inference analysis on the input physical topology images, output detection results, and end the entire process. Figure 8 It can clearly demonstrate the entire process from data preparation and model building to output results.

[0035] In one example, such as Figure 9 The diagram illustrates the core reasoning process for route anomaly detection. First, preprocessed physical topology image data is read and used as model input. Then, for this physical topology image data, six specific anomaly detection methods are performed in parallel: flying line detection, detour detection, missing line detection, reversed AZ (Average-Zone) detection, closure detection, and clustering detection, enabling the simultaneous identification of multiple anomaly types. Each detection module focuses on identifying its unique visual pattern, such as straight long line segments of flying lines, detours, missing breakpoints, reversed AZ curves, closed loops, and disordered clustered distributions. The outputs of all detection modules are aggregated to a problem detection result aggregation node, which integrates information such as the type, location, and confidence level of various anomalies to form a complete structured detection result. Figure 9 It can clearly demonstrate how the model processes multiple anomaly checks in parallel during inference and outputs the results.

[0036] In one example, such as Figure 10 The diagram illustrates the overall architecture corresponding to an embodiment of the present invention. This architecture may include an image input module, supporting single, batch, or BASE64 encoded access to physical topology routing images. Input data can enter a preprocessing module to perform data cleaning and resolution conversion operations, ensuring data quality. The preprocessing module connects to a data storage and update submodule for persistent storage. Then, it can enter a data annotation module, which uses tools to annotate based on six predefined anomaly rules, generating training labels. A data augmentation module applies geometric transformations and color adjustment methods, such as translation, rotation, and brightness modification, to improve sample diversity. The augmented data is input into a model training module, where it is trained, saved, and updated using the YOLO algorithm. Finally, a routing problem detection output module loads the model, performs streaming inference, and encapsulates the results for external provision via an interface. The various modules of the architecture work collaboratively to ensure the efficiency and scalability of the detection process.

[0037] In this embodiment of the invention, physical topology routing images of the transmission network can be acquired first. These images include routing start points, end points, transfer points, and routing segment information between connection points, covering at least one scenario of optical path topology routing and circuit topology routing, ensuring the comprehensiveness and representativeness of the input data. Then, these physical topology routing images are input into a pre-trained target detection model. This model is trained using sample images labeled with routing anomaly types and anomaly location information, enabling automatic identification of potential problems in the routing. Finally, the routing anomaly detection results output by the target detection model are obtained, including the specific anomaly type and the anomaly location information in the physical topology routing images, thereby achieving accurate location and classification of routing vulnerabilities.

[0038] This invention, through the construction of an intelligent identification process based on target detection, effectively solves the limitations of existing technologies that rely on hardware deployment or predefined rules. This invention does not rely on external hardware devices such as RFID or NFC, avoiding the applicability issues of dumb resource digitalization solutions in existing network scenarios, and can cover topology anomaly detection at the logical resource layer, including optical paths and circuits. Simultaneously, by replacing complex rule design with a target detection model, it overcomes the identification blind spots and inefficiencies caused by ambiguous business logic and incomplete rules in predefined rule verification schemes, thus improving the verification scope and accuracy.

[0039] Figure 11 The routing anomaly detection device 1100 shown can achieve Figure 1 The method described in the embodiment achieves the same technical effect, and can be specifically referred to in the above description. Figure 1 The routing anomaly detection method of the illustrated embodiment will not be described in detail here. The routing anomaly detection device 1100 includes: The acquisition unit 1101 is used to acquire a physical topology routing image of the transmission network; the physical topology routing image includes route start point, end point, transfer point, and routing line segments between connection points; the physical topology routing includes at least one of optical path topology routing and circuit topology routing. The input unit 1102 is used to input the physical topology routing image into a pre-trained target detection model; wherein the target detection model is trained by sample images labeled with routing anomaly types and anomaly region locations; The detection unit 1103 is used to obtain the routing anomaly detection result output by the target detection model. The routing anomaly detection result includes the anomaly type of the detected abnormal route and the abnormal location in the physical topology routing image.

[0040] Optionally, the acquisition unit 1101 is used for: Log in to the corresponding network management system page based on the optical path identifier or circuit identifier; Capture the physical topology routing view displayed on the network management system page to generate the physical topology routing image.

[0041] Optionally, the device further includes ( Figure 11 (not shown in the image) The acquisition unit 1104 is used to acquire sample images containing physical topology routes before inputting the physical topology route image into the pre-trained target detection model; The enhancement unit 1105 is used to perform data enhancement processing on the acquired sample image. The data enhancement processing includes at least one of translation, rotation, mirroring, flipping, or brightness adjustment on the sample image. The training unit 1106 is used to train the constructed initial target detection model based on the data augmented sample images to obtain the pre-trained target detection model.

[0042] Optionally, the training unit 1106 is used for: The training process is optimized; the optimization includes at least one of the following: Randomly erase samples that have undergone data augmentation. The learning rate is dynamically adjusted during the training process using a learning rate warm-up and cosine annealing strategy. Early stop technique is used to control the number of training rounds.

[0043] Optionally, the route anomaly detection result may further include at least one of the following: the name of the detected optical path or circuit, the identifier of the detected optical path or circuit, and the confidence level of the detected abnormal route.

[0044] Optionally, the anomaly type includes at least one of the following: flying wire, detour, missing, reverse laying, closure, or clustering.

[0045] In this embodiment of the invention, physical topology routing images of the transmission network can be acquired first. These images include routing start points, end points, transfer points, and routing segment information between connection points, covering at least one scenario of optical path topology routing and circuit topology routing, ensuring the comprehensiveness and representativeness of the input data. Then, these physical topology routing images are input into a pre-trained target detection model. This model is trained using sample images labeled with routing anomaly types and anomaly location information, enabling automatic identification of potential problems in the routing. Finally, the routing anomaly detection results output by the target detection model are obtained, including the specific anomaly type and the anomaly location information in the physical topology routing images, thereby achieving accurate location and classification of routing vulnerabilities.

[0046] This invention, through the construction of an intelligent identification process based on target detection, effectively solves the limitations of existing technologies that rely on hardware deployment or predefined rules. This invention does not rely on external hardware devices such as RFID or NFC, avoiding the applicability issues of dumb resource digitalization solutions in existing network scenarios, and can cover topology anomaly detection at the logical resource layer, including optical paths and circuits. Simultaneously, by replacing complex rule design with a target detection model, it overcomes the identification blind spots and inefficiencies caused by ambiguous business logic and incomplete rules in predefined rule verification schemes, thus improving the verification scope and accuracy.

[0047] Figure 12 This is a schematic diagram of the structure of an electronic device provided in one embodiment of the present invention. Please refer to it. Figure 12At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and memory. The memory may include main memory, such as high-speed random-access memory (RAM), or non-volatile memory, such as at least one disk drive. Of course, the electronic device may also include other hardware required for other business operations.

[0048] The processor, network interface, and memory can be interconnected via an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 12 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0049] Memory is used to store programs. Specifically, programs may include program code, which includes computer operation instructions. Memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0050] The processor reads the corresponding computer program from non-volatile memory into main memory and then executes it, forming a non-contiguous transfer configuration at the logical level. The processor executes the program stored in memory and specifically performs the following operations: Obtain a physical topology routing image of the transmission network; the physical topology routing image includes route start point, end point, transfer point, and routing line segments between connection points; the physical topology routing includes at least one of optical path topology routing and circuit topology routing; The physical topology routing image is input into a pre-trained target detection model; wherein the target detection model is trained using sample images labeled with routing anomaly types and anomaly region locations; Obtain the routing anomaly detection results output by the target detection model. The routing anomaly detection results include the anomaly type of the detected abnormal routes and the abnormal location in the physical topology routing image.

[0051] The above is as described in the present invention. Figure 1The routing anomaly detection method disclosed in the embodiments described above can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in one or more embodiments of the present invention. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in one or more embodiments of the present invention can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0052] The electronic device can also perform Figure 1 The method for detecting routing anomalies described herein will not be elaborated upon here.

[0053] This invention also provides a computer-readable storage medium that stores one or more programs, the programs including instructions that, when executed by a portable electronic device including multiple applications, enable the portable electronic device to perform... Figure 1 The method of the illustrated embodiment will not be described in detail here.

[0054] This invention also provides a computer program product stored in a storage medium and executed by at least one processor to implement... Figure 1 The method of the illustrated embodiment will not be described in detail here.

[0055] Of course, in addition to the software implementation, the electronic device of the present invention does not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc. That is to say, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0056] In summary, the above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of one or more embodiments of the present invention should be included within the scope of protection of one or more embodiments of the present invention.

[0057] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0058] Computer-readable media, including both permanent and non-permanent, removable and non-removable media, can store information using any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined in this embodiment of the invention, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0059] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0060] The various embodiments in this invention are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

Claims

1. A method for detecting routing anomalies, characterized in that, include: Obtain a physical topology routing image of the transmission network; the physical topology routing image includes route start point, end point, transfer point, and routing line segments between connection points; The physical topology routing includes at least one of optical path topology routing and circuit topology routing; The physical topology routing image is input into a pre-trained target detection model; wherein the target detection model is trained using sample images labeled with routing anomaly types and anomaly region locations; Obtain the routing anomaly detection results output by the target detection model. The routing anomaly detection results include the anomaly type of the detected abnormal routes and the abnormal location in the physical topology routing image.

2. The method according to claim 1, characterized in that, The process of obtaining the physical topology routing image of the transmission network includes: Log in to the corresponding network management system page based on the optical path identifier or circuit identifier; Capture the physical topology routing view displayed on the network management system page to generate the physical topology routing image.

3. The method according to claim 1, characterized in that, Before inputting the physical topology routing image into the pre-trained object detection model, the method further includes: Collect sample images containing physical topology routes; The collected sample images are subjected to data augmentation processing, which includes at least one of translation, rotation, mirroring, flipping, or brightness adjustment of the sample images. The initial target detection model is trained based on the data-augmented sample images to obtain the pre-trained target detection model.

4. The method according to claim 3, characterized in that, The step of training the initial object detection model based on the data-augmented sample images includes: The training process is optimized; the optimization includes at least one of the following: Randomly erase samples that have undergone data augmentation. The learning rate is dynamically adjusted during the training process using a learning rate warm-up and cosine annealing strategy. Early stop technique is used to control the number of training rounds.

5. The method according to claim 1, characterized in that, The route anomaly detection result also includes at least one of the following: the name of the detected optical path or circuit, the identifier of the detected optical path or circuit, and the confidence level of the detected abnormal route.

6. The method according to any one of claims 1-5, characterized in that, The abnormality type includes at least one of the following: flying wire, detour, missing, reverse laying, closure, or clustering.

7. A routing anomaly detection device, characterized in that, include: The acquisition unit is used to acquire a physical topology routing image of the transmission network; the physical topology routing image includes the route start point, end point, transfer point, and routing line segments between the connection points; The physical topology routing includes at least one of optical path topology routing and circuit topology routing; The input unit is used to input the physical topology routing image into a pre-trained target detection model; wherein the target detection model is trained using sample images labeled with routing anomaly types and anomaly region locations; The detection unit is used to obtain the routing anomaly detection results output by the target detection model. The routing anomaly detection results include the anomaly type of the detected abnormal routes and the abnormal location in the physical topology routing image.

8. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store computer-executable instructions that, when executed by a processor, implement the steps of the method described in any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the method described in any one of claims 1 to 6.

Images