A power distribution network wire state identification method and device, a terminal device, and a storage medium

Abstract

The application discloses a power distribution network wire state recognition method and device, terminal equipment and storage medium, and belongs to the power distribution network field. The method is: inputting a flight image of a wire to be recognized into a preset backbone network to obtain a preliminary feature map output by the backbone network; inputting the preliminary feature map into an MDPE module to enable the MDPE module to perform multi-dimensional feature perception processing and noise suppression processing on the preliminary feature map to obtain a deep semantic feature map; inputting the deep semantic feature map into a preset HSFPN module to enable the HSFPN module to perform attention feature fusion on the deep semantic feature map to obtain a multi-scale fusion feature map; inputting the multi-scale fusion feature map into a detection head to obtain key point coordinates and contour segmentation masks of the wire to be recognized, and determining a wire state of the wire to be recognized according to the key point coordinates and the contour segmentation masks. Through implementation of the application, the problem of low power distribution network wire state recognition efficiency in the prior art can be solved.

Description

Technical Field

[0001] This invention relates to the field of power distribution networks, and in particular to a method, apparatus, terminal equipment, and storage medium for identifying the status of power distribution network conductors. Background Technology

[0002] As a core component of the power distribution network, the operating status of distribution network conductors directly affects the stability and reliability of the power supply network, and is closely related to the electricity safety of a large number of power users and the normal operation of social production and life. Distribution network conductors are mostly erected in outdoor open environments, widely covering complex areas such as mountainous areas, suburbs, and urban-rural fringe areas. They are exposed to natural climate erosion, external environmental interference, and high-load operating conditions for a long time, making them prone to faults such as conductor strand breakage and insulation aging. If these faults are not detected in time, they may cause local power outages, or even short circuits, discharges, or personal injury accidents. Therefore, condition monitoring of distribution network conductors is a key link in ensuring the safe and stable operation of the distribution network.

[0003] For a long time, the condition inspection of distribution network conductors has mainly relied on manual on-site inspections. Manual inspections require maintenance personnel to go to the line site and judge the condition of the conductors by visual observation and handheld testing equipment. This method is not only inefficient and costly in terms of manpower and time, but also limited by objective conditions such as complex terrain and inclement weather, making it difficult to achieve comprehensive coverage of distribution network conductors in mountainous areas and wilderness areas that are difficult for personnel to reach. At the same time, the test results are highly dependent on the work experience and subjective judgment of the inspectors, which is prone to missed detections and false detections, and cannot meet the development needs of intelligent and refined distribution network operation and maintenance. Summary of the Invention

[0004] This invention provides a method, apparatus, terminal equipment, and storage medium for identifying the status of power distribution network conductors. The method can solve the problem of low efficiency in identifying the status of power distribution network conductors in the prior art.

[0005] To address the aforementioned technical problems, an embodiment of the present invention provides a method for identifying the status of conductors in a distribution network, comprising: Acquire aerial images of the conductor to be identified; The aerial images are input into a preset backbone network to obtain a preliminary feature map output by the backbone network; The preliminary feature map is input into a preset MDPE module, so that the MDPE module performs multi-dimensional feature perception processing and noise suppression processing on the preliminary feature map to obtain a deep semantic feature map. The deep semantic feature map is input into a preset HSFPN module, so that the HSFPN module performs attention feature fusion on the deep semantic feature map to obtain a multi-scale fused feature map; The multi-scale fused feature map is input into a preset detection head to obtain the key point coordinates and contour segmentation mask of the wire to be identified output by the detection head; The conductor state is determined based on the key point coordinates and contour segmentation mask of the conductor to be identified.

[0006] Furthermore, the MDPE module includes a smoothing network, a channel attention submodule, a self-attention submodule, and a spatial attention submodule; The step of inputting the preliminary feature map into a preset MDPE module, so that the MDPE module performs multi-dimensional feature perception processing and noise suppression processing on the preliminary feature map to obtain a deep semantic feature map, includes: The preliminary feature map is input into a smoothing network, which uses grouped convolution and Gaussian smoothing operations to filter out high-frequency redundant information in the preliminary feature map, resulting in a denoised base feature map. The denoised base feature map is input into the channel attention submodule, so that the channel attention submodule calculates the weight coefficient of each feature channel in the denoised base feature map, and divides each feature channel into high importance feature channels and low importance feature channels based on each weight coefficient. The self-attention submodule is used to enhance the representation of detailed features in the high-importance feature channel, resulting in an enhanced high-importance feature channel; The spatial attention submodule is used to suppress noise in the low importance feature channel to obtain the noise-suppressed low importance feature channel; The enhanced high-importance feature channels and the suppressed low-importance feature channels are concatenated to obtain the deep semantic feature map.

[0007] Furthermore, the HSFPN module includes an edge sensing unit, a spatial selection unit, and a frequency selection unit; The step of inputting the deep semantic feature map into a preset HSFPN module, so that the HSFPN module performs attention feature fusion on the deep semantic feature map to obtain a multi-scale fused feature map, includes: The deep semantic feature map is input into the edge perception unit, so that the edge perception unit performs edge enhancement processing on the deep semantic feature map to obtain an edge-enhanced fused feature map. The fused feature map is input to the spatial selection unit, so that after the spatial selection unit performs a pooling operation on the fused feature map, it generates a spatial feature map through a preset first convolutional layer. The spatial feature map is input to the frequency selection unit, so that the frequency selection unit enhances the low-frequency semantic information of the spatial feature map through mean filtering, and then generates high-frequency edge features through a preset residual method. Based on the high-frequency edge features and the spatial selection features, a dual-domain optimized feature is generated; Combining the fused feature map and the dual-domain optimized features, the final multi-scale fused features are generated through a preset second convolutional layer.

[0008] Furthermore, the coordinates of key points of the power distribution network conductors include conductor endpoints, conductor inflection points, conductor intersections, and conductor branch points; The step of inputting the multi-scale fused feature map into a preset detection head to obtain the key point coordinates and contour segmentation mask of the conductor to be identified output by the detection head includes: The multi-scale fused feature map is input to the detection head so that the detection head outputs the wire endpoints, wire inflection points, wire intersections, wire branch points, and contour segmentation masks of the wire to be identified. The detection head is obtained by joint training and optimization of a multi-task loss function based on aerial images of several power distribution network conductors and corresponding historical key point annotation data and historical contour segmentation mask annotation data of the power distribution network conductors.

[0009] Further, determining the conductor state of the conductor to be identified based on the key point coordinates and contour segmentation mask includes: The overall outline shape of the conductor to be identified is determined based on the outline segmentation mask of the conductor to be identified; Based on the conductor endpoints, inflection points, intersection points, and branch points of the conductor to be identified, determine the topological connection relationship and topological direction of the conductor to be identified; The overall contour shape of the conductor to be identified is subjected to contour integrity detection, and the topological continuity of the conductor to be identified is verified based on the topological connection relationship and topological direction of the conductor to be identified. If either a contour break or a topological continuity interruption is detected in the conductor to be identified, the conductor's state is determined to be abnormal. Conversely, it is determined that the wire to be identified is in a normal state.

[0010] Furthermore, before acquiring aerial images of the conductor to be identified, the process also includes: Obtain the original aerial image of the conductor to be identified; The original aerial image is preprocessed to obtain an aerial image of the conductor to be identified; The preprocessing includes any one or a combination of the following: image grayscale conversion, illumination equalization adjustment, and scale normalization.

[0011] An embodiment of the present invention also provides a power distribution network conductor status identification device, comprising: an image acquisition module, a feature extraction module, a feature perception processing module, a feature fusion module, a detection module, and a conductor status identification module; The image acquisition module is used to acquire aerial images of the conductor to be identified; The feature extraction module is used to input the aerial image into a preset backbone network to obtain a preliminary feature map output by the backbone network; The feature perception processing module is used to input the preliminary feature map into a preset MDPE module, so that the MDPE module performs multi-dimensional feature perception processing and noise suppression processing on the preliminary feature map to obtain a deep semantic feature map. The feature fusion module is used to input the deep semantic feature map into a preset HSFPN module, so that the HSFPN module performs attention feature fusion on the deep semantic feature map to obtain a multi-scale fused feature map; The detection module is used to input the multi-scale fused feature map into a preset detection head to obtain the key point coordinates and contour segmentation mask of the wire to be identified output by the detection head; The conductor state recognition module is used to determine the conductor state of the conductor to be recognized based on the key point coordinates and contour segmentation mask of the conductor to be recognized.

[0012] Furthermore, the power distribution network conductor status identification device further includes: an image preprocessing module; The image preprocessing module is used to acquire the original aerial image of the conductor to be identified; The original aerial image is preprocessed to obtain an aerial image of the conductor to be identified; The preprocessing includes any one or a combination of the following: image grayscale conversion, illumination equalization adjustment, and scale normalization.

[0013] This application also provides a terminal device, including: One or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the power distribution network conductor status identification method as described in the above embodiments of the invention.

[0014] This application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the power distribution network conductor status identification method as described in the above embodiments.

[0015] The following benefits can be obtained by implementing the present invention: This invention provides a method, apparatus, terminal device, and storage medium for identifying the state of power distribution network conductors. The method inputs an aerial photograph into a preset backbone network to obtain a preliminary feature map output by the backbone network, thus completing the initial feature extraction of the aerial photograph. The preliminary feature map is then input into a preset MDPE module, which performs multi-dimensional feature perception processing and noise suppression processing on the preliminary feature map to obtain a deep semantic feature map. This enhances the feature representation of slender linear targets such as power distribution network conductors, effectively suppresses redundant noise and interference information in complex backgrounds, and solves the problem of insufficient capture of local details of small targets. Through dynamic channel segmentation and multiple attention enhancement, the deep semantic feature map retains key details of the conductors while possessing stronger semantic discrimination capabilities. The deep semantic feature map is then input into a preset HSFPN module, which performs attention processing on the deep semantic feature map. Force feature fusion is used to obtain a multi-scale fused feature map, thereby solving the problems of semantic misalignment, target boundary ambiguity, and weak semantic expression of small targets in the deep and shallow feature fusion process. The generated multi-scale fused feature map is adapted to the detection requirements of conductors at different scales, further improving the integrity and reliability of conductor features. Then, the multi-scale fused feature map is input into a preset detection head to obtain the key point coordinates and contour segmentation mask of the conductor to be identified. Based on the key point coordinates and contour segmentation mask of the conductor to be identified, the conductor state of the conductor to be identified is determined. Thus, by detecting the integrity of the conductor contour and verifying the topological continuity, it is possible to accurately determine whether there are abnormalities such as contour breakage and topological continuity interruption of the conductor. This enables accurate differentiation between the normal and abnormal states of the conductor to be identified, providing a reliable basis for fault judgment for distribution network conductor inspection, helping power operation and maintenance personnel to promptly detect faults such as conductor strand breakage, and ensuring the safe and stable operation of distribution network lines. Attached Figure Description

[0016] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a method for identifying the status of power distribution network conductors according to a certain embodiment of this application; Figure 2 This is a schematic diagram of the structure of a power distribution network conductor status identification device provided in a certain embodiment of this application; Figure 3 This is a schematic diagram of the structure of a power distribution network conductor status identification device provided in another embodiment of this application; Figure 4 This is a schematic diagram of the structure of a terminal device provided in a certain embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] 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 pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0020] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0021] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0022] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0023] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0024] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0025] See Figure 1 To address the problem of low efficiency in identifying the status of power distribution network conductors in existing technologies, an embodiment of the present invention provides a method for identifying the status of power distribution network conductors, comprising: S1. Obtain aerial images of the conductor to be identified; Specifically, aerial images of the conductors to be identified are acquired, and feature extraction and status identification operations of the distribution network conductors will be carried out based on these aerial images.

[0026] In a preferred embodiment, before acquiring aerial images of the conductor to be identified, the method further includes: Obtain the original aerial image of the conductor to be identified; The original aerial image is preprocessed to obtain an aerial image of the conductor to be identified; The preprocessing includes any one or a combination of the following: image grayscale processing, illumination equalization adjustment, and scale normalization processing; As an illustration, when carrying out distribution network conductor inspection operations, it is necessary to use a drone equipped with high-definition imaging equipment to take original aerial images of the conductors to be identified, and then perform image preprocessing on the original aerial images; Specifically, during the shooting process, the external environment and shooting equipment can cause various interferences to the original aerial images, resulting in problems such as blurred features, inconsistent scales, and uneven lighting. Therefore, targeted preprocessing operations are required on the original aerial images of the guide lines to be identified, so as to obtain aerial images suitable for subsequent feature extraction. Among them, image grayscale processing is used to remove redundant information in the color dimension of the original aerial images, reduce the complexity of subsequent calculations, and focus on the contour and texture features of the guide lines to be identified; illumination equalization adjustment is used to improve the uneven lighting problems caused by strong light, backlight, shadows, etc. in the original aerial images, improve the contrast between the guide lines to be identified and the background, and enhance the visual features of the guide lines to be identified; scale normalization processing is used to unify the original aerial images of different shooting distances and resolutions to a preset size specification, ensure the consistency of input image data dimensions, and improve the consistency and accuracy of detection results.

[0027] S2. Input the aerial image into a preset backbone network to obtain a preliminary feature map output by the backbone network; As an illustration, in order to provide a semantic and spatial feature foundation for subsequent feature fusion and target detection, this embodiment requires preliminary feature extraction of aerial images; Specifically, the aerial images are input into the backbone network, which then performs layer-by-layer downsampling and feature enhancement on the aerial images through stacked convolutional layers, pooling layers, and residual connection structures. The shallow texture features, mid-level contour features, and deep semantic features of the aerial images are extracted sequentially, and finally, a multi-scale preliminary feature map is output. This preliminary feature map not only retains the spatial location information of the power distribution network conductors, but also has the semantic representation ability to distinguish conductors from complex backgrounds.

[0028] S3. Input the preliminary feature map into the preset MDPE module so that the MDPE module performs multi-dimensional feature perception processing and noise suppression processing on the preliminary feature map to obtain a deep semantic feature map. To illustrate, in order to improve the ability to capture the features of small wires and solve the problems of insufficient local details and noise interference in small target detection, it is necessary to perform channel-level dynamic segmentation and multi-dimensional attention enhancement processing on the preliminary feature map to obtain a deep semantic feature map.

[0029] In a preferred embodiment, the MDPE module includes a smoothing network, a channel attention submodule, a self-attention submodule, and a spatial attention submodule; The step of inputting the preliminary feature map into a preset MDPE module, so that the MDPE module performs multi-dimensional feature perception processing and noise suppression processing on the preliminary feature map to obtain a deep semantic feature map, includes: The preliminary feature map is input into a smoothing network, which uses grouped convolution and Gaussian smoothing operations to filter out high-frequency redundant information in the preliminary feature map, resulting in a denoised base feature map. The denoised base feature map is input into the channel attention submodule, so that the channel attention submodule calculates the weight coefficient of each feature channel in the denoised base feature map, and divides each feature channel into high importance feature channels and low importance feature channels based on each weight coefficient. The self-attention submodule is used to enhance the representation of detailed features in the high-importance feature channel, resulting in an enhanced high-importance feature channel; The spatial attention submodule is used to suppress noise in the low importance feature channel to obtain the noise-suppressed low importance feature channel; The enhanced high-importance feature channels and the suppressed low-importance feature channels are concatenated to obtain the deep semantic feature map; Indicatively, the MDPE module is used to enhance the feature representation of slender linear small targets such as distribution network conductors and suppress complex background noise. The MDPE module includes a smoothing network, a channel attention submodule, a self-attention submodule, and a spatial attention submodule.

[0030] Specifically, firstly, the preliminary feature map is input into a smoothing network, which is a lightweight network structure containing 1×1 convolutional layers and batch normalization (BN) layers to filter high-frequency redundant noise in the preliminary feature map and stabilize the feature distribution. Thus, the smoothing network uses grouped convolution operations to transform the channel dimension and perform weighted summation of features on the preliminary feature map to extract more abstract basic features. Then, Gaussian smoothing operations are used to filter high-frequency noise in the spatial dimension of the preliminary feature map, preserving key details such as wire outlines, thereby obtaining a denoised basic feature map.

[0031] In a schematic manner, the denoised base feature map is input into the channel attention submodule so that the channel attention submodule calculates the weight coefficient of each feature channel in the denoised base feature map; Specifically, the channel attention submodule performs global average pooling on the denoised base feature map, compressing the spatial information of each feature channel into a single value. Then, it learns the importance weights of each channel through two fully connected layers, outputting weight coefficients consistent with the number of feature channels. Assuming... For the first The weighting coefficients of each characteristic channel represent the contribution of each channel to the characteristics of the distribution network conductors. This is an illustration of how feature channels are divided into high-importance feature channels based on their respective weighting coefficients. and low importance feature channels ; Specifically, the dynamic partitioning of channels is achieved through the following formula: ; In the formula, This represents the mean of the channel weights; The standard deviation of the channel weights; This represents a hyperparameter used to control the dynamic range of the threshold. This represents the channel threshold, used to distinguish between high-importance feature channels and low-importance feature channels; That is, the weighting coefficient is greater than The feature channels are divided into high-importance feature channels, and the weight coefficients are no greater than 1. The feature channels are divided into low importance feature channels.

[0032] Then, the self-attention submodule is used to enhance the expression of detailed features in the high-importance feature channel. That is, the self-attention submodule calculates the correlation weight between pixels in the high-importance feature channel, focuses on key detailed features such as wire endpoints, inflection points, and edge textures, strengthens the semantic representation ability of fine wires, and obtains the enhanced high-importance feature channel. Furthermore, the Spatial Attention (SA) submodule is used to suppress noise in the low-importance feature channels. Specifically, the Spatial Attention submodule performs local feature weighting on the low-importance feature channels to weaken the spatial response of redundant information such as background and large targets, retaining only effective local details, thereby obtaining the noise-suppressed low-importance feature channels. Finally, the enhanced high-importance feature channels and the suppressed low-importance feature channels are concatenated to obtain the deep semantic feature map.

[0033] S4. Input the deep semantic feature map into the preset HSFPN module so that the HSFPN module performs attention feature fusion on the deep semantic feature map to obtain a multi-scale fused feature map; To illustrate, after obtaining the deep semantic feature map, it is necessary to further perform multi-scale feature alignment, edge enhancement, and dual-domain optimization on the deep semantic feature map to strengthen the feature representation of slender small targets such as power distribution network conductors, suppress background redundant noise, and provide a data foundation for the subsequent detection head to output accurate key point coordinates and contour segmentation masks.

[0034] In a preferred embodiment, the HSFPN module includes an edge sensing unit, a spatial selection unit, and a frequency selection unit; The step of inputting the deep semantic feature map into a preset HSFPN module, so that the HSFPN module performs attention feature fusion on the deep semantic feature map to obtain a multi-scale fused feature map, includes: The deep semantic feature map is input into the edge perception unit, so that the edge perception unit performs edge enhancement processing on the deep semantic feature map to obtain an edge-enhanced fused feature map. The fused feature map is input to the spatial selection unit, so that after the spatial selection unit performs a pooling operation on the fused feature map, it generates a spatial feature map through a preset first convolutional layer. The spatial feature map is input to the frequency selection unit, so that the frequency selection unit enhances the low-frequency semantic information of the spatial feature map through mean filtering, and then generates high-frequency edge features through a preset residual method. Based on the high-frequency edge features and the spatial selection features, a dual-domain optimized feature is generated; Combining the fused feature map and the dual-domain optimized features, the final multi-scale fused features are generated through a preset second convolutional layer; Schematic representation: The HSFPN module includes an edge-aware unit, a spatial selection unit, and a frequency selection unit. The edge-aware unit is used to dynamically capture and enhance high-frequency edge information in the deep semantic feature map, achieving accurate semantic and geometric aggregation of shallow and deep features. The spatial selection unit is used to select and enhance fused features in the spatial domain, highlighting the spatial location information of small targets and suppressing background spatial redundancy. The frequency selection unit is used to separate and enhance features in the frequency domain, strengthening low-frequency semantic information and high-frequency edge features, further improving the semantic expression capability of small targets and the distinguishability of target boundaries. Specifically, the deep semantic feature map is input to the edge perception unit. The edge perception unit adopts a bidirectional fusion path, adaptively weighting and complementing the high-importance feature channels and low-importance feature channels (including background noise and redundant information) in the deep semantic feature map. After weighting and upsampling, the deep semantic features are injected into the shallow branch, while the shallow features inject fine edge information into the high-level path to improve the spatial positioning accuracy of small targets. A preset edge enhancement algorithm is introduced into the bidirectional fusion path to suppress low-frequency background information and highlight high-frequency edge information, thereby improving the boundary discrimination of small targets. The specific calculation formula is as follows: ; ; In the formula, This represents the first weighted guiding map obtained after linear mapping and Sigmoid activation, which is used for weighted modulation of high-importance feature channels; This represents a 3×3 convolution with a Sigmoid activation function, used to enhance the saliency of edge features; This represents a convolution operation with a kernel size of 1×1, which serves as a linear mapping process used to adjust the channel dimension and transform features in highly important feature channels. , Represents high-importance feature channels in the deep semantic feature map; This represents the first weighted guiding map obtained after linear mapping and Sigmoid activation, used for weighted modulation of low-importance feature channels; This represents a convolution operation with a kernel size of 1×1, which serves as a linear mapping process used to adjust the channel dimension and transform features in low-importance feature channels. , This represents low-importance feature channels in the deep semantic feature map; This represents the features after edge enhancement; This represents the intermediate features of the bidirectional fusion path in the edge sensing unit; This represents the edge enhancement factor, used to control the intensity of edge enhancement; This represents the average pooling operation, used to extract the global mean information of features, which helps to enhance edge features; express; Indicates the height of the deep semantic feature map; Represents the width of the deep semantic feature map; This represents the number of channels in the deep semantic feature map.

[0035] Specifically, although the edge perception unit achieves complementarity between deep and shallow features through bidirectional weighted fusion, reducing redundant information and improving the localization accuracy of small targets, its semantic enhancement capability still has certain limitations. On the one hand, the global semantic information of small targets is inherently weak, and even if shallow branches are injected, it is difficult to fully express it. On the other hand, the edge perception unit is more inclined to edge compensation during feature fusion. Although it can highlight the outline of small targets, it is difficult to guarantee the semantic expression of small targets in complex backgrounds. Therefore, the fused feature map is input into the spatial selection unit, which generates a spatial attention map through local spatial structure, focusing limited semantic features on the target region, and then generates a spatial feature map through the first convolutional layer. The specific formula is as follows: ; ; In the formula, This represents the intermediate feature map after pooling and convolution processing; This represents a convolution operation with a kernel size of 3×3; This represents the max pooling operation, used to compress features along the channel dimension and highlight salient information; This represents the average pooling operation, used to compress features along the channel dimension while preserving global statistics. Represents spatial feature maps; This indicates a depthwise separable convolution operation with a kernel size of 5×7, used to enhance the local expressive power of features; This indicates a depthwise separable convolution operation with a kernel size of 3×3; This indicates a channel dimension transformation operation, used to transform intermediate feature maps. The number of channels was adjusted to match Consistency is required to achieve point-by-point multiplication; This indicates a pointwise multiplication operation, used to weightedly fuse the features enhanced by depthwise separable convolution with the spatial attention map; Indicates the target number of channels; Specifically, the spatial feature map is input to the frequency selection unit, which performs a mean filtering operation, i.e., local averaging, to enhance the low-frequency semantic information of the spatial feature map, and then generates high-frequency edge features using a preset residual method. Finally, based on the high-frequency edge features and the spatial selection features, dual-domain optimized features are generated. The specific formula is as follows: ; Combining the fused feature map and the dual-domain optimized features, the final multi-scale fused features are generated through a pre-defined second convolutional layer. The specific formula is as follows: ; In the formula, This represents a convolution operation with a kernel size of 3×3; express; Indicates the function name declaration command; This represents the fused feature map output by the edge sensing unit; Indicates the spatial domain transformation parameters; Indicates the frequency domain transformation parameters; This indicates a feature splicing operation.

[0036] S5. Input the multi-scale fused feature map into the preset detection head to obtain the key point coordinates and contour segmentation mask of the wire to be identified output by the detection head; Indicatively, after image preprocessing, feature extraction, MDPE module enhancement, and HSFPN module feature fusion in steps S1-S4, the generated multi-scale fused feature map possesses clear conductor edge details, enhanced small target semantic expression, and suppressed background noise. Therefore, it provides reliable data support for subsequent conductor state determination. The multi-scale fused feature map is input into a preset detection head to determine the key point coordinates and contour segmentation mask of the conductor to be identified. The key point coordinates are used to clarify the specific spatial locations of the conductor endpoints, inflection points, intersections, and branch points, and to clarify the topological connection relationship and direction of the conductor. The contour segmentation mask is used to outline the pixel-level complete contour of the conductor to be identified, distinguish the conductor region from the background region, and provide a basis for subsequent contour integrity detection.

[0037] In a preferred embodiment, the key point coordinates of the power distribution network conductors include conductor endpoints, conductor inflection points, conductor intersections, and conductor branch points; The step of inputting the multi-scale fused feature map into a preset detection head to obtain the key point coordinates and contour segmentation mask of the conductor to be identified output by the detection head includes: The multi-scale fused feature map is input to the detection head so that the detection head outputs the wire endpoints, wire inflection points, wire intersections, wire branch points, and contour segmentation masks of the wire to be identified. The detection head is obtained by joint training and optimization of a multi-task loss function based on aerial images of several power distribution network conductors and corresponding historical power distribution network conductor key point annotation data and historical contour segmentation mask annotation data. Schematic representation: The key point coordinates of the distribution network conductors include conductor endpoints, conductor inflection points, conductor intersections, and conductor branch points. The conductor endpoints are used to determine the start and end range of the conductor to be identified, clarify the spatial extension boundary of the conductor, and provide a basic positional reference for judging whether there are abnormalities such as conductor breakage or detachment. The conductor inflection points are used to capture changes in the spatial orientation of the conductor, accurately characterize the bending shape of the conductor in complex terrain, and assist in identifying abnormal bending of the conductor caused by stress or aging. The conductor intersections are used to identify the intersection positions between different conductors and investigate safety hazards such as excessive conductor crossings or contact. The conductor branch points are used to distinguish the main and branch structures of the conductor, clarify the topological connection relationship of the distribution network conductors, and provide a structural basis for subsequent judgment of conductor abnormalities. Since the detection head needs to perform two core tasks simultaneously—accurately outputting the coordinates of four types of key points of the distribution network conductors (regression task) and the conductor contour segmentation mask (segmentation task)—and the two tasks need to work together to meet the high-precision detection requirements of slender and small targets of the distribution network conductors, the detection head needs to be optimized through joint training of a multi-task loss function based on aerial images of several distribution network conductors and corresponding historical distribution network conductor key point annotation data and historical contour segmentation mask annotation data. Specifically, aerial images of power distribution network conductors covering different inspection scenarios such as mountainous areas, forest areas, and urban-rural fringe areas are selected as training samples. The endpoints, inflection points, intersections, and branch points of the conductors in each sample image are accurately labeled with coordinates to obtain historical power distribution network conductor key point labeling data. At the same time, the conductor regions in each sample image are labeled with pixel-level contours to obtain historical contour segmentation mask labeling data. The training sample images are input into the initial detection head, and the CIoU loss of the key point regression task and the cross-entropy loss of the contour segmentation task are calculated respectively. The two types of losses are fused by the multi-task loss function to obtain the total loss value. Based on the total loss value, the network parameters of the detection head are adjusted synchronously through backpropagation of the gradient descent optimization algorithm. The training is iterated until the total loss value converges and the detection accuracy reaches the preset threshold, and finally a detection head that can be adapted to the power distribution network conductor detection scenario is obtained. Specifically, the multi-scale fused feature map is input into the trained detection head so that the detection head outputs the wire endpoints, wire inflection points, wire intersections, wire branch points, and contour segmentation masks of the wire to be identified.

[0038] S6. Determine the conductor state of the conductor to be identified based on the key point coordinates and contour segmentation mask of the conductor to be identified; Indicatively, the key point coordinates and contour segmentation mask of the conductor to be identified reflect the topological structure and complete shape of the conductor. The key point coordinates clearly show the node distribution of the conductor, while the contour segmentation mask outlines the complete pixel-level contour of the conductor. They can work together to capture abnormal features such as conductor breakage and abnormal bending. Therefore, by combining the two for analysis, the operating status of the conductor to be identified can be accurately determined, avoiding the problems of missed detection and false detection caused by single feature analysis.

[0039] In a preferred embodiment, determining the conductor state of the conductor to be identified based on the key point coordinates and contour segmentation mask includes: The overall outline shape of the conductor to be identified is determined based on the outline segmentation mask of the conductor to be identified; Based on the conductor endpoints, inflection points, intersection points, and branch points of the conductor to be identified, determine the topological connection relationship and topological direction of the conductor to be identified; The overall contour shape of the conductor to be identified is subjected to contour integrity detection, and the topological continuity of the conductor to be identified is verified based on the topological connection relationship and topological direction of the conductor to be identified. If either a contour break or a topological continuity interruption is detected in the conductor to be identified, the conductor's state is determined to be abnormal. Conversely, it is determined that the wire to be identified is in a normal state; Specifically, based on the contour segmentation mask of the conductor to be identified, the overall contour shape of the conductor to be identified is determined. For example, the single-pixel skeleton line of the conductor is extracted by the contour segmentation mask to clarify the overall extension trajectory, thickness distribution and spatial coverage of the conductor. It is possible to intuitively observe whether there are abnormal signs such as local pixel loss or contour break gaps in the conductor. Furthermore, based on the conductor endpoints, conductor inflection points, conductor intersection points and conductor branch points of the conductor to be identified, the topological connection relationship and topological direction of the conductor to be identified are determined. For example, the start and end positions of a single conductor segment are determined by the conductor endpoints, the complete trajectory of the conductor is formed by connecting the endpoints by the conductor inflection points, the spatial intersection relationship between different conductors is clarified by the conductor intersection points, and the connection nodes between the conductor trunk and branches are distinguished by the conductor branch points, thereby sorting out the overall topological network structure of the distribution network conductor. Then, contour integrity detection is performed on the overall contour shape of the conductor to be identified. For example, the number of connected components of the conductor contour is calculated. If two or more independent connected components are detected in the conductor contour and the distance between the connected components exceeds a preset threshold (set according to the normal thickness of the conductor and the aerial photography resolution), or if there are obvious breaks or gaps in the single-pixel skeleton line of the conductor contour, it can be preliminarily determined that the conductor has a suspected contour break. At the same time, local pixel analysis is performed on the contour break area to eliminate false positive break points caused by background occlusion and sudden changes in lighting during drone aerial photography, so as to ensure the accuracy of contour integrity detection.

[0040] Based on the topological connection relationship and topological direction of the conductor to be identified, a topological continuity verification is performed on the conductor. For example, according to the preset topological logic of the distribution network conductor, the rationality of the connection between various key points and the trajectory coherence are verified, and problems such as disordered distribution of key points and abnormal connection relationship are investigated. Specifically, conductor endpoints, conductor inflection points, conductor crossing points, and conductor branching points, as core nodes of the conductor topology, are key indicators for determining discontinuities in topological continuity, and their specific functions are as follows: (1) Wire endpoints: As the core of the start and end of a single wire segment, under normal circumstances, a single wire segment has only two endpoints, and the two endpoints need to be connected by continuous wire outlines and inflection points to form a complete trajectory; if it is detected that the wire to be identified has three or more endpoints, or there are no continuous inflection points and wire outlines connecting the endpoints, it is determined that the topological continuity of the wire to be identified is interrupted. For example, after the wire strand is broken, a new endpoint will be formed at the break, resulting in an abnormal number of endpoints of the wire to be identified, and there is no effective connection between the new endpoint and the original endpoint; (2) Conductor inflection points: used to connect different directions of conductors. Under normal circumstances, inflection points should be evenly distributed on the conductor extension trajectory. The conductor outline should be continuous between adjacent inflection points and between inflection points and endpoints. The angle distribution of inflection points should conform to the normal bending law of conductors (the normal bending angle of distribution network conductors is usually not less than 30°). If it is detected that there is no continuous conductor outline between adjacent inflection points of the conductor to be identified, the inflection point angle is abnormal, such as approaching 0° or 180°, which does not conform to the actual conductor direction, or there is a trajectory discontinuity between the inflection point and the endpoint, it is determined that the topological continuity is interrupted. (3) Crossing point of conductors: Under normal circumstances, it is only a spatial intersection node between different conductors and does not affect the topological continuity of a single conductor. If a discontinuity is detected in the outline and trajectory of the conductors on both sides of the crossing point, it is determined that the topological continuity of the conductor to be identified is interrupted in the crossing point area. For example, if the conductor to be identified breaks or falls off at the crossing point, the conductors on both sides of the crossing point cannot form a continuous trajectory. (4) Wire branch point: As the core of the connection between the main trunk and the branch, under normal circumstances, the branch point needs to connect the trajectory of the main trunk wire and the trajectory of the branch wire at the same time. The endpoints and inflection points of the main trunk and the branch need to form a continuous topological connection around the branch point. If it is detected that the branch point only connects to the main trunk or only connects to the branch, without a corresponding branch or main trunk trajectory connection, or there is a trajectory discontinuity between the branch point and the endpoints of the main trunk and the branch, it is determined that the topological continuity of the wire to be identified is interrupted in the branch point area, such as the branch wire breaking or falling off, which causes the topological connection between the branch and the main trunk to be broken.

[0041] When either a contour break or a topological continuity interruption is detected in the conductor to be identified, the conductor state is determined to be in an abnormal state. Specifically, when the conductor contour has a verified connection component anomaly, the conductor to be identified is determined to have a contour break. When the above four types of key point verifications reveal an abnormal number of endpoints, a discontinuous inflection point connection, or disordered connection of intersections or branch points, the conductor to be identified is determined to have a topological continuity interruption. Conversely, it is determined that the wire to be identified is in a normal state.

[0042] See Figure 2This invention provides a power distribution network conductor status identification device according to an embodiment of the present invention, comprising: an image acquisition module, a feature extraction module, a feature perception processing module, a feature fusion module, a detection module, and a conductor status identification module; The image acquisition module is used to acquire aerial images of the conductor to be identified; The feature extraction module is used to input the aerial image into a preset backbone network to obtain a preliminary feature map output by the backbone network; The feature perception processing module is used to input the preliminary feature map into a preset MDPE module, so that the MDPE module performs multi-dimensional feature perception processing and noise suppression processing on the preliminary feature map to obtain a deep semantic feature map. The feature fusion module is used to input the deep semantic feature map into a preset HSFPN module, so that the HSFPN module performs attention feature fusion on the deep semantic feature map to obtain a multi-scale fused feature map; The detection module is used to input the multi-scale fused feature map into a preset detection head to obtain the key point coordinates and contour segmentation mask of the wire to be identified output by the detection head; The conductor state recognition module is used to determine the conductor state of the conductor to be recognized based on the key point coordinates and contour segmentation mask of the conductor to be recognized.

[0043] In a preferred embodiment, the MDPE module includes a smoothing network, a channel attention submodule, a self-attention submodule, and a spatial attention submodule; The feature perception and processing module is specifically used for: The preliminary feature map is input into a smoothing network, which uses grouped convolution and Gaussian smoothing operations to filter out high-frequency redundant information in the preliminary feature map, resulting in a denoised base feature map. The denoised base feature map is input into the channel attention submodule, so that the channel attention submodule calculates the weight coefficient of each feature channel in the denoised base feature map, and divides each feature channel into high importance feature channels and low importance feature channels based on each weight coefficient. The self-attention submodule is used to enhance the representation of detailed features in the high-importance feature channel, resulting in an enhanced high-importance feature channel; The spatial attention submodule is used to suppress noise in the low importance feature channel to obtain the noise-suppressed low importance feature channel; The enhanced high-importance feature channels and the suppressed low-importance feature channels are concatenated to obtain the deep semantic feature map.

[0044] In a preferred embodiment, the HSFPN module includes an edge sensing unit, a spatial selection unit, and a frequency selection unit; The feature fusion module is specifically used for: The deep semantic feature map is input into the edge perception unit, so that the edge perception unit performs edge enhancement processing on the deep semantic feature map to obtain an edge-enhanced fused feature map. The fused feature map is input to the spatial selection unit, so that after the spatial selection unit performs a pooling operation on the fused feature map, it generates a spatial feature map through a preset first convolutional layer. The spatial feature map is input to the frequency selection unit, so that the frequency selection unit enhances the low-frequency semantic information of the spatial feature map through mean filtering, and then generates high-frequency edge features through a preset residual method. Based on the high-frequency edge features and the spatial selection features, a dual-domain optimized feature is generated; Combining the fused feature map and the dual-domain optimized features, the final multi-scale fused features are generated through a preset second convolutional layer.

[0045] In a preferred embodiment, the key point coordinates of the power distribution network conductors include conductor endpoints, conductor inflection points, conductor intersections, and conductor branch points; The detection module is specifically used for: The multi-scale fused feature map is input to the detection head so that the detection head outputs the wire endpoints, wire inflection points, wire intersections, wire branch points, and contour segmentation masks of the wire to be identified. The detection head is obtained by joint training and optimization of a multi-task loss function based on aerial images of several power distribution network conductors and corresponding historical key point annotation data and historical contour segmentation mask annotation data of the power distribution network conductors.

[0046] In a preferred embodiment, the conductor state identification module is specifically used for: The overall outline shape of the conductor to be identified is determined based on the outline segmentation mask of the conductor to be identified; Based on the conductor endpoints, inflection points, intersection points, and branch points of the conductor to be identified, determine the topological connection relationship and topological direction of the conductor to be identified; The overall contour shape of the conductor to be identified is subjected to contour integrity detection, and the topological continuity of the conductor to be identified is verified based on the topological connection relationship and topological direction of the conductor to be identified. If either a contour break or a topological continuity interruption is detected in the conductor to be identified, the conductor's state is determined to be abnormal. Conversely, it is determined that the wire to be identified is in a normal state.

[0047] See Figure 3 In a preferred embodiment, the power distribution network conductor status identification device further includes: an image preprocessing module; The image preprocessing module is used to acquire the original aerial image of the conductor to be identified; The original aerial image is preprocessed to obtain an aerial image of the conductor to be identified; The preprocessing includes any one or a combination of the following: image grayscale conversion, illumination equalization adjustment, and scale normalization.

[0048] See Figure 4 One embodiment of this application also provides a terminal device, including: One or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the distribution network conductor status identification method as described above.

[0049] The processor controls the overall operation of the terminal device to complete all or part of the steps of the aforementioned power distribution network conductor status identification method. The memory stores various types of data to support the operation of the terminal device. This data may include, for example, instructions for any application or method used to operate on the terminal device, as well as application-related data. The memory can be implemented using any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0050] In an exemplary embodiment, the terminal device may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the power distribution network conductor status identification method as described in any of the above embodiments and achieve the same technical effect as the above method.

[0051] In another exemplary embodiment, a computer-readable storage medium including a computer program is also provided. When executed by a processor, the computer program implements the steps of the distribution network conductor state identification method as described in any of the above embodiments. For example, the computer-readable storage medium may be the aforementioned memory including the computer program, which may be executed by a processor of a terminal device to complete the distribution network conductor state identification method as described in any of the above embodiments and achieve the same technical effects as the aforementioned method.

[0052] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for identifying the status of conductors in a power distribution network, characterized in that, include: Acquire aerial images of the conductor to be identified; The aerial images are input into a preset backbone network to obtain a preliminary feature map output by the backbone network; The preliminary feature map is input into a preset MDPE module, so that the MDPE module performs multi-dimensional feature perception processing and noise suppression processing on the preliminary feature map to obtain a deep semantic feature map. The deep semantic feature map is input into a preset HSFPN module, so that the HSFPN module performs attention feature fusion on the deep semantic feature map to obtain a multi-scale fused feature map; The multi-scale fused feature map is input into a preset detection head to obtain the key point coordinates and contour segmentation mask of the wire to be identified output by the detection head; The conductor state is determined based on the key point coordinates and contour segmentation mask of the conductor to be identified.

2. The method for identifying the status of power distribution network conductors as described in claim 1, characterized in that, The MDPE module includes a smoothing network, a channel attention submodule, a self-attention submodule, and a spatial attention submodule. The step of inputting the preliminary feature map into a preset MDPE module, so that the MDPE module performs multi-dimensional feature perception processing and noise suppression processing on the preliminary feature map to obtain a deep semantic feature map, includes: The preliminary feature map is input into a smoothing network, which uses grouped convolution and Gaussian smoothing operations to filter out high-frequency redundant information in the preliminary feature map, resulting in a denoised base feature map. The denoised base feature map is input into the channel attention submodule, so that the channel attention submodule calculates the weight coefficient of each feature channel in the denoised base feature map, and divides each feature channel into high importance feature channels and low importance feature channels based on each weight coefficient. The self-attention submodule is used to enhance the representation of detailed features in the high-importance feature channel, resulting in an enhanced high-importance feature channel; The spatial attention submodule is used to suppress noise in the low importance feature channel to obtain the noise-suppressed low importance feature channel; The enhanced high-importance feature channels and the suppressed low-importance feature channels are concatenated to obtain the deep semantic feature map.

3. The method for identifying the status of power distribution network conductors as described in claim 1, characterized in that, The HSFPN module includes an edge sensing unit, a spatial selection unit, and a frequency selection unit; The step of inputting the deep semantic feature map into a preset HSFPN module, so that the HSFPN module performs attention feature fusion on the deep semantic feature map to obtain a multi-scale fused feature map, includes: The deep semantic feature map is input into the edge perception unit, so that the edge perception unit performs edge enhancement processing on the deep semantic feature map to obtain an edge-enhanced fused feature map. The fused feature map is input to the spatial selection unit, so that after the spatial selection unit performs a pooling operation on the fused feature map, it generates a spatial feature map through a preset first convolutional layer. The spatial feature map is input to the frequency selection unit, so that the frequency selection unit enhances the low-frequency semantic information of the spatial feature map through mean filtering, and then generates high-frequency edge features through a preset residual method. Based on the high-frequency edge features and the spatial selection features, a dual-domain optimized feature is generated; Combining the fused feature map and the dual-domain optimized features, the final multi-scale fused features are generated through a preset second convolutional layer.

4. The method for identifying the status of power distribution network conductors as described in claim 1, characterized in that, The coordinates of key points of the distribution network conductors include conductor endpoints, conductor inflection points, conductor intersections, and conductor branch points; The step of inputting the multi-scale fused feature map into a preset detection head to obtain the key point coordinates and contour segmentation mask of the conductor to be identified output by the detection head includes: The multi-scale fused feature map is input to the detection head so that the detection head outputs the wire endpoints, wire inflection points, wire intersections, wire branch points, and contour segmentation masks of the wire to be identified. The detection head is obtained by joint training and optimization of a multi-task loss function based on aerial images of several power distribution network conductors and corresponding historical key point annotation data and historical contour segmentation mask annotation data of the power distribution network conductors.

5. The method for identifying the status of power distribution network conductors as described in claim 1, characterized in that, The step of determining the conductor state based on the key point coordinates and contour segmentation mask of the conductor to be identified includes: The overall outline shape of the conductor to be identified is determined based on the outline segmentation mask of the conductor to be identified; Based on the conductor endpoints, inflection points, intersection points, and branch points of the conductor to be identified, determine the topological connection relationship and topological direction of the conductor to be identified; The overall contour shape of the conductor to be identified is subjected to contour integrity detection, and the topological continuity of the conductor to be identified is verified based on the topological connection relationship and topological direction of the conductor to be identified. If either a contour break or a topological continuity interruption is detected in the conductor to be identified, the conductor's state is determined to be abnormal. Conversely, it is determined that the wire to be identified is in a normal state.

6. The method for identifying the status of power distribution network conductors as described in claim 1, characterized in that, Before acquiring aerial images of the conductor to be identified, the process also includes: Obtain the original aerial image of the conductor to be identified; The original aerial image is preprocessed to obtain an aerial image of the conductor to be identified; The preprocessing includes any one or a combination of the following: image grayscale conversion, illumination equalization adjustment, and scale normalization.

7. A power distribution network conductor status identification device, characterized in that, include: The system includes an image acquisition module, a feature extraction module, a feature perception and processing module, a feature fusion module, a detection module, and a conductor state recognition module. The image acquisition module is used to acquire aerial images of the conductor to be identified; The feature extraction module is used to input the aerial image into a preset backbone network to obtain a preliminary feature map output by the backbone network; The feature perception processing module is used to input the preliminary feature map into a preset MDPE module, so that the MDPE module performs multi-dimensional feature perception processing and noise suppression processing on the preliminary feature map to obtain a deep semantic feature map. The feature fusion module is used to input the deep semantic feature map into a preset HSFPN module, so that the HSFPN module performs attention feature fusion on the deep semantic feature map to obtain a multi-scale fused feature map; The detection module is used to input the multi-scale fused feature map into a preset detection head to obtain the key point coordinates and contour segmentation mask of the wire to be identified output by the detection head; The conductor state recognition module is used to determine the conductor state of the conductor to be recognized based on the key point coordinates and contour segmentation mask of the conductor to be recognized.

8. The power distribution network conductor status identification device as described in claim 7, characterized in that, Also includes: Image preprocessing module; The image preprocessing module is used to acquire the original aerial image of the conductor to be identified; The original aerial image is preprocessed to obtain an aerial image of the conductor to be identified; The preprocessing includes any one or a combination of the following: image grayscale conversion, illumination equalization adjustment, and scale normalization.

9. A terminal device, characterized in that, include: One or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the power distribution network conductor status identification method as described in any one of claims 1-6.

10. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the distribution network conductor status identification method as described in any one of claims 1-6.

Images