Unmanned aerial vehicle transmission, transformation and distribution integrated intelligent inspection method and system based on image processing algorithm, and medium

Abstract

The invention discloses an unmanned aerial vehicle transmission, transformation and distribution integrated intelligent inspection method and system based on an image processing algorithm, and a medium. The method comprises the following steps: constructing a gridding intelligent inspection network, and configuring an unmanned aerial vehicle nest integrating charging, meteorological monitoring and 5G communication; the central control platform integrates transmission, transformation and distribution inspection requirements and dynamically optimizes a task sequence; the unmanned aerial vehicle collects an equipment image, and an optimized image is generated through multi-dimensional feature fusion; an improved MFB-Otsu algorithm is combined with a Gaussian process to model and position an optimal segmentation threshold; intelligent defect identification is realized by using a multi-scale deep convolutional neural network and a cross-professional attention mechanism, and a three-dimensional evaluation result including types, positions and severity is output; and edge real-time processing and central platform closed-loop optimization are realized through a two-stage architecture. According to the method, efficient collaborative inspection of the transmission, transformation and distribution equipment is realized, the defect identification precision and efficiency are improved, and the method has the characteristics of adaptive optimization and cross-professional fusion.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent inspection of power equipment, and specifically to an integrated intelligent inspection method, system, and medium for transmission, transformation, and distribution using unmanned aerial vehicles (UAVs) based on image processing algorithms. The method is particularly suitable for the automated inspection of power grid transmission lines, substation equipment, and distribution lines. Background Art

[0002] With the rapid development of the power system, the scale of power transmission, transformation, and distribution equipment continues to expand. Traditional manual inspection methods can no longer meet the needs of efficient and accurate operation and maintenance. Drone inspection technology has gradually been applied due to its flexibility and efficiency, but the following problems still exist in actual operation: Insufficient multi-disciplinary collaboration: Existing inspection systems are usually designed for single power equipment (such as transmission lines or substations), lacking an integrated coordinated scheduling mechanism for transmission, transformation, and distribution, resulting in low resource utilization and limited inspection efficiency. Low image processing accuracy: In complex industrial environments, equipment images are subject to significant noise interference. Traditional segmentation algorithms (such as Otsu) rely on a single grayscale feature for threshold selection, making it difficult to adapt to changing scenarios, resulting in decreased defect recognition accuracy. Insufficient real-time and intelligence: Drone edge computing capabilities are limited, and existing methods mostly rely on cloud-based processing, making it difficult to achieve real-time defect detection. At the same time, model parameters are fixed and cannot be dynamically optimized according to equipment type and environment, affecting inspection adaptability. Lack of closed-loop optimization: Most systems only complete defect detection and do not feed back recognition results to the scheduling and image processing links, making it difficult to achieve adaptive improvement of inspection strategies.

[0003] In response to the above problems, there is an urgent need for an integrated intelligent inspection method, system, and medium for transmission, transformation, and distribution using drones based on image processing algorithms to improve the intelligence level of power equipment operation and maintenance. Summary of the Invention

[0004] The present invention proposes a method, system, and medium for integrated intelligent inspection of power transmission, transformation, and distribution using drones based on image processing algorithms. By integrating multi-dimensional feature perception with mathematical acceleration models, the accuracy and computational efficiency of image segmentation of power equipment are significantly improved, thus realizing intelligent joint inspection of power transmission, transformation, and distribution equipment.

[0005] In a first aspect, an embodiment of the present application provides a method for integrated intelligent inspection of transmission, transformation and distribution using a drone based on an image processing algorithm, the method comprising: S1. Build an intelligent inspection network, divide the inspection area into grids, and determine the grid unit size based on equipment distribution density and inspection frequency requirements. Deploy an intelligent drone nest within each grid unit, integrating an autonomous charging module, a weather monitoring unit, and a 5G communication module. Use the communication module to establish a real-time data connection between the nest and the central control platform, forming an intelligent inspection network covering the entire area. S2. The central control platform receives and integrates inspection requirements from various disciplines, including power transmission, transformation, and distribution. It uses an intelligent scheduling algorithm based on equipment status assessment to dynamically calculate task priorities. It also optimizes and generates a cross-disciplinary joint inspection task sequence based on the drone's endurance and grid coverage. S3. Use the industrial camera carried by the drone to collect images of the transmission and distribution equipment, and simultaneously perform multi-dimensional image feature fusion on the collected images to generate an optimized fused image; S4. Segmenting the optimized fused image based on the MFB-Otsu algorithm, and quickly locating the optimal segmentation threshold through Gaussian process modeling and expected improvement function; S5. Construct a deep convolutional neural network containing a multi-scale convolutional kernel group, and optimize the feature extraction strategy for transmission lines, substation equipment, and distribution equipment respectively through a cross-disciplinary attention mechanism; perform multi-scale defect intelligent identification on the segmented images through the deep convolutional neural network, and output a three-dimensional assessment result including defect type, location coordinates, and severity score; S6: Optimize and adjust the recognition results. Dynamically adjust the feature fusion weight coefficient in S2 according to the device type. Optimize the segmentation parameters in S3 based on real-time meteorological data. Update the defect recognition model parameters in S4 through an online learning mechanism based on historical segmentation results, performing adaptive closed-loop optimization. Among them, steps S3-S6 are implemented through a two-level architecture deployed on the drone edge computing node and the central control platform. The drone node is responsible for the real-time processing of S3-S5, and the central platform is responsible for the model optimization and result verification of S6, forming a complete intelligent inspection closed-loop system.

[0006] Optionally, in an implementation of the first aspect of the present invention, step S2 specifically includes: S2-1. Receive inspection requirements from power transmission, transformation, and distribution professionals, and establish a multi-dimensional inspection requirement matrix based on equipment type, defect level, and operating data; S2-2. Adopt an intelligent scheduling algorithm based on equipment status assessment, combining the real-time health of the equipment, historical failure probability, and operation and maintenance urgency to dynamically calculate the comprehensive priority weight of each inspection task; S2-3. Based on the real-time endurance of the UAV, the distribution of charging stations, and the coverage of the grid partition, a task allocation optimization model is constructed to minimize the total inspection time and maximize resource utilization, generating the optimal cross-disciplinary joint inspection task sequence.

[0007] Optionally, in an implementation of the first aspect of the present invention, the synchronously performing multi-dimensional image feature fusion on the collected image in step S3 includes: Grayscale histogram equalization is used to extract grayscale value features and maintain the global brightness distribution of the image; Apply the Sobel operator to calculate the gradient amplitude characteristics and enhance the edge contours of key parts of equipment insulators and conductor connections; Extract LBP texture features to suppress background noise in complex industrial environments; The three features are weightedly fused according to the weight coefficient of 0.4:0.4:0.2 to generate the optimized fused image.

[0008] Optionally, in an implementation of the first aspect of the present invention, the step S4 of segmenting the optimized fused image based on the MFB-Otsu algorithm and quickly locating the optimal segmentation threshold through Gaussian process modeling and expected improvement function includes: S4-1. Adopt the improved Otsu algorithm with multi-feature fusion, and combine the grayscale distribution, edge gradient and texture features of the image to construct a multi-dimensional segmentation evaluation function; S4-2, probabilistically fitting the segmentation threshold search space through Gaussian process modeling, and dynamically evaluating the segmentation performance of candidate thresholds using the expected improvement function (EI); S4-3. Based on the Bayesian optimization strategy, the Gaussian process model is iteratively updated to quickly converge to the global optimal segmentation threshold, achieving high-precision extraction of inspection targets.

[0009] Optionally, in an implementation of the first aspect of the present invention, the step S4-1, using an improved Otsu algorithm with multi-feature fusion to construct a multidimensional segmentation evaluation function in combination with the grayscale distribution, edge gradient, and texture features of the image, includes: S4-1-1. Calculate the grayscale inter-class variance based on the recursive Otsu algorithm: Use the recursive formula to dynamically update the statistical parameters of the foreground and background: , , in, For the threshold When the pixel ratio of the background area is Indicated as the threshold When , the grayscale mean of the background area is Grayscale The probability of the pixel appearing; S4-1-2, Combine edge gradient features to enhance segmentation boundaries: Calculate the image gradient magnitude using the Sobel operator , and construct the edge energy term: , in is the threshold The set of segmented foreground and background pixels, Represents the gradient magnitude of the pixel in the image, represents the edge energy term, which is used to measure the sum of the edge strength at the segmentation boundary; S4-1-3. Introducing texture features to optimize segmentation robustness: Calculate texture contrast features based on gray-level co-occurrence matrix GLCM: , in, is the threshold Next, the gray level in the gray level co-occurrence matrix GLCM and The joint probability of S4-1-4. Construct a multidimensional evaluation function and optimize the threshold: Search range in grayscale In the internal structure, the three types of features are integrated to solve the optimal threshold : , in, Represents the inter-class variance of traditional Otsu, which measures the separation between foreground and background, and the weight coefficient The contribution weights corresponding to grayscale, edge, and texture features respectively satisfy , and dynamically adjusted according to the image signal-to-noise ratio, Represents the average gray value of the entire image, which is used to limit the threshold search range. Represents the optimal threshold, and the optimal segmentation threshold is determined by maximizing the fusion evaluation function. Contrast(T) represents the texture contrast feature calculated based on the gray-level co-occurrence matrix under a given segmentation threshold T.

[0010] Optionally, in an implementation of the first aspect of the present invention, the step S4-2, performing probabilistic fitting on the segmentation threshold search space through Gaussian process modeling, and dynamically evaluating the segmentation performance of the candidate threshold using the expected improvement function EI, includes: S4-2-1. Establish a Gaussian process proxy model for segmentation performance: The threshold T is used as the input variable and the corresponding multidimensional evaluation function value F(T) is used as the output to construct a Gaussian process regression model: , in, is the multidimensional evaluation function value, indicating the threshold The corresponding segmentation performance, is the mean function of the Gaussian process, which represents the The initial prediction of is the covariance kernel function based on the radial basis function RBF, which measures the threshold and The similarities between Represents a Gaussian process model, which is used to fit the nonlinear relationship between threshold and segmentation performance. Represents the Gaussian process function in the Gaussian process model; S4-2-2. Design an adaptive sampling strategy: a) Uniformly sample N candidate thresholds in the initial stage , calculate its evaluation function value ; b) Update the Gaussian process hyperparameters based on the sampled data to obtain the posterior probability distribution; S4-2-3. Using the expected improvement function Guidance thresholds: Define the improvement function , indicating the current threshold Relative to the known best of which is the current optimal evaluation function value, Calculating expected improvement : , in, Used to select the most promising candidate threshold, and Forecast the mean and standard deviation for a Gaussian process, To balance the parameters and control the weight of exploration and utilization, They are the cumulative distribution function CDF and probability density function PDF of the standard normal distribution respectively; S4-2-4, iterative optimization until convergence: a) Select Maximum candidate threshold Conduct practical assessments; b) Add the training set to update the Gaussian process model; c) Repeat the above process until Less than the set threshold or the maximum number of iterations is reached.

[0011] Optionally, in an implementation of the first aspect of the present invention, S4-3, iteratively updating the Gaussian process model based on a Bayesian optimization strategy to quickly converge to a global optimal segmentation threshold to achieve high-precision extraction of inspection targets, includes: S4-3-1. Build an adaptive optimization framework: Establish a Bayesian optimization framework with image segmentation evaluation index as the objective function and threshold as the optimization variable, and establish the mapping relationship between threshold and segmentation performance through the Gaussian process model; S4-3-2. Implement intelligent iterative optimization: a) Generate a set of candidate thresholds based on the current model predictions; b) Select the most promising evaluation points using an acquisition function that balances exploration and exploitation; c) Dynamically adjust the search strategy, focusing on global exploration in the early stages and shifting towards local, refined search in the later stages; S4-3-3. Dynamic model update mechanism: After each evaluation of a new threshold, the parameters and hyperparameters of the Gaussian process model are updated in real time to gradually improve the model's fitting accuracy to the objective function; S4-3-4. Set intelligent termination conditions: When the stability of the optimal solution and sufficient convergence of the search interval are met at the same time, the optimization process will be automatically terminated; S4-3-5. Output the optimal segmentation solution: Output the optimal segmentation threshold determined by Bayesian optimization to achieve accurate extraction of inspection targets.

[0012] Optionally, in an implementation of the first aspect of the present invention, S5 specifically includes: S5-1. Construct a multi-scale feature extraction network: A parallel convolution group consisting of 3×3, 5×5, and 7×7 convolution kernels is used to extract local detail features and global structural features of transmission lines, substation equipment, and distribution equipment, respectively. S5-2. Design of a cross-disciplinary attention module: The channel attention mechanism dynamically adjusts the feature weights of the three professional fields of power transmission, power transformation, and power distribution to achieve adaptive fusion of professional features; S5-3. Implement graded defect identification: a) The primary network layer identifies the defect location coordinates and outputs a two-dimensional positioning heat map; b) The intermediate network layer classifies the defect type and generates a classification probability distribution; c) The high-level network layer assesses the severity of defects and outputs a score of 1-5; S5-4. Generate three-dimensional evaluation results: Fusion of location coordinates, defect type, and severity score into structured data to produce a three-dimensional output containing spatial location, attribute information, and condition assessment; S5-5. Optimize network training strategy: A multi-task joint loss function is used to balance the three indicators of positioning accuracy, classification accuracy and scoring reliability.

[0013] In a second aspect, an embodiment of the present application provides a drone-based intelligent inspection system for integrated transmission, transformation, and distribution based on an image processing algorithm, which is applied to the drone-based intelligent inspection method for integrated transmission, transformation, and distribution based on an image processing algorithm as described in the first aspect. The system includes: The intelligent inspection network module is used to build an intelligent inspection network, dividing the inspection area into a grid. The grid unit size is determined based on the equipment distribution density and inspection frequency requirements. An intelligent drone nest is configured in each grid unit. The nest integrates an autonomous charging module, a weather monitoring unit, and a 5G communication module. The communication module establishes a real-time data connection between the nest and the central control platform, forming an intelligent inspection network covering the entire area. A multi-disciplinary collaborative inspection scheduling module is used by the central control platform to receive and integrate inspection requirements from the power transmission, transformation, and distribution sectors. It uses an intelligent scheduling algorithm based on equipment status assessment to dynamically calculate task priorities. It also optimizes and generates a cross-disciplinary joint inspection task sequence based on the drone's endurance and grid coverage. The multi-dimensional feature fusion processing module is used to collect images of transmission and distribution equipment through the industrial camera carried by the drone, and synchronously perform multi-dimensional image feature fusion on the collected images to generate an optimized fused image; An improved image segmentation processing module is used to segment the optimized fused image based on the MFB-Otsu algorithm, and quickly locate the optimal segmentation threshold through Gaussian process modeling and expected improvement function; The multi-scale defect recognition module is used to construct a deep convolutional neural network containing a multi-scale convolutional kernel group. Through a cross-disciplinary attention mechanism, it optimizes feature extraction strategies for transmission lines, substation equipment, and distribution equipment. The deep convolutional neural network performs intelligent multi-scale defect recognition on the segmented images and outputs a three-dimensional assessment result including defect type, location coordinates, and severity score. The adaptive optimization processing module is used to perform optimization adjustments based on the recognition results. It dynamically adjusts the feature fusion weight coefficient in S2 according to the equipment type; optimizes the segmentation parameters in S3 based on real-time meteorological data; and updates the defect recognition model parameters in S4 through an online learning mechanism for adaptive closed-loop optimization. Among them, steps S3-S6 are implemented through a two-level architecture deployed on the drone edge computing node and the central control platform. The drone node is responsible for the real-time processing of S3-S5, and the central platform is responsible for the model optimization and result verification of S6, forming a complete intelligent inspection closed-loop system.

[0014] In a third aspect, an embodiment of the present application provides an electronic device, including: processor; a memory for storing processor-executable instructions; Among them, the processor is configured to implement the drone integrated transmission, transformation and distribution intelligent inspection method based on image processing algorithm as described in the first aspect when executing the instructions.

[0015] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium storing a program, wherein the program instructs a device to execute the drone-based intelligent inspection method for integrated transmission, transformation and distribution based on an image processing algorithm as described in the first aspect.

[0016] The present invention discloses a method and system for integrated intelligent inspection of transmission, transformation and distribution using drones based on image processing algorithms. The method includes: constructing a grid-based intelligent inspection network, configuring drone nests with integrated charging, weather monitoring and 5G communication; a central control platform integrating transmission, transformation and distribution inspection requirements, and dynamically optimizing task sequences; drones collecting equipment images, generating optimized images through multi-dimensional feature fusion; using an improved MFB-Otsu algorithm combined with Gaussian process modeling to locate the optimal segmentation threshold; utilizing a multi-scale deep convolutional neural network and a cross-disciplinary attention mechanism to achieve intelligent defect recognition, and outputting a three-dimensional assessment result including type, location and severity; and achieving real-time edge processing and closed-loop optimization of the central platform through a two-level architecture. The present invention realizes efficient collaborative inspection of transmission, transformation and distribution equipment, improves the accuracy and efficiency of defect identification, and has the characteristics of adaptive optimization and cross-disciplinary integration.

[0017] Beneficial effects: 1. Multi-disciplinary collaborative and efficient inspection: By dividing the inspection area into grids and configuring intelligent drone nests, combined with the dynamic task scheduling algorithm of the central control platform, cross-disciplinary joint inspection of transmission, transformation, and distribution equipment can be achieved, significantly improving resource utilization and operation and maintenance efficiency.

[0018] 2. High-precision image processing and defect recognition. Multi-dimensional feature fusion technology (grayscale, edge, texture) is used to optimize image quality. Combined with the improved MFB-Otsu algorithm and Gaussian process modeling, it quickly locates the optimal segmentation threshold and improves the accuracy of target extraction in complex environments.

[0019] 3. Real-time processing and intelligent optimization: Through a two-level architecture of drone edge computing nodes and a central platform, real-time processing of image acquisition, segmentation, and recognition is achieved, reducing data transmission latency.

[0020] 4. Enhanced adaptability and robustness. By combining real-time meteorological data and equipment status feedback, the system dynamically optimizes inspection task sequences and image processing parameters to ensure system stability and reliability in changing environments.

[0021] 5. Reduced operation and maintenance costs. Automated inspections and intelligent analysis reduce manual intervention, shorten defect detection cycles, and reduce repeated inspection costs through cross-disciplinary task integration, providing efficient and economical technical support for power system operation and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1A flow chart of a method for integrated intelligent inspection of transmission, transformation and distribution using drones based on image processing algorithms is provided in one embodiment of the present application.

[0023] Figure 2 This is a structural diagram of the improved image segmentation processing module provided in one embodiment of the present application.

[0024] Figure 3 This is an architecture diagram of an integrated intelligent inspection system for transmission, transformation and distribution using drones based on image processing algorithms, provided in one embodiment of the present application.

[0025] Figure 4 A schematic diagram of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments.

[0027] It should be noted that, in the embodiments of the present application, "at least one" refers to one or more, and "more" refers to two or more. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art in the art to which this application relates. The terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application.

[0028] It should be noted that, in the embodiments of the present application, words such as "first" and "second" are only used for the purpose of distinguishing descriptions, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying an order. Features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or design schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a concrete way.

[0029] Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0030] Example 1

[0031] The core of this invention is to build an intelligent inspection network, which realizes intelligent inspection and defect identification of transmission and distribution equipment through the collaborative work of drones, edge computing nodes and a central control platform.

[0032] Figure 1 This is a flow chart of a method for intelligent inspection of transmission, transformation and distribution integration using drones based on image processing algorithms according to an embodiment of the present application. Figure 1 As shown, the specific steps include: S1. Build an intelligent inspection network, divide the inspection area into grids, and determine the grid unit size based on equipment distribution density and inspection frequency requirements; configure an intelligent drone nest in each grid unit, which integrates an autonomous charging module, a weather monitoring unit, and a 5G communication module; establish a real-time data connection between the nest and the central control platform through the communication module, forming an intelligent inspection network covering the entire area.

[0033] Building an intelligent inspection network is a core component of the entire intelligent inspection system. By dividing the inspection area into a grid and determining the grid unit size based on equipment density and inspection frequency requirements, refined management of the inspection area is achieved. Within each grid unit, intelligent drone nests are deployed. These nests integrate autonomous charging modules, weather monitoring units, and 5G communication modules. These modules establish real-time data connections with the central control platform, forming an intelligent inspection network covering the entire area.

[0034] Among them, dynamic grid partitioning technology can be used to develop a dynamic partitioning algorithm based on the Voronoi diagram. The grid shape can be automatically adjusted according to the device distribution density. A grid parameter optimization model is established: the objective function is min (inspection blind area) + max (resource balance), and the constraints are: device density gradient, terrain complexity, and airspace restrictions. An improved K-means clustering algorithm is used to divide the area based on the device criticality as the weight. At the same time, a grid self-healing mechanism is designed to dynamically adjust the grid boundaries when the number of devices increases or decreases.

[0035] In addition, according to the intelligent machine nest configuration plan, the modular machine nest design includes: energy modules, such as fast charging piles + battery swap compartments, which support 15-minute ultra-fast energy replenishment; perception modules, such as six-element weather stations + lidar obstacle avoidance; communication modules, such as 5G private networks + LoRa dual-channel redundant transmission.

[0036] Adaptive network topology management is used to build a hierarchical networking architecture. For example, it is divided into the following functional categories: backbone layer (fixed machine nest nodes); mobile layer (inspection drone nodes); and emergency layer (mobile supply nodes).

[0037] It also includes using a real-time data fusion system to build a unified data middle platform to integrate: equipment inventory data, real-time telemetry data, meteorological environment data, and drone status data; developing an edge-cloud collaborative processing framework, such as edge nodes: responsible for data preprocessing and real-time alarms; central platform: responsible for in-depth analysis and decision support.

[0038] This intelligent inspection network breaks through the limitations of traditional fixed deployment. Through dynamic grid division and intelligent machine nest configuration, it achieves precise coverage and efficient operation and maintenance of complex power equipment, providing a solid network foundation for integrated inspection.

[0039] Specifically, in this embodiment, S2. The central control platform receives and integrates the inspection requirements of power transmission, transformation, and distribution; uses an intelligent scheduling algorithm based on equipment status assessment to dynamically calculate task priorities; and optimizes and generates a cross-disciplinary joint inspection task sequence based on the drone's endurance and grid coverage.

[0040] Specifically, in this embodiment, step S2 specifically includes: S2-1. Receive inspection requests from transmission, transformation, and distribution professionals and establish a multidimensional inspection demand matrix based on equipment type, defect level, and operational data. Specifically, using multidimensional demand integration technology, this matrix integrates the differentiated requirements of each of the transmission, transformation, and distribution professionals. Matrix dimensions include: equipment type (conductor / insulator / transformer, etc.), defect level (emergency / major / general), and operational data (temperature / load / historical failure rate). Tensor decomposition technology is used to extract the associated features of each professional's requirements, enabling effective integration of these requirements.

[0041] S2-2. Utilize an intelligent scheduling algorithm based on equipment status assessment, combining real-time equipment health, historical failure probability, and O&M urgency to dynamically calculate the comprehensive priority weight of each inspection task. Specifically, a dynamic priority assessment model can be established, designing a three-tiered evaluation system: a) Foundational layer: real-time equipment health (based on sensor data); b) Historical layer: equipment failure probability (based on O&M big data analysis); and c) Strategic layer: O&M urgency (based on grid operating conditions). Simultaneously, a fuzzy comprehensive evaluation algorithm can be developed, quantifying the influence weight of each factor through a membership function. A time decay factor can also be introduced to dynamically adjust the contribution of historical data.

[0042] S2-3. Based on the real-time endurance of the UAV, the distribution of charging stations, and the coverage of the grid partition, a task allocation optimization model is constructed to minimize the total inspection time and maximize resource utilization, generating the optimal cross-disciplinary joint inspection task sequence.

[0043] Specifically, a multi-objective optimization model was established using an intelligent optimization scheduling engine. The objective function was min (total inspection time) + max (resource utilization). Constraints included drone endurance, charging station distribution, and weather restrictions. An improved NSGA-II algorithm was used to find the Pareto optimal solution set. An adaptive mutation operator was designed to enhance the algorithm's convergence in complex scenarios.

[0044] Leveraging a real-time dynamic adjustment mechanism, a digital twin-based simulation verification module was developed to pre-evaluate the feasibility of the mission sequence. Multiple feedback channels were established: a) real-time monitoring of drone status; b) early warning of sudden defects; and c) early warning of weather changes. A rolling horizon optimization strategy was employed to replan the mission sequence every 15 minutes.

[0045] This intelligent scheduling method effectively solves problems such as uneven resource allocation, delayed response, and professional barriers in traditional inspections, and provides core support for integrated transmission, transformation, and distribution inspections.

[0046] S3. Using an industrial camera mounted on a drone to capture images of the transmission and distribution equipment, multi-dimensional image feature fusion is performed on the captured images to generate an optimized fused image. Specifically, in this embodiment, the image processing steps include grayscale histogram equalization, Sobel operator, LBP texture feature extraction, and feature fusion based on specific weight coefficients.

[0047] Specifically, the implementation steps include: Grayscale histogram equalization (GHE) is used to extract grayscale value features and maintain the global brightness distribution of the image. GHE is an image enhancement technique used to improve image contrast and brightness distribution. By adjusting the grayscale value distribution of an image, the image's grayscale values ​​become more uniform, thereby enhancing the image's visual quality and the effectiveness of subsequent processing. In this method, GHE calculates the image's grayscale histogram and normalizes it to enhance the image's contrast and brightness distribution.

[0048] The Sobel operator is used to calculate gradient magnitude features to enhance the edge contours of key areas such as equipment insulators and conductor connections. The Sobel operator is an edge detection operator that extracts edge information by calculating the horizontal and vertical gradients of an image. In this method, the Sobel operator is used to calculate gradient magnitude features to enhance the edge contours of key areas such as equipment insulators and conductor connections. The Sobel operator uses two 3×3 convolution kernels to calculate the horizontal and vertical gradients, respectively, to extract edge information from the image.

[0049] Extract LBP texture features to suppress background noise in complex industrial environments. LBP (Local Binary Pattern) is a texture feature extraction method that generates a binary pattern by comparing the grayscale values ​​of each pixel with its neighboring pixels to extract image texture features. In this method, LBP texture features are used to suppress background noise in complex industrial environments.

[0050] The LBP feature generates a binary pattern by comparing the grayscale values ​​of the central pixel with its neighboring pixels, thereby extracting the texture features of the image.

[0051] The three features are weighted and fused according to the weight coefficients of 0.4:0.4:0.2 to generate an optimized fused image. This weighted fusion method can combine the advantages of different features to improve the image processing effect.

[0052] The S4 is to segment the optimized fused image based on the MFB-Otsu algorithm (Multi-Feature Boosted Otsu, i.e., the multi-feature enhanced Otsu algorithm), and quickly locate the optimal segmentation threshold through Gaussian process modeling and expected improvement function.

[0053] Figure 2 This is a structural diagram of an improved image segmentation processing module provided in one embodiment of the present application. Figure 2 As shown, the improved image segmentation processing module includes an input module, a core processing module, a dynamic optimization module, an image segmentation result output module, and a multi-scale defect intelligent recognition module. Specifically, the S4 includes: S4-1. The improved Otsu algorithm with multi-feature fusion is used to construct a multidimensional segmentation evaluation function by combining the grayscale distribution, edge gradient and texture features of the image.

[0054] Specifically, the S4-1, using the improved Otsu algorithm with multi-feature fusion, combines the grayscale distribution, edge gradient and texture features of the image to construct a multi-dimensional segmentation evaluation function, including: S4-1-1. Calculate the grayscale inter-class variance based on the recursive Otsu algorithm: Use the recursive formula to dynamically update the statistical parameters of the foreground and background: , , in, For the threshold When the pixel ratio of the background area is Indicated as the threshold When , the grayscale mean of the background area is Grayscale This step involves recursively applying the Otsu algorithm to dynamically update the statistical parameters of the foreground and background to calculate the inter-class variance. The core idea of ​​the Otsu algorithm is to determine the optimal segmentation threshold by maximizing the inter-class variance.

[0055] S4-1-2, Combine edge gradient features to enhance segmentation boundaries: Calculate the image gradient magnitude using the Sobel operator , and construct the edge energy term: , in is the threshold The set of segmented foreground and background pixels, Represents the gradient magnitude of the pixel in the image, Represents the edge energy term, which is used to measure the sum of the edge strengths at the segmentation boundary. This step calculates the image gradient magnitude using the Sobel operator and constructs the edge energy term to enhance the segmentation boundary.

[0056] S4-1-3. Introducing texture features to optimize segmentation robustness: Calculate texture contrast features based on gray-level co-occurrence matrix GLCM: , in, is the threshold Next, the gray level in the gray level co-occurrence matrix GLCM and The joint probability of ; This step calculates the texture contrast feature through the gray level co-occurrence matrix (GLCM) to enhance the robustness of segmentation.

[0057] S4-1-4. Construct a multidimensional evaluation function and optimize the threshold: Search range in grayscale In the internal structure, the three types of features are integrated to solve the optimal threshold : , in, Represents the inter-class variance of traditional Otsu, which measures the separation between foreground and background, and the weight coefficient The contribution weights corresponding to grayscale, edge, and texture features respectively satisfy , and dynamically adjusted according to the image signal-to-noise ratio, Represents the average gray value of the entire image, which is used to limit the threshold search range. represents the optimal threshold, and the best segmentation threshold is determined by maximizing the fusion evaluation function. contrast(T) represents the texture contrast feature calculated based on the gray-level co-occurrence matrix under the given segmentation threshold T. This step constructs a multidimensional evaluation function by fusing grayscale, edge, and texture features, and determines the optimal segmentation result by optimizing the threshold.

[0058] The S4-1 method combines the Otsu algorithm, edge detection, texture features and multi-feature fusion to construct a multidimensional segmentation evaluation function to optimize the image segmentation effect.

[0059] S4-2. Probabilistic fitting of the segmentation threshold search space is performed through Gaussian process modeling, and the expected improvement function EI is used to dynamically evaluate the segmentation performance of the candidate threshold.

[0060] Specifically, the S4-2, performing probabilistic fitting on the segmentation threshold search space through Gaussian process modeling, and dynamically evaluating the segmentation performance of the candidate threshold using the expected improvement function EI, includes: S4-2-1. Establish a Gaussian process proxy model for segmentation performance: The threshold T is used as the input variable and the corresponding multidimensional evaluation function value F(T) is used as the output to construct a Gaussian process regression model: , in, is the multidimensional evaluation function value, indicating the threshold The corresponding segmentation performance, is the mean function of the Gaussian process, which represents the The initial prediction of is the covariance kernel function based on the radial basis function RBF, which measures the threshold and The similarities between Represents a Gaussian process model, which is used to fit the nonlinear relationship between threshold and segmentation performance. Represents the Gaussian process function in the Gaussian process model; constructs a Gaussian process regression model to fit the nonlinear relationship between threshold and segmentation performance. Gaussian process is a nonlinear modeling method widely used in machine learning and optimization problems.

[0061] S4-2-2. Design an adaptive sampling strategy. This step involves an adaptive sampling strategy, including initial sampling and updating Gaussian process hyperparameters. The specific steps are as follows: a) Uniformly sample N candidate thresholds in the initial stage , calculate its evaluation function value ; b) Update the Gaussian process hyperparameters based on the sampled data to obtain the posterior probability distribution.

[0062] S4-2-3. Using the expected improvement function Guide threshold. Use the expected improvement function (EI) to guide the selection of candidate thresholds. The expected improvement function is a collection function used to select the next sampling point in Bayesian optimization. The specific steps are as follows: Define the improvement function , indicating the current threshold Relative to the known best of which is the current optimal evaluation function value, Calculating expected improvement : , in, Used to select the most promising candidate threshold, and Forecast the mean and standard deviation for a Gaussian process, To balance the parameters and control the weight of exploration and utilization, They are the cumulative distribution function CDF and probability density function PDF of the standard normal distribution respectively; S4-2-4, iterative optimization until convergence. By continuously updating the Gaussian process model and selecting the optimal threshold until convergence. The specific steps are as follows: a) Select Maximum candidate threshold Conduct a practical assessment; b) Add the training set to update the Gaussian process model; c) Repeat the above process until Less than the set threshold or the maximum number of iterations is reached.

[0063] The S4-2 method combines Gaussian process modeling, adaptive sampling strategy, expected improvement function and iterative optimization to construct an efficient threshold search method.

[0064] S4-3. Based on the Bayesian optimization strategy, the Gaussian process model is iteratively updated to quickly converge to the global optimal segmentation threshold, achieving high-precision extraction of inspection targets.

[0065] Specifically, the S4-3, iteratively updating the Gaussian process model based on the Bayesian optimization strategy, quickly converging to the global optimal segmentation threshold, and achieving high-precision extraction of inspection targets, includes: S4-3-1. Build an adaptive optimization framework: Establish a Bayesian optimization framework with the image segmentation evaluation index as the objective function and the threshold as the optimization variable, and establish a mapping relationship between the threshold and segmentation performance through a Gaussian process model. This step involves establishing a Bayesian optimization framework with the image segmentation evaluation index as the objective function and the threshold as the optimization variable, and establishing a mapping relationship between the threshold and segmentation performance through a Gaussian process model.

[0066] S4-3-2. Implement intelligent iterative optimization: a) Generate a set of candidate thresholds based on the current model predictions; b) Select the most promising evaluation points using an acquisition function that balances exploration and exploitation; c) Dynamically adjust the search strategy, focusing on global exploration in the early stages and shifting to local, refined search later. This step includes generating candidate thresholds, selecting the most promising evaluation points, and dynamically adjusting the search strategy.

[0067] S4-3-3. Dynamic Model Update Mechanism: After each evaluation of a new threshold, the parameters and hyperparameters of the Gaussian process model are updated in real time to gradually improve the model's fit to the objective function. This step involves updating the parameters and hyperparameters of the Gaussian process model in real time to improve the model's fit to the objective function.

[0068] S4-3-4, Set intelligent termination conditions: When the stability of the optimal solution and sufficient convergence of the search interval are met at the same time, the optimization process is automatically terminated. This step involves setting the termination conditions for the stability of the optimal solution and the convergence of the search interval.

[0069] S4-3-5. Output the optimal segmentation solution: Output the optimal segmentation threshold determined by Bayesian optimization to achieve accurate extraction of inspection targets. This step involves outputting the optimal segmentation threshold determined by Bayesian optimization to achieve accurate extraction of inspection targets.

[0070] S5. Construct a deep convolutional neural network containing a multi-scale convolutional kernel group, and optimize the feature extraction strategy for transmission lines, substation equipment, and distribution equipment respectively through a cross-disciplinary attention mechanism; perform multi-scale defect intelligent identification on the segmented images through the deep convolutional neural network, and output a three-dimensional assessment result including the defect type, location coordinates, and severity score.

[0071] Specifically, in this embodiment, S5 specifically includes: S5-1. Construct a multi-scale feature extraction network: Parallel convolution groups containing 3×3, 5×5, and 7×7 kernels are used to extract local detail features and global structural features of transmission lines, substation equipment, and distribution equipment, respectively. Multi-scale feature fusion plays an important role in image recognition and object detection, improving the model's feature representation capabilities.

[0072] S5-2. Design of cross-disciplinary attention modules: The channel attention mechanism dynamically adjusts the feature weights of the three professional fields of power transmission, transformation, and distribution to achieve adaptive fusion of professional features. The channel attention mechanism dynamically adjusts the feature weights of different professional fields to achieve adaptive fusion of features.

[0073] S5-3. Implement hierarchical defect identification to identify defect location, type and severity: a) The primary network layer identifies the defect location coordinates and outputs a two-dimensional positioning heat map; b) The intermediate network layer classifies the defect type and generates a classification probability distribution; c) The high-level network layer assesses the severity of the defect and outputs a score of 1-5.

[0074] S5-4. Generate 3D assessment results. Simultaneously, the positioning, classification, and scoring results are integrated into structured data. The positioning coordinates, defect type, and severity score are integrated into structured data to form a 3D output containing spatial location, attribute information, and status assessment. S5-5. Optimize network training strategies. Use a multi-task joint loss function to balance positioning, classification, and scoring metrics. Specifically, use a multi-task joint loss function to balance positioning accuracy, classification accuracy, and scoring reliability.

[0075] S6 optimizes and adjusts the recognition results, dynamically adjusting the feature fusion weight coefficients in S2 based on the device type; optimizing the segmentation parameters in S3 using real-time meteorological data; and updating the defect recognition model parameters in S4 through an online learning mechanism for adaptive closed-loop optimization. This combination of adaptive feature fusion, online learning, real-time data optimization, and closed-loop optimization demonstrates high technical feasibility and promising application prospects.

[0076] For example, dynamic feature fusion optimization driven by device type can build a device feature knowledge base and establish a multi-dimensional device feature matrix, including structural features (such as conductor diameter and number of insulators), material features (such as metal, ceramic, and composite materials), and failure mode features (such as cracks, corrosion, and looseness). Graph neural network-based device association analysis can also be developed. An adaptive weight adjustment engine and real-time weight adjustment algorithm can also be used, with a frequency of, for example, 1,000 dynamic weight adjustments per second.

[0077] For image segmentation enhancement of weather perception, a multimodal weather impact model can be established, a weather-image quality mapping matrix can be established, and a real-time parameter optimization pipeline can be used. The matrix structure is shown in Table 1 below: Table 1 Weather perception parameter mapping and compensation strategy

[0078] This solution builds an intelligent inspection system with autonomous evolution capabilities by deeply integrating equipment characteristics, environmental factors and continuous learning mechanisms, achieving breakthrough improvements in adaptability, accuracy and reliability compared to traditional methods.

[0079] Among them, steps S3-S6 are implemented through a two-level architecture deployed on the drone edge computing node and the central control platform. The drone node is responsible for the real-time processing of S3-S5, and the central platform is responsible for the model optimization and result verification of S6, forming a complete intelligent inspection closed-loop system.

[0080] Specifically, in this embodiment, the collaborative work solution can be presented in a function allocation matrix as shown in Table 2: Table 2 Functional coordination of edge-central two-level architecture

[0081] Example 2

[0082] like Figure 3 As shown, the present application provides an architecture diagram of an integrated intelligent inspection system for transmission, transformation and distribution using drones based on image processing algorithms, which is applied to the integrated intelligent inspection system for transmission, transformation and distribution using drones based on image processing algorithms as described in Example 1, including an intelligent inspection network module 11, a multi-professional collaborative inspection scheduling module 12, a multi-dimensional feature fusion processing module 13, an improved image segmentation processing module 14, a multi-scale defect recognition module 15, and an adaptive optimization processing module 16.

[0083] The intelligent inspection network module 11 is used to build an intelligent inspection network, divide the inspection area into grids, and determine the size of the grid units based on the equipment distribution density and inspection frequency requirements; configure an intelligent drone nest in each grid unit, and the nest integrates an autonomous charging module, a meteorological monitoring unit and a 5G communication module; establish a real-time data connection between the nest and the central control platform through the communication module, forming an intelligent inspection network covering the entire area.

[0084] The multi-disciplinary collaborative inspection scheduling module 12 is used for the central control platform to receive and integrate the inspection requirements of the transmission, transformation, and distribution disciplines; adopts an intelligent scheduling algorithm based on equipment status assessment to dynamically calculate task priorities; and optimizes and generates a cross-disciplinary joint inspection task sequence based on the drone's endurance and grid coverage.

[0085] The multi-dimensional feature fusion processing module 13 is used to collect images of transmission and distribution equipment through the industrial camera carried by the drone, and synchronously perform multi-dimensional image feature fusion on the collected images to generate an optimized fused image.

[0086] The improved image segmentation processing module 14 is used to segment the optimized fused image based on the MFB-Otsu algorithm, and quickly locate the optimal segmentation threshold through Gaussian process modeling and expected improvement function.

[0087] The multi-scale defect recognition module 15 is used to construct a deep convolutional neural network containing a multi-scale convolutional kernel group, and optimize the feature extraction strategy for transmission lines, substation equipment, and distribution equipment respectively through a cross-disciplinary attention mechanism; the segmented image is subjected to multi-scale defect intelligent recognition through the deep convolutional neural network, and a three-dimensional assessment result including defect type, location coordinates, and severity score is output.

[0088] The adaptive optimization processing module 16 is used to perform optimization adjustments based on the recognition results, dynamically adjust the feature fusion weight coefficient in S2 according to the equipment type; optimize the segmentation parameters in S3 in combination with real-time meteorological data; and update the defect recognition model parameters in S4 through an online learning mechanism to perform adaptive closed-loop optimization.

[0089] Among them, steps S3-S6 are implemented through a two-level architecture deployed on the drone edge computing node and the central control platform. The drone node is responsible for the real-time processing of S3-S5, and the central platform is responsible for the model optimization and result verification of S6, forming a complete intelligent inspection closed-loop system.

[0090] Figure 4 This is an electronic device provided by an embodiment of the present application. Figure 4 As shown, the electronic device includes at least the following parts: a processor 101 and a memory 100 , a communication interface 103 , and a bus 102 .

[0091] In an embodiment of the present application, the memory 100 is used to store instructions executable by the processor 101, and the processor 101 is configured to implement the method of the first aspect when executing the instructions.

[0092] In an embodiment of the present application, a computer-readable storage medium includes instructions, and the instructions instruct a device to execute the method of the first aspect. For example, the instructions instruct the device to execute Figure 1 The method is shown in the process steps.

[0093] The program running in the electronic device involved in one embodiment of the present application may be a program that controls a central processing unit (CPU) and the like to implement the functions of the above-mentioned embodiment involved in one embodiment of the present invention (a program that causes a computer to function). The information processed by these devices is temporarily stored in random access memory (RAM) while being processed, and is then stored in various ROMs such as read-only memory (Flash ROM) and a hard disk drive (HDD), where it is read, modified, and written as needed by the CPU.

[0094] It should be noted that a portion of the electronic device of the above embodiment may also be implemented by a computer. In this case, a program for implementing the control function may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read into a computer and executed.

[0095] It should be noted that the "computer" mentioned here refers to a computer built into an electronic device, employing hardware including an operating system (OS) and peripheral devices. Furthermore, "computer-readable recording medium" refers to removable media such as floppy disks, magneto-optical disks, ROMs, and CD-ROMs, as well as storage devices such as hard disks built into computers.

[0096] Furthermore, "computer-readable recording media" may include: media that dynamically store programs for a short period of time, such as communication lines when transmitting programs via networks such as the Internet or communication lines such as telephone lines; and media that store programs for a fixed period of time, such as volatile memory within computers acting as servers or clients in this context. Furthermore, the aforementioned program may be a program for implementing a portion of the aforementioned functions, or a program that can achieve the aforementioned functions by combining with a program already stored in a computer.

[0097] Furthermore, the electronic device in the above-described embodiments can also be implemented as a collection of multiple devices (a device group). Each device comprising the device group may include some or all of the functions or functional blocks of the electronic device in the above-described embodiments. A device group only needs to include all of the functions or functional blocks of the electronic device.

[0098] Those skilled in the art should recognize that the above embodiments are merely intended to illustrate the present application and are not intended to limit the present application. As long as they are within the spirit of the present application, appropriate changes and modifications to the above embodiments are within the scope of protection claimed in the present application.

Claims

1. An intelligent inspection method for integrated transmission, transformation and distribution using drones based on image processing algorithms, characterized in that: The method comprises: S1. Build an intelligent inspection network, divide the inspection area into grids, and determine the grid unit size based on equipment distribution density and inspection frequency requirements. Deploy an intelligent drone nest within each grid unit, integrating an autonomous charging module, a weather monitoring unit, and a 5G communication module. Use the communication module to establish a real-time data connection between the nest and the central control platform, forming an intelligent inspection network covering the entire area. S2. The central control platform receives and integrates inspection requirements from various disciplines, including power transmission, transformation, and distribution. It uses an intelligent scheduling algorithm based on equipment status assessment to dynamically calculate task priorities. It also optimizes and generates a cross-disciplinary joint inspection task sequence based on the drone's endurance and grid coverage. S3. Use the industrial camera carried by the drone to collect images of the transmission and distribution equipment, and simultaneously perform multi-dimensional image feature fusion on the collected images to generate an optimized fused image; S4. Segmenting the optimized fused image based on the MFB-Otsu algorithm, and quickly locating the optimal segmentation threshold through Gaussian process modeling and expected improvement function; S5. Construct a deep convolutional neural network containing a multi-scale convolutional kernel group, and optimize the feature extraction strategy for transmission lines, substation equipment, and distribution equipment respectively through a cross-disciplinary attention mechanism; perform multi-scale defect intelligent identification on the segmented images through the deep convolutional neural network, and output a three-dimensional assessment result including defect type, location coordinates, and severity score; S6: Perform optimization and adjustment based on the recognition results, dynamically adjust the feature fusion weight coefficient in S2 according to the equipment type; optimize the segmentation parameters in S3 based on real-time meteorological data; update the defect recognition model parameters in S4 through the online learning mechanism to perform adaptive closed-loop optimization; Among them, steps S3-S6 are implemented through a two-level architecture deployed on the drone edge computing node and the central control platform. The drone node is responsible for the real-time processing of S3-S5, and the central platform is responsible for the model optimization and result verification of S6, forming a complete intelligent inspection closed-loop system.

2. The method for integrated intelligent inspection of transmission, transformation and distribution by drones based on image processing algorithm according to claim 1 is characterized in that: The step S2 specifically includes: S2-1. Receive inspection requirements from power transmission, transformation, and distribution professionals, and establish a multi-dimensional inspection requirement matrix based on equipment type, defect level, and operating data; S2-2. Adopt an intelligent scheduling algorithm based on equipment status assessment, combining the real-time health of the equipment, historical failure probability, and operation and maintenance urgency to dynamically calculate the comprehensive priority weight of each inspection task; S2-3. Based on the real-time endurance of the UAV, the distribution of charging stations, and the coverage of the grid partition, a task allocation optimization model is constructed to minimize the total inspection time and maximize resource utilization, generating the optimal cross-disciplinary joint inspection task sequence.

3. The method for integrated intelligent inspection of transmission, transformation and distribution by drones based on image processing algorithm according to claim 1 is characterized in that: In step S3, multi-dimensional image feature fusion is synchronously performed on the collected image, including: Grayscale histogram equalization is used to extract grayscale value features and maintain the global brightness distribution of the image; Apply the Sobel operator to calculate the gradient amplitude characteristics and enhance the edge contours of key parts of equipment insulators and conductor connections; Extract LBP texture features to suppress background noise in complex industrial environments; The three features are weightedly fused according to the weight coefficient of 0.4:0.4:0.2 to generate the optimized fused image.

4. The method for integrated intelligent inspection of transmission, transformation and distribution using drones based on image processing algorithms according to claim 2 is characterized in that: S4, segmenting the optimized fused image based on the MFB-Otsu algorithm, and quickly locating the optimal segmentation threshold through Gaussian process modeling and expected improvement function, including: S4-1. Adopt the improved Otsu algorithm with multi-feature fusion, and combine the grayscale distribution, edge gradient and texture features of the image to construct a multi-dimensional segmentation evaluation function; S4-2, probabilistically fitting the segmentation threshold search space through Gaussian process modeling, and dynamically evaluating the segmentation performance of candidate thresholds using the expected improvement function (EI); S4-3. Based on the Bayesian optimization strategy, the Gaussian process model is iteratively updated to quickly converge to the global optimal segmentation threshold, achieving high-precision extraction of inspection targets.

5. The method for integrated intelligent inspection of transmission, transformation and distribution using drones based on image processing algorithms according to claim 4 is characterized in that: S4-1, using the improved Otsu algorithm with multi-feature fusion, combining the grayscale distribution, edge gradient and texture features of the image to construct a multi-dimensional segmentation evaluation function, including: S4-1-1. Calculate the grayscale inter-class variance based on the recursive Otsu algorithm: Use the recursive formula to dynamically update the statistical parameters of the foreground and background: ， ， in, For the threshold When the pixel ratio of the background area is Indicated as the threshold When , the grayscale mean of the background area is Grayscale The probability of the pixel appearing; S4-1-2, Combine edge gradient features to enhance segmentation boundaries: Calculate the image gradient magnitude using the Sobel operator , and construct the edge energy term: ， in is the threshold The set of segmented foreground and background pixels, Represents the gradient magnitude of the pixel in the image, represents the edge energy term, which is used to measure the sum of the edge strength at the segmentation boundary; S4-1-3. Introducing texture features to optimize segmentation robustness: Calculate texture contrast features based on gray-level co-occurrence matrix GLCM: ， in, is the threshold Next, the gray level in the gray level co-occurrence matrix GLCM and The joint probability of S4-1-4. Construct a multidimensional evaluation function and optimize the threshold: Search range in grayscale In the internal structure, the three types of features are integrated to solve the optimal threshold : ， in, Represents the inter-class variance of traditional Otsu, which measures the separation between foreground and background, and the weight coefficient The contribution weights corresponding to grayscale, edge, and texture features respectively satisfy , and dynamically adjusted according to the image signal-to-noise ratio, Represents the average gray value of the entire image, which is used to limit the threshold search range. Represents the optimal threshold, and the optimal segmentation threshold is determined by maximizing the fusion evaluation function. Contrast(T) represents the texture contrast feature calculated based on the gray-level co-occurrence matrix under a given segmentation threshold T.

6. The method for integrated intelligent inspection of transmission, transformation and distribution using drones based on image processing algorithms according to claim 5 is characterized in that: S4-2, performing probabilistic fitting on the segmentation threshold search space through Gaussian process modeling, and dynamically evaluating the segmentation performance of the candidate threshold using the expected improvement function EI, includes: S4-2-1. Establish a Gaussian process proxy model for segmentation performance: The threshold T is used as the input variable and the corresponding multidimensional evaluation function value F(T) is used as the output to construct a Gaussian process regression model: ， in, is the multidimensional evaluation function value, indicating the threshold The corresponding segmentation performance, is the mean function of the Gaussian process, which represents the The initial prediction of is the covariance kernel function based on the radial basis function RBF, which measures the threshold and The similarities between Represents a Gaussian process model, which is used to fit the nonlinear relationship between threshold and segmentation performance. Represents the Gaussian process function in the Gaussian process model; S4-2-2. Design an adaptive sampling strategy: a) Uniformly sample N candidate thresholds in the initial stage , calculate its evaluation function value ; b) Update the Gaussian process hyperparameters based on the sampled data to obtain the posterior probability distribution; S4-2-3. Using the expected improvement function Guidance thresholds: Define the improvement function , indicating the current threshold Relative to the known best of which is the current optimal evaluation function value, Calculating expected improvement : ， in, Used to select the most promising candidate threshold, and Forecast the mean and standard deviation for a Gaussian process, To balance the parameters and control the weight of exploration and utilization, They are the cumulative distribution function CDF and probability density function PDF of the standard normal distribution respectively; S4-2-4, iterative optimization until convergence: a) Select Maximum candidate threshold Conduct practical assessments; b) Add the training set to update the Gaussian process model; c) Repeat the above process until Less than the set threshold or the maximum number of iterations is reached.

7. The method for integrated intelligent inspection of transmission, transformation and distribution using drones based on image processing algorithms according to claim 6 is characterized in that: S4-3, iteratively updating the Gaussian process model based on the Bayesian optimization strategy, quickly converging to the global optimal segmentation threshold, and achieving high-precision extraction of inspection targets, includes: S4-3-1. Build an adaptive optimization framework: Establish a Bayesian optimization framework with image segmentation evaluation index as the objective function and threshold as the optimization variable, and establish the mapping relationship between threshold and segmentation performance through the Gaussian process model; S4-3-2. Implement intelligent iterative optimization: a) Generate a set of candidate thresholds based on the current model predictions; b) Select the most promising evaluation points using an acquisition function that balances exploration and exploitation; c) Dynamically adjust the search strategy, focusing on global exploration in the early stages and shifting towards local, refined search in the later stages; S4-3-3. Dynamic model update mechanism: After each evaluation of a new threshold, the parameters and hyperparameters of the Gaussian process model are updated in real time to gradually improve the model's fitting accuracy to the objective function; S4-3-4. Set intelligent termination conditions: When the stability of the optimal solution and sufficient convergence of the search interval are met at the same time, the optimization process will be automatically terminated; S4-3-5. Output the optimal segmentation solution: Output the optimal segmentation threshold determined by Bayesian optimization to achieve accurate extraction of inspection targets.

8. The method for integrated intelligent inspection of transmission, transformation and distribution using drones based on image processing algorithms according to claim 5 is characterized in that: The S5 specifically includes: S5-1. Construct a multi-scale feature extraction network: A parallel convolution group consisting of 3×3, 5×5, and 7×7 convolution kernels is used to extract local detail features and global structural features of transmission lines, substation equipment, and distribution equipment, respectively. S5-2. Design of cross-disciplinary attention modules: The channel attention mechanism dynamically adjusts the feature weights of the three professional fields of power transmission, power transformation, and power distribution to achieve adaptive fusion of professional features; S5-3. Implement graded defect identification: a) The primary network layer identifies the defect location coordinates and outputs a two-dimensional positioning heat map; b) The intermediate network layer classifies the defect type and generates a classification probability distribution; c) The high-level network layer assesses the severity of defects and outputs a score of 1-5; S5-4. Generate three-dimensional evaluation results: Fusion of location coordinates, defect type, and severity score into structured data to produce a three-dimensional output containing spatial location, attribute information, and condition assessment; S5-5. Optimize network training strategy: A multi-task joint loss function is used to balance the three indicators of positioning accuracy, classification accuracy and scoring reliability.

9. An unmanned aerial vehicle (UAV) intelligent inspection system for transmission, transformation and distribution integration based on an image processing algorithm, applied to the unmanned aerial vehicle (UAV) intelligent inspection method for transmission, transformation and distribution integration based on an image processing algorithm as claimed in any one of claims 1 to 8, characterized in that: The system comprises: The intelligent inspection network module is used to build an intelligent inspection network, dividing the inspection area into a grid. The grid unit size is determined based on the equipment distribution density and inspection frequency requirements. An intelligent drone nest is configured in each grid unit. The nest integrates an autonomous charging module, a weather monitoring unit, and a 5G communication module. The communication module establishes a real-time data connection between the nest and the central control platform, forming an intelligent inspection network covering the entire area. A multi-disciplinary collaborative inspection scheduling module is used by the central control platform to receive and integrate inspection requirements from the power transmission, transformation, and distribution sectors. It uses an intelligent scheduling algorithm based on equipment status assessment to dynamically calculate task priorities. It also optimizes and generates a cross-disciplinary joint inspection task sequence based on the drone's endurance and grid coverage. The multi-dimensional feature fusion processing module is used to collect images of transmission and distribution equipment through the industrial camera carried by the drone, and synchronously perform multi-dimensional image feature fusion on the collected images to generate an optimized fused image; An improved image segmentation processing module is used to segment the optimized fused image based on the MFB-Otsu algorithm, and quickly locate the optimal segmentation threshold through Gaussian process modeling and expected improvement function; The multi-scale defect recognition module is used to construct a deep convolutional neural network containing a multi-scale convolutional kernel group. Through a cross-disciplinary attention mechanism, it optimizes feature extraction strategies for transmission lines, substation equipment, and distribution equipment. The deep convolutional neural network performs intelligent multi-scale defect recognition on the segmented images and outputs a three-dimensional assessment result including defect type, location coordinates, and severity score. The adaptive optimization processing module is used to perform optimization adjustments based on the recognition results. It dynamically adjusts the feature fusion weight coefficient in S2 according to the equipment type; optimizes the segmentation parameters in S3 based on real-time meteorological data; and updates the defect recognition model parameters in S4 through an online learning mechanism for adaptive closed-loop optimization. Among them, steps S3-S6 are implemented through a two-level architecture deployed on the drone edge computing node and the central control platform. The drone node is responsible for the real-time processing of S3-S5, and the central platform is responsible for the model optimization and result verification of S6, forming a complete intelligent inspection closed-loop system.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program, and the program instructs the device to execute the UAV transmission, transformation and distribution integrated intelligent inspection method based on image processing algorithm as described in any one of claims 1 to 8.

Images