Electric power inspection method and system, computing device and storage medium

Abstract

The invention discloses an electric power inspection method and system, computing equipment and a storage medium, and relates to the technical field of electric power equipment inspection, and the method comprises the steps: carrying out the multispectral image collection of electric power equipment through a ULP camera; the collected multispectral image is preprocessed; fusing the preprocessed multispectral images, and analyzing an abnormal area on the surface of the power equipment; performing feature extraction on the abnormal area on the surface of the power equipment; the extracted features are compared with standard parameters in a fault feature library, and defect classification is carried out; and generating an inspection route based on the defect classification result. According to the electric power inspection method provided by the invention, the hardware complexity and the system power consumption are reduced, and the optimal inspection path is generated by using the A * algorithm, so that the inspection task can preferentially cover a high-priority fault point, and the responsiveness of inspection is improved. Finally, full-process automation from image acquisition to polling task generation is realized, and the overall efficiency of the polling task is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment inspection, and in particular to a power equipment inspection method and system. Background Art

[0002] Power equipment inspection is an important part of ensuring the safe operation of the power system. Traditional power inspection methods rely on manual inspection and conventional image acquisition equipment, such as ordinary cameras or single-spectrum cameras. These devices can usually only collect single-spectrum information, and have limited recognition effects on equipment surface faults (such as cracks, corrosion, deformation, etc.), and cannot fully and accurately detect multiple surface anomalies.

[0003] Most existing power inspection systems rely on a variety of sensor devices such as infrared thermal imagers and sound sensors to collect different types of inspection data. However, this multi-sensor combined inspection method has problems such as complex hardware, high power consumption, and high equipment maintenance costs, making it difficult to meet the needs of long-term, large-scale inspections. In addition, in terms of path planning, conventional inspection methods usually generate inspection paths based on the geographical location of the equipment, but lack an effective mechanism for planning paths based on the priority of equipment failures, making it difficult to ensure that inspection tasks respond quickly to high-priority failures.

[0004] Therefore, a power inspection method is needed to reduce the burden on equipment, improve the quality and efficiency of obtaining inspection data, optimize the inspection path according to the severity of the fault, and improve the responsiveness and accuracy of the inspection. Summary of the invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is: the existing power equipment inspection method has the problems of complex hardware, high power consumption, and difficulty in effectively optimizing the inspection path according to the severity of the fault, as well as the optimization problem of how to improve the responsiveness of the inspection task.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: a power inspection method, comprising:

[0008] Use ULP camera to collect multispectral images of power equipment; preprocess the collected multispectral images; fuse the preprocessed multispectral images to analyze abnormal areas on the surface of power equipment; extract features of abnormal areas on the surface of power equipment; compare the extracted features with standard parameters in the fault feature library to classify defects; generate inspection routes based on the defect classification results.

[0009] As a preferred solution of the power inspection method described in the present invention, the multispectral image acquisition includes:

[0010] Use ULP cameras to collect visible light images of the surface of power equipment and record the color and brightness information of the equipment surface;

[0011] Switch the ULP camera to the near-infrared light channel, collect near-infrared images of the same target area, and record the infrared reflection characteristics of the device surface;

[0012] The collected visible light and near-infrared light images are stored after being numbered by region.

[0013] As a preferred solution of the power inspection method described in the present invention, the preprocessing includes performing Gaussian filtering on the stored visible light and near-infrared light images to obtain a noise-reduced image;

[0014] Apply median filtering to the image after Gaussian filtering to remove speckle noise;

[0015] The image processed by median filtering is input into the Canny edge detection algorithm to generate an image containing the edge information of the device surface;

[0016] The image containing edge information is processed by Laplacian operator to obtain an image with enhanced details.

[0017] As a preferred solution of the power inspection method described in the present invention, wherein: the analysis of the abnormal area on the surface of the power equipment includes aligning the visible light image with the near infrared image through a registration algorithm based on image feature points, and combining the aligned images to form a composite image;

[0018] Perform weighted fusion on the visible light and near infrared spectrum data in the composite image to obtain a fused image;

[0019] The fused image is segmented using a threshold segmentation algorithm to extract areas where the brightness and color exceed the set threshold, generating a binary image containing abnormal areas on the surface of the power equipment.

[0020] As a preferred solution of the power inspection method described in the present invention, wherein: the feature extraction includes extracting the RGB color histogram of the abnormal area marked in the binary image and generating a color feature vector;

[0021] The grayscale image of the abnormal area is analyzed using the gray-level co-occurrence matrix to obtain texture features, including contrast and uniformity parameters;

[0022] Calculate the average reflection intensity value of the abnormal area to generate brightness features;

[0023] The color features, texture features and brightness features are integrated into a set of feature parameters of the abnormal area to form a complete feature vector.

[0024] As a preferred solution of the power inspection method described in the present invention, wherein: the defect classification includes inputting the feature vector of the abnormal area into the defect classification model, performing similarity comparison with the standard features in the fault feature library, and calculating the similarity;

[0025] Based on the similarity results, the abnormal area is identified as a specific defect type;

[0026] Use the preset severity criteria, combined with defect type and similarity, to grade the severity of defects;

[0027] Combine equipment number, defect type, severity, and equipment geographic location information to generate a complete failure data record.

[0028] As a preferred solution of the power inspection method of the present invention, wherein: the generation of the inspection route includes taking the fault data records in the fault database as input, prioritizing each fault point according to the severity of the fault, assigning a priority weight, and using the A* algorithm to generate an optimal path covering all fault points;

[0029] The generated inspection paths are arranged in order of priority of the fault points to form an inspection task list, which is then transmitted to the mobile devices of the inspectors.

[0030] Another object of the present invention is to provide a ULP camera-based power inspection system, which can solve the problems of high hardware complexity, high data acquisition cost, and lack of priority mechanism for inspection paths in existing power equipment inspection systems by constructing a ULP camera-based power inspection system.

[0031] To solve the above technical problems, the present invention provides the following technical solutions: an electric power inspection system based on a ULP camera, comprising: an image acquisition module, a data processing module, an anomaly recognition module, a feature extraction module, a defect classification module and an inspection planning module; the image acquisition module is used to use the ULP camera to perform multi-spectral image acquisition on the electric power equipment; the data processing module is used to pre-process the collected multi-spectral image; the anomaly recognition module is used to fuse the pre-processed multi-spectral image and analyze the abnormal area on the surface of the electric power equipment; the feature extraction module is used to extract features from the abnormal area on the surface of the electric power equipment; the defect classification module is used to compare the extracted features with the standard parameters in the fault feature library to classify the defects; the inspection planning module is used to generate an inspection route based on the defect classification results.

[0032] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned power inspection method when executing the computer program.

[0033] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the power inspection method described above.

[0034] Beneficial effects of the present invention: The power inspection method provided by the present invention realizes low-power multi-spectral data acquisition through a ULP camera, covering visible light and near-infrared spectral information, significantly reducing hardware complexity and system power consumption. Combined with multi-spectral image fusion and feature extraction technology, it can efficiently identify abnormal areas on the surface of the equipment, and compare them with the fault feature library to complete defect classification and severity grading. Based on the defect classification results and severity, the A* algorithm is used to generate the optimal inspection path, so that the inspection task can preferentially cover high-priority fault points and improve the responsiveness of the inspection. Finally, the full process automation from image acquisition to inspection task generation is realized, which improves the overall efficiency of the inspection task. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0036] Figure 1 An overall flow chart of a power inspection method provided by an embodiment of the present invention.

[0037] Figure 2 An overall structural diagram of a power inspection system based on a ULP camera provided in accordance with a second embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0039] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0040] Example 1, reference Figure 1 , as an embodiment of the present invention, provides a power inspection method, comprising:

[0041] S1: Multispectral image acquisition of power equipment using ULP camera;

[0042] S2: preprocessing the collected multispectral images;

[0043] S3: Fusing the pre-processed multispectral images to analyze the abnormal areas on the surface of the power equipment;

[0044] S4: Extract features of abnormal areas on the surface of power equipment;

[0045] S5: Compare the extracted features with the standard parameters in the fault feature library to classify the defects;

[0046] S6: Generate an inspection route based on the defect classification results.

[0047] The multispectral image acquisition comprises:

[0048] Use ULP cameras to collect visible light images of the surface of power equipment and record the color and brightness information of the equipment surface;

[0049] Switch the ULP camera to the near-infrared light channel, collect near-infrared images of the same target area, and record the infrared reflection characteristics of the device surface;

[0050] The collected visible light and near-infrared light images are stored after being numbered by region.

[0051] The preprocessing includes performing Gaussian filtering on the stored visible light and near infrared light images to obtain a denoised image;

[0052] Apply median filtering to the image after Gaussian filtering to remove speckle noise;

[0053] The image processed by median filtering is input into the Canny edge detection algorithm to generate an image containing the edge information of the device surface;

[0054] The image containing edge information is processed by Laplacian operator to obtain an image with enhanced details.

[0055] The analyzing of the abnormal area on the surface of the power equipment includes aligning the visible light image with the near infrared image by a registration algorithm based on image feature points, and combining the aligned images to form a composite image;

[0056] Perform weighted fusion on the visible light and near infrared spectrum data in the composite image to obtain a fused image;

[0057] The fused image is segmented using a threshold segmentation algorithm to extract areas where the brightness and color exceed the set threshold, generating a binary image containing abnormal areas on the surface of the power equipment.

[0058] The feature extraction includes extracting the RGB color histogram of the abnormal area marked in the binary image and generating a color feature vector;

[0059] The grayscale image of the abnormal area is analyzed using the gray-level co-occurrence matrix to obtain texture features, including contrast and uniformity parameters;

[0060] Calculate the average reflection intensity value of the abnormal area to generate brightness features;

[0061] The color features, texture features and brightness features are integrated into a set of feature parameters of the abnormal area to form a complete feature vector.

[0062] The defect classification includes inputting the feature vector of the abnormal area into the defect classification model, performing similarity comparison with the standard features in the fault feature library, and calculating the similarity;

[0063] Based on the similarity results, the abnormal area is identified as a specific defect type, such as cracks, corrosion, or deformation;

[0064] Use the preset severity criteria, combined with defect type and similarity, to grade the severity of defects;

[0065] Combine equipment number, defect type, severity, and equipment geographic location information to generate a complete failure data record.

[0066] The generation of the inspection route includes taking the fault data records in the fault database as input, prioritizing each fault point according to the severity of the fault, assigning a priority weight, and using the A* algorithm to generate an optimal path covering all fault points;

[0067] The generated inspection paths are arranged in order of priority of the fault points to form an inspection task list, which is then transmitted to the mobile devices of the inspectors.

[0068] The visible light and near-infrared light images of power equipment are collected using the ULP camera, which are expressed as:

[0069]

[0070] in, Represents multispectral image pixel data, Represents pixels The visible light image matrix of the position, Represents pixels The near-infrared image matrix of the position.

[0071] Data is stored as datasets by region number , providing input for subsequent steps.

[0072] Image preprocessing:

[0073] For multispectral images Perform preprocessing.

[0074] Gaussian filtering, expressed as:

[0075]

[0076] in, represents the image after Gaussian filtering, represents a Gaussian filter with a standard deviation of , Denotes the convolution operation.

[0077] Median filtering, expressed as:

[0078]

[0079] in, represents the image after median filtering, Represents the filter window.

[0080] Edge detection, expressed as:

[0081]

[0082] in, represents the edge detection image, Represents the Canny edge detection operator.

[0083] Detail enhancement, expressed as:

[0084]

[0085] in, represents the image after detail enhancement, represents the Laplace operator, Represents the enhancement factor.

[0086] Multispectral image fusion and abnormal area recognition:

[0087] Using multispectral imagery Perform multi-spectral fusion and identify abnormal areas.

[0088] Image fusion, expressed as:

[0089]

[0090] in, represents the fused image, represents the weight coefficient of visible light and near infrared, Represents visible and near-infrared images after detail enhancement.

[0091] Threshold segmentation is expressed as:

[0092]

[0093] in, represents a binary image, Indicates the brightness threshold.

[0094] Feature extraction: Extract features of abnormal regions .

[0095] Color features, expressed as:

[0096]

[0097] in, Represents an RGB color histogram.

[0098] Texture features, expressed as:

[0099]

[0100] in, represents the gray-level co-occurrence matrix, Indicates contrast, Indicates uniformity.

[0101] Brightness characteristics:

[0102]

[0103] in, represents the average brightness, represents the number of pixels in the abnormal area, and the eigenvector is .

[0104] Defect classification:

[0105] Will Fault signature library Comparison, similarity Expressed as:

[0106] Similarity calculation, expressed as:

[0107]

[0108] in, Indicates similarity, represents the dot product, represents the eigenvector norm.

[0109] According to the severity factor , get the classification , expressed as:

[0110]

[0111] in, Indicates the severity of the defect, Indicates the preset coefficient.

[0112] The generated data record is represented as:

[0113]

[0114] in, Indicates Data records of each fault point, Indicates the fault location.

[0115] Inspection path generation:

[0116] According to all records , depending on the severity and geographic location Generate priority weights .

[0117] Priority weight calculation, expressed as:

[0118]

[0119] in, Indicates The priority weight of each fault point, Indicates the fault point Distance from the starting position.

[0120] use Algorithm, calculate the inspection path. Path cost Expressed as:

[0121]

[0122] in, Represents the total path cost to the current node, Represents the cumulative cost from the starting point to the current node, which is defined as the severity weight of the fault point.

[0123] ,in To estimate the distance to the target node.

[0124] Each fault point is calculated After forming a task list ,according to Sequential sorting, expressed as:

[0125]

[0126] The data is transmitted to the inspectors’ equipment to form an optimized inspection route.

[0127] Example 2, reference Figure 2 , as an embodiment of the present invention, provides a power inspection system based on a ULP camera, comprising:

[0128] Image acquisition module 100, data processing module 200, anomaly recognition module 300, feature extraction module 400, defect classification module 500 and inspection planning module 600;

[0129] The image acquisition module 100 is used to acquire multi-spectral images of power equipment using a ULP camera;

[0130] The data processing module 200 is used to pre-process the collected multispectral images;

[0131] The abnormality identification module 300 is used to fuse the pre-processed multi-spectral images and analyze the abnormal areas on the surface of the power equipment;

[0132] The feature extraction module 400 is used to extract features from abnormal areas on the surface of the power equipment;

[0133] The defect classification module 500 is used to compare the extracted features with the standard parameters in the fault feature library to perform defect classification;

[0134] The inspection planning module 600 is used to generate an inspection route based on the defect classification result.

[0135] Embodiment 3, an embodiment of the present invention, is different from the first two embodiments in that:

[0136] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0137] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0138] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0139] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0140] Example 4 is an embodiment of the present invention, which provides a power inspection method. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0141] By comparing the performance of the method of the present invention with traditional method 1: UAV multi-device inspection system and traditional method 2: ground robot inspection system in terms of power consumption, abnormal area recognition accuracy and inspection path response time in power equipment inspection, the superiority of the power inspection method based on ULP camera of the present invention in terms of low power consumption, multi-spectral data acquisition, abnormal recognition accuracy and inspection responsiveness is verified.

[0142] The experiment was set up on an overhead power line support, where three groups of power equipment were installed. Each group of equipment contained preset surface defect areas (cracks, corrosion, and deformation). Three methods were used to inspect these devices and record various parameters.

[0143] Traditional method 1 uses drones equipped with visible light cameras and infrared thermal imagers to collect visible light images and infrared temperature data on the surface of power equipment, respectively. During the flight of the drone, visible light images are used to detect the color and brightness characteristics of the equipment surface, and the infrared thermal imager is used to identify abnormal temperature areas. During the inspection process, the drone flight needs to be managed, and the data is transmitted to the post-processing software for analysis to identify abnormal areas on the equipment surface. The flight of the drone and the operation of multiple devices result in high overall system power consumption.

[0144] Traditional method 2 uses a ground robot inspection system, which is equipped with a visible light camera, ultrasonic sensor, and lidar system. The visible light camera is used to collect images of the equipment surface, the ultrasonic sensor detects surface defects by measuring echoes, and the lidar is mainly used to generate a three-dimensional map of the inspection area. The system relies on a combination of multiple sensor modules to complete data collection, and the data fusion process is complex. Data comparison is mainly completed through manual intervention or preset software algorithms to determine the location and nature of abnormal areas. This multi-sensor inspection method consumes high power, and the inspection path generation depends on the path preset.

[0145] In the experiment, data collection and analysis were performed based on three comparison dimensions: power consumption, abnormal area recognition accuracy, and inspection path response time. The experimental results are shown in Table 1.

[0146] Table 1 Comparison of experimental results

[0147]

[0148] In the power consumption test, under the same inspection conditions, three methods were used to complete the surface data collection of power equipment and record the power consumption. The method of the present invention collects multi-spectral images through a single ULP camera device, which significantly reduces power consumption compared to traditional methods one and two. In traditional method one, the drone needs to maintain a flight state and carry a visible light camera and an infrared thermal imager at the same time, and the overall power consumption is relatively high. In traditional method two, the ground robot has high power consumption due to the combined operation of multiple sensors and navigation control, resulting in high overall power consumption of the inspection equipment. The low-power design of the present invention greatly reduces the energy consumption of the equipment, making it more suitable for long-term and large-scale inspection tasks.

[0149] The method of the present invention uses a ULP camera to independently complete the inspection task, and the power consumption is lower than the traditional multi-device combination solution. The drone of the traditional method one needs to maintain flight and carry multiple devices, and the overall power consumption is relatively high. The ground robot of the traditional method two is equipped with multiple sensor modules and self-navigation systems, which is complicated to operate and consumes a lot of power. The present invention realizes multi-spectral image acquisition of a single device based on the ULP camera, which greatly reduces the power consumption during the inspection process and reflects the advantages of the low-power inspection solution.

[0150] In terms of the accuracy of abnormal area recognition, the surfaces of the three groups of power equipment were preset with known abnormal areas (cracks, corrosion and deformation) to test the recognition of the three methods. The method of the present invention uses a ULP camera to collect visible light and near-infrared light images, and performs data analysis through multi-spectral fusion technology, which greatly improves the recognition accuracy. Traditional method 1 requires post-processing analysis of visible light and infrared thermal images, and is relatively sensitive to color and temperature changes, but due to the lack of multi-spectral fusion, it cannot capture subtle changes in surface structure. Traditional method 2 relies on ultrasonic sensors and lidar data, and has low recognition of structural anomalies, especially when detecting cracks and corrosion. The accuracy is insufficient.

[0151] The method of the present invention effectively improves the recognition accuracy of subtle surface defects such as cracks and corrosion by fusing visible light and near-infrared spectra. In contrast, although traditional method one has temperature information, it is slightly insufficient in the recognition of subtle structures. It mainly relies on temperature difference judgment and fails to effectively fuse spectral data. Although the ultrasonic sensor and lidar in traditional method two can detect surface distance and depth, their ability to recognize surface defects is insufficient, especially when detecting smaller surface anomalies, the accuracy is low. The method of the present invention achieves more comprehensive feature extraction based on multi-spectral data, significantly improving the recognition accuracy of abnormal areas.

[0152] In terms of inspection path response time, the method of the present invention directly uses the A* algorithm to generate the optimal inspection path based on the severity of the fault after the defect is detected. Thanks to the real-time processing and automated path planning of multispectral images, the method of the present invention does not require subsequent data fusion or manual intervention, significantly shortening the path response time. Traditional method one relies on drones and post-processing analysis software, which takes a long time in the data transmission and processing process. In traditional method two, the ground robot needs to complete the path preset first, and relies on the synchronous fusion of multi-sensor data, and the path generation time is relatively long. Overall, the solution of the present invention significantly improves the inspection response efficiency in path planning.

[0153] The A* algorithm is a heuristic search algorithm that combines the optimality guarantee of the Dijkstra algorithm with a heuristic estimation function. The heuristic estimation function is used to estimate the cost from the current node to the target node. In inspection path planning, this estimation function can take into account the severity of defects. For example, for inspection points with serious defects, the estimation function will give higher priority, making the algorithm more inclined to search towards these points during the search process. It maintains an open list and a closed list to record the nodes to be explored and the nodes that have been explored, and continuously selects the nodes in the open list with the minimum comprehensive cost (the actual cost from the starting node to the current node plus the estimated cost to the target node) for expansion.

[0154] For example, in a power inspection system, for power towers with fault alarms, the A* algorithm can use heuristic functions to estimate the priority of reaching these fault points based on the location of the power towers and the fault alarm information, and quickly generate an efficient inspection path passing through these fault points.

[0155] The present invention realizes automatic inspection path generation through the A* algorithm, combines multi-spectral data processing and defect severity classification to generate the optimal inspection path. Traditional method one relies on subsequent processing and analysis, and needs to wait for the drone data transmission to be completed, and the path generation response is slow. Traditional method two requires the completion of multi-sensor data fusion before generating according to the preset path, and the overall response speed is limited by data synchronization and transmission. In contrast, the solution of the present invention realizes the real-time generation of the inspection path, greatly reduces the response time, and demonstrates superior inspection efficiency.

[0156] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A power inspection method, characterized in that: include: Use ULP cameras to collect multispectral images of power equipment; Preprocess the collected multispectral images; Fuse the pre-processed multispectral images to analyze abnormal areas on the surface of power equipment; Extract features from abnormal areas on the surface of power equipment; Compare the extracted features with the standard parameters in the fault feature library to classify the defects; Generate inspection routes based on defect classification results.

2. The power inspection method according to claim 1, characterized in that: The multispectral image acquisition comprises: Use ULP cameras to collect visible light images of the surface of power equipment and record the color and brightness information of the equipment surface; Switch the ULP camera to the near-infrared light channel, collect near-infrared images of the same target area, and record the infrared reflection characteristics of the device surface; The collected visible light and near-infrared light images are stored after being numbered by region.

3. The power inspection method according to claim 2, characterized in that: The pre-processing comprises: Perform Gaussian filtering on the stored visible light and near infrared light images to obtain denoised images; Apply median filtering to the image after Gaussian filtering to remove speckle noise; The image processed by median filtering is input into the Canny edge detection algorithm to generate an image containing the edge information of the device surface; The image containing edge information is processed by Laplacian operator to obtain an image with enhanced details.

4. The power inspection method according to claim 3, characterized in that: The analysis of abnormal areas on the surface of the power equipment includes: Visible light images and near infrared images are aligned through a registration algorithm based on image feature points, and the aligned images are combined to form a composite image; Perform weighted fusion on the visible light and near infrared spectrum data in the composite image to obtain a fused image; The fused image is segmented using a threshold segmentation algorithm to extract areas where the brightness and color exceed the set threshold, generating a binary image containing abnormal areas on the surface of the power equipment.

5. The power inspection method according to claim 4, characterized in that: The feature extraction includes extracting the RGB color histogram of the abnormal area marked in the binary image and generating a color feature vector; The grayscale image of the abnormal area is analyzed using the gray-level co-occurrence matrix to obtain texture features, including contrast and uniformity parameters; Calculate the average reflection intensity value of the abnormal area to generate brightness features; The color features, texture features and brightness features are integrated into a set of feature parameters of the abnormal area to form a complete feature vector.

6. The power inspection method according to claim 5, characterized in that: The defect classification includes inputting the feature vector of the abnormal area into the defect classification model, performing similarity comparison with the standard features in the fault feature library, and calculating the similarity; Based on the similarity results, the abnormal area is identified as a specific defect type; Use the preset severity criteria, combined with defect type and similarity, to grade the severity of defects; Combine equipment number, defect type, severity, and equipment geographic location information to generate a complete failure data record.

7. The power inspection method according to claim 6, characterized in that: The generation of the inspection route includes taking the fault data records in the fault database as input, prioritizing each fault point according to the severity of the fault, assigning a priority weight, and using the A* algorithm to generate an optimal path covering all fault points; The generated inspection paths are arranged in order of priority of the fault points to form an inspection task list, which is then transmitted to the mobile devices of the inspectors.

8. A system using the power inspection method according to any one of claims 1 to 7, characterized in that: include: An image acquisition module (100), a data processing module (200), an abnormality recognition module (300), a feature extraction module (400), a defect classification module (500), and an inspection planning module (600); The image acquisition module (100) is used to use a ULP camera to acquire multispectral images of electric power equipment; The data processing module (200) is used to pre-process the collected multispectral image; The abnormality identification module (300) is used to fuse the pre-processed multi-spectral images and analyze the abnormal area on the surface of the power equipment; The feature extraction module (400) is used to extract features from abnormal areas on the surface of the power equipment; The defect classification module (500) is used to compare the extracted features with standard parameters in the fault feature library to perform defect classification; The inspection planning module (600) is used to generate an inspection route based on the defect classification result.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the power inspection method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the power inspection method according to any one of claims 1 to 7 are implemented.

Images