A method, system, device and storage medium for classifying electric power image data

Abstract

The present invention provides a method, system, device and storage medium for classifying power image data, and relates to the technical field of image data classification. The method comprises obtaining a grayscale image template set representing the edge of a typical power equipment by acquiring a typical power equipment target image; obtaining a grayscale image representing the edge of the image through an edge detection algorithm of the power equipment image to be classified, calculating the similarity between the template set and the grayscale image of the image edge and extracting descriptors; making similarity judgments on the descriptors and classifying them into the same category; naming them according to the classification results and image naming rules to obtain a classified sub-data set; and obtaining the degree of registration in the sub-image set by using the SIFT algorithm between the images, and marking the images with a registration rate greater than a set threshold as the same power equipment target. The effective classification of image data is achieved, the classification efficiency and accuracy are improved, and the classified and marked image data facilitates the management and query of images.

Description

Technical Field

[0001] The present invention relates to the technical field of image data classification, and in particular to a method, system, device and storage medium for power image data classification. Background Art

[0002] The research and application of imaging technology and related devices have greatly facilitated the backup and data retention of power equipment image data. However, with the exponential growth of terminal-side visualization devices, power image data faces the problem of massive classification. In the subsequent image acquisition process, information such as image naming and image shooting device attributes are synchronously marked to better play the value of data. However, there is still a large amount of historical image data that cannot be automatically classified, which requires a lot of manpower. At the same time, inspection personnel forget the shooting details, etc., resulting in a large amount of historical image data being shelved in the storage for a long time and unable to play its application value. Summary of the invention

[0003] To this end, an embodiment of the present invention provides a method, system, device and storage medium for classifying electric power image data, which is used to solve the problems in the prior art that a large amount of historical image data cannot be automatically classified, requires a lot of manpower, and the inspection personnel forget the shooting details, resulting in a large amount of historical image data being shelved in the memory for a long time and unable to play its application value.

[0004] In order to solve the above problems, an embodiment of the present invention provides a method for classifying power image data, the method comprising:

[0005] S1: Acquire a typical power equipment target image, and obtain a grayscale image template set representing the edge of the typical power equipment based on an edge detection algorithm;

[0006] S2: The power equipment image to be classified is subjected to an edge detection algorithm to obtain a grayscale image representing the edge of the image, the template set and the grayscale image of the image edge are similarly calculated, the SIFT algorithm is used to detect and locate the feature points of the template set and the grayscale image, and descriptors are extracted for the detected feature points;

[0007] S3: Using the distance between the descriptors after standardization as the similarity value between the feature points, the descriptors are judged for similarity and classified into the same category;

[0008] S4: Calculate the support vector for each image based on the similarity value, select the first n categories of templates with the largest histogram to name the image, and if there are multiple support vectors, copy the image to multiple categories and name them according to the targets in the template to obtain multiple sub-image sets named after typical power equipment;

[0009] S5: In the sub-image set, the degree of registration is obtained by using the SIFT algorithm between the images, and the images with a registration rate greater than a set threshold are marked as the same power equipment target.

[0010] Preferably, the edge detection algorithm includes Canny edge detection and Sobel gradient operator.

[0011] Preferably, the method for calculating the similarity between the template set and the grayscale image of the edge of the image is:

[0012]

[0013] Among them, T′(x′,y′) and I′(x′,y′) represent the average grayscale values ​​of the template image and the matching image at (x′,y′), respectively; R(x,y) represents the similarity; x, y represent the coordinates of the grayscale image of the edge of the image; x′, y′ represent the coordinates of the grayscale image of the template set.

[0014] Preferably, the method for extracting descriptors for the detected feature points is:

[0015] A Gaussian pyramid is constructed in the local area around the feature point to obtain information of three scales A, B, and C. The three scale information is fused to obtain A&B, A&C, B&C, and A&B&C information. The scale and accuracy complement each other. In a total of seven scale combinations of A, B, C, A&B, A&C, B&C, and A&B&C, the gradient amplitude and direction are calculated, and the gradient information of each sub-area is histogram counted. The histograms of all sub-areas are connected into a vector to obtain the descriptor.

[0016] Preferably, the method for calculating the gradient magnitude and direction comprises:

[0017] First, use the gradient operator to calculate the horizontal gradient G of the image x and the vertical gradient G y , the calculation formula is as follows:

[0018]

[0019] In the formula, I represents the image, * represents the convolution operation;

[0020] Then calculate the gradient magnitude G and gradient direction θ of each pixel, the calculation formula is as follows:

[0021]

[0022] Among them, G represents the gradient amplitude, θ represents the gradient direction, and G x and G y Represent the horizontal gradient and vertical gradient of the image respectively.

[0023] Preferably, Z-score standardization is used to process the feature vectors of the template set and the grayscale image, wherein the Z-score standardization calculation formula is:

[0024] Ds=(D-μ) / σ

[0025] Where μ is the mean of the descriptor feature vector, σ is the standard deviation of the descriptor feature vector, D is the descriptor feature vector, and Ds is the standardized descriptor feature vector.

[0026] Preferably, the distance between the descriptors after standardization is used as the similarity value between the feature points, and the method for determining the similarity of the descriptors is:

[0027] Descriptors with a distance between μ-2σ and μ+2σ are judged to be similar and classified into the same category.

[0028] The embodiment of the present invention further provides a power image data classification system, which is used to implement the power image data classification method described above, and specifically includes:

[0029] A template set acquisition module is used to acquire a typical power equipment target image and obtain a grayscale image template set representing the edge of the typical power equipment based on an edge detection algorithm;

[0030] A descriptor extraction module is used to obtain a grayscale image representing the edge of the power equipment image to be classified through an edge detection algorithm, perform similarity calculation between the template set and the grayscale image of the image edge, use the SIFT algorithm to detect and locate feature points of the template set and the grayscale image, and extract descriptors for the detected feature points;

[0031] A similarity judgment module is used to use the distance between the descriptors after standardization as the similarity value between the feature points, to judge the similarity of the descriptors, and to classify them into the same category;

[0032] The image naming module is used to calculate the support vector of each image according to the similarity value, select the first n categories of templates with the largest histogram to name the image, and if there are multiple support vectors, copy the image to multiple categories and name them according to the targets in the template to obtain multiple sub-image sets named after typical power equipment;

[0033] The registration module is used to obtain the registration degree between the sub-image sets by using the SIFT algorithm between the images, and mark the images with a registration rate greater than a set threshold as the same power equipment target.

[0034] An embodiment of the present invention also provides an electronic device, which includes a processor, a memory and a bus system, wherein the processor and the memory are connected through the bus system, the memory is used to store instructions, and the processor is used to execute the instructions stored in the memory to implement the above-mentioned power image data classification method.

[0035] An embodiment of the present invention further provides a computer storage medium, wherein the computer storage medium stores a computer software product, wherein the computer software product includes several instructions for enabling a computer device to execute the above-mentioned power image data classification method.

[0036] It can be seen from the above technical solutions that the present invention has the following advantages:

[0037] The embodiment of the present invention provides a method, system, device and storage medium for classifying power image data, which obtains a typical power equipment target image and obtains a grayscale image template set representing the edge of the typical power equipment based on an edge detection algorithm; obtains a grayscale image representing the edge of the image through an edge detection algorithm for the power equipment image to be classified, performs similarity calculation between the template set and the grayscale image of the image edge, uses the SIFT algorithm to detect and locate the feature points of the template set and the grayscale image, and extracts descriptors for the detected feature points; uses the distance between the descriptors based on standardization as the similarity value between the feature points, performs similarity judgment on the descriptors, and classifies them into the same category; names them according to the image naming rules based on the classification results to obtain a classified sub-data set; and is used to obtain the degree of registration in the sub-image set by using the SIFT algorithm between the images, and marks the images with a registration rate greater than a set threshold as the same power equipment target. The present invention realizes the effective classification of image data, improves the classification efficiency and accuracy, and the classified and marked image data facilitates the management and query of images, and can also be used as a test set for model testing. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the implementation cases of the present invention or the technical solutions in the prior art, the following is a brief description of the drawings required for use in the embodiments. By referring to the drawings, the features and advantages of the present invention will be more clearly understood. The drawings are schematic and should not be understood as limiting the present invention in any way. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0039] Figure 1 is a flow chart of a method for classifying power image data provided in an embodiment;

[0040] Figure 2 A block diagram of a power image data classification system provided according to an embodiment. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] Embodiment 1

[0043] There is a large amount of historical power image data that cannot be automatically classified, which requires a lot of manpower. At the same time, the inspection personnel forget the shooting details, etc., resulting in a large amount of historical power image data being shelved in the memory for a long time and unable to play its application value. Figure 1 As shown, an embodiment of the present invention provides a method for classifying power image data, the method comprising:

[0044] S1: Acquire a typical power equipment target image, and obtain a grayscale image template set representing the edge of the typical power equipment based on an edge detection algorithm;

[0045] S2: The power equipment image to be classified is subjected to an edge detection algorithm to obtain a grayscale image representing the edge of the image, the template set and the grayscale image of the image edge are similarly calculated, the SIFT algorithm is used to detect and locate the feature points of the template set and the grayscale image, and descriptors are extracted for the detected feature points;

[0046] S3: Using the distance between the descriptors after standardization as the similarity value between the feature points, the descriptors are judged for similarity and classified into the same category;

[0047] S4: Calculate the support vector for each image based on the similarity value, select the first n categories of templates with the largest histogram to name the image, and if there are multiple support vectors, copy the image to multiple categories and name them according to the targets in the template to obtain multiple sub-image sets named after typical power equipment;

[0048] S5: In the sub-image set, the degree of registration is obtained by using the SIFT algorithm between the images, and the images with a registration rate greater than a set threshold are marked as the same power equipment target.

[0049] As can be seen from the above technical scheme, the present invention provides a method for classifying power image data, by obtaining a typical power equipment target image, and obtaining a grayscale image template set representing the edge of the typical power equipment based on an edge detection algorithm; the power equipment image to be classified is subjected to an edge detection algorithm to obtain a grayscale image representing the edge of the image, the template set and the grayscale image of the image edge are similarly calculated, the SIFT algorithm is used to detect and locate the feature points of the template set and the grayscale image, and descriptors are extracted for the detected feature points; the distance between the descriptors based on standardization is used as the similarity value between the feature points, the descriptors are similarly judged, and the same category is classified; according to the classification result, the image is named according to the image naming rule to obtain the classified sub-data set; in the sub-image set, the SIFT algorithm between the images is used to obtain the degree of registration, and the images with a registration rate greater than a set threshold are marked as the same power equipment target. The present invention realizes the effective classification of image data, improves the classification efficiency and accuracy, and the classified and marked image data facilitates the management and query of images, and can also be used as a test set for model testing. In this embodiment, in step S1, a typical power equipment target image is obtained, and a grayscale image template set representing the edge of the typical power equipment is obtained based on an edge detection algorithm. In this embodiment, the edge detection algorithm preferably uses Canny edge detection and Sobel gradient operator, but is not limited to this algorithm.

[0050] In this embodiment, in step S2, the power equipment image to be classified is subjected to an edge detection algorithm to obtain a grayscale image representing the edge of the image, the similarity between the template set and the grayscale image of the image edge is calculated, the SIFT algorithm is used to detect and locate the feature points of the template set and the grayscale image, and descriptors are extracted for the detected feature points.

[0051] Specifically, in this embodiment, the template set and the grayscale image of the edge of the image are similar, and the similarity calculation method used is the standard correlation matching method, but is not limited to this method. In this embodiment, the grayscale of the matching image and the template image is used to calculate the matching degree between the two through a normalized correlation measurement formula. The formula is:

[0052]

[0053] Among them, T′(x′,y′) and I′(x′,y′) represent the average grayscale values ​​of the template image and the matching image at (x′,y′), respectively; R(x,y) represents the similarity; x, y represent the coordinates of the grayscale image of the edge of the image; x′, y′ represent the coordinates of the grayscale image of the template set.

[0054] Furthermore, in this embodiment, the SIFT algorithm is used to detect and locate the feature points of the template set and the grayscale image. The SIFT algorithm can detect the key points in the image, and the location is to detect the position of the key points in the image. The image coordinate system can be established with the upper left corner of the image as the origin to locate the position of the key points.

[0055] Furthermore, the method for extracting descriptors for the detected feature points is as follows: construct a Gaussian pyramid in the local area around the feature points, wherein the Gaussian pyramid is a data structure for representing image scale space, and is mainly used to detect image features at different scales. Its structure is the original image - Gaussian smoothing of the bottom layer image - Gaussian smoothing of the previous layer image, and each layer contains a smoothed version of the original image and multiple image pyramids at different scales. The constructed Gaussian pyramid is used to obtain the information of the three scales A, B, and C, and the three scale information is fused to obtain the information of A&B, A&C, B&C, and A&B&C, and the scale and accuracy are mutually compensated. In a total of 7 scale combinations of A, B, C, A&B, A&C, B&C, and A&B&C, the gradient amplitude and direction are calculated, and the gradient information of each sub-region is histogram-statisticed, and the histograms of all sub-regions are connected into a vector to obtain the descriptor.

[0056] The methods for calculating the gradient magnitude and direction include:

[0057] First, use the gradient operator to calculate the horizontal gradient G of the image x and the vertical gradient G y , the calculation formula is as follows:

[0058]

[0059] In the formula, I represents the image, * represents the convolution operation;

[0060] Then calculate the gradient magnitude G and gradient direction θ of each pixel, the calculation formula is as follows:

[0061]

[0062] Among them, G represents the gradient amplitude, θ represents the gradient direction, and G x and G y Represent the horizontal gradient and vertical gradient of the image respectively.

[0063] In this embodiment, in step S3, the distance between the descriptors after standardization is used as the similarity value between the feature points, and the descriptors are judged for similarity and classified into the same category.

[0064] In the descriptor generation process, in this embodiment, the feature vectors of the template set and the grayscale image are processed by Z-score standardization:

[0065] Ds=(D-μ) / σ

[0066] Where μ is the mean of the descriptor feature vector, σ is the standard deviation of the descriptor feature vector, D is the descriptor feature vector, and Ds is the standardized descriptor feature vector.

[0067] Furthermore, in this embodiment, the distance between the descriptors after standardization is used as the similarity value between the feature points, and the method for determining the similarity of the descriptors is as follows:

[0068] Descriptors with a distance between μ-2σ and μ+2σ are judged to be similar and classified into the same category.

[0069] In this embodiment, in step S4, after classification, support vectors are calculated for each image according to the similarity value, and the first n templates with the largest histograms are selected to name the image. If there are multiple support vectors, the image is copied to multiple categories and named according to the targets in the templates to obtain multiple sub-images named after typical power equipment. In this embodiment, the first 5 templates with the largest histograms are selected to name the image.

[0070] The support vector calculation for each image based on the similarity value specifically includes: encoding the template for comparison of each image as the horizontal axis; the similarity between the image and the template as the vertical axis, presenting a scattered distribution, using the scattered points as a histogram, and using the histogram to sort by size.

[0071] In this embodiment, in step S5, in the sub-image set, the SIFT algorithm between the images is used to obtain the degree of registration, and the images with a registration rate greater than a set threshold are marked as the same power equipment target. Specifically, images with registration feature points reaching 90% or more of the image feature points are marked as the same power equipment (including different phases) target; images with a registration rate lower than the threshold may be non-standardized equipment or have occlusions, which are identified and stored for personnel to identify.

[0072] The images classified and marked by the present invention can realize the effective classification of image data, facilitate the management and query of images, and can also be used as a test set to test the model.

[0073] Embodiment 2

[0074] like Figure 2 As shown, the present invention provides a power image data classification system, which is used to implement the power image data classification method described in Example 1, and specifically includes:

[0075] A template set acquisition module is used to acquire a typical power equipment target image and obtain a grayscale image template set representing the edge of the typical power equipment based on an edge detection algorithm;

[0076] A descriptor extraction module is used to obtain a grayscale image representing the edge of the power equipment image to be classified through an edge detection algorithm, perform similarity calculation between the template set and the grayscale image of the image edge, use the SIFT algorithm to detect and locate feature points of the template set and the grayscale image, and extract descriptors for the detected feature points;

[0077] A similarity judgment module is used to use the distance between the descriptors after standardization as the similarity value between the feature points, to judge the similarity of the descriptors, and to classify them into the same category;

[0078] The image naming module is used to calculate the support vector of each image according to the similarity value, select the first n categories of templates with the largest histogram to name the image, and if there are multiple support vectors, copy the image to multiple categories and name them according to the targets in the template to obtain multiple sub-image sets named after typical power equipment;

[0079] The registration module is used to obtain the registration degree between the sub-image sets by using the SIFT algorithm between the images, and mark the images with a registration rate greater than a set threshold as the same power equipment target.

[0080] A power image data classification system of the present embodiment is used to implement the aforementioned power image data classification method. Therefore, the specific implementation method of the power image data classification system can be seen in the embodiment part of the aforementioned power image data classification method. For example, the template set acquisition module 10, the descriptor extraction module 20, the similarity judgment module 30, the image naming module 40, and the registration module 50 are respectively used to implement steps S1, S2, S3, S4, and S5 in the aforementioned power image data classification method. Therefore, its specific implementation method can refer to the description of the corresponding embodiments of each part. In order to avoid redundancy, it will not be repeated here.

[0081] Embodiment 3

[0082] An embodiment of the present invention further provides an electronic device, comprising a processor, a memory and a bus system, wherein the processor and the memory are connected via the bus system, the memory is used to store instructions, and the processor is used to execute the instructions stored in the memory to implement the above-mentioned power image data classification method. To avoid redundancy, it will not be described here.

[0083] Embodiment 4

[0084] The embodiment of the present invention further provides a computer storage medium, wherein the computer storage medium stores a computer software product, wherein the computer software product includes several instructions for enabling a computer device to execute the above-mentioned power image data classification method. To avoid redundancy, it will not be described here.

[0085] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

[0086] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0087] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0088] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1The steps for the functions specified in one or more boxes.

[0089] Obviously, the above embodiments are merely examples for clear explanation and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived from these are still within the protection scope of the invention.

Claims

1. A method for classifying electric power image data, characterized in that: include: S1: Acquire a typical power equipment target image, and obtain a grayscale image template set representing the edge of the typical power equipment based on an edge detection algorithm; S2: The power equipment image to be classified is subjected to an edge detection algorithm to obtain a grayscale image representing the edge of the image, the template set and the grayscale image of the image edge are similarly calculated, the SIFT algorithm is used to detect and locate the feature points of the template set and the grayscale image, and descriptors are extracted for the detected feature points; The method for extracting descriptors for the detected feature points is: A Gaussian pyramid is constructed in the local area around the feature point, where the Gaussian pyramid is a data structure used for image scale space representation and is used to detect image features at different scales; Its structure is original image - Gaussian smoothing of the bottom layer image - Gaussian smoothing of the previous layer image. Each layer contains a smoothed version of the original image and image pyramids at multiple scales. The constructed Gaussian pyramid is used to obtain the information of three scales A, B, and C. The three scale information is fused to obtain A&B, A&C, B&C, and A&B&C information. The scale and accuracy complement each other. In a total of 7 scale combinations of A, B, C, A&B, A&C, B&C, and A&B&C, the gradient amplitude and direction are calculated, and the gradient information of each sub-region is histogram-statisticed. The histograms of all sub-regions are connected into a vector to obtain a descriptor. S3: Using the distance between the descriptors after standardization as the similarity value between the feature points, the descriptors are judged for similarity and classified into the same category; S4: Calculate the support vector for each image based on the similarity value, select the first n categories of templates with the largest histogram to name the image, and if there are multiple support vectors, copy the image to multiple categories and name them according to the targets in the template to obtain multiple sub-image sets named after typical power equipment; The method for calculating the support vector for each image according to the similarity value is: The template for each image comparison is encoded as the horizontal axis; the similarity between the image and the template is used as the vertical axis, showing a scattered distribution, and the scattered points are used as a histogram, and the size is sorted by the histogram; S5: In the sub-image set, the degree of registration is obtained by using the SIFT algorithm between the images, and the images with a registration rate greater than a set threshold are marked as the same power equipment target.

2. The method for classifying electric power image data according to claim 1, characterized in that: The edge detection algorithm includes Canny edge detection and Sobel gradient operator.

3. The method for classifying electric power image data according to claim 1, characterized in that: The method for calculating the similarity between the template set and the grayscale image of the edge of the image is: Among them, T′(x′,y′) and I′(x′,y′) represent the average grayscale values ​​of the template image and the matching image at (x′,y′), respectively; R(x,y) represents the similarity; x, y represent the coordinates of the grayscale image of the edge of the image; x′, y′ represent the coordinates of the grayscale image of the template set.

4. The method for classifying electric power image data according to claim 1, characterized in that: The method for calculating the gradient magnitude and direction comprises: First, use the gradient operator to calculate the horizontal gradient G of the image x and the vertical gradient G y , the calculation formula is as follows: In the formula, I represents the image, * represents the convolution operation; Then calculate the gradient magnitude G and gradient direction θ of each pixel, and the calculation formula is as follows: Among them, G represents the gradient amplitude, θ represents the gradient direction, and G x and G y Represent the horizontal gradient and vertical gradient of the image respectively.

5. The method for classifying electric power image data according to claim 1, characterized in that: In the descriptor generation process, Z-score normalization is used to process the feature vectors of the template set and the grayscale image, where the Z-score normalization calculation formula is: Ds=(D-μ) / σ Where μ is the mean of the descriptor feature vector, σ is the standard deviation of the descriptor feature vector, D is the descriptor feature vector, and Ds is the standardized descriptor feature vector.

6. The method for classifying electric power image data according to claim 5, characterized in that: The distance between the descriptors after standardization is used as the similarity value between the feature points. The method for judging the similarity of the descriptors is as follows: Descriptors with a distance between μ-2σ and μ+2σ are judged to be similar and classified into the same category.

7. A power image data classification system, characterized in that: The system is used to implement the power image data classification method according to any one of claims 1 to 6, specifically comprising: A template set acquisition module is used to acquire a typical power equipment target image and obtain a grayscale image template set representing the edge of the typical power equipment based on an edge detection algorithm; A descriptor extraction module is used to obtain a grayscale image representing the edge of the power equipment image to be classified through an edge detection algorithm, perform similarity calculation between the template set and the grayscale image of the image edge, use the SIFT algorithm to detect and locate feature points of the template set and the grayscale image, and extract descriptors for the detected feature points; A similarity judgment module is used to use the distance between the descriptors after standardization as the similarity value between the feature points, to judge the similarity of the descriptors, and to classify them into the same category; The image naming module is used to calculate the support vector of each image according to the similarity value, select the first n categories of templates with the largest histogram to name the image, and if there are multiple support vectors, copy the image to multiple categories and name them according to the targets in the template to obtain multiple sub-image sets named after typical power equipment; The registration module is used to obtain the registration degree between the sub-image sets by using the SIFT algorithm between the images, and mark the images with a registration rate greater than a set threshold as the same power equipment target.

8. An electronic device, characterized in that: The electronic device includes a processor, a memory and a bus system, the processor and the memory are connected through the bus system, the memory is used to store instructions, and the processor is used to execute the instructions stored in the memory to implement the power image data classification method described in any one of claims 1 to 6.

9. A computer storage medium, characterized in that: The computer storage medium stores a computer software product, and the computer software product includes a number of instructions for enabling a computer device to execute the power image data classification method according to any one of claims 1 to 6.

Images