Power equipment image extraction system

Abstract

The embodiment of the invention provides a power equipment image extraction system, and belongs to the technical field of power equipment image recognition. Comprising an unmanned aerial vehicle used for shooting equipment images in a transformer substation; the image uploading module is in communication connection with the unmanned aerial vehicle and is used for receiving the equipment image shot by the unmanned aerial vehicle; and the image processing module is in communication connection with the image uploading module and is used for receiving the equipment image and processing the equipment image. According to the electrical equipment image extraction system provided by the invention, image shooting is performed on electrical equipment in a transformer substation through the unmanned aerial vehicle, the equipment image is uploaded to the image processing module through the image uploading module, and the image processing module processes the equipment image; and the processed equipment image is input into the background server for feature recognition, so that intelligent inspection of the power equipment is realized, the inspection efficiency is improved, subjective errors are reduced, and the inspection reliability is improved.

Description

Technical Field

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

[0002] Substation equipment inspection is a vital link in the power system, which ensures the stable operation of substation equipment and the reliability of power supply in the store.

[0003] In the prior art, the inspection and recording method is generally adopted manually at fixed time and fixed point. However, this method needs to consume a lot of manpower and material resources, and there are subjective errors in manual inspection, which leads to poor inspection reliability and low inspection efficiency.

[0004] The inventors of the present application discovered during the process of implementing the present invention that the above-mentioned solutions in the prior art have the defects of low inspection efficiency and poor reliability. Summary of the invention

[0005] An object of an embodiment of the present invention is to provide a system for extracting images of electric power equipment, which has the functions of high inspection efficiency and high reliability.

[0006] In order to achieve the above object, an embodiment of the present invention provides a system for extracting an image of a power device, comprising:

[0007] Drones, used to capture images of equipment inside substations;

[0008] An image uploading module, which is in communication with the drone and is used to receive the device image taken by the drone;

[0009] An image processing module, which is in communication with the image uploading module and is used to receive the device image and process the device image;

[0010] The backend server is connected to the image processing module for receiving the processed device image output by the image processing module and performing feature recognition.

[0011] Optionally, the background server includes a controller, and the controller is used to execute:

[0012] Obtain image data of each power device in the power grid system;

[0013] Construct an image extraction model for power equipment;

[0014] Using the image data of each electric power device in the power grid system to train the electric power device image extraction model;

[0015] Get the current power equipment image;

[0016] The current electric power equipment image is input into the trained electric power equipment image extraction model to obtain relevant information in the electric power equipment image.

[0017] Optionally, acquiring image data of each electric power device in the power grid system includes:

[0018] Obtain images of components in each power device in the power grid system;

[0019] Classifying and labeling each of the component images;

[0020] The component images and categories of each of the electric power devices are aggregated to form image data of each of the electric power devices.

[0021] Optionally, using the image data of each electric power device in the power grid system to train the electric power device image extraction model includes:

[0022] Preprocessing the image in the image data;

[0023] Performing feature extraction on the preprocessed image;

[0024] The features are input into a classification model to train the classification model.

[0025] Optionally, preprocessing the image in the image data includes processing the image using a spatial domain denoising method or a frequency domain denoising method.

[0026] Optionally, extracting features from the preprocessed image includes:

[0027] Grayscale the image;

[0028] According to formula (1), the horizontal gradient of each pixel in the image is obtained:

[0029] G x (x,y)=α1(H(x+2,y)-H(x+1,y))+α2(H(x-1,y)-H(x-2,y,(1)

[0030] Among them, G x (x, y) is the horizontal gradient at the pixel point (x, y), H(x+1, y) is the pixel value at the pixel point (x+1, y), H(x-1, y) is the pixel value at the pixel point (x-1, y), H(x+2, y) is the pixel value at the pixel point (x+2, y), H(x-2, y) is the pixel value at the pixel point (x-2, y), α1 is the first horizontal weight, α2 is the second horizontal weight, and α1+α2=1;

[0031] According to formula (2), the vertical gradient of each pixel in the image is obtained:

[0032] G y (x,y)=β1(H(x,y+2)-H(x,y+1))+β2(H(x,y-1)-H(x,y-2,(2)

[0033] Among them, G y (x, y) is the vertical gradient at the pixel point (x, y), H(x, y+1) is the pixel value at the pixel point (x, y+1), H(x, y-1) is the pixel value at the pixel point (x, y-1), H(x, y+2) is the pixel value at the pixel point (x, y+2), H(x, y-2) is the pixel value at the pixel point (x, y-2), β1 is the first vertical weight, β2 is the second vertical weight, and β1+β2=1.

[0034] Optionally, extracting features from the preprocessed image further includes:

[0035] The gradient magnitude of each pixel in the image is calculated according to formula (3):

[0036]

[0037] Among them, G(x,y) is the gradient amplitude at the pixel point (x,y);

[0038] The gradient direction of each pixel in the image is calculated according to formula (4):

[0039]

[0040] Among them, α(x,y) is the gradient direction at the pixel point (x,y).

[0041] Optionally, extracting features from the preprocessed image further includes:

[0042] Dividing the image into a plurality of windows of preset sizes;

[0043] Obtaining the pixel gradient of each of the windows and forming a gradient histogram;

[0044] The features of the image are obtained according to the multiple gradient histograms.

[0045] Optionally, the classification model includes a SVM classifier.

[0046] Optionally, the classification model includes a random forest model.

[0047] Through the above technical scheme, the power equipment image extraction system provided by the present invention uses a drone to capture images of power equipment inside the substation, and uploads the equipment images to the image processing module through the image upload module. The image processing module processes the equipment images, and then inputs the processed equipment images into the background server for feature recognition, thereby realizing intelligent inspection of power equipment, improving inspection efficiency, reducing subjective errors, and thereby improving inspection reliability.

[0048] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0050] Figure 1 is a structural block diagram of a system for extracting an image of electric power equipment according to an embodiment of the present invention;

[0051] Figure 2 is a flow chart of obtaining relevant information of an electric power equipment image in an electric power equipment image extraction system according to an embodiment of the present invention;

[0052] Figure 3 is a flow chart of acquiring power equipment image data in a power equipment image extraction system according to one embodiment of the present invention;

[0053] Figure 4 is a flowchart of model training in a power equipment image extraction system according to an embodiment of the present invention;

[0054] Figure 5 The present invention is a flowchart of extracting features from an image in a power equipment image extraction system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0055] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0056] Figure 1 FIG. 1 is a structural block diagram of a system for extracting an image of an electric power device according to an embodiment of the present invention. Figure 1 In the present invention, the extraction system may include a drone, an image uploading module, an image processing module and a background server.

[0057] The drone is used to take images of the equipment inside the substation. The image upload module is connected to the drone for receiving the equipment images taken by the drone. The image processing module is connected to the image upload module for receiving the equipment images and processing them. The background server is connected to the image processing module for receiving the processed equipment images output by the image processing module and performing feature recognition.

[0058] When it is necessary to extract the features of the power equipment image, a drone is first used to capture images of the power equipment inside the substation, and the equipment images are uploaded to the image processing module through the image upload module. The image processing module processes the equipment images, and then the processed equipment images are input into the background server for feature recognition, so as to realize intelligent inspection of the power equipment.

[0059] Traditional inspection of power equipment generally adopts the method of manual inspection and recording at fixed time and fixed point. However, this method requires a lot of manpower and material resources, and there are subjective errors in manual inspection, which leads to poor inspection reliability and low inspection efficiency. In this embodiment of the present invention, the method of automatically extracting and identifying the features of the equipment image can realize intelligent inspection of power equipment, improve inspection efficiency, reduce subjective errors, and thus improve inspection reliability.

[0060] In this embodiment of the present invention, the backend server may include a controller for executing: Figure 2 Specifically, Figure 2 In the controller, the following is done:

[0061] In step S10, image data of each electric power device in the power grid system is obtained, wherein the image data of the electric power device includes device images, device classifications, and the like.

[0062] In step S11, a power equipment image extraction model is constructed.

[0063] In step S12, the image data of each electric power device in the power grid system is used to train the electric power device image extraction model.

[0064] In step S13, the current electric power equipment image is acquired.

[0065] In step S14, the current power equipment image is input into the trained power equipment image extraction model to obtain relevant information in the power equipment image.

[0066] In step S10 to step S14, the image data of each power device in the power grid system is first obtained, and then a power device image extraction model is constructed, and the power device image extraction model is trained using the image data to obtain a trained device image extraction model. The current power device image taken by the drone is obtained, and the current power device image is input into the trained power device image extraction model to extract the power device related information in the power device image.

[0067] In this embodiment of the present invention, the steps for acquiring image data of power equipment in the power grid system may be as follows: Figure 3 Specifically, Figure 3 The controller is also used to perform:

[0068] In step S20, images of components in each electric power device in the power grid system are obtained.

[0069] In step S21, each component image is classified and labeled.

[0070] In step S22 , the component images and categories of each electric device are aggregated to form image data of each electric device.

[0071] In this embodiment of the present invention, the training steps for the power equipment image extraction model can be as follows: Figure 4 Specifically, Figure 4 The controller is also used to perform:

[0072] In step S30, the image in the image data is preprocessed, wherein the preprocessing of the image may include processing the image using a spatial domain denoising method or a frequency domain denoising method.

[0073] In step S31, feature extraction is performed on the preprocessed image.

[0074] In step S32, the features are input into the classification model to train the classification model. The specific form of the classification model may include a binary SVM classifier as the classification model, or a multi-classification random forest model as the classification model.

[0075] In step S30 to step S32, after the image of the power equipment is acquired, the image is preprocessed, and feature extraction is performed on the preprocessed image. After the feature in the image is extracted, the feature is input into the classification model for classification training.

[0076] In this embodiment of the present invention, after acquiring the image, it is necessary to extract the features of the image. The specific extraction step can be as follows: Figure 5 Specifically, Figure 5 The controller is also used to perform:

[0077] In step S40, the image is grayscaled.

[0078] In step S41, the horizontal gradient of each pixel in the image is obtained according to formula (1):

[0079] G x (x,y)=α1(H(x+2,y)-H(x+1,y))+α2(H(x-1,y)-H(x-2,y,(1)

[0080] Among them, G x (x, y) is the horizontal gradient at the pixel point (x, y), H(x+1, y) is the pixel value at the pixel point (x+1, y), H(x-1, y) is the pixel value at the pixel point (x-1, y), H(x+2, y) is the pixel value at the pixel point (x+2, y), H(x-2, y) is the pixel value at the pixel point (x-2, y), α1 is the first horizontal weight, α2 is the second horizontal weight, and α1+α2=1, x, y are integer numbers. Specifically, the size of the first horizontal weight and the second horizontal weight can be determined according to the size of the pixel value of the pixel point (x, y) and the two adjacent horizontal pixel points (x+1, y) and (x-1, y). Specifically, the pixel value of the pixel point (x, y) is subtracted from the pixel value of the two adjacent horizontal pixel points (x+1, y) and (x-1, y). The smaller the difference, the larger the corresponding horizontal weight. By adopting this method, the accuracy of subsequent feature extraction can be further improved.

[0081] In step S42, the vertical gradient of each pixel in the image is obtained according to formula (2):

[0082] G y (x,y)=β1(H(x,y+2)-H(x,y+1))+β2(H(x,y-1)-H(x,y-2,(2)

[0083] Among them, G y(x, y) is the vertical gradient at the pixel point (x, y), H(x, y+1) is the pixel value at the pixel point (x, y+1), H(x, y-1) is the pixel value at the pixel point (x, y-1), H(x, y+2) is the pixel value at the pixel point (x, y+2), H(x, y-2) is the pixel value at the pixel point (x, y-2), β1 is the first vertical weight, β2 is the second vertical weight, and β1+β2=1. Specifically, the size of the first vertical weight and the second vertical weight can be determined according to the size of the pixel value of the pixel point (x, y) and the two adjacent vertical pixel points (x, y+1) and (x, y-1). Specifically, the pixel value of the pixel point (x, y) is subtracted from the pixel value of the two adjacent vertical pixel points (x, y+1) and (x, y-1). The smaller the difference, the larger the corresponding horizontal weight. In this way, the accuracy of subsequent feature extraction can be further improved.

[0084] In step S43, the gradient magnitude of each pixel in the image is calculated according to formula (3):

[0085]

[0086] Among them, G(x,y) is the gradient amplitude at the pixel point (x,y).

[0087] In step S44, the gradient direction of each pixel in the image is calculated according to formula (4):

[0088]

[0089] Among them, α(x,y) is the gradient direction at the pixel point (x,y).

[0090] In step S45, the image is divided into a plurality of windows of preset sizes.

[0091] In step S46, the pixel gradient of each window is obtained and a gradient histogram is formed, wherein each window corresponds to a gradient histogram.

[0092] In step S47, the features of the image are obtained based on multiple gradient histograms. The multiple windows are then combined into a block, and the descriptor of the block is obtained. Finally, all the blocks in the image are spliced ​​to obtain the features of the image. The features are input into the SVM classifier to classify the image.

[0093] In step S40 to step S47, the image is first grayed, and then the gradient, gradient amplitude and gradient direction of each pixel in the image are calculated. The image is divided into a plurality of windows of preset sizes, and a gradient histogram is formed according to the pixel gradient of each window, and then the features in the image are extracted according to the gradient histogram. Specifically, the method of extracting features from the image using the gradient histogram can more accurately express the image information using fewer features.

[0094] Through the above technical scheme, the power equipment image extraction system provided by the present invention uses a drone to capture images of power equipment inside the substation, and uploads the equipment images to the image processing module through the image upload module. The image processing module processes the equipment images, and then inputs the processed equipment images into the background server for feature recognition, thereby realizing intelligent inspection of power equipment, improving inspection efficiency, reducing subjective errors, and thereby improving inspection reliability.

[0095] 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.

[0096] 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.

[0097] 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 including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0098] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0099] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0100] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0101] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0102] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0103] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A system for extracting images of electric power equipment, characterized in that: include: Drones, used to capture images of equipment inside substations; An image uploading module, which is in communication with the drone and is used to receive the device image taken by the drone; An image processing module, which is in communication with the image uploading module and is used to receive the device image and process the device image; The backend server is connected to the image processing module for receiving the processed device image output by the image processing module and performing feature recognition.

2. The extraction system according to claim 1, characterized in that The backend server includes a controller, and the controller is used to execute: Obtain image data of each power device in the power grid system; Construct an image extraction model for power equipment; Using the image data of each electric power device in the power grid system to train the electric power device image extraction model; Get the current power equipment image; The current electric power equipment image is input into the trained electric power equipment image extraction model to obtain relevant information in the electric power equipment image.

3. The extraction system according to claim 2, characterized in that Acquiring image data of each power device in the power grid system includes: Obtain images of components in each power device in the power grid system; Classify and label each of the component images; The component images and categories of each of the electric power devices are aggregated to form image data of each of the electric power devices.

4. The extraction system according to claim 2, characterized in that Using the image data of each electric power device in the power grid system to train the electric power device image extraction model includes: Preprocessing the image in the image data; Performing feature extraction on the preprocessed image; The features are input into a classification model to train the classification model.

5. The extraction system according to claim 4, characterized in that Preprocessing the image in the image data includes processing the image using a spatial domain denoising method or a frequency domain denoising method.

6. The extraction system according to claim 4, characterized in that Extracting features from the preprocessed image includes: Grayscale the image; According to formula (1), the horizontal gradient of each pixel in the image is obtained: G x (x,y)=α1(H(x+2,y)-H(x+1,y))+α2(H(x-1,y)-H(x-2,y, (1) Among them, G x (x, y) is the horizontal gradient at the pixel point (x, y), H(x+1, y) is the pixel value at the pixel point (x+1, y), H(x-1, y) is the pixel value at the pixel point (x-1, y), H(x+2, y) is the pixel value at the pixel point (x+2, y), H(x-2, y) is the pixel value at the pixel point (x-2, y), α1 is the first horizontal weight, α2 is the second horizontal weight, and α1+α2=1; According to formula (2), the vertical gradient of each pixel in the image is obtained: G y (x,y)=β1(H(x,y+2)-H(x,y+1))+β2(H(x,y-1)-H(x,y-2, (2) Among them, G y (x, y) is the vertical gradient at the pixel point (x, y), H(x, y+1) is the pixel value at the pixel point (x, y+1), H(x, y-1) is the pixel value at the pixel point (x, y-1), H(x, y+2) is the pixel value at the pixel point (x, y+2), H(x, y-2) is the pixel value at the pixel point (x, y-2), β1 is the first vertical weight, β2 is the second vertical weight, and β1+β2=1.

7. The extraction system according to claim 6, characterized in that Extracting features from the preprocessed image further comprises: The gradient magnitude of each pixel in the image is calculated according to formula (3): Where G(x, y) is the gradient amplitude at the pixel point (x, y); The gradient direction of each pixel in the image is calculated according to formula (4): Among them, α(x, y) is the gradient direction at the pixel point (x, y).

8. The extraction system according to claim 7, characterized in that Extracting features from the preprocessed image further comprises: Dividing the image into a plurality of windows of preset sizes; Obtaining the pixel gradient of each of the windows and forming a gradient histogram; The features of the image are obtained according to the multiple gradient histograms.

9. The extraction system according to claim 4, characterized in that The classification model includes a SVM classifier.

10. The extraction system according to claim 4, characterized in that The classification model includes a random forest model.

Images