Fault detection method, system and equipment for power equipment and medium

Abstract

The invention discloses a fault detection method and system for power equipment, equipment and a medium, and the method comprises the steps: obtaining an infrared image, position information and environment temperature of each piece of power equipment, sorting the infrared images of the equipment at the same environment temperature based on a position sequence, and obtaining a to-be-detected infrared image sequence; selecting an image with the highest clustering repetition degree from the sequence according to a preset rule as a target infrared image, performing processing by using a preset model, obtaining a first fault area image in the target image, performing de-duplication processing on other images in the sequence by using the first fault area image, and obtaining a second fault area image in the target image; and after each time of de-weighting, using the model again for other images after de-weighting, identifying a new fault area image and updating the first fault area image, and taking the last updated first fault area image as a fault area identification result of the power equipment corresponding to the sequence. According to the invention, the data volume of model processing is reduced, and the speed and stability of model processing are improved.

Description

Technical Field

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

[0002] With the massive access of UHV power equipment and AC / DC hybrid power equipment, the coverage of the power grid has been greatly expanded. Faced with a large number of diverse power equipment, relying solely on traditional manual detection methods to detect power equipment faults is increasingly unable to meet the current maintenance needs.

[0003] At present, the image recognition and other technologies in the field of artificial intelligence can be used to process and analyze the collected infrared images, which can assist or even replace manual fault diagnosis of power equipment, greatly improve inspection efficiency and reduce maintenance costs; however, if all the collected image data is input into the model for identification one by one, the excessive amount of data may affect the speed of fault detection and even easily cause the system to crash. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present application provides a method, system, device and medium for fault detection of electric power equipment, which reduces the amount of data processed by the model and improves the speed and stability of model processing.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a method for detecting a fault of an electric device, comprising:

[0006] Acquire the infrared image to be detected, the location information and the ambient temperature of each electric device at the location, and sort the infrared images to be detected of each electric device at the same ambient temperature based on the order of the locations to obtain at least one sequence of infrared images to be detected;

[0007] According to a preset extraction rule, the infrared image sequence to be detected is divided into a target infrared image and a first subsequence; wherein the target infrared image is the infrared image to be detected with the largest repetition of cluster regions in the infrared image sequence to be detected;

[0008] According to a preset target detection model, the target infrared image is detected to obtain a first fault area image corresponding to the target infrared image;

[0009] According to the first fault region image, deduplication processing is sequentially performed on each infrared image to be detected in the first subsequence, and during each deduplication processing, the deduplication-free infrared image to be detected is detected according to the target detection model to obtain a second fault region image corresponding to the deduplication-free infrared image to be detected, and during each deduplication processing, the first fault region image is updated according to the current second fault region image;

[0010] The most recently updated first fault area image is used as the fault area recognition result of each electric power device corresponding to the infrared image sequence to be detected.

[0011] Compared with the prior art, the embodiments of the present application have the following beneficial effects: by grouping infrared images of power equipment under the same ambient temperature together and sorting them according to the location information of the equipment, the interference caused by ambient temperature differences can be reduced; by selecting images with the highest clustering area repetition, it is convenient to determine fault-frequently occurring areas; the fault area in the target image is identified by a trained model to obtain a first fault area image; the first fault area image is used to deduplicate other images in the subsequence to avoid repeated detection of the same fault area, thereby reducing the data processing amount of the model and improving the speed and stability of model detection; the model is then used to detect the deduplicated image to ensure comprehensive detection and enhance the fault detection capability of the entire system.

[0012] In some embodiments of the first aspect of the present application, dividing the infrared image sequence to be detected into target infrared images and a first subsequence according to a preset extraction rule includes:

[0013] Acquire pixel values ​​of all feature points of the first infrared image to be detected in the sequence of infrared images to be detected, and cluster all feature points based on the magnitude of the pixel values ​​to obtain at least one clustering region;

[0014] The clustering area repetition of each infrared image to be detected in the sequence of infrared images to be detected is calculated, and the infrared image to be detected with the largest clustering area repetition in the sequence of infrared images to be detected is selected as the target infrared image, and the target infrared image is removed from the sequence of infrared images to be detected to obtain a first subsequence corresponding to the target infrared image.

[0015] The above embodiment has the following beneficial effects: by extracting the target infrared image, it is convenient to determine the fault-frequently occurring area, and then using the fault-frequently occurring area image to deduplicate other images in the subsequence, and then identifying each fault area on the deduplicated image, thereby reducing the amount of data processed by the model during subsequent fault detection and improving detection efficiency.

[0016] In some embodiments of the first aspect of the present application, detecting the target infrared image according to a preset target detection model to obtain a first fault area image corresponding to the target infrared image includes:

[0017] Input the target infrared image into a pre-built image detection model and output the corresponding one or more first regions; the first region is composed of feature points whose pixel values ​​are greater than a first preset threshold;

[0018] The center points of the one or more first areas are obtained respectively, and circles are drawn with each center point as the center and the closest distance from each center point to the edge of each first area as the radius, so as to obtain a first fault area image corresponding to the target infrared image; wherein the first fault area image contains at least one fault area.

[0019] The above embodiment has the following beneficial effects: the fault area in the target image is identified by the trained model, and a first fault area image is obtained, which provides a basis for further analysis and troubleshooting.

[0020] In some embodiments of the first aspect of the present application, the deduplication processing is sequentially performed on each infrared image to be detected in the first subsequence according to the first fault area image, including:

[0021] Acquire a fault area in the first fault area image, and slide on the edge of the fault area based on a preset sliding window to acquire a transition area connected to the fault area;

[0022] Matching the feature points in the transition area with the feature points of the next infrared image to be detected in the first subsequence, and aligning the first fault area image with the next infrared image to be detected according to the matching result;

[0023] The aligned target area is removed from the next infrared image to be detected to obtain an infrared sub-image to be detected corresponding to the next infrared image to be detected, wherein the target area includes the fault area and the transition area.

[0024] The above embodiment has the following beneficial effects: before using the model to detect other images in the subsequence, the first fault area image is used to deduplicate other images in the subsequence, thereby avoiding repeated detection of the same fault area, reducing the amount of data for model detection, and improving the speed and stability of model processing.

[0025] In some embodiments of the first aspect of the present application, each time deduplication processing is performed, the deduplication-free infrared image to be detected is detected according to the target detection model to obtain a second fault area image corresponding to the deduplication-free infrared image to be detected, and each time deduplication processing is performed, the first fault area image is updated according to the current second fault area image, including:

[0026] Input the infrared sub-image to be detected into the image detection model and output one or more corresponding second regions; the second region is composed of each feature point whose pixel value is greater than a second preset threshold;

[0027] Respectively obtain the center points of the one or more second areas, and draw circles with the center points as the center and the closest distances from the center points to the edges of the second areas as the radius, to obtain a second fault area image corresponding to the target infrared image; wherein the second fault area image contains at least one fault area;

[0028] The second fault area image is merged into the first fault area image to update the first fault area image.

[0029] The above-mentioned embodiment has the following beneficial effects: by reapplying the target detection model to the image to be detected with the known fault area removed, new or hidden fault areas can be discovered, ensuring comprehensive detection and enhancing the fault detection capability of the entire system. By using the new fault area image to update the first fault area image, a more comprehensive fault distribution situation can be obtained, and at the same time, a better basis is provided for cyclically repeating the detection of the next image to be detected.

[0030] In a second aspect, an embodiment of the present application further provides a fault detection system for electric power equipment, comprising: an acquisition module, an extraction module, a first identification module and a circulation module;

[0031] The acquisition module is used to acquire the infrared image to be detected, the location information and the ambient temperature of each power device at the location, and to sort the infrared images to be detected of each power device at the same ambient temperature based on the order of the locations to obtain at least one infrared image sequence to be detected;

[0032] The extraction module is used to divide the infrared image sequence to be detected into a target infrared image and a first subsequence according to a preset extraction rule; wherein the target infrared image is the infrared image to be detected with the largest cluster area repetition in the infrared image sequence to be detected;

[0033] The first recognition module is used to detect the target infrared image according to a preset target detection model to obtain a first fault area image corresponding to the target infrared image;

[0034] The loop module is used to perform deduplication processing on each infrared image to be detected in the first subsequence in turn according to the first fault area image, and in each deduplication processing, detect the deduplicated infrared image to be detected according to the target detection model to obtain the second fault area image corresponding to the deduplicated infrared image to be detected, and in each deduplication processing, update the first fault area image according to the current second fault area image.

[0035] In some embodiments of the second aspect of the present application, the extraction module includes: a clustering unit and a calculation unit;

[0036] The clustering unit is used to obtain pixel values ​​of all feature points of the first infrared image to be detected in the infrared image sequence to be detected, and cluster all feature points based on the size of the pixel values ​​to obtain at least one clustering area;

[0037] The calculation unit is used to calculate the cluster area repetition of each infrared image to be detected in the sequence of infrared images to be detected, and select the infrared image to be detected with the largest cluster area repetition in the sequence of infrared images to be detected as the target infrared image, and remove the target infrared image from the sequence of infrared images to be detected to obtain a first subsequence corresponding to the target infrared image.

[0038] In some embodiments of the second aspect of the present application, the first recognition module includes: a first model recognition unit and a first image generation unit;

[0039] The first model recognition unit is used to input the target infrared image into a pre-built image detection model and output one or more corresponding first regions; the first region is composed of each feature point whose pixel value is greater than a first preset threshold;

[0040] The first image recognition unit is used to respectively obtain the center points of the one or more first areas, and draw circles with each center point as the center and the closest distance from each center point to the edge of each first area as the radius, to obtain a first fault area image corresponding to the target infrared image; wherein the first fault area image contains at least one fault area.

[0041] In some embodiments of the second aspect of the present application, the circulation module includes: a deduplication unit;

[0042] The deduplication unit is used to obtain the fault area in the first fault area image, and slide on the edge of the fault area based on a preset sliding window to obtain a transition area connected to the fault area, match the feature points in the transition area with the feature points of the next infrared image to be detected in the first subsequence, and align the first fault area image with the next infrared image to be detected according to the matching result, remove the aligned target area in the next infrared image to be detected, and obtain the infrared sub-image to be detected corresponding to the next infrared image to be detected, wherein the target area includes the fault area and the transition area.

[0043] In some embodiments of the second aspect of the present application, the loop module further comprises: a repeat identification unit;

[0044] The repeated recognition unit is used to input the infrared sub-image to be detected into the image detection model and output the corresponding one or more second areas; the second area is composed of each feature point whose pixel value is greater than a second preset threshold, and the center points of the one or more second areas are respectively obtained, and circles are drawn with each center point as the center and the closest distance from each center point to the edge of each second area as the radius to obtain a second fault area image corresponding to the target infrared image, wherein the second fault area image contains at least one fault area, and the second fault area image is merged into the first fault area image to update the first fault area image.

[0045] Compared with the prior art, the above embodiment has the following beneficial effects: by grouping the infrared images of power equipment under the same ambient temperature together and sorting them according to the location information of the equipment, the interference caused by the difference in ambient temperature can be reduced; by selecting the image with the highest clustering area repetition, it is convenient to determine the fault-frequently occurring areas; the fault area in the target image is identified by the trained model to obtain the first fault area image; the first fault area image is used to deduplicate other images in the subsequence to avoid repeated detection of the same fault area, thereby reducing the data processing amount of the model and improving the speed and stability of the model detection; the model is then used to detect the deduplicated image to ensure comprehensive detection and enhance the fault detection capability of the entire system.

[0046] In a third aspect, an embodiment of the present application also provides a fault detection device for an electric power device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when loaded into the processor, implements the steps of a fault detection method for an electric power device.

[0047] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for fault detection of an electric power equipment are implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 : A flow chart of a fault detection method for an electric power device provided in some embodiments of the present invention.

[0049] Figure 2 : A structural schematic diagram of a fault detection system for an electric power equipment provided in some embodiments of the present invention.

[0050] Figure 3 : A structural diagram of a fault detection device for an electric power device provided in some embodiments of the present invention. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0052] Embodiment 1:

[0053] Please refer to Figure 1 , a fault detection method for electric power equipment provided by an embodiment of the present invention comprises steps S1 to S5:

[0054] Step S1: Obtain the infrared image to be detected, location information and ambient temperature of each power device, and sort the infrared images to be detected of each power device at the same ambient temperature based on the order of location to obtain at least one infrared image sequence to be detected.

[0055] In this embodiment, images of various power devices are captured by infrared image capture equipment, and the location information corresponding to each infrared image and the ambient temperature of the power equipment during capture are recorded. The infrared images of power devices at the same ambient temperature are grouped together and sorted according to the location information of the devices. This can reduce interference caused by differences in ambient temperature and ensure the reliability of subsequent fault detection.

[0056] Step S2: dividing the infrared image sequence to be detected into a target infrared image and a first subsequence according to a preset extraction rule; wherein the target infrared image is the infrared image to be detected with the largest cluster region repetition in the infrared image sequence to be detected.

[0057] In some embodiments of the present application, step S2 can be implemented by the following preferred implementations:

[0058] S21: acquiring pixel values ​​of all feature points of the first infrared image to be detected in the sequence of infrared images to be detected, and clustering all feature points based on the magnitude of the pixel values ​​to obtain at least one clustering region;

[0059] S22: calculating the clustering region repetition of each infrared image to be detected in the sequence of infrared images to be detected, selecting the infrared image to be detected with the largest clustering region repetition in the sequence of infrared images to be detected as the target infrared image, and removing the target infrared image from the sequence of infrared images to be detected to obtain a first subsequence corresponding to the target infrared image;

[0060] Further, as a preference, the formula for calculating the cluster region repetition degree may be as follows:

[0061] where f i is the clustering area repetition degree of the i-th infrared image to be detected, a1 is the number of infrared images to be detected that contain the first clustering area after removing the i-th infrared image to be detected in the infrared image sequence to be detected, a2 is the number of infrared images to be detected that contain the second clustering area after removing the i-th infrared image to be detected in the infrared image sequence to be detected, an is the number of infrared images to be detected that contain the n-th clustering area after removing the i-th infrared image to be detected in the infrared image sequence to be detected, t is the number of other infrared images to be detected in the infrared image sequence to be detected except the i-th infrared image to be detected, T i is the total number of clustered regions in the i-th infrared image to be detected;

[0062] For example, in a certain infrared image sequence to be detected, there are cluster regions a, b, c and d in the infrared image A to be detected; there are cluster regions a, e, f and h in the infrared image B to be detected; there are cluster regions a and e in the infrared image C to be detected. Then, the cluster region repetition of the infrared image A to be detected is: (1+0+0+0) / 4=0.25, the cluster region repetition of the infrared image B to be detected is: (1+0.5+0+0) / 4=0.375, and the cluster region repetition of the infrared image C to be detected is: (1+0.5) / 2=0.75, then the infrared image C to be detected is selected as the target infrared image.

[0063] In this embodiment, by extracting the target infrared image, it is convenient to determine the fault-frequently occurring areas. The images of the fault-frequently occurring areas are subsequently used to deduplicate other images in the subsequence, and then the deduplicated images are used to identify each fault area, thereby reducing the amount of data processed by the model during subsequent fault detection and improving detection efficiency.

[0064] Step S3: Detect the target infrared image according to a preset target detection model to obtain a first fault area image corresponding to the target infrared image.

[0065] In some embodiments of the present application, step S3 can be implemented by the following preferred implementations:

[0066] S31: inputting the target infrared image into a pre-built image detection model and outputting one or more corresponding first regions; the first region is composed of feature points whose pixel values ​​are greater than a first preset threshold;

[0067] S32: Obtain the center points of the one or more first areas respectively, and draw circles with each center point as the center and the closest distance from each center point to the edge of each first area as the radius, to obtain a first fault area image corresponding to the target infrared image; wherein the first fault area image contains at least one fault area.

[0068] In this embodiment, the fault area in the target image is identified by using the trained model to obtain a first fault area image, which provides a basis for further analysis and troubleshooting.

[0069] Step S4: according to the first fault area image, deduplication processing is sequentially performed on each infrared image to be detected in the first subsequence.

[0070] In some embodiments of the present application, step S4 can be implemented by the following preferred implementations:

[0071] S41: acquiring a fault area in the first fault area image, and sliding on the edge of the fault area based on a preset sliding window to acquire a transition area connected to the fault area;

[0072] S42: matching feature points in the transition area with feature points of the next infrared image to be detected in the first subsequence, and aligning the first fault area image with the next infrared image to be detected according to the matching result;

[0073] S43: removing the aligned target area from the next infrared image to be detected to obtain an infrared sub-image to be detected corresponding to the next infrared image to be detected, wherein the target area includes the fault area and the transition area.

[0074] In this embodiment, before using the model to detect other images in the subsequence, the first fault area image is used to match other images in the subsequence, and then the same parts as those in the first fault area image are removed from the other images, thereby avoiding repeated detection of the same fault area, reducing the amount of data detected by the model, and improving the speed and stability of model processing.

[0075] Step S5: During each deduplication process, the deduplicated infrared image to be detected is detected according to the target detection model to obtain the second fault area image corresponding to the deduplicated infrared image to be detected, and each deduplication process is performed, the first fault area image is updated according to the current second fault area image.

[0076] In some embodiments of the present application, step S5 can be implemented by the following preferred implementations:

[0077] S51: inputting the infrared sub-image to be detected into the image detection model and outputting one or more corresponding second regions; the second region is composed of feature points whose pixel values ​​are greater than a second preset threshold;

[0078] S52: respectively obtaining the center points of the one or more second areas, and drawing circles with the center points as the center and the shortest distances from the center points to the edges of the second areas as the radius, to obtain a second fault area image corresponding to the target infrared image; wherein the second fault area image contains at least one fault area;

[0079] S53: Merge the second fault area image into the first fault area image to update the first fault area image.

[0080] In this embodiment, the target detection model is applied again to the image to be detected with the known fault area removed, so that new or hidden fault areas can be discovered, ensuring comprehensive detection and enhancing the fault detection capability of the entire system. Using the new fault area image to update the first fault area image can obtain a more comprehensive fault distribution situation, and also provide a better basis for cyclically repeating the detection of the next image to be detected.

[0081] In summary, compared with the prior art, the embodiments of the present application have the following beneficial effects: by grouping infrared images of power equipment under the same ambient temperature together and sorting them according to the location information of the equipment, the interference caused by differences in ambient temperature can be reduced; by selecting images with the highest degree of clustering region repetition, it is convenient to determine areas where faults frequently occur; the fault area in the target image is identified by a trained model to obtain a first fault area image; the first fault area image is used to deduplicate other images in the subsequence to avoid repeated detection of the same fault area, thereby reducing the amount of data processing for the model and improving the speed and stability of model detection; the model is then used to detect the deduplicated image to ensure comprehensive detection and enhance the fault detection capability of the entire system.

[0082] Embodiment 2:

[0083] Please refer to Figure 2 Based on the same inventive concept, an embodiment of the present invention discloses a fault detection system for electric power equipment, including: an acquisition module M1, an extraction module M2, a first identification module M3 and a circulation module M4.

[0084] The acquisition module M1 is used to acquire the infrared image to be detected, the location information and the ambient temperature of each power device, and to sort the infrared images to be detected of each power device at the same ambient temperature based on the order of location to obtain at least one infrared image sequence to be detected.

[0085] The acquisition module M1 of this embodiment can reduce interference caused by ambient temperature differences and ensure the reliability of subsequent fault detection by grouping infrared images of power equipment under the same ambient temperature and sorting them according to the location information of the equipment.

[0086] The extraction module M2 is used to divide the infrared image sequence to be detected into a target infrared image and a first subsequence according to a preset extraction rule; wherein the target infrared image is the infrared image to be detected with the largest cluster area repetition in the infrared image sequence to be detected;

[0087] The extraction module M2 includes: a clustering unit and a calculation unit; the clustering unit is used to obtain the pixel values ​​of all feature points of the first infrared image to be detected in the infrared image sequence to be detected, and cluster all feature points based on the size of the pixel values ​​to obtain at least one clustering area; the calculation unit is used to calculate the repetition of the clustering areas of each infrared image to be detected in the infrared image sequence to be detected, and select the infrared image to be detected with the largest repetition of the clustering areas in the infrared image sequence to be detected as the target infrared image, and remove the target infrared image from the infrared image sequence to be detected to obtain a first subsequence corresponding to the target infrared image;

[0088] Further, as a preference, the formula for calculating the cluster region repetition degree may be as follows:

[0089] where f i is the clustering area repetition degree of the i-th infrared image to be detected, a1 is the number of infrared images to be detected that contain the first clustering area after removing the i-th infrared image to be detected in the infrared image sequence to be detected, a2 is the number of infrared images to be detected that contain the second clustering area after removing the i-th infrared image to be detected in the infrared image sequence to be detected, an is the number of infrared images to be detected that contain the n-th clustering area after removing the i-th infrared image to be detected in the infrared image sequence to be detected, t is the number of other infrared images to be detected in the infrared image sequence to be detected except the i-th infrared image to be detected, T i is the total number of clustered regions in the i-th infrared image to be detected.

[0090] The extraction module M2 of this embodiment extracts the target infrared image to facilitate the determination of the fault-frequently occurring areas, and subsequently uses the images of the fault-frequently occurring areas to deduplicate other images in the subsequence, and then uses the deduplicated images to identify each fault area, thereby reducing the amount of data processed by the model during subsequent fault detection and improving detection efficiency.

[0091] The first recognition module M3 is used to detect the target infrared image according to a preset target detection model to obtain a first fault area image corresponding to the target infrared image;

[0092] The first recognition module M3 includes: a first model recognition unit and a first image generation unit; the first model recognition unit is used to input the target infrared image into a pre-built image detection model and output the corresponding one or more first areas; the first area is composed of each feature point whose pixel value is greater than a first preset threshold; the first image recognition unit is used to obtain the center point of the one or more first areas respectively, and draw a circle with each center point as the center and the closest distance from each center point to the edge of each first area as the radius to obtain a first fault area image corresponding to the target infrared image; wherein the first fault area image contains at least one fault area.

[0093] The first recognition module M3 of this embodiment recognizes the fault area in the target image through the trained model to obtain a first fault area image, which provides a basis for further analysis and troubleshooting.

[0094] The loop module M4 is used to perform deduplication processing on each infrared image to be detected in the first subsequence in turn according to the first fault area image, and in each deduplication processing, detect the deduplicated infrared image to be detected according to the target detection model to obtain the second fault area image corresponding to the deduplicated infrared image to be detected, and in each deduplication processing, update the first fault area image according to the current second fault area image;

[0095] The loop module M4 includes: a deduplication unit; the deduplication unit is used to obtain the fault area in the first fault area image, and slide on the edge of the fault area based on a preset sliding window to obtain a transition area connected to the fault area, match the feature points in the transition area with the feature points of the next infrared image to be detected in the first subsequence, and align the first fault area image with the next infrared image to be detected according to the matching result, remove the aligned target area in the next infrared image to be detected, and obtain the infrared sub-image to be detected corresponding to the next infrared image to be detected, wherein the target area includes the fault area and the transition area.

[0096] The loop module M4 of this embodiment uses the first fault area image to deduplicate other images in the subsequence before using the model to detect other images in the subsequence, thereby avoiding repeated detection of the same fault area, reducing the amount of data detected by the model, and improving the speed and stability of model processing.

[0097] The loop module M4 also includes: a repeated recognition unit; the repeated recognition unit is used to input the infrared sub-image to be detected into the image detection model and output the corresponding one or more second areas; the second area is composed of each feature point whose pixel value is greater than a second preset threshold, and the center points of the one or more second areas are respectively obtained, and circles are drawn with each center point as the center and the closest distance from each center point to the edge of each second area as the radius to obtain a second fault area image corresponding to the target infrared image; wherein the second fault area image contains at least one fault area; the second fault area image is merged into the first fault area image to update the first fault area image.

[0098] The loop module M4 of this embodiment reapplies the target detection model to the image to be detected with the known fault area removed, so as to discover new or hidden fault areas, ensure comprehensive detection, and enhance the fault detection capability of the entire system. By using the new fault area image to update the first fault area image, a more comprehensive fault distribution situation can be obtained, and at the same time, a better basis is provided for cyclically repeating the detection of the next image to be detected.

[0099] In summary, compared with the prior art, the embodiments of the present application have the following beneficial effects: by grouping infrared images of power equipment under the same ambient temperature together and sorting them according to the location information of the equipment, the interference caused by differences in ambient temperature can be reduced; by selecting images with the highest degree of clustering region repetition, it is convenient to determine areas where faults frequently occur; the fault area in the target image is identified by a trained model to obtain a first fault area image; the first fault area image is used to deduplicate other images in the subsequence to avoid repeated detection of the same fault area, thereby reducing the amount of data processing for the model and improving the speed and stability of model detection; the model is then used to detect the deduplicated image to ensure comprehensive detection and enhance the fault detection capability of the entire system.

[0100] The division of the modules described above is only a logical function division. There may be other division methods in actual implementation. For example, multiple modules may be combined or integrated into another system.

[0101] Embodiment three:

[0102] Figure 3 The structure diagram of a fault detection device for electric power equipment of the present application is shown. Figure 3 As shown, the fault detection device may include: a processor N1, a memory N2, a data interface N3 and a communication bus N4.

[0103] Among them: the processor N1, the memory N2, and the data interface N3 communicate with each other through the communication bus N4; the data interface N3 is used for data communication with other devices such as input devices or output devices; the processor N1 is used to execute the program N5, which can specifically execute the relevant steps in the above-mentioned embodiment of the fault detection method of the power equipment.

[0104] Specifically, the program N5 may include program codes, and the program codes include computer executable instructions.

[0105] The processor N1 may be a central processing unit CPU, or an application-specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application. A fault detection device for an electric power device includes one or more processors, which may be processors of the same type, such as one or more CPUs, or processors of different types, such as one or more CPUs and one or more ASICs.

[0106] The memory N2 is used to store the program N5. The memory N2 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0107] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system or other device. In addition, the embodiments of the present application are not directed to any particular programming language.

[0108] Embodiment 4:

[0109] An embodiment of the present invention also provides a computer-readable storage medium, which stores at least one executable instruction. When the executable instruction is executed on a fault detection device / system of an electric power device, the fault detection device / system of the electric power device enables the fault detection method of the electric power device in any of the above-mentioned method embodiments.

[0110] In the specification provided herein, a large number of specific details are described. However, it is understood that the embodiments of the present application can be practiced without these specific details. Similarly, in order to simplify the application and help understand one or more of the various inventive aspects, in the above description of the exemplary embodiments of the present application, the various features of the embodiments of the present application are sometimes grouped together into a single embodiment, figure, or description thereof. Wherein, the claims that follow the specific embodiment are thus clearly incorporated into the specific embodiment, wherein each claim itself is used as a separate embodiment of the application.

[0111] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and further may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive.

Claims

1. A method for detecting a fault of an electric power device, characterized in that: include: Acquire the infrared image to be detected, the location information and the ambient temperature of each electric device at the location, and sort the infrared images to be detected of each electric device at the same ambient temperature based on the order of the locations to obtain at least one sequence of infrared images to be detected; According to a preset extraction rule, the infrared image sequence to be detected is divided into a target infrared image and a first subsequence; wherein the target infrared image is the infrared image to be detected with the largest repetition of cluster regions in the infrared image sequence to be detected; According to a preset target detection model, the target infrared image is detected to obtain a first fault area image corresponding to the target infrared image; According to the first fault region image, deduplication processing is sequentially performed on each infrared image to be detected in the first subsequence, and during each deduplication processing, the deduplication-free infrared image to be detected is detected according to the target detection model to obtain a second fault region image corresponding to the deduplication-free infrared image to be detected, and during each deduplication processing, the first fault region image is updated according to the current second fault region image; The most recently updated first fault area image is used as the fault area recognition result of each electric power device corresponding to the infrared image sequence to be detected.

2. A method for detecting a fault in an electric power device according to claim 1, characterized in that: The step of dividing the infrared image sequence to be detected into target infrared images and a first subsequence according to a preset extraction rule includes: Acquire pixel values ​​of all feature points of the first infrared image to be detected in the sequence of infrared images to be detected, and cluster all feature points based on the magnitude of the pixel values ​​to obtain at least one clustering region; The clustering area repetition of each infrared image to be detected in the sequence of infrared images to be detected is calculated, and the infrared image to be detected with the largest clustering area repetition in the sequence of infrared images to be detected is selected as the target infrared image, and the target infrared image is removed from the sequence of infrared images to be detected to obtain a first subsequence corresponding to the target infrared image.

3. A method for detecting a fault in an electric power device according to claim 2, characterized in that: The detecting the target infrared image according to the preset target detection model to obtain the first fault area image corresponding to the target infrared image includes: Input the target infrared image into a pre-built image detection model and output the corresponding one or more first regions; the first region is composed of feature points whose pixel values ​​are greater than a first preset threshold; The center points of the one or more first areas are obtained respectively, and circles are drawn with each center point as the center and the closest distance from each center point to the edge of each first area as the radius, so as to obtain a first fault area image corresponding to the target infrared image; wherein the first fault area image contains at least one fault area.

4. A method for detecting a fault in an electric power device as claimed in claim 3, characterized in that: The step of sequentially performing deduplication processing on each infrared image to be detected in the first subsequence according to the first fault area image includes: Acquire a fault area in the first fault area image, and slide on the edge of the fault area based on a preset sliding window to acquire a transition area connected to the fault area; Matching the feature points in the transition area with the feature points of the next infrared image to be detected in the first subsequence, and aligning the first fault area image with the next infrared image to be detected according to the matching result; The aligned target area is removed from the next infrared image to be detected to obtain an infrared sub-image to be detected corresponding to the next infrared image to be detected, wherein the target area includes the fault area and the transition area.

5. A method for detecting a fault in an electric power device as claimed in claim 4, characterized in that: During each deduplication process, the deduplication-free infrared image to be detected is detected according to the target detection model to obtain a second fault area image corresponding to the deduplication-free infrared image to be detected, and during each deduplication process, the first fault area image is updated according to the current second fault area image, including: Input the infrared sub-image to be detected into the image detection model and output one or more corresponding second regions; the second region is composed of each feature point whose pixel value is greater than a second preset threshold; Respectively obtain the center points of the one or more second areas, and draw circles with the center points as the center and the closest distances from the center points to the edges of the second areas as the radius, to obtain a second fault area image corresponding to the target infrared image; wherein the second fault area image contains at least one fault area; The second fault area image is merged into the first fault area image to update the first fault area image.

6. A fault detection system for electric power equipment, characterized in that: include: An acquisition module, an extraction module, a first recognition module and a circulation module; The acquisition module is used to acquire the infrared image to be detected, the location information and the ambient temperature of each power device at the location, and to sort the infrared images to be detected of each power device at the same ambient temperature based on the order of the locations to obtain at least one infrared image sequence to be detected; The extraction module is used to divide the infrared image sequence to be detected into a target infrared image and a first subsequence according to a preset extraction rule; wherein the target infrared image is the infrared image to be detected with the largest cluster area repetition in the infrared image sequence to be detected; The first recognition module is used to detect the target infrared image according to a preset target detection model to obtain a first fault area image corresponding to the target infrared image; The loop module is used to perform deduplication processing on each infrared image to be detected in the first subsequence in turn according to the first fault area image, and in each deduplication processing, detect the deduplicated infrared image to be detected according to the target detection model to obtain the second fault area image corresponding to the deduplicated infrared image to be detected, and in each deduplication processing, update the first fault area image according to the current second fault area image.

7. A fault detection system for electric power equipment as claimed in claim 6, characterized in that: The extraction module includes: a clustering unit and a calculation unit; The clustering unit is used to obtain pixel values ​​of all feature points of the first infrared image to be detected in the infrared image sequence to be detected, and cluster all feature points based on the size of the pixel values ​​to obtain at least one clustering area; The calculation unit is used to calculate the cluster area repetition of each infrared image to be detected in the sequence of infrared images to be detected, and select the infrared image to be detected with the largest cluster area repetition in the sequence of infrared images to be detected as the target infrared image, and remove the target infrared image from the sequence of infrared images to be detected to obtain a first subsequence corresponding to the target infrared image.

8. A fault detection system for electric power equipment as claimed in claim 7, characterized in that: The first recognition module includes: a first model recognition unit and a first image generation unit; The first model recognition unit is used to input the target infrared image into a pre-built image detection model and output one or more corresponding first regions; the first region is composed of each feature point whose pixel value is greater than a first preset threshold; The first image recognition unit is used to respectively obtain the center points of the one or more first areas, and draw circles with each center point as the center and the closest distance from each center point to the edge of each first area as the radius, to obtain a first fault area image corresponding to the target infrared image; wherein the first fault area image contains at least one fault area.

9. A fault detection system for electric power equipment as claimed in claim 8, characterized in that: The circulation module includes: a deduplication unit; The deduplication unit is used to obtain the fault area in the first fault area image, and slide on the edge of the fault area based on a preset sliding window to obtain a transition area connected to the fault area, match the feature points in the transition area with the feature points of the next infrared image to be detected in the first subsequence, and align the first fault area image with the next infrared image to be detected according to the matching result, remove the aligned target area in the next infrared image to be detected, and obtain the infrared sub-image to be detected corresponding to the next infrared image to be detected, wherein the target area includes the fault area and the transition area.

10. A fault detection system for electric power equipment according to claim 9, characterized in that: The loop module further comprises: a repetition recognition unit; The repeated recognition unit is used to input the infrared sub-image to be detected into the image detection model and output the corresponding one or more second areas; the second area is composed of each feature point whose pixel value is greater than a second preset threshold, and the center points of the one or more second areas are respectively obtained, and circles are drawn with each center point as the center and the closest distance from each center point to the edge of each second area as the radius to obtain a second fault area image corresponding to the target infrared image, wherein the second fault area image contains at least one fault area, and the second fault area image is merged into the first fault area image to update the first fault area image.

11. A fault detection device for electric power equipment, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is loaded into a processor, the steps of a fault detection method for electric power equipment according to claims 1 to 5 are implemented.

12. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of a fault detection method for electric power equipment according to claims 1 to 5 are implemented.

Images