Power transformation equipment defect identification method and system based on video stream analysis

Abstract

The embodiment of the invention provides a power transformation equipment defect identification method and system based on video stream analysis, and belongs to the technical field of power transformation equipment monitoring. The method comprises the following steps: acquiring an infrared image frame in a current video stream; determining the actual position of the power transformation equipment based on the infrared image frame; calculating a distortion ratio of the actual position; judging whether the distortion ratio is smaller than or equal to a preset distortion ratio threshold value or not; under the condition that it is judged that the distortion ratio is larger than the distortion ratio threshold value, the intersection-union ratio of the actual position and the default anchoring frame position is calculated; judging whether the intersection-union ratio is smaller than a preset threshold value or not; under the condition that the intersection-to-union ratio is smaller than the threshold value, adjusting the angle of the current camera, and returning to execute the step of obtaining the infrared image frame in the current video stream; and under the condition that the intersection-to-union ratio is judged to be greater than or equal to the threshold value, performing defect identification operation based on the infrared image frame and the visible light image.

Description

Technical Field

[0001] This invention relates to the field of monitoring technology for power equipment, and more specifically to a method and system for identifying defects in power equipment based on video stream analysis. Background Technology

[0002] As a core node in the power system, the reliability of substation equipment operation directly affects grid security. Traditional manual inspection methods suffer from low efficiency, strong subjectivity, and the inability to monitor in real time, making it difficult to meet the needs of modern intelligent power systems. With the development of computer vision and deep learning technologies, image recognition-based automatic monitoring systems for substations are becoming increasingly common. These systems use cameras to capture images of equipment (such as circuit breakers, transformers, and insulators) and utilize target detection algorithms to identify defects, detect abnormal temperatures, and provide early warnings of mechanical faults. This type of technology significantly improves inspection efficiency. For example, by fusing visible light and infrared images, it can simultaneously capture the structural features and thermodynamic characteristics of equipment, thereby accurately identifying defects such as localized overheating, leakage, or mechanical damage.

[0003] However, the accuracy of image recognition models is highly dependent on the camera's alignment. In the outdoor environment of substations, cameras are constantly exposed to complex conditions such as wind, rain, temperature variations, mechanical vibrations, or external impacts, making them prone to issues like loose mounting bases, physical lens displacement, or rotation angle shifts. This can lead to image distortion or viewpoint warping, specifically radial distortion (such as barrel or pincushion deformation) and tangential distortion (such as image tilt). When the camera angle deviates, the position, scale, or shape of the target equipment in the image may be distorted, directly reducing the reliability of the recognition algorithm. For example, for meter readings or switch opening / closing status detection that require precise positioning, even slight angular deviations can cause misjudgments or missed detections, ultimately affecting the accuracy of fault warnings.

[0004] While existing research has attempted to improve recognition robustness through multimodal data fusion (such as combining infrared, acoustic, or vibration signals), most technologies still assume the camera angle is ideal and lack effective solutions for automatic camera angle calibration. Currently, only a few fields (such as vehicle-mounted panoramic imaging systems) involve dynamic calibration techniques, which correct images through distortion coefficient calculation or pixel-level fusion. However, these methods are difficult to directly transfer to substation environments: on the one hand, substation equipment has a complex structure, and cameras are deployed in various locations (such as at high altitudes or in confined spaces), requiring adaptation to coverage needs from different perspectives; on the other hand, power monitoring has extremely high requirements for image geometric accuracy, while traditional calibration methods focus more on visual presentation and neglect the geometric accuracy assurance of the preceding stages of the recognition model. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for identifying defects in substation equipment based on video stream analysis. This method and system can automatically adjust the shooting angle of substation cameras.

[0006] To achieve the above objectives, embodiments of the present invention provide a method for identifying defects in substation equipment based on video stream analysis, comprising: Retrieve infrared image frames from the current video stream; The actual location of the substation is determined based on the infrared image frame; Calculate the torsion ratio at the actual location; Determine whether the twist ratio is less than or equal to a preset twist ratio threshold; If the distortion ratio is determined to be greater than the distortion ratio threshold, the intersection-union ratio of the actual position and the default anchor box position is calculated; Determine whether the crossover-union ratio is less than a preset threshold; If the cross-union ratio is determined to be less than the threshold, the angle of the current camera is adjusted, and the process returns to the step of acquiring infrared image frames in the current video stream. If the cross-union ratio is determined to be greater than or equal to the threshold, a defect identification operation is performed based on the infrared image frame and the visible light image.

[0007] Optionally, calculating the distortion ratio at the actual location includes: The torsion ratio is calculated according to formula (1): (1) in, The twist ratio is... , Let be the pixel width and pixel length of the largest inscribed rectangle at the actual location. , The pixel width and pixel length of the maximum bounding rectangle at the actual location.

[0008] Optionally, acquiring infrared image frames from the current video stream includes: Infrared image frames are extracted from the video stream at preset frame intervals; Determine the overlap between any two infrared image frames; Determine whether the overlap degree is greater than a preset overlap degree threshold; If the overlap is determined to be greater than the overlap threshold, any one of the infrared image frames is output.

[0009] Optionally, determining the actual location of the substation based on the infrared image frame includes: The infrared image frame is binarized. The actual position is determined by using a proportional threshold operation based on the result of the binarization operation.

[0010] Optionally, the actual position is determined by applying a proportional thresholding operation based on the result of the binarization operation, including: Determine the maximum and minimum grayscale values ​​in the current infrared image frame; The suspected region is determined in the current infrared image frame by multiplying a preset ratio value by the maximum value; The suspected area with the largest area is selected as the actual location.

[0011] Optionally, the camera has a built-in image recognition model, which includes a skip connection layer for performing feature transformation operations according to formulas (2) and (3): (2) (3) in, This is the exponential moving average of the gradient of the network parameters. The hidden state vector of the previous time step. Given a function for a learnable weight vector, To skip the number of neurons in the connection layer, For kernel function, The weights of the loss function for positive samples. This is the exponential moving average of the squared gradient. The weights are the weights of the loss function for negative samples.

[0012] Optionally, adjust the current camera angle, including: Calculate the crossover ratio difference between the previous adjustment and the two previous adjustments; Determine whether the crossover-union ratio difference is greater than 0; If the cross-union ratio difference is greater than 0, the camera is adjusted using the previous camera adjustment angle; If the cross-union ratio is determined to be less than 0, the previous camera adjustment angle is used as the reverse solution to generate a new adjustment angle to adjust the camera.

[0013] Optionally, a new adjustment angle is generated, including: The inverse solution interval is determined based on the inverse solution; The angle outside the reverse solution interval is selected as the adjustment angle.

[0014] On the other hand, the present invention also provides a defect identification system for substation equipment based on video stream analysis, the system including a processor configured to perform any of the methods described above.

[0015] In another aspect, the present invention also provides a computer-readable storage medium storing instructions for being read by a machine to cause the machine to perform any of the methods described above.

[0016] Through the above technical solutions, the embodiments of the present invention provide a method and system for identifying defects in substation equipment based on video stream analysis. The method and system improve the accuracy of monitoring and identifying substation equipment and the stability of camera operation by combining frame images of the video stream to autonomously correct the shooting angle of the camera.

[0017] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a method for identifying defects in substation equipment based on video stream analysis according to an embodiment of the present invention; Figure 2 This is a flowchart of a method for acquiring infrared image frames according to an embodiment of the present invention; Figure 3 This is a flowchart of a method for adjusting a camera according to an embodiment of the present invention. Detailed Implementation

[0019] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0020] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application all comply with relevant laws and regulations. In the embodiments of this application, certain existing industry solutions such as software, components, and models may be mentioned. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used such solutions.

[0021] like Figure 1The diagram shows a flowchart of a method for identifying defects in substation equipment based on video stream analysis, according to an embodiment of the present invention. Figure 1 In this method, the steps may include: In step S10, infrared image frames from the current video stream are acquired; In step S11, the actual location of the substation equipment is determined based on the infrared image frame; In step S12, the torsion ratio at the actual position is calculated; In step S13, it is determined whether the twist ratio is less than or equal to a preset twist ratio threshold. In step S14, if the distortion ratio is greater than the distortion ratio threshold, the intersection-union ratio of the actual position and the default anchor box position is calculated. In step S15, it is determined whether the crossover-union ratio is less than a preset threshold; In step S16, if the cross-union ratio is less than the threshold, the angle of the current camera is adjusted, and the process returns to the step of acquiring infrared image frames in the current video stream. In step S17, if the crossover ratio is greater than or equal to the threshold, a defect identification operation is performed based on the infrared image frame and the visible light image.

[0022] In such Figure 1 In the method shown, step S10 can be used to acquire an infrared image frame of the current video stream. This current video stream can be obtained directly through an infrared imaging device. The method for acquiring the infrared image frame can be of various forms known to those skilled in the art. In one example of the present invention, the method for acquiring the infrared image frame may include, for example... Figure 2 The steps shown are described in this. Figure 2 The method for acquiring the infrared image frame may include the following steps: In step S20, infrared image frames are extracted from the video stream according to a preset frame interval; In step S21, the overlap between every two infrared image frames is determined; In step S22, it is determined whether the overlap degree is greater than the preset overlap degree threshold. In step S23, if the overlap is greater than the overlap threshold, any infrared image frame is output. In step S24, if any overlap degree is less than or equal to the overlap degree threshold, it is determined that an infrared image frame cannot be obtained at present.

[0023] Step S11 can be used to determine the actual location of the substation equipment based on the infrared image frame. Specifically, in one example of the present invention, step S11 may involve first performing a binarization operation on the infrared image frame, and then determining the actual location using a proportional thresholding operation based on the result of the binarization operation. Furthermore, considering that there may be noisy areas in the infrared image frame, in one example of the present invention, the method for determining the actual location may involve first determining the maximum and minimum grayscale values ​​in the current infrared image frame, then using the product of a preset proportional value and the maximum value to determine a suspected area in the current infrared image frame, and finally selecting the suspected area with the largest area as the actual location.

[0024] Step S12 can be used to calculate the distortion ratio at the actual location. The specific calculation method for this distortion ratio can be of various forms known to those skilled in the art. In one example of the present invention, the distortion ratio can be calculated using the following formula (1): (1) in, For the twist ratio, , The pixel width and pixel length of the largest inscribed rectangle at the actual location. , The pixel width and pixel length of the largest bounding rectangle at the actual location.

[0025] Step S13 can be used to determine whether the distortion ratio is less than or equal to a preset distortion ratio threshold. Since the area of ​​a transformer in a video image is generally rectangular, and the distortion ratio is the ratio of the area of ​​the largest inscribed rectangle to the area of ​​the largest bounding rectangle, the distortion ratio of the area where the transformer is located in the video image is close to 1. When determining whether the distortion ratio is less than or equal to the preset distortion ratio threshold, if the distortion ratio is less than the threshold, it indicates that the current image's angle offset is already large, and therefore, step S16 is needed to correct the camera's angle. Otherwise, no correction is needed, and the process returns to step S10 to continue monitoring.

[0026] Step S14 can be used to calculate the intersection-over-union (IoU) ratio between the actual position and the default anchor box. This IoU ratio represents the difference between the current camera's angle and the standard angle. A larger IoU ratio indicates a smaller difference between the current camera's angle and the standard angle; conversely, a smaller IoU ratio indicates a larger difference. Therefore, in this embodiment, step S15 is needed to determine whether the IoU ratio is less than a preset threshold. If it is less than the threshold, it indicates a large difference between the current camera's angle and the standard angle.

[0027] Step S16 can be used to adjust the current camera angle. In this embodiment, the method for adjusting the camera can be of various forms known to those skilled in the art. In one example of the present invention, the method for adjusting the camera may include, for example... Figure 3 The steps shown are described in this. Figure 3 In this context, adjusting the camera may include the following steps: In step S30, the crossover ratio difference between the previous adjustment and the previous two adjustments is calculated; In step S31, it is determined whether the difference in the crossover ratio is greater than 0; In step S32, if the crossover ratio difference is greater than 0, the camera is adjusted using the previous camera adjustment angle. In step S33, if the cross-union ratio is less than 0, the previous camera adjustment angle is used as the reverse solution to generate a new adjustment angle to adjust the camera. Specifically, in this example, the method for generating the adjustment angle may be to first determine the reverse solution interval based on the reverse solution, and then select an angle outside the reverse solution interval as the adjustment angle. Furthermore, in this embodiment, the method for detecting defects in substation equipment using a camera can, for example, incorporate an image recognition model within the camera. This image recognition model can include an input layer, a feature fusion layer, a skip connection layer, a feature stitching layer, and a classification layer. The input layer can be used to input an image. The feature fusion layer can split and combine features from the input image. The skip connection layer can be used to perform feature transformation on the output of the feature fusion layer. The feature stitching layer can be used to stitch together the result of the feature transformation operation and the result before feature transformation. The classification layer can be used to predict the result of the feature stitching, thereby obtaining the recognition result. The feature transformation operation can be implemented using the following formulas (2) and (3): (2) (3) in, This is the exponential moving average of the gradient of the network parameters. The hidden state vector of the previous time step. Given a function for a learnable weight vector, To skip the number of neurons in the connection layer, For kernel function, The weights of the loss function for positive samples. This is the exponential moving average of the squared gradient. The weights are the weights of the loss function for negative samples.

[0028] On the other hand, the present invention also provides a substation equipment defect identification system based on video stream analysis, the system including a processor configured to perform any of the methods described above. Specifically, the system... Figure 1 In this method, the steps may include: In step S10, infrared image frames from the current video stream are acquired; In step S11, the actual location of the substation equipment is determined based on the infrared image frame; In step S12, the torsion ratio at the actual position is calculated; In step S13, it is determined whether the twist ratio is less than or equal to a preset twist ratio threshold. In step S14, if the distortion ratio is greater than the distortion ratio threshold, the intersection-union ratio of the actual position and the default anchor box position is calculated. In step S15, it is determined whether the crossover-union ratio is less than a preset threshold; In step S16, if the cross-union ratio is less than the threshold, the angle of the current camera is adjusted, and the process returns to the step of acquiring infrared image frames in the current video stream. In step S17, if the crossover ratio is greater than or equal to the threshold, a defect identification operation is performed based on the infrared image frame and the visible light image.

[0029] In such Figure 1 In the method shown, step S10 can be used to acquire an infrared image frame of the current video stream. This current video stream can be obtained directly through an infrared imaging device. The method for acquiring the infrared image frame can be of various forms known to those skilled in the art. In one example of the present invention, the method for acquiring the infrared image frame may include, for example... Figure 2 The steps shown are described in this. Figure 2 The method for acquiring the infrared image frame may include the following steps: In step S20, infrared image frames are extracted from the video stream according to a preset frame interval; In step S21, the overlap between every two infrared image frames is determined; In step S22, it is determined whether the overlap degree is greater than the preset overlap degree threshold. In step S23, if the overlap is greater than the overlap threshold, any infrared image frame is output. In step S24, if any overlap degree is less than or equal to the overlap degree threshold, it is determined that an infrared image frame cannot be obtained at present.

[0030] Step S11 can be used to determine the actual location of the substation equipment based on the infrared image frame. Specifically, in one example of the present invention, step S11 may involve first performing a binarization operation on the infrared image frame, and then determining the actual location using a proportional thresholding operation based on the result of the binarization operation. Furthermore, considering that there may be noisy areas in the infrared image frame, in one example of the present invention, the method for determining the actual location may involve first determining the maximum and minimum grayscale values ​​in the current infrared image frame, then using the product of a preset proportional value and the maximum value to determine a suspected area in the current infrared image frame, and finally selecting the suspected area with the largest area as the actual location.

[0031] Step S12 can be used to calculate the distortion ratio at the actual location. The specific calculation method for this distortion ratio can be of various forms known to those skilled in the art. In one example of the present invention, the distortion ratio can be calculated using the following formula (1): (1) in, For the twist ratio, , The pixel width and pixel length of the largest inscribed rectangle at the actual location. , The pixel width and pixel length of the largest bounding rectangle at the actual location.

[0032] Step S13 can be used to determine whether the distortion ratio is less than or equal to a preset distortion ratio threshold. Since the area of ​​a transformer in a video image is generally rectangular, and the distortion ratio is the ratio of the area of ​​the largest inscribed rectangle to the area of ​​the largest bounding rectangle, the distortion ratio of the area where the transformer is located in the video image is close to 1. When determining whether the distortion ratio is less than or equal to the preset distortion ratio threshold, if the distortion ratio is less than the threshold, it indicates that the current image's angle offset is already large, and therefore, step S16 is needed to correct the camera's angle. Otherwise, no correction is needed, and the process returns to step S10 to continue monitoring.

[0033] Step S14 can be used to calculate the intersection-over-union (IoU) ratio between the actual position and the default anchor box. This IoU ratio represents the difference between the current camera's angle and the standard angle. A larger IoU ratio indicates a smaller difference between the current camera's angle and the standard angle; conversely, a smaller IoU ratio indicates a larger difference. Therefore, in this embodiment, step S15 is needed to determine whether the IoU ratio is less than a preset threshold. If it is less than the threshold, it indicates a large difference between the current camera's angle and the standard angle.

[0034] Step S16 can be used to adjust the current camera angle. In this embodiment, the method for adjusting the camera can be of various forms known to those skilled in the art. In one example of the present invention, the method for adjusting the camera may include, for example... Figure 3 The steps shown are described in this. Figure 3 In this context, adjusting the camera may include the following steps: In step S30, the crossover ratio difference between the previous adjustment and the previous two adjustments is calculated; In step S31, it is determined whether the difference in the crossover ratio is greater than 0; In step S32, if the crossover ratio difference is greater than 0, the camera is adjusted using the previous camera adjustment angle. In step S33, if the cross-union ratio is less than 0, the previous camera adjustment angle is used as the reverse solution to generate a new adjustment angle to adjust the camera. Specifically, in this example, the method for generating the adjustment angle may be to first determine the reverse solution interval based on the reverse solution, and then select an angle outside the reverse solution interval as the adjustment angle. Furthermore, in this embodiment, the method for detecting defects in substation equipment using a camera can, for example, incorporate an image recognition model within the camera. This image recognition model can include an input layer, a feature fusion layer, a skip connection layer, a feature stitching layer, and a classification layer. The input layer can be used to input an image. The feature fusion layer can split and combine features from the input image. The skip connection layer can be used to perform feature transformation on the output of the feature fusion layer. The feature stitching layer can be used to stitch together the result of the feature transformation operation and the result before feature transformation. The classification layer can be used to predict the result of the feature stitching, thereby obtaining the recognition result. The feature transformation operation can be implemented using the following formulas (2) and (3): (2) (3) in, This is the exponential moving average of the gradient of the network parameters. The hidden state vector of the previous time step. Given a function for a learnable weight vector, To skip the number of neurons in the connection layer, For kernel function, The weights of the loss function for positive samples. This is the exponential moving average of the squared gradient. The weights are the weights of the loss function for negative samples.

[0035] In another aspect, the present invention also provides a computer-readable storage medium storing instructions for being read by a machine to cause the machine to perform any of the methods described above.

[0036] Through the above technical solutions, the embodiments of the present invention provide a method and system for identifying defects in substation equipment based on video stream analysis. The method and system improve the accuracy of monitoring and identifying substation equipment and the stability of camera operation by combining frame images of the video stream to autonomously correct the shooting angle of the camera.

[0037] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0038] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0039] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0040] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0041] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0042] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0043] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, 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 technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0044] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0045] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for identifying defects in substation equipment based on video stream analysis, characterized in that, include: Retrieve infrared image frames from the current video stream; The actual location of the substation is determined based on the infrared image frame; Calculate the torsion ratio at the actual location; Determine whether the twist ratio is less than or equal to a preset twist ratio threshold; If the distortion ratio is determined to be greater than the distortion ratio threshold, the intersection-union ratio of the actual position and the default anchor box position is calculated; Determine whether the crossover-union ratio is less than a preset threshold; If the cross-union ratio is determined to be less than the threshold, the angle of the current camera is adjusted, and the process returns to the step of acquiring infrared image frames in the current video stream. If the cross-union ratio is determined to be greater than or equal to the threshold, a defect identification operation is performed based on the infrared image frame and the visible light image.

2. The method according to claim 1, characterized in that, Calculating the distortion ratio at the actual location includes: The torsion ratio is calculated according to formula (1): ，（1） in, The twist ratio is... , Let be the pixel width and pixel length of the largest inscribed rectangle at the actual location. , The pixel width and pixel length of the maximum bounding rectangle at the actual location.

3. The method according to claim 1, characterized in that, Retrieve infrared image frames from the current video stream, including: Infrared image frames are extracted from the video stream at preset frame intervals; Determine the overlap between any two infrared image frames; Determine whether the overlap degree is greater than a preset overlap degree threshold; If the overlap is determined to be greater than the overlap threshold, any one of the infrared image frames is output.

4. The method according to claim 1, characterized in that, Determining the actual location of the substation based on the infrared image frame includes: The infrared image frame is binarized. The actual position is determined by using a proportional threshold operation based on the result of the binarization operation.

5. The method according to claim 4, characterized in that, The actual position is determined using a proportional thresholding operation based on the result of the binarization operation, including: Determine the maximum and minimum grayscale values ​​in the current infrared image frame; The suspected region is determined in the current infrared image frame by multiplying a preset ratio value by the maximum value; The suspected area with the largest area is selected as the actual location.

6. The method according to claim 1, characterized in that, The camera has a built-in image recognition model, which includes a skip connection layer. The skip connection layer is used to perform feature transformation operations according to formulas (2) and (3): ，（2） ，（3） in, This is the exponential moving average of the gradient of the network parameters. The hidden state vector of the previous time step. Given a function for a learnable weight vector, To skip the number of neurons in the connection layer, For kernel function, The weights of the loss function for positive samples. This is the exponential moving average of the squared gradient. The weights are the weights of the loss function for negative samples.

7. The method according to claim 1, characterized in that, Adjust the current camera angle, including: Calculate the crossover ratio difference between the previous adjustment and the two previous adjustments; Determine whether the crossover-union ratio difference is greater than 0; If the cross-union ratio difference is greater than 0, the camera is adjusted using the previous camera adjustment angle; If the cross-union ratio is determined to be less than 0, the previous camera adjustment angle is used as the reverse solution to generate a new adjustment angle to adjust the camera.

8. The method according to claim 7, characterized in that, Generate new adjustment angles, including: Determine the inverse solution interval based on the inverse solution; The angle outside the reverse solution interval is selected as the adjustment angle.

9. A defect identification system for substation equipment based on video stream analysis, characterized in that, The system includes a processor configured to perform the method as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that are read by a machine to cause the machine to perform the method as described in any one of claims 1 to 8.

Images