Oil level anomaly detection method and device for intelligent power station

Abstract

The invention discloses an oil level anomaly detection method and device for an intelligent power station, and the method comprises the steps: obtaining image data of a circular sleeve in a transformer, and enabling the circular sleeve to be filled with oil; performing semantic segmentation on the image data by using a semantic recognition model to obtain a semantic image of an oil region in the circular sleeve; processing the semantic image to obtain an oil level line, a starting line and a terminating line of the semantic image, the height of the oil level line and the height between the starting line and the terminating line; wherein the starting line and the ending line are parallel lines relative to the oil level line in the semantic image; the oil level proportion is calculated according to the height of the oil level line and the height between the starting line and the ending line, and whether the oil level is abnormal or not is judged according to the oil level proportion. Oil level anomaly detection of the intelligent power station is achieved through the image recognition technology, dependence on manual inspection experience is reduced, the anomaly detection cost is reduced, and the operation efficiency and safety of the intelligent power station are improved.

Description

Technical Field

[0001] The present invention relates to the field of oil level detection of an oil level gauge, and in particular to a method and device for detecting abnormal oil level in a smart power station. Background Art

[0002] With the vigorous development of power station infrastructure in recent decades, the oil level detection of the circular bushing in the transformer at high altitude has become increasingly frequent. A certain amount of oil in the circular bushing of the transformer can well protect the transformer from temperature drop, heat insulation and other effects. Therefore, how to identify the abnormal oil level of the circular bushing of the transformer to ensure the long-term effective operation of the power station has become an important issue. This not only helps to avoid transformer failures caused by abnormal oil level in the circular bushing, but also improves the operating efficiency and safety of the power station. Therefore, regular detection of the oil level in the circular bushing of the transformer is one of the important measures to ensure the reliability of the national power grid system.

[0003] The detection of oil level abnormality in the circular bushing of traditional power stations mainly relies on the staff to climb the transformer and make a qualitative assessment with the naked eye to determine whether the oil level in the circular bushing is abnormal. However, this operation mode has obvious safety hazards and efficiency problems. First, the staff climb the high-voltage transformer, which itself seriously threatens their life safety. Working in a high-voltage environment, a slight carelessness may cause a serious safety accident. Secondly, it takes a lot of manpower and material resources for personnel to inspect the oil level abnormality. This includes not only the labor input of the staff, but also the cost of various expensive insulating protective clothing. The investment of these costs undoubtedly increases the operating cost of the power station. The qualitative technology of the human eye relies heavily on the long-term experience judgment of the staff, which makes the detection accuracy of the circular bushing oil level not high, which in turn makes it impossible to ensure the stability and effective operation of the national power grid system. Summary of the invention

[0004] In response to the above technical problems, the present invention provides a method and device for detecting oil level anomalies in a smart power station, which can realize abnormal detection of the power station by detecting the oil level.

[0005] A first aspect of the present invention provides a method for detecting oil level abnormality in a smart power station, comprising:

[0006] Acquire image data of a circular bushing in a transformer, wherein the circular bushing contains oil;

[0007] Using a semantic recognition model to perform semantic segmentation on the image data to obtain a semantic image of the oil region in the circular casing;

[0008] Processing the semantic image to obtain an oil level line, a start line and an end line of the semantic image, as well as a height of the oil level line and a height between the start line and the end line; wherein the start line and the end line are parallel lines in the semantic image relative to the oil level line;

[0009] The oil level ratio is calculated according to the height of the oil level line and the height between the start line and the end line, and whether the oil level is abnormal is determined according to the oil level ratio.

[0010] In an optional implementation, the semantic segmentation of the image data using a semantic recognition model to obtain a semantic image of the oil region in the circular casing includes:

[0011] An ellipse fitting edge processing is performed on the semantic image using an image ellipse fitting algorithm to obtain a regular and smooth first semantic image, wherein the outer contour of the first semantic image is circular.

[0012] In an optional implementation, the processing of the semantic image to obtain the oil level line, the start line and the end line of the semantic image, as well as the height of the oil level line and the height between the start line and the end line includes:

[0013] Processing the first semantic image using a gradient local zero-crossing algorithm to obtain a start line and an end line tangent to the first semantic image;

[0014] The height between the start line and the end line is obtained according to the pixel value between the start line and the end line.

[0015] In an optional implementation, the using a gradient local zero-crossing algorithm to process the first semantic image to obtain a start line and an end line tangent to the first semantic image includes:

[0016] Sum the pixels in the row dimension to obtain the total number of pixels in the row dimension of the first semantic image;

[0017] Gradient processing is performed according to the total number of pixels in the row dimension of the first semantic image, and the starting line and the ending line are obtained based on a gradient descent algorithm.

[0018] In an optional implementation, the processing of the semantic image to obtain the oil level line, the start line and the end line of the semantic image, as well as the height of the oil level line and the height between the start line and the end line, further includes:

[0019] Sum the pixels in the row dimension to obtain the total number of pixels in the row dimension of the first semantic image;

[0020] Gradient processing is performed according to the total number of pixels in the row dimension of the first semantic image, and the oil level line is obtained based on a gradient maximum algorithm.

[0021] In an optional implementation manner, the calculating the oil level ratio according to the height of the oil level line and the height between the start line and the end line, and judging whether the oil level is abnormal according to the oil level ratio, includes:

[0022] The ratio of the height of the oil level line to the height between the start line and the end line is calculated, and if the height ratio is less than a preset threshold, it is determined that the oil level is abnormal.

[0023] In an optional implementation manner, if the height ratio is less than a preset threshold, determining that the oil level is abnormal includes:

[0024] If the height ratio is less than 0.5, it is determined that the oil level is abnormal.

[0025] A second aspect of the present invention provides an oil level abnormality detection device for a smart power station, comprising:

[0026] An acquisition module, used for acquiring image data of a circular bushing in a transformer, wherein the circular bushing contains oil;

[0027] A processing module, used for performing semantic segmentation on the image data using a semantic recognition model to obtain a semantic image of the oil area in the circular casing;

[0028] A positioning module, used for processing the semantic image to obtain an oil level line, a start line and an end line of the semantic image, as well as a height of the oil level line and a height between the start line and the end line; wherein the start line and the end line are parallel lines in the semantic image relative to the oil level line;

[0029] The judgment module is used to calculate the oil level ratio according to the height of the oil level line and the height between the start line and the end line, and judge whether the oil level is abnormal according to the oil level ratio.

[0030] According to a third aspect of the present invention, there is provided an electronic device, comprising:

[0031] At least one processor; and at least one memory communicatively connected to the processor, wherein: the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute the method described in the first aspect of the embodiment of the present invention.

[0032] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a computer, the method according to the first aspect of the embodiment of the present invention is executed.

[0033] The present invention uses image recognition technology to realize oil level anomaly detection in smart power stations, reduces reliance on human inspection experience, reduces the cost of anomaly detection, and improves the operating efficiency and safety of smart power stations. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 The figure is a flow chart of an oil level abnormality detection method for a smart power station according to an embodiment of the present invention.

[0035] Figure 2 Schematic diagram of a circular sleeve image in an embodiment of the present invention.

[0036] Figure 3 It is a schematic diagram of obtaining a semantic image of an oil region in a circular casing in an embodiment of the present invention.

[0037] Figure 4 For Figure 3 Schematic diagram of the image after ellipse fitting.

[0038] Figure 5 For Figure 4 Schematic diagram of obtaining the start line and the end line through image processing.

[0039] Figure 6 For Figure 4 Schematic diagram of obtaining the oil level line through image processing.

[0040] Figure 7 The present invention is a schematic diagram of a module of an oil level abnormality detection device for a smart power station according to an embodiment of the present invention. DETAILED DESCRIPTION

[0041] 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 those skilled in the art without creative work are within the scope of protection of the present invention.

[0042] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0043] See also Figure 1 The present invention provides an oil level abnormality detection method for a smart power station, comprising the following steps.

[0044] S10: Acquire image data of a circular bushing in a transformer, wherein the circular bushing is filled with oil.

[0045] The circular casing image can be taken by a preset camera point, or by a drone cruise.

[0046] In one embodiment, a drone is used to take real-time photos of the circular casing in the high-altitude substation to collect video or image data thereof. If the collected data is in video format, the images of the circular casing need to be cropped or intercepted frame by frame to obtain a picture file for detection.

[0047] like Figure 2 As shown, the acquired image is an image of the circular bushing of the transformer, and after preprocessing such as cropping, a partial image of the circular bushing filled with oil is obtained.

[0048] S20: Using a semantic recognition model to perform semantic segmentation on the image data to obtain a semantic image of the oil area in the circular casing.

[0049] When the semantic recognition model is trained, labelme can be used to annotate all collected image data to obtain a semantic image of the circular casing. The annotated data is divided into several groups of training image data, each group of training image data includes semantic feature information of the circular casing and semantic feature information of the oil level line of the circular casing.

[0050] The training data is input into the semantic feature extraction model of the circular sleeve until the training gradient loss floats small and the precision recall rate and accuracy rate are basically kept at the maximum, at which time the circular sleeve semantic feature extraction model is obtained. The sleeve semantic feature extraction model adopts the yolov8-seg model.

[0051] The image data of the circular sleeve with detection is input into the sleeve semantic feature extraction model to obtain the semantic image of the circular sleeve semantic feature information, such as Figure 3 As shown, a semantic graph is obtained.

[0052] S30: Processing the semantic image to obtain the oil level line, the start line and the end line of the semantic image, as well as the height of the oil level line and the height between the start line and the end line; wherein the start line and the end line are parallel lines relative to the oil level line in the semantic image.

[0053] Due to the angle of drone photography, the recognized shape of the circular casing may be irregular, similar to an ellipse, and the original circular shape needs to be restored during calculation.

[0054] Therefore, the present invention uses an image ellipse fitting algorithm to perform ellipse fitting edge processing on the semantic image to obtain a regular and smooth first semantic image.

[0055] In the present invention, an ellipse fitting algorithm is performed on the semantic feature information of the circular sleeve through an image ellipse fitting algorithm to obtain a regular and smooth first semantic image of the circular sleeve, wherein the outer contour of the first semantic image is circular, such as Figure 4 shown.

[0056] In some embodiments, the pixels of the oil region in the semantic image containing the semantic feature information of the circular casing are set to 125, and the pixels of the air above the oil level region in the circular casing are set to 255. The maximum circumscribed contour point of the oil region and the maximum circumscribed contour point of the air region are found through the contour retrieval mode. The maximum circumscribed contour point of the oil region and the maximum circumscribed contour point of the air region are subjected to the ellipse fitting algorithm to obtain the coordinates of the center point of the major axis and the minor axis of the ellipse and the ellipse rotation angle information. The coordinates of the center point of the major axis and the minor axis of the ellipse and the ellipse rotation angle information are regressed to the semantic image containing the semantic feature information of the circular casing to obtain a regular and smooth first semantic image. The first semantic image (oil region) includes the oil region and the air region.

[0057] Then the starting line and the ending line are obtained in the following way.

[0058] In some embodiments, the first semantic image is processed using a local gradient zero-crossing algorithm to obtain a start line and an end line tangent to the first semantic image, and the height between the start line and the end line is obtained according to the pixel value between the start line and the end line.

[0059] For example, the pixels in the oil area in the first semantic image are summed according to the row dimension to obtain S1. Then, a horizontal one-dimensional convolution is constructed to perform gradient calculation on S1 to obtain a series of horizontal lines L1, and the convolution kernel is [1,1,1,1,0,-1,-1,-1,-1]. Then, a local gradient search is performed on L1 to find the first gradient greater than 0. At this time, the starting line L tangent to the circular casing is obtained. start .

[0060] Similarly, the pixels in the row dimension are summed to obtain the total number of pixels in the row dimension of the first semantic image; gradient processing is performed according to the total number of pixels in the row dimension of the first semantic image, and the starting line and the ending line are obtained based on the gradient descent algorithm, such as Figure 5 shown.

[0061] For example, the air area in the first semantic image is summed according to the row dimension pixels to obtain S2. Then a horizontal one-dimensional convolution is constructed to perform gradient on S2 to obtain a series of horizontal lines L2, and the convolution kernel is [1,1,1,1,0,-1,-1,-1,-1]. Then a local gradient search is performed on L2 to find the first gradient greater than 0. At this time, the termination line L tangent to the circular sleeve is obtained. end .

[0062] Furthermore, if Figure 6 As shown, the method for determining the oil level line includes the following steps.

[0063] The row dimension pixels of the first semantic image are summed to obtain the total number of row dimension pixels of the first semantic image. For example, the row dimension pixels of the oil region in the first semantic image of the regular and smooth circular casing are summed to obtain S3. A horizontal one-dimensional convolution is constructed to perform gradient on S3 to obtain a series of horizontal lines L3, and the convolution kernel is [1, 1, 1, 0, -1, -1, -1].

[0064] Then, gradient processing is performed according to the total number of pixels in the row dimension of the first semantic image, and the oil level line is obtained based on the gradient maximum algorithm. For example, a maximum search is performed on L3 to find the location of the maximum value of L3. At this time, the oil level line L of the circular casing is obtained. oil .

[0065] S40: Calculate the oil level ratio according to the height of the oil level line and the height between the start line and the end line, and determine whether the oil level is abnormal according to the oil level ratio.

[0066] In this step, the height h of the oil level line can be calculated (h is the height of the oil level line L oil Height and starting line L start The difference between the heights) and the starting line L start and the termination line L endIf the height ratio is less than a preset threshold, the oil level is judged to be abnormal. start and the termination line L end The height between is the height H of the first semantic image. The oil level ratio r=h / H.

[0067] In some embodiments, if the height ratio is less than 0.5, the oil level is judged to be abnormal. When r is less than 0.5, the oil level is abnormal and the system will issue a reminder message; when r is greater than 0.5, the oil level is normal.

[0068] Through the above method, it is possible to use image recognition technology to complete the oil level anomaly detection of the smart power station, reduce the dependence on human inspection experience, reduce the cost of anomaly detection, and improve the operating efficiency and safety of the smart power station.

[0069] like Figure 7 As shown, the present invention also provides an oil level abnormality detection device for a smart power station, including: an acquisition module 71, a processing module 72, a positioning module 73, and a judgment module 74.

[0070] The acquisition module 71 is used to acquire image data of a circular bushing in a transformer, wherein the circular bushing is filled with oil.

[0071] The processing module 72 is used to perform semantic segmentation on the image data using a semantic recognition model to obtain a semantic image of the oil area in the circular casing.

[0072] In one implementation, an image ellipse fitting algorithm is used to perform ellipse fitting edge processing on the semantic image to obtain a regular and smooth first semantic image, wherein the outer contour of the first semantic image is circular.

[0073] The positioning module 73 is used to process the semantic image to obtain the oil level line, the starting line and the ending line of the semantic image, as well as the height of the oil level line and the height between the starting line and the ending line; wherein the starting line and the ending line are parallel lines relative to the oil level line in the semantic image.

[0074] In one embodiment, the first semantic image is processed using a gradient local zero-crossing algorithm to obtain a start line and an end line tangent to the first semantic image; the height between the start line and the end line is obtained according to the pixel value between the start line and the end line. The pixels in the row dimension are summed to obtain the total number of pixels in the row dimension of the first semantic image; gradient processing is performed based on the total number of pixels in the row dimension of the first semantic image, and the start line and the end line are obtained based on a gradient descent algorithm.

[0075] Furthermore, the pixels in the row dimension are summed to obtain the total number of pixels in the row dimension of the first semantic image; gradient processing is performed according to the total number of pixels in the row dimension of the first semantic image, and the oil level line is obtained based on a gradient maximum algorithm.

[0076] The judgment module 74 is used to calculate the oil level ratio according to the height of the oil level line and the height between the start line and the end line, and judge whether the oil level is abnormal according to the oil level ratio.

[0077] In one embodiment, the ratio of the height of the oil level line to the height between the start line and the end line is calculated, and if the height ratio is less than a preset threshold, the oil level is judged to be abnormal. For example, if the height ratio is less than 0.5, the oil level is judged to be abnormal.

[0078] Corresponding to the above method, such as Figure 7 As shown, the present invention also provides an oil level abnormality detection device for a smart power station, comprising:

[0079] The functions or methods performed by the above modules can refer to the description of the oil level abnormality detection method of the above smart power station, which will not be repeated here.

[0080] The present invention also provides an electronic device, comprising:

[0081] At least one processor; and at least one memory in communication with the processor, wherein: the memory stores program instructions that can be executed by the processor, and the processor calls the program instructions to execute the above-mentioned oil level abnormality detection method of the smart power station.

[0082] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the oil level abnormality detection method of the above-mentioned smart power station is implemented.

[0083] It can be understood that computer-readable storage media may include: any entity or device capable of carrying a computer program, recording media, USB flash drives, mobile hard disks, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media, etc. A computer program includes computer program code. The computer program code may be in source code form, object code form, executable file, or some intermediate form, etc. A computer-readable storage medium may include: any entity or device capable of carrying a computer program code, recording media, USB flash drives, mobile hard disks, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), and software distribution media, etc.

[0084] In certain embodiments of the present invention, the device may include a controller, which is a single-chip microcomputer chip that integrates a processor, a memory, a communication module, etc. The processor may refer to the processor included in the controller. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc.

[0085] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

[0086] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0087] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting oil level abnormality in a smart power station, characterized in that: include: Acquire image data of a circular bushing in a transformer, wherein the circular bushing contains oil; Using a semantic recognition model to perform semantic segmentation on the image data to obtain a semantic image of the oil region in the circular casing; Processing the semantic image to obtain an oil level line, a start line and an end line of the semantic image, as well as a height of the oil level line and a height between the start line and the end line; wherein the start line and the end line are parallel lines in the semantic image relative to the oil level line; The oil level ratio is calculated according to the height of the oil level line and the height between the start line and the end line, and whether the oil level is abnormal is determined according to the oil level ratio.

2. The oil level abnormality detection method of a smart power station according to claim 1, characterized in that: The method of performing semantic segmentation on the image data by using a semantic recognition model to obtain a semantic image of the oil area in the circular casing includes: An ellipse fitting edge processing is performed on the semantic image using an image ellipse fitting algorithm to obtain a regular and smooth first semantic image, wherein the outer contour of the first semantic image is circular.

3. The oil level abnormality detection method of a smart power station according to claim 2, characterized in that: The processing of the semantic image to obtain the oil level line, the start line and the end line of the semantic image, as well as the height of the oil level line and the height between the start line and the end line, includes: Processing the first semantic image using a gradient local zero-crossing algorithm to obtain a start line and an end line tangent to the first semantic image; The height between the start line and the end line is obtained according to the pixel value between the start line and the end line.

4. The oil level abnormality detection method of a smart power station according to claim 3, characterized in that: The step of processing the first semantic image by using a local gradient zero-crossing algorithm to obtain a start line and an end line tangent to the first semantic image includes: Sum the pixels in the row dimension to obtain the total number of pixels in the row dimension of the first semantic image; Gradient processing is performed according to the total number of pixels in the row dimension of the first semantic image, and the starting line and the ending line are obtained based on a gradient descent algorithm.

5. The oil level abnormality detection method of a smart power station according to claim 2 or 3, characterized in that: The processing of the semantic image to obtain the oil level line, the start line and the end line of the semantic image, as well as the height of the oil level line and the height between the start line and the end line, further includes: Sum the pixels in the row dimension to obtain the total number of pixels in the row dimension of the first semantic image; Gradient processing is performed according to the total number of pixels in the row dimension of the first semantic image, and the oil level line is obtained based on a gradient maximum algorithm.

6. The oil level abnormality detection method of a smart power station according to claim 1, characterized in that: The calculating the oil level ratio according to the height of the oil level line and the height between the start line and the end line, and judging whether the oil level is abnormal according to the oil level ratio, includes: The ratio of the height of the oil level line to the height between the start line and the end line is calculated, and if the height ratio is less than a preset threshold, it is determined that the oil level is abnormal.

7. The oil level abnormality detection method of a smart power station according to claim 6, characterized in that: If the height ratio is less than a preset threshold, determining that the oil level is abnormal includes: If the height ratio is less than 0.5, it is determined that the oil level is abnormal.

8. An oil level abnormality detection device for a smart power station, characterized in that: include: An acquisition module, used for acquiring image data of a circular bushing in a transformer, wherein the circular bushing contains oil; A processing module, used for performing semantic segmentation on the image data using a semantic recognition model to obtain a semantic image of the oil area in the circular casing; A positioning module, used for processing the semantic image to obtain the oil level line, the start line and the end line of the semantic image, as well as the height of the oil level line and the height between the start line and the end line; wherein the starting line and the ending line are parallel lines relative to the oil level line in the semantic image; The judgment module is used to calculate the oil level ratio according to the height of the oil level line and the height between the start line and the end line, and judge whether the oil level is abnormal according to the oil level ratio.

9. An electronic device, characterized in that: include: at least one processor; And at least one memory communicatively connected to the processor, wherein: the memory stores program instructions executable by the processor, and the processor calls the program instructions to execute the oil level abnormality detection method for the smart power station as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a computer, the oil level abnormality detection method for a smart power station as claimed in any one of claims 1 to 7 is executed.

Images