A method and system for evaluating the operating state of a mutual inductor

Abstract

The present application relates to the technical field of measurement, in particular to a kind of mutual inductor operating state evaluation method and system.The method includes: obtaining the gray scale diagram of the ultraviolet image of mutual inductor after being irradiated by ultraviolet light source;The skeleton isolated function of each pixel point in the gray scale diagram is obtained by analyzing the gray scale diagram;The neighborhood structure consistency of each pixel point is obtained;The enhancement coefficient of each pixel point is calculated;The product of enhancement coefficient and the gray scale value of each pixel point is used as the enhanced gray scale value, and then the enhanced gray scale diagram is obtained;The enhanced gray scale diagram is binarized and segmented, and the light spot region is obtained to determine the discharge condition of mutual inductor.The scheme of the present application can accurately and effectively identify the discharge of mutual inductor, so as to accurately evaluate the operating state of mutual inductor.

Description

Technical Field

[0001] This invention relates to the field of measurement technology. More specifically, this invention relates to a method and system for evaluating the operating status of a current transformer. Background Technology

[0002] An instrument transformer is a special type of transformer that reduces high voltage and large current to low voltage and small current signals through a set ratio to transmit information to measuring instruments, meters, and protection devices; at the same time, instrument transformers are also the most widely used primary equipment.

[0003] Current transformers can be classified into three main categories: electromagnetic, capacitive, and electronic. Currently, capacitive voltage transformers are difficult to verify according to prescribed periods. There are three main reasons for this:

[0004] 1. Power outages at substations are difficult to manage. Even if there are plans for power outage maintenance, the focus is mainly on testing the insulation performance of equipment, and it is difficult to schedule time for testing the error characteristics of instrument transformers.

[0005] 2. As voltage levels increase, the size and weight of equipment for on-site testing of instrument transformer error characteristics increase, significantly increasing the difficulty, labor intensity, and implementation costs of on-site work.

[0006] 3. The rate of increase in relevant personnel is far lower than the rate of expansion of the power grid, making it difficult to continue implementing the traditional working mode of testing the error characteristics of instrument transformers.

[0007] Therefore, the method of using ultraviolet radiation to assess the operating status of current transformers has the advantages of saving manpower and being convenient and intuitive.

[0008] However, traditional image recognition algorithms cannot effectively identify small spots or pixels near the ultraviolet light map of a current transformer in a discharged state. They are easily confused with background noise, and during filtering, small spots are easily filtered out as well, affecting the recognition effect and making it impossible to accurately assess the operating status of the current transformer.

[0009] Therefore, there is a need for an image processing method that can effectively distinguish between small light spots and noise, and only enhance the light spots, in order to accurately identify whether the current transformer is discharging, so as to assess the operating status of the current transformer. Summary of the Invention

[0010] The purpose of this invention is to provide a method and system for evaluating the operating status of current transformers, in order to solve the problem that the current technology cannot accurately and effectively identify the discharge of current transformers; to this end, this invention provides solutions in the following two aspects.

[0011] In a first aspect, the present invention provides a method for evaluating the operating status of a current transformer, comprising:

[0012] Obtain a grayscale image of the current transformer after it has been irradiated by an ultraviolet light source.

[0013] The grayscale image is analyzed to obtain the skeleton isolation function of each pixel in the grayscale image; the skeleton isolation function is negatively correlated with the normalized value of the minimum skeleton distance of the corresponding pixel and the isolation, and positively correlated with the grayscale value of the corresponding pixel; the isolation characterizes the difference between any pixel and the pixels in its neighborhood.

[0014] Obtain the neighborhood structure consistency of each pixel; the neighborhood structure consistency is negatively correlated with the difference between any two pixels in the neighborhood centered on any pixel.

[0015] Calculate the enhancement coefficient for each pixel. The enhancement coefficient is positively correlated with the skeleton isolation function and negatively correlated with the neighborhood structure consistency.

[0016] The product of the enhancement coefficient and the gray value of each pixel is used as the enhanced gray value, thus obtaining the enhanced grayscale image;

[0017] The enhanced grayscale image is binarized and segmented to obtain the spot region, so as to determine the discharge status of the transformer.

[0018] The above scheme analyzes the grayscale image of the ultraviolet image captured by the current transformer to obtain the skeleton isolation function and neighborhood structure consistency of each pixel in the grayscale image. Based on the skeleton isolation function and neighborhood structure consistency, the enhancement coefficient of each pixel is determined to enhance the pixels in the grayscale image. Then, the spot area is extracted from the enhanced grayscale image, which can accurately identify the discharge status of the current transformer.

[0019] Optionally, the enhancement coefficient for:

[0020] ;

[0021] in, It is a pixel. The skeleton isolation function, It is a pixel. Neighborhood structure consistency It is the first coefficient. It is the second coefficient. In pixels The neighborhood centered on it.

[0022] The above scheme proposes a method for accurately calculating the enhancement coefficient.

[0023] Optionally, the skeleton isolation function for:

[0024] ;

[0025] in, It is a pixel. The skeleton isolation function, It is a pixel. The normalized value of the minimum skeleton distance. , These are setting parameters, It is a pixel. grayscale value, It is the maximum gray value of a pixel in a grayscale image. It's about adjusting parameters. For pixels Due to the isolation of exp(), it is an exponential function.

[0026] The above scheme proposes a method for accurately calculating the skeleton isolation function.

[0027] Optionally, the neighborhood structure is consistent. for:

[0028] ;

[0029] in, Based on pixels The neighborhood centered on, and This represents the grayscale values ​​of the first and second pixels in the i-th pair of pixels in the neighborhood. and This represents the coordinates of the first pixel and the second pixel. It is an adjustment parameter for the tolerance of grayscale differences. It is the total number of pixel pairs in the neighborhood, and exp() is an exponential function; the pixel pair is composed of any two pixels selected without repetition from all pixels in the neighborhood.

[0030] Optionally, the isolation is the ratio of the number of pixels with differences greater than a difference threshold to the total number of pixels in the neighborhood; the difference is the absolute value of the difference between the grayscale values ​​of any pixel and all pixels in its neighborhood.

[0031] The aforementioned neighborhood structure consistency can characterize the differences between any pixel and its neighboring pixels.

[0032] Optionally, the neighborhood can be an eight-neighborhood or a circular region centered on any pixel with a neighborhood radius of r.

[0033] Optionally, the step of binarizing and segmenting the enhanced grayscale image to obtain the spot region, in order to determine the discharge status of the transformer, includes:

[0034] Set the segmentation threshold;

[0035] The gray values ​​of pixels with values ​​greater than the segmentation threshold are set to 1, and the gray values ​​of pixels with values ​​less than or equal to the segmentation threshold are set to 0, thus obtaining a binary image.

[0036] Edge detection is performed on the binary image to output a spot image;

[0037] If the area of ​​the spot region in the spot image is greater than the area threshold, then the transformer is discharging.

[0038] Optionally, the skeleton distance is obtained using a skeleton extraction algorithm based on distance transformation.

[0039] Optionally, the process of acquiring the ultraviolet image is as follows:

[0040] The current transformer is measured using an infrared thermometer. When the temperature difference between the next moment and the temperature of the adjacent previous moment is greater than the set difference temperature, an ultraviolet imager is used to capture an ultraviolet image of the current transformer.

[0041] The above scheme uses temperature to perform preliminary screening of the ultraviolet images of the current transformer, which can improve the efficiency of discharge identification.

[0042] In the second aspect, a current transformer operating status assessment system includes:

[0043] processor;

[0044] The memory stores computer instructions for evaluating the operating status of the current transformer. When the computer instructions are executed by the processor, the system performs the aforementioned method for evaluating the operating status of the current transformer.

[0045] The beneficial effects of this invention are as follows:

[0046] The solution of the present invention effectively enhances the spot area and does not enhance or slightly enhances the background and its noise, so that the enhanced grayscale image can accurately identify the discharge status of the transformer, thereby effectively evaluating the operating status of the transformer. Attached Figure Description

[0047] Figure 1 This schematically illustrates a flowchart of the steps in a current transformer operation status assessment method according to this embodiment;

[0048] Figure 2 The schematic diagram illustrates the structural block diagram of a current transformer operation status assessment system in this embodiment. Detailed Implementation

[0049] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0050] like Figure 1 As shown, a method for evaluating the operating status of a current transformer in this embodiment includes the following steps:

[0051] Step S1: Obtain the grayscale image of the current transformer after it has been irradiated by an ultraviolet light source.

[0052] In this embodiment, by turning on the "sun-blind mode" of the ultraviolet imager (e.g., OFIL SuperB) to automatically filter sunlight, and pointing the lens at the current transformer at a distance of 5-10 meters, the front, left, and right sides of the current transformer's ceramic sleeve and umbrella skirt can be captured for imaging.

[0053] During the shooting process, when small blue-white spots appear on the screen, it is suspected to be a discharge. At this time, the temperature of the current transformer is measured using an infrared thermometer. When the temperature difference between the next moment and the adjacent previous moment is greater than 0.5℃, it is confirmed to be a discharge. The ultraviolet image at this time is recorded. It should be noted that a camera with a higher resolution should be used as much as possible, as it has an important impact on the subsequent spot detection.

[0054] In this embodiment, the selected ultraviolet images are converted into grayscale images, and Gaussian filtering is used to filter random noise from the grayscale images to obtain filtered grayscale images.

[0055] Step S2: Enhance the grayscale values ​​of each pixel in the grayscale image to obtain an enhanced grayscale image.

[0056] In this embodiment, enhancement coefficients are obtained based on a pre-constructed skeleton isolation function and neighborhood structure consistency. The enhancement coefficients are then used to enhance the grayscale values ​​of each pixel in the grayscale image to obtain enhanced grayscale values.

[0057] The skeleton isolation function is derived from the minimum skeleton distance, pixel isolation, and grayscale features obtained by the distance-transform-based skeleton extraction algorithm.

[0058] Specifically, the process of obtaining the skeleton isolation function is as follows:

[0059] First, obtain the isolation of each pixel in the grayscale image.

[0060] In this embodiment, the difference between the gray values ​​of each pixel and its neighboring pixels is obtained, the difference is compared with the difference threshold, and the ratio of the number of differences greater than the difference threshold to the total number of neighboring pixels is used as the isolation of the corresponding pixel.

[0061] It should be noted that small and large light spots will appear in the ultraviolet image when the transformer is discharging. Small light spots are areas with fewer than 4 pixels, and large light spots are areas with more than or equal to 4 pixels. Generally speaking, isolated small light spots and the edges of light spots are often difficult to detect. Therefore, it is necessary to analyze the isolation of each pixel.

[0062] The aforementioned neighborhood can be an eight-neighborhood or a circular region with a radius r centered on any pixel. Pixels within the neighborhood are either pixels within the eight-neighborhood or pixels within the circular region with a radius r centered on any pixel. The neighborhood radius can be a length of 5 or more pixels, specifically 5 or 7.

[0063] The difference mentioned above is the absolute value of the difference between the gray values ​​of any pixel and all pixels in its neighborhood; the difference threshold can be 15, or it can be determined according to the actual situation.

[0064] Next, calculate the minimum skeleton distance.

[0065] When the spot size is particularly small (small spot size), the signal characteristics of the spot itself are interfered with by noise and artifacts from dust on the camera equipment. The feature values ​​of these three are all high and similar, making it difficult to accurately distinguish between the spot and the interference. However, in reality, small spots and noise are still different. Small spots have a distinct central bright spot with high central brightness and a large gray value, while noise has a gray value similar to the background, low central brightness, and a small gray value.

[0066] Therefore, in this embodiment, a skeleton extraction algorithm based on distance transform is introduced. A grayscale threshold is set and the image is binarized. A median transform is performed on the binary image to obtain the median distance field and the centerline image. Then, the centerline image is thinned to obtain the skeleton region. The coordinates of the center point of the skeleton region are used as the coordinates of the skeleton region. Subsequent operations can be performed on the grayscale image, that is, the minimum skeleton distance from each pixel to the nearest skeleton is calculated, and the normalized value of the minimum skeleton distance is obtained, denoted as […]. .

[0067] The normalized value is the ratio of the minimum skeleton distance to the set neighborhood radius. The set neighborhood radius is the radius of the circular region mentioned above; it is introduced to eliminate the influence of the dimensions of the skeleton distance.

[0068] The coordinates of the skeleton mentioned above are the center point of the "skeleton" in the extracted grayscale image. For example, the skeleton of a circular light spot can be the center of the circular light spot.

[0069] It should be noted that since the skeleton is generated by the central axis of the bright area connected structure in the grayscale image, it only exists in obvious discharge structures (such as the area of ​​a large light spot) and not in the background. Therefore, for a large light spot (with 4 or more pixels), each light spot will form a skeleton; however, when the number of pixels in a light spot is less than 4 (smaller light spot), the light spot will not form a skeleton. In this case, the pixels in the smaller light spot will be mixed with the pixels in the background area and noise pixels, making them difficult to distinguish.

[0070] Among them, the skeleton extraction algorithm based on distance transform is a technique based on mathematical morphology theory. Its specific steps include: (i) performing initial binarization on the image; (ii) classifying the pixels in the binarized image to obtain interior points, boundary points, and isolated points; (iii) calculating the distance from each interior point to a non-interior point and assigning the distance to the pixel; (iv) binarizing the assigned image until the binarization result of the next assigned image is all 0.

[0071] Since the algorithm for skeleton extraction based on distance transformation is existing technology, it will not be described in detail here.

[0072] In this embodiment, the distance between pixels belonging to the background, noise, and smaller light spots and the nearest skeleton, i.e., the smallest skeleton distance, must be relatively large.

[0073] Then, construct the skeleton isolation function.

[0074] This embodiment constructs a skeleton isolation function based on the minimum skeleton distance between pixels, pixel isolation, and grayscale features. The aim is to increase the difference between the light spot, background, and noise.

[0075] Specifically, the skeleton isolation function is:

[0076] ;

[0077] in, It is a pixel. The skeleton isolation function, It is a pixel. The normalized value of the minimum skeleton distance. , These are setting parameters, It is a pixel. grayscale value, It is the maximum gray value of a pixel in a grayscale image. It's about adjusting parameters. For pixels Due to the isolation of exp(), it is an exponential function.

[0078] in, Its function is to control The influence It is a constant greater than 0, with an empirical value of 0.7. This is a setting parameter for adjusting the influence of isolation; an empirical value of 0.6 is used. Used for magnification Take 2 experience points.

[0079] It should be noted that since the skeleton only forms in the area where the larger light spot is located, for pixels in the background area, the distance from that pixel to the nearest skeleton is relatively large. The size is relatively large, and due to the low isolation of the background area, Smaller, therefore, formula In this case, both the numerator and denominator are relatively large. This will result in slight changes (i.e., a slight increase or decrease); at the same time, the grayscale value of the background is significantly lower, therefore... It was also lower, resulting in The value is also low.

[0080] For pixels belonging to noise, the distance from the nearest skeleton is also relatively large. In this case, the skeleton distance is large, and the isolation of noise pixels is generally high. The value is the largest, therefore The noise level decreased significantly; and the grayscale value of the noise was also significantly lower (common noises in ultraviolet images include thermal noise from sensor thermal current fluctuations, reflections from the inner surface of the ultraviolet lens, optical scattering interference noise from stray light, and background ultraviolet stray light noise from non-discharge light sources in the environment; these noises are relatively common and their grayscale values ​​are lower than those of the light spot), therefore... It is also lower, thus... The value will be the lowest.

[0081] For pixels within a larger light spot area (when the number of pixels is greater than or equal to 4), the resulting skeleton is located on or near the pixel itself. Approaching 0, It tends to 1, and its brightness is almost the highest across the entire image, therefore tending towards 1, Maximum, output Its value is also the highest.

[0082] For pixels in a smaller spot area (less than 4 pixels), because there are too few pixels in this smaller spot area to form a skeleton, the distance from the pixel to the nearest skeleton is relatively large. The light spot is relatively large, and the isolation of the light spot area is relatively high. Larger, therefore The value is the largest, therefore A significant decrease. However, because the gray values ​​of pixels within a smaller spot area are larger than the gray values ​​of pixels in the noise or background, at this time... It acted as a correction factor, and the high brightness made... Larger, resulting in The value is actually between the background value and the background value. and areas with larger light spots Between, therefore, utilizing It can increase the contrast between light spot and noise.

[0083] For the above four regions The order is as follows: noise < background < smaller spot < larger spot. Among them, Setting it to 2 makes the pixels in the smaller light spot area correspond to... Corresponding to pixels in the background area The differences are significant.

[0084] It should be noted that the purpose of introducing the skeleton extraction algorithm based on distance transformation is to distinguish between pixels belonging to the light spot and the background and noise, especially for the small light spot pixels and the background and noise, so as to achieve better distinction and facilitate the subsequent identification of transformer discharge.

[0085] The process of obtaining neighborhood structure consistency in this embodiment is as follows:

[0086] First, obtain the neighborhood of each pixel. Pair all pixels in the neighborhood to obtain different pixel pairs. Calculate the grayscale difference between the two pixels in each pixel pair. Based on the grayscale difference, obtain the neighborhood structure consistency.

[0087] The pairwise pairing mentioned above involves selecting any two distinct pixels from all pixels without repetition, which is all possible combinations of selecting two pixels from all pixels.

[0088] Specifically, each pixel is known to have a circular region centered on that pixel. If this neighborhood contains a discharge spot, the high brightness of the spot relative to the background will disrupt the structural continuity of the neighborhood. Therefore, calculating the structural consistency of a neighborhood may also reflect whether the neighborhood contains a spot, and it is necessary to calculate the neighborhood structural consistency. Specifically:

[0089] ;

[0090] in, It is the neighborhood centered on pixel I(x,y). and This represents the grayscale values ​​of the first and second pixels in the i-th pair of pixels in the neighborhood. and This represents the coordinates of the first pixel and the second pixel. It is a parameter representing the tolerance for grayscale differences. It is the total number of pixel pairs in the neighborhood.

[0091] In the background area, the grayscale changes of pixels in the neighborhood are relatively smooth and regular; however, if there is a spot of discharge, the grayscale changes will be inconsistent and chaotic. Therefore, by traversing several pixel pairs from a neighborhood, their grayscale differences will show large fluctuations. The grayscale difference of the selected i-th pair of traversed pixels; It is an adjustment setting parameter that controls the tolerance for grayscale differences.

[0092] The aforementioned neighborhood structure consistency is negatively correlated with the grayscale difference among all pixel pairs within the neighborhood; that is, the greater the grayscale difference, the smaller the neighborhood structure consistency of the corresponding pixel. Specifically, as the number of pixels in the neighborhood of any given pixel that have significant differences from that given pixel increases, The value first decreases and then increases (from a value close to 1 to close to 0, and then increases back to close to 1). The specific reason is that if the gray value of any pixel's neighboring pixels differs from its gray value, then any pixel's neighboring pixels may contain pixels belonging to the light spot; as the number of pixels belonging to the light spot increases, the gray value difference between pixel pairs increases and then decreases again.

[0093] The aforementioned significant difference can be determined by comparing the difference between two pixels with a difference threshold. When the difference exceeds the difference threshold, the grayscale values ​​of the two pixels are considered to have a significant difference. The difference threshold can be set to 15, but it can also be determined based on the specific circumstances.

[0094] In this embodiment, the enhancement coefficient for:

[0095] ;

[0096] in, It is a pixel. The skeleton isolation function, It is a pixel. Neighborhood structure consistency It is the first coefficient. It is the second coefficient.

[0097] in, It is the first coefficient of the skeleton isolation function, controlling The empirical coefficient for enhancing the isolation function of pixel skeletons is 2.0; yes The second coefficient controls the overall enhancement of the disturbance region of the control structure, with an empirical coefficient of 1.2.

[0098] The aforementioned enhancement coefficient increases as the number of pixels in the neighborhood of any given pixel (pixels in this neighborhood are those that differ significantly from any given pixel, i.e., those that may belong to the spot area) increases. The value of is getting larger and larger. The value first decreases and then increases. The value first increases and then decreases. At this point, the pixels that are enhanced more strongly are those close to the skeleton and those within the skeleton region; while for pixels that are far from the skeleton region, the enhancement is smaller.

[0099] The enhancement level of pixels near the skeleton differs from that of pixels within the skeleton region. Pixels near the skeleton may be edge pixels of the skeleton region, in which case their grayscale value is significantly enhanced, enabling identification of whether the pixel truly belongs to the skeleton region (which corresponds to the spot area).

[0100] For example, when pixel When the neighborhood of a pixel does not contain pixels with significant differences (i.e., the pixels in the neighborhood are background and / or noise), then... The value is small. A value close to 1 indicates no enhancement or only a slight enhancement for that pixel.

[0101] Based on the degree of isolation of pixels and the consistency of neighborhood structure, the above method enhances each pixel in the grayscale image to different degrees, achieving a significant enhancement of pixels near the spot area, while not enhancing or only slightly enhancing pixels belonging to the background and / or noise. This increases the difference between the spot, background, and noise, highlighting the purpose of the spot.

[0102] Specifically, the enhanced grayscale value of the pixel for:

[0103] ;

[0104] in, It is a pixel. grayscale value, It is a pixel. The enhancement coefficient.

[0105] Step S3: The enhanced grayscale image is binarized and segmented to obtain the spot area, so as to determine the discharge status of the transformer.

[0106] For the enhanced grayscale image, if ≥ If so, the corresponding grayscale value will be assigned the value 1. < If the grayscale value is 0, then the corresponding grayscale value will be assigned to 0, and a binary image will be output.

[0107] Wherein, the segmentation threshold is set ,in It is the average gray level of the enhanced image. It is the standard deviation of grayscale. yes The coefficient is empirically 1.5.

[0108] Edge detection is performed on the binary image to output a spot image. If the area of ​​all spots in the spot image is greater than the area threshold, the transformer has a discharge abnormality and needs to be shut down for maintenance immediately.

[0109] The edge detection uses the Canny edge detection algorithm, which can detect many small circular regions, i.e., light spots, in a binary image.

[0110] The total area of ​​the aforementioned circular regions can characterize abnormal discharge conditions in the instrument transformer. When the area is small, the transformer exhibits low-level discharge, which may correspond to normal operation or an early warning state, requiring only regular monitoring. When the area is large, it may indicate deteriorated insulation performance or approaching a fault, necessitating maintenance measures or immediate shutdown for repair. Therefore, this embodiment sets an area threshold to determine normal or abnormal discharge points in the instrument transformer, accurately assessing its operating status.

[0111] The area threshold mentioned above can be obtained by enhancing the grayscale image of a current transformer in normal operation (without discharge) and obtaining the corresponding spot image, and then using the average area of ​​the spot in the spot image as the area threshold.

[0112] The solution of the present invention can significantly enhance pixels near the spot area by obtaining the skeleton isolation function and neighborhood structure consistency, while not enhancing or only slightly enhancing other pixels (pixels belonging to the background and / or noise), thereby achieving the purpose of highlighting the spot and avoiding the problem that small spots are not easy to detect. It can effectively identify the discharge status of subsequent transformers.

[0113] This invention also provides a system for evaluating the operating status of a current transformer. For example... Figure 2 As shown, the system includes a processor and a memory, the memory storing computer program instructions, which, when executed by the processor, implement the current transformer operating status evaluation method according to the present invention.

[0114] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and therefore will not be described in detail here.

[0115] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this invention can be implemented by computer-readable / executable instructions stored or otherwise maintained on such a computer-readable medium.

[0116] In the description of this specification, "multiple" means at least two, such as two, three or more, etc., unless otherwise expressly and specifically defined.

[0117] While various embodiments of the invention have been shown and described in this specification, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention.

Claims

1. A method for evaluating the operating status of a current transformer, characterized in that, include: Obtain a grayscale image of the current transformer after it has been irradiated by an ultraviolet light source. Analysis of the grayscale image yields the skeleton isolation function for each pixel: ;in, It is a pixel. The skeleton isolation function, It is a pixel. The normalized value of the minimum skeleton distance. , These are setting parameters, It is a pixel. grayscale value, It is the maximum gray value of a pixel in a grayscale image. It's about adjusting parameters. For pixels The isolation of the skeleton is represented by exp(), which is an exponential function. The skeleton isolation function is negatively correlated with the normalized value of the minimum skeleton distance of the corresponding pixel and the isolation, and positively correlated with the gray value of the corresponding pixel. The isolation characterizes the difference between any pixel and the pixels in its neighborhood. To obtain the neighborhood structure consistency of each pixel: ;in, Based on pixels The neighborhood centered on, , This represents the grayscale values ​​of the first and second pixels in the i-th pair of pixels in the neighborhood. , This represents the coordinates of the first pixel and the second pixel. It is an adjustment parameter for the tolerance of grayscale differences. It is the total number of pixel pairs in the neighborhood; neighborhood structure consistency is negatively correlated with the difference between any two pixels in the neighborhood centered on any given pixel. The enhancement coefficient of each pixel is calculated. The enhancement coefficient is positively correlated with the skeleton isolation function and negatively correlated with the neighborhood structure consistency. The product of the enhancement coefficient and the gray value of each pixel is used as the enhanced gray value, thus obtaining the enhanced grayscale image; The enhanced grayscale image is binarized and segmented to obtain the spot region, so as to determine the discharge status of the transformer.

2. The method for evaluating the operating status of a current transformer according to claim 1, characterized in that, The enhancement coefficient for: ； in, It is a pixel. The skeleton isolation function, It is a pixel. Neighborhood structure consistency It is the first coefficient. It is the second coefficient. In pixels The neighborhood centered on it.

3. The method for evaluating the operating status of a current transformer according to claim 1, characterized in that, The pixel pair consists of any two pixels selected without repetition from all pixels in the neighborhood.

4. The method for evaluating the operating status of a current transformer according to claim 1, characterized in that, The isolation is the ratio of the number of pixels with differences greater than the difference threshold to the total number of pixels in the neighborhood; the difference is the absolute value of the difference between the grayscale values ​​of any pixel and all pixels in its neighborhood.

5. The method for evaluating the operating status of a current transformer according to claim 1, characterized in that, The neighborhood can be an eight-neighborhood or a circular region centered on any pixel with a radius of r.

6. The method for evaluating the operating status of a current transformer according to claim 1, characterized in that, The step of binarizing and segmenting the enhanced grayscale image to obtain the spot region, in order to determine the discharge status of the transformer, includes: Set the segmentation threshold; The gray values ​​of pixels with values ​​greater than the segmentation threshold are set to 1, and the gray values ​​of pixels with values ​​less than or equal to the segmentation threshold are set to 0, thus obtaining a binary image. Edge detection is performed on the binary image to output a spot image; If the area of ​​all spot regions in the spot image is greater than the area threshold, the transformer has a discharge abnormality and needs to be shut down for maintenance immediately.

7. The method for evaluating the operating status of a current transformer according to claim 1, characterized in that, The skeleton distance is obtained using a skeleton extraction algorithm based on distance transformation.

8. The method for evaluating the operating status of a current transformer according to claim 7, characterized in that, The process of acquiring the ultraviolet image is as follows: The current transformer is measured using an infrared thermometer. When the temperature difference between the next moment and the temperature of the adjacent previous moment is greater than the set difference temperature, an ultraviolet imager is used to capture an ultraviolet image of the current transformer.

9. A current transformer operating status evaluation system, characterized in that, include: processor; A memory storing computer instructions for evaluating the operating status of a current transformer, which, when executed by the processor, cause the system to perform a method for evaluating the operating status of a current transformer according to any one of claims 1-8.

Images