A method, device and medium for positioning a hot area of an electric energy metering box based on operating state recognition

Abstract

The application discloses a kind of based on operating state identification electric energy metering box hot area positioning method, comprising the following steps: according to the historical operation data of electric energy metering box, clustering algorithm is realized to divide working state;According to the working state of division, operating state identification model is constructed;The working state of electric energy metering box in real-time operating state is identified by operating state identification model, and the reliability evaluation of real-time operating state is carried out;When the reliability evaluation of real-time operating state changes, the infrared image of electric energy metering box in the time range before and after the change of reliability evaluation is acquired, and the infrared image is positioned in hot area;According to the hot area of positioning, its possible area of thermal fault is judged.The present application can find the change of reliability evaluation in time, and analyze infrared image when change occurs, and accurately locate the area of thermal fault.

Description

Technical Field

[0001] This invention relates to the field of electricity metering box technology, and in particular to a method, device, and medium for locating the thermal zone of an electricity metering box based on operating status identification. Background Technology

[0002] With the rapid development of the power industry, the requirements for the accuracy and reliability of electricity metering devices are constantly increasing. As an indispensable basic device in the power market technical support system, the monitoring of the operating status and fault diagnosis of electricity metering boxes are particularly important. Currently, the inspection of electricity metering boxes mainly relies on periodic inspections, but this method has shortcomings such as unsatisfactory results and inability to detect problems in a timely manner. During long-term operation, electricity metering boxes are prone to faults such as localized overheating due to factors such as ambient temperature, humidity, voltage, and current. Localized overheating not only affects the accuracy of electricity metering but may also lead to equipment damage, posing a threat to the safe operation of the power grid. However, existing technologies cannot promptly detect and locate the thermal fault areas of electricity metering boxes, resulting in the inability to take timely measures and affecting the operational safety and reliability of the electricity metering boxes.

[0003] Therefore, a new method is urgently needed to monitor the operating status of the electricity metering box in real time, promptly detect changes in reliability evaluation, analyze the infrared images when changes occur, accurately locate the thermal fault area, thereby ensuring the accuracy and impartiality of electricity metering and improving the operational reliability of the electricity metering box. Summary of the Invention

[0004] Existing technologies suffer from problems such as the inability to monitor the operating status of electricity metering boxes in real time, the inability to detect changes in reliability assessments in a timely manner, and the inability to accurately locate thermal fault areas. Therefore, to solve the above problems, this invention provides a method, device, and medium for locating thermal areas of electricity metering boxes based on operating status identification.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for locating the hot zone of an energy metering box based on operational status identification, characterized by comprising the following steps:

[0007] S100. Based on the historical operating data of the electricity metering box, the working status is divided using a clustering algorithm; an operating status identification model is constructed based on the divided working status.

[0008] S200: The operating status of the power metering box under real-time operating conditions is identified by the operating status identification model, and the reliability of the real-time operating status is evaluated.

[0009] S300: When the reliability evaluation of the real-time operating status changes, acquire infrared images of the power metering box within the time range before and after the change in reliability evaluation, and locate the thermal area in the infrared images.

[0010] S400: Determine the area where a thermal failure may occur based on the located thermal area.

[0011] A method for locating hot zones of an energy metering box based on operational status identification according to claim 1, characterized in that, when conducting reliability evaluation of the real-time operational status, a reliability evaluation model is constructed, which includes the following steps:

[0012] S200-1, Feature Selection: Select several features of the electricity metering box, calculate the importance of the features to the operating status using the ReliefF algorithm, sort them according to importance, and calculate the correlation between importance. Select data with high importance as input features for features with strong correlation.

[0013] S200-2, an operational reliability benchmark model is constructed based on historical operational data and a Gaussian mixture model;

[0014] S200-3 uses the selected features in the real-time operating state as feature vectors, calculates the distance between the current feature vector and the benchmark Gaussian mixture model based on Mahalanobis distance, obtains the degree of deviation between the feature vector and the Gaussian mixture model, and maps it to [0,1] as the reliability evaluation index of the power metering box.

[0015] The present invention further includes obtaining parameters from the historical operating data of the power metering box, including voltage, current, power, temperature, humidity, operating time, and ambient temperature. The voltage, current, power, temperature, humidity, operating time, and ambient temperature are clustered using the DPC algorithm and then trained using the XGBOOST algorithm to construct an operating status identification model.

[0016] The present invention further provides the following processing method for the infrared image after acquisition: low-frequency features and high-frequency features are acquired from the infrared image; adaptive weighting processing is performed on the low-frequency features; edge processing is performed on the high-frequency features; after completion, the low-frequency features and high-frequency features are fused to obtain a new image; inverse wavelet transform is performed on the new image to obtain the final image; then, noise reduction and correction processing is performed on the image to obtain the final image of the power metering box; and the NiBlack algorithm is used to detect hot areas in the final image.

[0017] The invention further incorporates a weighted compression method for the final image, dividing it into non-overlapping high- and low-frequency matrix regions to ensure uniform illumination in each region. A grayscale segmentation threshold is calculated for each region based on the standard deviation and grayscale mean formula. Subsequently, a particle swarm optimization algorithm is used to calculate the optimal segmentation threshold for each matrix region. The optimal clustering threshold is then used to binarize adjacent regions, calculating the average, minimum, and maximum grayscale values ​​of the entire final image. Temperature change parameters are then calculated using the temperature difference, and feature parameters are extracted by inputting the grayscale values ​​of infrared images from all regions of the constructed image sample feature space. The resulting data is input into the decision-maker model, outputting the thermal region detection result.

[0018] The present invention further specifies that the extraction of low-frequency features uses region energy, and its calculation formula is as follows: Where F represents the region energy of the image; x and y represent the coordinates of the pixel; m and n represent the coordinates of adjacent points; M represents the original image pixel; and t represents the weighting factor.

[0019] The present invention further includes an adaptive weighting process for the regional energy, as follows: ;

[0020] The present invention further specifies that, in performing edge smoothing processing on high-frequency features, a binary edge image is generated using high-frequency sub-band coefficients, and F... A and F B This indicates that edge feature information is calculated, denoted as L. A and L B Then for L A and L B Perform an expansion morphology to obtain the region containing edge features, denoted as Z. A and Z B , where X OR The calculation is as follows: ;

[0021] The present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the described method for locating the thermal zone of an energy metering box based on operating state identification.

[0022] The present invention also provides a readable storage medium having a program stored thereon, which, when executed by a processor, implements the aforementioned method for locating the thermal zone of an energy metering box based on operating status identification.

[0023] This invention includes at least one of the following beneficial effects:

[0024] Compared with existing technologies, this invention provides a method for locating hot areas of an energy metering box based on operational status identification. 1. It can monitor the operational status of the energy metering box in real time, promptly detect changes in reliability evaluation, and improve the operational safety and reliability of the energy metering box; 2. When the reliability evaluation changes, it can analyze infrared images within that time period to accurately locate the thermal fault area and take timely measures, ensuring the accuracy and impartiality of energy metering; 3. It improves the accuracy of operational status identification and reliability evaluation by employing methods such as the ReliefF algorithm to calculate feature importance, the Gaussian mixture model to construct a reliability benchmark model, and Mahalanobis distance to calculate the deviation of feature vectors; 4. It improves the accuracy and efficiency of hot area location by performing adaptive weighted processing, edge processing, wavelet inverse transform, and noise reduction correction on infrared images; 5. It accurately detects the hot areas of the energy metering box by employing methods such as the Niblack algorithm, weighted compression, and particle swarm optimization, providing strong support for fault diagnosis. Attached Figure Description

[0025] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0026] Figure 1 This is a flowchart illustrating Embodiment 1 of the present invention. Detailed Implementation

[0027] The following will describe in detail the implementation of this application with reference to the accompanying drawings and embodiments, so that the implementation process of how this application uses technical means to solve technical problems and achieve technical effects can be fully understood and implemented accordingly.

[0028] Example 1

[0029] like Figure 1 As shown, this embodiment provides a method for locating the hot zone of an energy metering box based on operating status identification, including the following steps:

[0030] S100. Based on the historical operating data of the electricity metering box, the working status is divided using a clustering algorithm; an operating status identification model is constructed based on the divided working status; in this process, the data of the electricity metering box is first clustered using the DPC clustering algorithm, that is, the parameters obtained from the historical operating data of the electricity metering box include voltage, current, power, temperature, humidity, operating time, and ambient temperature. The voltage, current, power, temperature, humidity, operating time, and ambient temperature are clustered using the DPC algorithm and then trained using the XGBOOST algorithm to construct the operating status identification model.

[0031] Relevant parameters, such as voltage, current, power, temperature, humidity, operating time, and ambient temperature, are obtained from historical operating data of the electricity metering box. This data undergoes preprocessing, including missing value handling, outlier handling, and normalization or standardization, to ensure comparability of different parameters. Based on the selected distance metric (e.g., Euclidean distance), the local density around each data point is calculated. A Gaussian kernel function is used to define the density, i.e., for each data point... i Its local density , where exp(·) is a logical judgment function, d ij It is data point x i and x j The distance between them d c It is a cutoff distance used to control the neighborhood range for density calculation. By calculating data point x i Minimum distance to other higher density points Defined as: The specific steps are as follows:

[0032] Input: Dataset X = {x1, x2, ... x} i ,…,x n}

[0033] Output: Cluster centers C = {c1, c2, ..., c3} n}

[0034] Step 1. Calculate the dataset X = {x1, x2, ... x} i ,…,x n Calculate the Euclidean distances between all sample points in the matrix, construct a similarity matrix, sort these distances in descending order, and select the top 2% of distances as the cutoff distance d. c ;

[0035] Step 2. Based on the similarity matrix and the cutoff distance d* c And combine the local density and relative distance formulas to calculate the local density ρ and relative distance δ of all sample points in dataset X;

[0036] Step 3. Draw a decision map based on the local density ρ and relative distance δ:

[0037] Step 4. Select cluster centers based on the decision map;

[0038] Step 5. Sort the remaining sample points in descending order of local density, and assign them to the clusters containing the nearest points with local densities greater than their own.

[0039] Subsequently, the XGBoost algorithm was used to train a historical dataset containing work state divisions to construct a running state recognition model; the XGBoost model is defined as follows: ,in This represents the Kth decision tree;

[0040] The objective function of XGBoost is defined as: ; ;

[0041] In the above formula, For loss function, For regularization terms, , ω is a hyperparameter that controls the degree of penalty; T is the number of leaf nodes; ω is the weight score of each leaf node.

[0042] S200: The operating status of the power metering box under real-time operating conditions is identified by the operating status identification model, and the reliability of the real-time operating status is evaluated.

[0043] S300: When the reliability evaluation of the real-time operating status changes, acquire infrared images of the power metering box within the time range before and after the change in reliability evaluation, and locate the thermal area in the infrared images.

[0044] S400: Determine the area where a thermal failure may occur based on the located thermal area.

[0045] In this embodiment, a reliability evaluation model is constructed when evaluating the real-time operating status, which includes the following steps:

[0046] S200-1, Feature Selection: Select several features of the power metering box, namely voltage, current, power, temperature, humidity, running time, and ambient temperature. Calculate the importance of each feature to the operating status using the ReliefF algorithm, sort them according to importance, and calculate the correlation between the importance values. Select data with high importance as input features for features with strong correlation.

[0047] The ReliefF algorithm is as follows:

[0048] Suppose D is a training sample set, and N is the number of feature attributes. diff(A, R1, R2) represents the distance between samples R1 and R2 under feature attribute A.

[0049] diff(A,R1,R2)= Input: Training dataset X, number of neighboring samples k for each region, number of sampling times m, number of regions n, number of repeated calculations N, feature weight vector W=0;

[0050] Output: Average feature weight vector (1) Randomly select a sample R from the sample set X and calculate the distance d between the sample point and samples of the same and different classes. hi d mi Then, sort the samples by distance from smallest to largest, and divide similar samples into n regions from nearest to farthest, with each unit size... ;

[0051] (2) Among samples of the same type, select the k samples that are closest to the selected sample points in each distance cell. If the number of sample points in a region is less than k, then all of them are selected. If the sample points in a region are an empty set, then the region is not selected. The total number of selected sample points is k. Among samples of different types, select the K sample points that are closest to the selected sample points with a number equal to the number of samples of the same type.

[0052] (3) Calculate the weight matrix W(A): p(C) represents the probability of class C occurring; class(R) represents the class to which random sample R belongs; p[class(R)] represents the probability of random sample R.

[0053] (4) Repeat N times to obtain the average feature weight vector: ;

[0054] The voltage, current, power, temperature, humidity, running time, and ambient temperature are sorted based on the calculated average feature weight vector. Then, the correlation between each feature is calculated using the Pearson algorithm. For two parameters with strong correlation, only the parameter with higher importance is retained, and redundant features are removed. In this embodiment, voltage, current, temperature, humidity, and running time are selected as feature inputs.

[0055] S200-2, an operational reliability benchmark model is constructed based on historical operational data and a Gaussian mixture model;

[0056] S200-3 uses the selected features from the real-time operating status as feature vectors, and calculates the distance between the current feature vector and the benchmark Gaussian mixture model based on Mahalanobis distance. It then obtains the degree of deviation between the feature vector and the Gaussian mixture model and maps it to the range [0,1] as the reliability evaluation index of the energy metering box. The Mahalanobis distance used to calculate the distance between the current feature vector and the benchmark Gaussian mixture model is as follows: S -1 It is the inverse of the covariance matrix.

[0057] Then calculate the degree of deviation. ;

[0058] ω 1+ ω 2+⋯+ ωk =1, and values ​​are assigned from largest to smallest based on the importance of the feature;

[0059] Then, based on the Mahalanobis distance mapping to the range [0, 1], it is used as the equipment operational reliability evaluation index, as shown in the following formula:

[0060] ORI=exp[-α·D(x)]; where α is the adjustment coefficient, and we take α=0.012.

[0061] In this embodiment, the infrared image is processed as follows after acquisition: low-frequency and high-frequency features are acquired from the infrared image; adaptive weighting is performed on the low-frequency features; edge processing is performed on the high-frequency features; after completion, the low-frequency and high-frequency features are fused to obtain a new image; inverse wavelet transform is performed on the new image to obtain the final image; then, noise reduction and correction processing are performed on the image to obtain the final image of the power metering box; the Niblack algorithm is used to detect hot areas in the final image.

[0062] Low-frequency features are extracted using region energy, and its calculation formula is as follows: Where F represents the region energy of the image; x and y represent the coordinates of the pixel; m and n represent the coordinates of adjacent points; M represents the original image pixel; and t represents the weighting factor. The adaptive weighting process for the region energy is as follows: ;

[0063] The present invention further specifies that, in performing edge smoothing processing on high-frequency features, a binary edge image is generated using high-frequency sub-band coefficients, and F... A and F B This indicates that edge feature information is calculated, denoted as L. A and L B Then for L A and L B Perform an expansion morphology to obtain the region containing edge features, denoted as Z. A and Z B , where X OR The calculation is as follows: ;

[0064] Based on this, the high and low frequency features of the electricity metering box image are fused using inverse wavelet transform: the specific process is as follows: In the formula, T(A,B) represents all directions of the image; D A and D B This represents a high-frequency image. After fusion, the processed low-frequency and high-frequency features are combined to obtain a new image. Furthermore, to ensure a more comprehensive representation of the new image, pixel values ​​are further matched and filtered using the following formula: Inverse wavelet transform is superior in preserving the main features of the original image and effectively improves computational efficiency.

[0065] This embodiment also includes noise reduction and correction processing for the image, the specific steps of which are as follows:

[0066] Step 1 - Calculate the mean of the four nearest color channels for the color difference in the image:

[0067] fn=(fa+fb+fc+fd) / 4, where fn represents the mixed brightness; fa, fb, fc, and fd are the standard image pixels representing the mixed information of the four colors.

[0068] Step 2 - Noise reduction uses the following method: the weight coefficients of fn are used to replace the weight coefficients of the image being denoised, thereby achieving a strong noise reduction effect on the image. The formula is as follows:

[0069] μ(a) = ∑θ(a,b)·μ(b), where μ(a) represents the pixel obtained after processing; θ(a,b) represents the weight coefficient; and μ(b) represents the surrounding pixels.

[0070] Step 3 - Correction Processing. The effective information of the noise-reduced electricity metering box image is extracted using the following formula: T = fa / fb;

[0071] Step 4: Perform brightness and chromaticity adjustments based on the information. Simultaneously, divide the filtered image to create a contrast image. Then normalize fa and fb, which means multiplying the proportional image by the new chromaticity. Convert the RGB image using the following formula:

[0072] R= fa+1.14fb, G=fa-0.39fc-0.58fb, B=fa+2.03fc;

[0073] Finally passed To achieve white balance correction of an image, where l R l G l B These represent the gain coefficients for the red, green, and blue channels, respectively; N RGB The normalized mean of the color channel.

[0074] The obtained images are compressed using a weighted method, dividing the final image into non-overlapping high- and low-frequency matrix regions to ensure uniform illumination in each region. Based on the standard deviation and gray-level mean formula, a gray-level segmentation threshold is calculated for each region; the formula is as follows: A represents the threshold formula; n represents the number of matrix regions;

[0075] Subsequently, the matrix region is calculated based on the particle swarm optimization algorithm to obtain the optimal image segmentation threshold. In this embodiment, s is assumed to represent the inter-class method and is used as the fitness. The optimal image segmentation threshold is as follows:

[0076] [A1, A2, ... A n ]=max[s1 2 (A1)+…s n 2 (A n )];

[0077] Binarization is performed on adjacent regions based on the optimal clustering threshold, and the average gray value, minimum gray value, and maximum gray value of the entire final image are calculated as follows: ;

[0078] U represents the template temperature; Umax represents the highest template temperature; Umin represents the lowest template temperature; h represents the template calculation; T represents the highest image temperature; n represents the number of binarization processes.

[0079] Subsequently, temperature change parameters are calculated using the temperature difference, and feature parameters are extracted by inputting the grayscale values ​​of infrared images of all regions in the constructed image sample feature space, as shown in the following formula: Where T represents the sample feature space; B represents the label image. A mapping function is then introduced to map the infrared sample image of the energy metering box to a high-dimensional space, achieving thermal region sample classification of the energy metering box image. A decision-maker plane model is set within this space to achieve the maximum classification margin for the samples. Considering that each decision-maker plane can classify two thermal region types, the hyperplane needs to be biased using the following formula: In the formula, R represents the hyperplane space; λ i λ represents the Lagrange multiplier; B represents the classification margin, ranging from [-1, 1]; D represents the penalty parameter; λ j This represents a general coefficient.

[0080] Combining the brightness factor method and cross-validation, the decision classification function for the sample points in the hot zone of the electricity metering box is calculated. The function is: Where g represents the hot region classification result, L represents the kernel function model, and e represents the kernel parameter. The kernel function and kernel parameter are selected based on the characteristics of the features. After classifying the hot regions according to the classification results, the hot regions are given special attention to quickly determine the areas where thermal faults may occur.

[0081] Example 2

[0082] The present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the described method for locating the thermal zone of an energy metering box based on operating state identification.

[0083] Example 3

[0084] The present invention also provides a readable storage medium having a program stored thereon, which, when executed by a processor, implements the aforementioned method for locating the thermal zone of an energy metering box based on operating status identification.

[0085] This invention provides a method for locating hot areas of an energy metering box based on operational status identification. 1. It can monitor the operational status of the energy metering box in real time, promptly detect changes in reliability evaluation, and improve the operational safety and reliability of the energy metering box; 2. When the reliability evaluation changes, it can analyze infrared images within that time period to accurately locate the thermal fault area and take timely measures, ensuring the accuracy and impartiality of energy metering; 3. It improves the accuracy of operational status identification and reliability evaluation by employing methods such as the ReliefF algorithm to calculate feature importance, the Gaussian mixture model to construct a reliability benchmark model, and Mahalanobis distance to calculate the deviation of feature vectors; 4. It improves the accuracy and efficiency of hot area location by performing adaptive weighted processing, edge processing, wavelet inverse transform, and noise reduction correction on the infrared images; 5. It accurately detects the hot areas of the energy metering box by employing methods such as the NiBlack algorithm, weighted compression, and particle swarm optimization, providing strong support for fault diagnosis.

[0086] As used in the specification and claims, certain terms refer to specific components. Those skilled in the art will understand that hardware manufacturers may use different names to refer to the same component. This specification and claims do not distinguish components based on differences in name, but rather on differences in function. The term "comprising" throughout the specification and claims is an open-ended term and should be interpreted as "comprising but not limited to." "Approximately" means that within an acceptable margin of error, those skilled in the art can solve the technical problem and substantially achieve the technical effect within a certain margin of error.

[0087] The foregoing description illustrates and describes several preferred embodiments of the present invention. However, as previously stated, it should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the inventive concept described herein through the foregoing teachings or techniques or knowledge in related fields. Any modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

Claims

1. A method for locating the thermal zone of an energy metering box based on operational status identification, characterized in that, Includes the following steps: S100. Based on the historical operating data of the electricity metering box, the working status is divided using a clustering algorithm; an operating status identification model is constructed based on the divided working status. S200: The operating status of the power metering box under real-time operating conditions is identified by the operating status identification model, and the reliability of the real-time operating status is evaluated. S300: When the reliability evaluation of the real-time operating status changes, acquire infrared images of the power metering box within the time range before and after the change in reliability evaluation, and locate the thermal area in the infrared images. S400. Determine the area where thermal failure may occur based on the located hot area; the reliability evaluation model is constructed when performing reliability evaluation on the real-time operating status, which includes the following steps: S200-1, Feature Selection: Select several features of the electricity metering box, calculate the importance of the features to the operating status using the ReliefF algorithm, sort them according to importance, and calculate the correlation between importance. Select data with high importance as input features for features with strong correlation. S200-2, an operational reliability benchmark model is constructed based on historical operational data and a Gaussian mixture model; S200-3 uses the selected features in the real-time operating state as feature vectors, calculates the distance between the current feature vector and the benchmark Gaussian mixture model based on Mahalanobis distance, obtains the degree of deviation between the feature vector and the Gaussian mixture model, and maps it to [0,1] as the reliability evaluation index of the power metering box.

2. The method for locating the thermal zone of an energy metering box based on operational status identification according to claim 1, characterized in that, The parameters obtained from the historical operating data of the power metering box include voltage, current, power, temperature, humidity, operating time, and ambient temperature. The voltage, current, power, temperature, humidity, operating time, and ambient temperature are clustered using the DPC algorithm and then trained using the XGBOOST algorithm to build an operating status identification model.

3. The method for locating the thermal zone of an energy metering box based on operational status identification according to claim 2, characterized in that, The infrared image processing method after acquisition is as follows: low-frequency and high-frequency features are acquired from the infrared image. Adaptive weighting is performed on the low-frequency features, and edge processing is performed on the high-frequency features. After completion, the low-frequency and high-frequency features are fused to obtain a new image. The new image is subjected to inverse wavelet transform to obtain the final image. Subsequently, the image is subjected to noise reduction and correction processing to obtain the final image of the power metering box. The Niblack algorithm is used to detect hot areas in the final image.

4. The method for locating the thermal zone of an energy metering box based on operational status identification according to claim 3, characterized in that, The final image is compressed using a weighted method, dividing it into non-overlapping high and low frequency matrix regions to ensure that each matrix region has uniform lighting conditions. Based on the standard deviation and gray-level mean formula, the gray-level segmentation threshold is calculated for each region. Subsequently, the matrix region is calculated based on the particle swarm optimization algorithm to obtain the optimal segmentation threshold of the image. The optimal segmentation threshold is used to perform binarization processing on adjacent regions to calculate the average gray value, minimum gray value, and maximum gray value of the entire final image. Then, the temperature change parameter is calculated by the temperature difference, and the feature parameters are extracted by inputting the gray values ​​of infrared images of all regions in the constructed image sample feature space. The obtained data is input into the decision-maker model, and the thermal region detection result is output.

5. The method for locating the thermal zone of an energy metering box based on operational status identification according to claim 4, characterized in that, Low-frequency features are extracted using region energy, and its calculation formula is as follows: Where F represents the region energy of the image; x and y represent the coordinates of the pixel; m and n represent adjacent coordinates; M represents the pixel value of the original image pixel; and t represents the weighting factor.

6. The method for locating the thermal zone of an energy metering box based on operational status identification according to claim 5, characterized in that, The adaptive weighting of regional energy is performed as follows: .

7. The method for locating the thermal zone of an energy metering box based on operational status identification according to claim 3, characterized in that, Edge smoothing of high-frequency features involves generating a binary edge image using high-frequency subband coefficients and then applying F... A and F B This indicates that edge feature information is calculated, denoted as L. A and L B Then for L A and L B Perform an expansion morphology to obtain the region containing edge features, denoted as Z. A and Z B , where X OR The calculation is as follows: .

8. An electronic device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the thermal zone location method for an energy metering box based on operating status identification as described in any one of claims 1-7.

9. A readable storage medium, characterized in that, It stores a program that, when executed by a processor, implements the method for locating the thermal zone of an energy metering box based on operating status identification as described in any one of claims 1-7.

Images