CN-122020548-A - Distribution lightning arrester aging defect identification method and device considering environment confidence, electronic equipment and storage medium

CN122020548ACN 122020548 ACN122020548 ACN 122020548ACN-122020548-A

Abstract

The invention discloses a method, a device, electronic equipment and a storage medium for identifying aging defects of a distribution arrester by considering environmental confidence, and belongs to the field of intelligent monitoring of power equipment; the method comprises the steps of mapping environment parameters to defect display conditions, quantifying the adaptation degree of an environment to ageing defect display to generate an environment confidence index, extracting visual features from inspection image data, carrying out similarity calculation on the visual features and standard defect features to obtain image similarity coefficients, weighting the image similarity coefficients based on the environment confidence index to form a fusion defect probability value, and outputting an ageing defect identification result of the power distribution arrester according to the fusion defect probability value. Therefore, the invention can solve the problems of low recognition confidence and missed detection erroneous judgment caused by the fact that the equipment is not free from defects and the environment is hidden due to the lack of quantitative evaluation on the environment in the prior art.

Inventors

LIU GAO
GUO JINCHAO
LIN JUNSHENG
LUO JINBIN
CHEN HAO
LI MING
YANG KE
XU JIE

Assignees

广东电网有限责任公司

Dates

Publication Date: 20260512
Application Date: 20260202

Claims (10)

1. The method for identifying the aging defect of the distribution lightning arrester by considering the environment confidence is characterized by comprising the following steps of: Collecting inspection image data of a power distribution lightning arrester and environment parameters in time sequence synchronization with the inspection image data; Comparing the environment parameters with preset defect display conditions, quantitatively evaluating the adaptation degree of the environment parameters to defect display, and generating an environment confidence index for inspection image data, wherein the defect display conditions represent conditions required to be met when the aging defect of the distribution arrester can be displayed; Extracting visual features of the inspection image data to generate a target visual feature vector, and calculating the similarity between the target visual feature vector and a preset standard defect feature vector to obtain an image similarity coefficient; according to the environment confidence index, carrying out weighted correction on the image similarity coefficient to generate a fusion defect probability value for representing the possibility of the aging defect of the distribution arrester; And generating a final distribution arrester aging defect identification result according to the fusion defect probability value.
2. The method for identifying the aging defect of the power distribution lightning arrester considering the environment confidence coefficient according to claim 1, wherein the defect display condition comprises the intensity value and the duration of action of the corresponding display environment factor when the aging defect of the power distribution lightning arrester can be displayed; Comparing the environmental parameter with a preset defect display condition, quantitatively evaluating the adaptation degree of the environmental parameter to defect display, and generating an environmental confidence index for inspection image data, wherein the method comprises the following steps: Analyzing phenomenon information for triggering the appearance of the aging defect and associated information existing along with the phenomenon information from environment parameters, wherein the phenomenon information is the intensity value and the duration of action of a gain environment factor which plays a gain role in the appearance of the defect; Comparing the intensity value of the gain environment with the intensity value of the display environment factor to generate an intensity consistency check result; comparing the environment duration value with the display duration threshold value to generate a duration consistency check result; according to the strength consistency check result and the time length consistency check result, calculating to obtain a defect activation index representing defect appearance probability; Determining a condition phenomenon according to the phenomenon information, calculating the similarity between the condition phenomenon and a preset standard defect phenomenon, and generating a phenomenon remarkable index representing the feature definition; Calculating the interference degree value caused to the phenomenon information according to the associated information to generate an associated interference index; And carrying out positive weighting calculation on the defect activation index and the phenomenon significance index, carrying out negative weighting calculation on the associated interference index, and generating an environment confidence index according to the result of the positive weighting calculation and the result of the negative weighting calculation.
3. The method for identifying aging defects of a distribution arrester with consideration of environmental confidence according to claim 2, wherein the calculating a defect activation index representing a defect appearance probability according to the strength consistency check result and the time consistency check result comprises: generating intensity occurrence probability according to an intensity consistency check result based on a preset intensity-probability mapping relation table; Generating a time occurrence probability according to a time consistency check result based on a preset time-probability mapping relation table; and calculating and generating a defect activation index according to the intensity occurrence probability and the time growth occurrence probability.
4. The method for identifying aging defects of a distribution arrester taking into account environmental confidence according to claim 3, wherein the negative environmental factors include wind factors and illumination factors; calculating the interference degree value caused to the phenomenon information according to the associated information to generate an associated interference index, including: Inputting the intensity value and the duration of action of the wind power factor into a preset wind power-shake blur conversion function, and calculating to obtain an image blur degree coefficient representing the motion blur degree of the image; Inputting the intensity value and the duration of action of the illumination factor into a preset illumination-contrast conversion function, and calculating to obtain an image overexposure coefficient representing the exposure deviation degree of the image; And calculating a weighted sum of the image ambiguity coefficient and the image overexposure coefficient, generating the interference degree value, and determining the interference degree value as an associated interference index.
5. The method for identifying aging defects of a distribution arrester with consideration of environmental confidence according to claim 4, wherein the performing a positive weighting calculation on the defect activation index and the phenomenon significance index, performing a negative weighting calculation on the associated interference index, and generating the environmental confidence index according to the result of the positive weighting calculation and the result of the negative weighting calculation comprises: Configuring corresponding forward gain weights for the defect activation indexes and the phenomenon significance indexes respectively; Based on the forward gain weight, performing weighted summation on the defect activation index and the phenomenon significance index to generate an environment dominant value; configuring negative attenuation weights for the associated interference indexes, and carrying out weighting processing on the associated interference indexes based on the negative attenuation weights to generate environment interference values; And correcting and calculating the environment dominant value by using the environment interference value to generate an environment confidence index, wherein the environment interference value and the environment confidence index are in negative correlation.
6. The method for identifying aging defects of a distribution arrester with consideration of environmental confidence according to claim 5, wherein the steps of performing visual feature extraction on the inspection image data to generate a target visual feature vector, calculating the similarity between the target visual feature vector and a preset standard defect feature vector, and obtaining an image similarity coefficient include: Performing target detection and region clipping on the inspection image data, and extracting a to-be-detected defect region image containing a power distribution lightning arrester component from the inspection image data; visual feature extraction is carried out on the image of the defect area to be detected, texture distribution feature data and edge shape feature data of the image of the defect area to be detected are generated, vectorization coding is carried out on the texture distribution feature data and the edge shape feature data, and a target visual feature vector is generated; And calculating cosine similarity values between the spatial mapping feature vector and a preset standard defect feature vector to generate an image similarity coefficient.
7. The method for identifying aging defects of a distribution arrester with consideration of environmental confidence according to claim 6, wherein the generating a final aging defect identification result of the distribution arrester according to the fused defect probability value comprises: Calculating a threshold drift compensation amount based on the associated interference index, and summing a preset reference judgment threshold and the threshold drift compensation amount to generate a dynamic effective threshold; Comparing the fusion defect probability value with the dynamic effective threshold value; if the probability value of the fusion defect is larger than the dynamic effective threshold value, generating a recognition result representing the defect with high confidence, otherwise, generating a recognition result representing suspected interference or normal.
8. The device for identifying the aging defect of the distribution arrester by considering the environment confidence is characterized by comprising a multi-source data acquisition module, an environment confidence evaluation module, a visual characteristic analysis module, a probability weighting fusion module and a defect identification judgment module; the multi-source data acquisition module is used for acquiring the inspection image data of the power distribution lightning arrester and the environment parameters which are synchronous with the timing sequence of the inspection image data; the environment confidence evaluation module is used for comparing the environment parameters with preset defect display conditions, quantitatively evaluating the adaptation degree of the environment parameters to defect display, and generating environment confidence indexes aiming at inspection image data, wherein the defect display conditions represent conditions required to be met when ageing defects of the distribution lightning arrester can be displayed; the visual feature analysis module is used for extracting visual features of the inspection image data, generating a target visual feature vector, calculating the similarity between the target visual feature vector and a preset standard defect feature vector, and obtaining an image similarity coefficient; the probability weighting fusion module is used for carrying out weighting correction on the image similarity coefficient according to the environment confidence index to generate a fusion defect probability value for representing the possibility of the aging defect of the distribution arrester; And the defect identification judging module is used for generating a final aging defect identification result of the distribution arrester according to the fusion defect probability value.
9. An electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, the processor implementing the method of identifying a power distribution arrester burn-in defect taking into account environmental confidence as claimed in any one of claims 1 to 7 when executing the computer program.
10. A storage medium comprising a stored computer program, wherein the computer program, when run, controls a device in which the storage medium is located to perform the method for identifying aging defects of a distribution arrester according to any one of claims 1 to 7, taking into account environmental confidence.

Description

Distribution lightning arrester aging defect identification method and device considering environment confidence, electronic equipment and storage medium Technical Field The invention relates to the technical field of intelligent monitoring of power equipment, in particular to a method and a device for identifying ageing defects of a distribution arrester by considering environment confidence, electronic equipment and a storage medium. Background The power distribution lightning arrester is used as a key overvoltage protection device in a power distribution network, and the running state of the power distribution lightning arrester is directly related to the safety and stability and the power supply reliability of a power grid. Because the lightning arrester is exposed in a complex outdoor natural environment for a long time, the lightning arrester is extremely easy to age due to the influence of factors such as moisture invasion, pollution deposition, electric stress impact and the like, and further the insulation performance is reduced, the thermal breakdown is even the explosion accident is caused. However, most of the existing visual recognition technologies focus on texture analysis or feature extraction of the acquired image itself, but neglect the decisive influence of "environmental conditions" on "defect appearance". In fact, many lightning arresters have significant "environmental dependencies" in that they are clearly visible only under certain weather conditions (such as certain humidity, light or temperature differences), but are "hidden" in non-suitable environments. The prior art lacks a quantitative evaluation mechanism for judging whether the current environment is suitable for defect display, and always forcibly judges whether the current environment is suitable for defect display or not according to the similarity of image features under the condition that the environmental conditions do not support clear display of defect features. The single-dimension recognition mode of 'seeing images off the environment' cannot distinguish the essential difference of 'no defect of the lightning arrester' and 'no appearance of defects caused by the environment', and the confidence coefficient of a recognition result is very easy to deviate from the actual situation, so that detection omission or erroneous judgment is caused. Disclosure of Invention The embodiment of the invention provides a method, a device, electronic equipment and a storage medium for identifying ageing defects of a distribution arrester by considering environmental confidence, which can solve the problems of low identification confidence and missed detection erroneous judgment caused by the fact that equipment is not free from defects and environment is hidden due to the lack of quantitative evaluation on environmental suitability in the prior art. The embodiment of the invention provides a method for identifying aging defects of a distribution arrester by considering environmental confidence, which comprises the following steps: Collecting inspection image data of a power distribution lightning arrester and environment parameters in time sequence synchronization with the inspection image data; Comparing the environment parameters with preset defect display conditions, quantitatively evaluating the adaptation degree of the environment parameters to defect display, and generating an environment confidence index for inspection image data, wherein the defect display conditions represent conditions required to be met when the aging defect of the distribution arrester can be displayed; Extracting visual features of the inspection image data to generate a target visual feature vector, and calculating the similarity between the target visual feature vector and a preset standard defect feature vector to obtain an image similarity coefficient; according to the environment confidence index, carrying out weighted correction on the image similarity coefficient to generate a fusion defect probability value for representing the possibility of the aging defect of the distribution arrester; And generating a final distribution arrester aging defect identification result according to the fusion defect probability value. Further, the defect display conditions comprise intensity values and duration of action of corresponding display environment factors when ageing defects of the distribution lightning arrester can be displayed; Comparing the environmental parameter with a preset defect display condition, quantitatively evaluating the adaptation degree of the environmental parameter to defect display, and generating an environmental confidence index for inspection image data, wherein the method comprises the following steps: Analyzing phenomenon information for triggering the appearance of the aging defect and associated information existing along with the phenomenon information from environment parameters, wherein the phenomenon information is the inte