Transformer substation equipment image detection method and system based on rapid fault identification

Abstract

The invention discloses a transformer substation equipment image detection method and system based on rapid fault recognition, and relates to the technical field of image recognition. The method comprises the following steps: acquiring transformer substation equipment images and preprocessing the images to obtain a transformer substation equipment image set, extracting transformer substation equipment image fault feature information from the transformer substation equipment image set and classifying and grading the transformer substation equipment image fault feature information to obtain a transformer substation equipment image classification and grading fault feature set; and constructing a rapid fault identification model according to the transformer substation equipment image classification and classification fault feature set, obtaining a transformer substation equipment fault image detection result, verifying the transformer substation equipment fault image detection result according to a manual detection result, and optimizing the rapid fault identification model according to a verification result. According to the invention, the fault related factors such as the fault equipment and position of the substation equipment and the influence degree of the fault on the operation of the substation can be quickly identified through the substation equipment image, so that the maintenance personnel can quickly obtain the fault related factors and process the fault.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular, to a method and system for detecting substation equipment images based on rapid fault recognition. Background Art

[0002] At the present stage of the continuous development of China's national economy, people's demand for electric power energy is increasing, which to a certain extent enhances the importance of substation equipment fault detection work. However, from the current situation of the development of substation equipment fault detection work in China, due to the influence of various factors, there are inevitably many deficiencies. Substation equipment fault recognition has become a practical problem to be solved in the relevant research fields, which not only plays a role in ensuring the safe and reliable operation of the power system, but also has very important practical significance for the sustainable development of society. Summary of the Invention

[0003] The present invention provides a method for detecting substation equipment images based on rapid fault recognition, including:

[0004] Step S1, the cloud platform collects substation equipment images through an image acquisition device and preprocesses the images to obtain a substation equipment image set;

[0005] Step S2, the cloud platform extracts substation equipment image fault feature information from the substation equipment image set and classifies and grades it to obtain a classified and graded fault feature set of substation equipment images;

[0006] Step S3, the cloud platform constructs a rapid fault recognition model according to the classified and graded fault feature set of substation equipment images to obtain the detection result of substation equipment fault images;

[0007] Step S4, the platform verifies the detection result of substation equipment fault images according to the manual detection result, and optimizes the rapid fault recognition model according to the verification result.

[0008] For the method for detecting substation equipment images based on rapid fault recognition as described above, wherein the sub-steps of the cloud platform collecting substation equipment images through an image acquisition device and preprocessing the images to obtain a substation equipment image set are as follows:

[0009] Step S11, the cloud platform collects substation equipment images through an image acquisition device to obtain substation equipment collected images;

[0010] Step S12, the cloud platform preprocesses the substation equipment collected images to obtain a substation equipment image set.

[0011] A substation equipment image detection method based on rapid fault identification as described above, wherein the cloud platform extracts fault feature information of substation equipment images from the substation equipment image set and classifies and grades them. The sub-steps of obtaining the classified and graded fault feature set of substation equipment images are as follows:

[0012] Step S21: The cloud platform extracts fault features from the substation equipment image set to obtain the substation equipment image fault feature set;

[0013] Step S22: The cloud platform classifies and grades the substation equipment image fault feature set to obtain the substation equipment image classified and graded fault feature set.

[0014] A substation equipment image detection method based on rapid fault identification as described above, wherein the cloud platform constructs a rapid fault identification model based on the classified and graded fault feature set of substation equipment images. The sub-steps of obtaining the substation equipment fault image detection result are as follows:

[0015] Step S31: The cloud platform constructs a rapid fault identification model based on the classified and graded fault feature set of substation equipment images;

[0016] Step S32: The platform verifies the substation equipment fault image detection result according to the manual detection result and optimizes the rapid fault identification model according to the verification result.

[0017] A substation equipment image detection method based on rapid fault identification as described above, wherein the cloud platform verifies the substation equipment fault image detection result according to the manual detection result and optimizes the rapid fault identification model according to the verification result. The sub-steps are as follows:

[0018] Step S41: The cloud platform verifies the accuracy of the substation equipment fault image detection result according to the manual detection result to obtain the accurate value of the substation equipment fault image detection result;

[0019] Step S42: The cloud platform judges the detection accuracy of the rapid fault identification model according to the accurate value of the substation equipment fault image monitoring result. If the detection accuracy is low, the rapid fault identification model is optimized according to the manual detection result.

[0020] The present invention also provides a substation equipment image detection system based on rapid fault identification, including:

[0021] An image acquisition and processing module, which acquires substation equipment images through an image acquisition device and preprocesses the images to obtain a substation equipment image set;

[0022] A classified and graded fault feature set acquisition module, which extracts fault feature information of substation equipment images from the substation equipment image set and classifies and grades them to obtain a classified and graded fault feature set of substation equipment images;

[0023] An image detection module constructs a fast fault recognition model based on the classified and graded fault feature set of substation equipment images, and obtains the detection results of substation equipment fault images;

[0024] A fast fault recognition model optimization module verifies the detection results of substation equipment fault images according to the manual detection results, and optimizes the fast fault recognition model according to the verification results.

[0025] An image detection system for substation equipment based on fast fault recognition as described above, wherein the image acquisition and processing module specifically includes:

[0026] An image acquisition sub-module acquires substation equipment images through an image acquisition device, and obtains the acquired images of substation equipment;

[0027] An image processing sub-module preprocesses the acquired images of substation equipment to obtain a set of substation equipment images.

[0028] An image detection system for substation equipment based on fast fault recognition as described above, wherein the classified and graded fault feature set acquisition module specifically includes:

[0029] A fault feature extraction sub-module extracts fault features from the set of substation equipment images to obtain a set of substation equipment image fault features;

[0030] A classification and grading sub-module classifies and grades the set of substation equipment image fault features on the platform to obtain a classified and graded fault feature set of substation equipment images.

[0031] An image detection system for substation equipment based on fast fault recognition as described above, wherein the image detection module specifically includes:

[0032] A fast fault recognition model construction sub-module constructs a fast fault recognition model based on the classified and graded fault feature set of substation equipment images;

[0033] An image detection result acquisition sub-module transmits the images of the substation equipment to be detected collected in real time to the fast fault recognition model, and obtains the detection results of substation equipment fault images.

[0034] An image detection system for substation equipment based on fast fault recognition as described above, wherein the fast fault recognition model optimization module specifically includes:

[0035] A result verification sub-module, and the cloud platform verifies the accuracy of the detection results of substation equipment fault images according to the manual detection results, and obtains the accurate value of the detection results of substation equipment fault images;

[0036] The model optimization sub-module determines the detection accuracy of the fast fault recognition model based on the accurate value of the monitoring result of the substation equipment fault image. If the detection accuracy is low, the fast fault recognition model is optimized according to the manual detection result.

[0037] The beneficial effects achieved by the present invention are as follows: The present invention can quickly identify fault-related factors such as the equipment with faults in the substation equipment, the location, and the impact of the fault on the operation of the substation through the substation equipment images, enabling the maintenance staff to quickly obtain the fault-related factors and handle the faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0039] Figure 1 It is a flowchart of a substation equipment image detection method based on fast fault recognition provided in Embodiment 1 of the present application;

[0040] Figure 2 It is a schematic diagram of a substation equipment image detection system based on fast fault recognition provided in Embodiment 2 of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0042] Embodiment 1

[0043] As Figure 1 shown, Embodiment 1 of the present application provides a substation equipment image detection method based on fast fault recognition, and the method includes the following steps:

[0044] Step S1: The cloud platform collects substation equipment images through an image acquisition device and preprocesses the images to obtain a substation equipment image set;

[0045] Further, the sub-steps of the cloud platform collecting substation equipment images through an image acquisition device and preprocessing the images to obtain a substation equipment image set are as follows:

[0046] Step S11: The cloud platform acquires images of substation equipment through an image acquisition device to obtain the acquired images of substation equipment.

[0047] Specifically, the acquired images of substation equipment include the acquired images of primary equipment and secondary equipment. Primary equipment refers to the equipment that directly generates, transmits, distributes, and uses electric energy, including transformers, high-voltage circuit breakers, disconnectors, busbars, lightning arresters, capacitors, reactors, etc. Secondary equipment refers to the equipment that measures, monitors, controls, and protects the operating conditions of primary equipment and systems, including relay protection devices, automatic devices, measurement and control devices, metering devices, automation systems, and DC devices that provide power for secondary equipment.

[0048] Step S12: The cloud platform preprocesses the acquired images of substation equipment to obtain the image set of substation equipment.

[0049] Specifically, preprocessing operations such as image denoising, size unification, and color normalization are performed on the acquired images of substation equipment to ensure the consistency and comparability of the data.

[0050] Step S2: The cloud platform extracts and classifies and grades the fault feature information of the images of substation equipment in the image set of substation equipment to obtain the classified and graded fault feature set of the images of substation equipment.

[0051] Furthermore, the sub-steps for the cloud platform to extract and classify and grade the fault feature information of the images of substation equipment in the image set of substation equipment to obtain the classified and graded fault feature set of the images of substation equipment are as follows:

[0052] Step S21: The cloud platform extracts fault features from the image set of substation equipment to obtain the fault feature set of the images of substation equipment.

[0053] Specifically, the images of substation equipment are processed through image processing techniques to improve the accuracy and efficiency of image feature extraction. Among them, image processing techniques include gray-scale transformation, filtering, edge detection, image segmentation, etc.; the fault feature information of the processed images of substation equipment is extracted through image extraction techniques. Among them, image feature extraction techniques include edge feature extraction techniques, color feature extraction techniques, texture feature extraction techniques, shape feature extraction techniques, etc.; the fault feature set of the images of substation equipment includes but is not limited to transformer fault features, disconnector fault features, capacitor fault features, and relay protection device fault features.

[0054] Step S22: The cloud platform classifies and grades the fault feature set of the images of substation equipment to obtain the classified and graded fault feature set of the images of substation equipment.

[0055] Specifically, classify the fault feature set of substation equipment images according to different characteristics of substation equipment to obtain a classified fault feature set of substation equipment images, which includes but is not limited to a fault feature subset of transformer equipment, a fault feature subset of switch equipment, a fault feature subset of protection equipment, a fault feature subset of reactive power compensation equipment, a fault feature subset of measurement equipment, a fault feature subset of control equipment, and a fault feature subset of protection and regulation equipment.

[0056] Grade the classified fault feature set of substation equipment images according to the fault impact degree caused by the faults of substation equipment. Based on the classified fault feature set of substation equipment images and combined with the operation elements of the substation, construct an importance vector set of substation equipment SB = {sb 1 , sb 2 … sb m}, an importance vector set of substation equipment fault factors YS = {ys 1 , ys 2 … ys n}, an importance vector set of fault location WZ = {wz 1 , wz 2 … wz n}, and an importance vector set of fault repair time XF = {xf 1 , xf 2 … xf n}. Determine their weights according to the influence degrees of fault factors, locations, and repair times on the operation of the substation. Among them, fault factors include but are not limited to equipment aging, lack of maintenance, overloading, environmental factors, and human operation errors.

[0057] Furthermore, the fault impact degree expression is as follows:

[0058]

[0059] Among them, GZD is the fault impact degree, sb j is the importance of the jth substation equipment, m is the total number of equipment in the substation, λ 1 is the weight value of the importance of fault factors, ys i is the importance of the ith fault factor in the jth substation equipment, λ 2 is the weight value of the importance of the fault location, wz i is the importance of the ith fault location in the jth substation equipment, λ 3 is the weight value of the importance of the fault repair time, xf i is the importance of the ith fault repair time in the jth substation equipment, n is the total number of faults of substation equipment in the jth substation equipment. Among them, λ 1 + λ2 +λ 3 = 1

[0060] Specifically, according to the fault impact degree, the substation equipment fault levels are divided into high-priority fault level, medium-priority fault level, low-priority fault level, and no-fault level. Among them, the high-priority fault level refers to the faults that have a serious impact on the operation of the substation; the medium-priority fault level refers to the faults that have a certain impact on the operation efficiency of the substation; the low-priority fault level refers to the faults that have a small impact on the operation of the substation; the no-fault level refers to the situation where the normal operation of the equipment will not affect the operation of the substation.

[0061] Step S3: The cloud platform constructs a fast fault recognition model based on the substation equipment image classification and grading fault feature set, and obtains the detection result of the substation equipment fault image;

[0062] Furthermore, the sub-steps for the cloud platform to construct a fast fault recognition model based on the substation equipment image classification and grading fault feature set and obtain the detection result of the substation equipment fault image are as follows:

[0063] Step S31: The cloud platform constructs a fast fault recognition model based on the substation equipment image classification and grading fault feature set;

[0064] Specifically, a fast fault recognition model training set and a fast fault recognition model test set are generated according to the substation equipment image classification and grading fault feature set. Among them, the fast fault recognition model training set constructs a model training input vector set according to the substation equipment image set

[0065] Where [(jd 1 , jd 2 …jd t …jd w ) 1 , (jd 1 , jd 2 …jd t …jd w ) 2 …(jd 1 , jd 2 …jd t …jd w ) j …(jd 1 , jd 2 …jd t …jd w ) m ) i is the i-th substation equipment image, (jd 1 , jd 2 …jd t …jd w )j is the j-th fault image of the i-th device, jd t is the t-th angle image of the j-th fault of the i-th device. The sub-recognition model ψ y (x) is trained through the input vector set of the fast fault recognition model training and the fault influence degree GZD. According to the formula Estimate the set of weights {γ 1 , γ 2 …γ Y} of the fault influence degree of substation equipment images. Through each sub-recognition model {ψ 1 (x), ψ 2 (x)…ψ Y (x)} and its corresponding weight values {γ 1 , γ 2 …γ Y}, determine the fault level of substation equipment through the formula ; Construct the model training output vector SC = {gz 1 , gz 2 …gz n} according to the substation equipment image classification and grading fault feature set and the substation fault level. Among them, gz 1 , gz 2 …gz n are information related to substation equipment faults, including but not limited to substation equipment types, fault factors, fault locations, and fault levels. Input the input vector set and output vector set of the fast fault recognition model training into the machine learning model, train the fast fault recognition model, and evaluate the performance of the fast fault recognition model through the fast fault recognition model test set, so that it can identify the substation equipment type, fault factors, fault location, fault level and other fault-related information of the image to be recognized according to the substation equipment image, enabling the staff to quickly locate the faulty equipment according to the substation image fault detection results, the impact on the operation of the substation, and the priority of fault repair.

[0066] Step S32: The cloud platform transmits the real-time collected substation equipment images to be detected to the fast fault recognition model to obtain the substation equipment fault image detection results;

[0067] Specifically, during the image transmission process, establish multiple transmission channels according to different characteristics of the images, so that the images can be efficiently transmitted to the fast fault recognition model, enabling the rapid generation of substation image detection results to discover the faults of substation equipment and handle the faults.

[0068] Step S4: The cloud platform verifies the substation equipment fault image detection results according to the manual detection results, and optimizes the fast fault recognition model according to the verification results;

[0069] Further, the cloud platform verifies the detection result of the substation equipment fault image according to the manual detection result. The sub-steps of optimizing the fast fault recognition model according to the verification result are as follows:

[0070] Step S41: The cloud platform verifies the accuracy of the detection result of the substation equipment fault image according to the manual detection result, and obtains the accurate value of the detection result of the substation equipment fault image;

[0071] Specifically, calculate the accurate value of the detection result of the substation equipment fault image according to the manual detection result, and construct a detection vector set according to the manual detection result, the detection result of the substation equipment fault image, and the importance of the detection information item where {rg 1 ,rg 2 …rg n} are different manual detection results, {mx 1 ,mx 2 …mx n} are different detection results of the substation equipment fault image, is the weight of different detection items, n is the total number of detection information items; calculate the accurate value of the detection result of the substation equipment fault image through the detection vector set by the formula

[0072] Step S42: The cloud platform judges the detection accuracy of the fast fault recognition model according to the accurate value of the monitoring result of the substation equipment fault image. If the detection accuracy is low, optimize the fast fault recognition model according to the manual detection result;

[0073] Specifically, the larger the accurate value JZ is, the lower the detection accuracy of the fast fault recognition model. Obtain the detection information items with low detection accuracy according to the accuracy, and obtain a large number of substation equipment images as the training set of the fast fault recognition model according to the detection information items, and optimize the training of the model, so that the accuracy of the detection result after the optimization of the fast fault recognition model is improved.

[0074] Embodiment 2

[0075] As Figure 2 shown, Embodiment 2 of the present application provides a substation equipment image detection system based on fast fault recognition, including:

[0076] An image acquisition and processing module 21, which acquires substation equipment images through an image acquisition device and preprocesses the images to obtain a substation equipment image set;

[0077] Further, the image acquisition and processing module 21 includes the following sub-modules:

[0078] An image acquisition sub-module, which acquires substation equipment images through an image acquisition device to obtain substation equipment acquired images;​

[0079] Specifically, the images collected by substation equipment include the images collected by primary equipment and the images collected by secondary equipment. Primary equipment refers to the equipment directly involved in the production, transmission, distribution, and use of electric energy, including transformers, high-voltage circuit breakers, disconnectors, busbars, lightning arresters, capacitors, reactors, etc. Secondary equipment refers to the equipment used to measure, monitor, control, and protect the operating conditions of primary equipment and systems, including relay protection devices, automatic devices, measurement and control devices, metering devices, automation systems, and DC devices that provide power for secondary equipment.

[0080] The image processing sub-module preprocesses the images collected by substation equipment to obtain a set of substation equipment images.

[0081] Specifically, operations such as image denoising, size unification, and color normalization are performed on the images collected by substation equipment to ensure the consistency and comparability of the data.

[0082] The classification and grading fault feature set acquisition module 22 extracts and classifies the fault feature information of the substation equipment images in the set of substation equipment images to obtain a classified and graded fault feature set of substation equipment images.

[0083] Furthermore, the classification and grading fault feature set acquisition module 22 includes the following sub-modules:

[0084] The feature extraction sub-module extracts the fault features of the substation equipment images in the set of substation equipment images to obtain a set of fault features of substation equipment images.

[0085] Specifically, the substation equipment images are processed through image processing techniques to improve the accuracy and efficiency of image feature extraction. Among them, image processing techniques include gray-scale transformation, filtering, edge detection, image segmentation, etc. The fault feature information of the processed substation equipment images is extracted through image extraction techniques. Among them, image feature extraction techniques include edge feature extraction techniques, color feature extraction techniques, texture feature extraction techniques, shape feature extraction techniques, etc. The set of fault features of substation equipment images includes, but is not limited to, transformer fault features, disconnector fault features, capacitor fault features, and relay protection device fault features.

[0086] The classification and grading sub-module classifies and grades the set of fault features of substation equipment images to obtain a classified and graded fault feature set of substation equipment images.

[0087] Specifically, the fault feature set of substation equipment images is classified according to different characteristics of substation equipment to obtain the classified fault feature set of substation equipment images, which includes but is not limited to the fault feature subset of transformer equipment, the fault feature subset of switch equipment, the fault feature subset of protection equipment, the fault feature subset of reactive power compensation equipment, the fault feature subset of measurement equipment, the fault feature subset of control equipment, and the fault feature subset of protection and regulation equipment.

[0088] The classified fault feature set of substation equipment images is graded according to the fault impact degree caused by the faults of substation equipment. According to the classified fault feature set of substation equipment images and combined with the operation elements of the substation, the importance vector set of substation equipment SB = {sb 1 , sb 2 … sb m} is constructed, the importance vector set of substation equipment fault factors YS = {ys 1 , ys 2 … ysn}, the importance vector set of fault location WZ = {wz 1 , wz 2 … wz n}, and the importance vector set of fault repair time XF = {xf 1 , xf 2 … xf n} are constructed. The weights are determined according to the influence degrees of fault factors, location, and repair time on the operation of the substation. Among them, fault factors include but are not limited to equipment aging, lack of maintenance, overloading, environmental factors, and human operation errors.

[0089] Furthermore, the fault impact degree expression is as follows:

[0090]

[0091] Among them, GZD is the fault impact degree, sb j is the importance of the j-th substation equipment, m is the total number of equipment in the substation, λ 1 is the weight value of the importance of fault factors, ys i is the importance of the i-th fault factor in the j-th substation equipment, λ 2 is the weight value of the importance of the fault location, wz i is the importance of the i-th fault location in the j-th substation equipment, λ 3 is the weight value of the importance of the fault repair time, xf i is the importance of the i-th fault repair time in the j-th substation equipment, n is the total number of faults of substation equipment in the j-th substation equipment. Among them, λ 1 + λ 2 + λ3 = 1。

[0092] Specifically, according to the fault impact degree, the substation equipment fault levels are divided into high-priority fault levels, medium-priority fault levels, low-priority fault levels, and no-fault levels. Among them, the high-priority fault level refers to a fault that seriously affects the operation of the substation; the medium-priority fault level refers to a fault that has a certain impact on the operation efficiency of the substation; the low-priority fault level refers to a fault that has a small impact on the operation of the substation; the no-fault level refers to that the normal operation of the equipment will not affect the operation of the substation.

[0093] The image detection module 23 constructs a fast fault recognition model based on the substation equipment image classification and grading fault feature set, and obtains the substation equipment fault image detection result;

[0094] Furthermore, the image detection module 23 includes the following sub-modules:

[0095] The fast fault recognition model construction sub-module, and the cloud platform constructs a fast fault recognition model based on the substation equipment image classification and grading fault feature set;

[0096] Specifically, a fast fault recognition model training set and a fast fault recognition model test set are generated according to the substation equipment image classification and grading fault feature set. Among them, the fast fault recognition model training set constructs a model training input vector set according to the substation equipment image set Where [(jd 1 , jd 2 …jd t …jd w ) 1 , (jd 1 , jd 2 …jd t …jd w ) 2 …(jd 1 , jd 2 …jd t …jd w ) j …(jd 1 , jd 2 …jd t …jd w ) m i is the i-th substation equipment image, (jd 1 , jd 2 …jd t …jd w ) j is the j-th fault image of the i-th equipment, jd t ​The t-th angular image of the j-th fault of the i-th device, training the input vector set and the fault impact degree GZD through the fast fault recognition model to train the sub-recognition model ψ y (x), according to the formula Estimate the set of weights {γ 1 , γ 2 … γ Y} of the fault impact degree of the images of substation equipment, and determine the fault level of substation equipment through each sub-recognition model {ψ 1 (x), ψ 2 (x)… ψ Y (x)} and its corresponding weight values {γ 1 , γ 2 … γ Y} through the formula ; Construct the model training output vector SC = {gz 1 , gz 2 … gz n} according to the substation equipment image classification and grading fault feature set and the substation fault level, where gz 1 , gz 2 … gz n are information related to substation equipment faults, including but not limited to substation equipment types, fault factors, fault locations, and fault levels. Input the fast fault recognition model training input vector set and output vector set into the machine learning model, train the fast fault recognition model, and evaluate the performance of the fast fault recognition model through the fast fault recognition model test set, so that it can identify the substation equipment type, fault factors, fault location, fault level and other fault-related information of the image to be recognized according to the substation equipment image, enabling the staff to quickly locate the faulty equipment according to the substation image fault detection result, the impact degree on the substation operation, and the fault repair priority.

[0097] The image detection result acquisition sub-module transmits the real-time collected substation equipment image to be detected to the fast fault recognition model to obtain the substation equipment fault image detection result;

[0098] Specifically, during the image transmission process, establish multiple transmission channels according to different characteristics of the image, so that the image can be efficiently transmitted to the fast fault recognition model, enabling the rapid generation of substation image detection results to discover the faults of substation equipment and process the faults.

[0099] The fast fault recognition model optimization module 24 verifies the substation equipment fault image detection result according to the manual detection result and optimizes the fast fault recognition model according to the verification result;

[0100] Furthermore, the fast fault recognition model optimization module 24 includes the following sub-modules:

[0101] The result verification sub-module verifies the accuracy of the substation equipment fault image detection result according to the manual detection result, and obtains the accurate value of the substation equipment fault image detection result;

[0102] Specifically, calculate the accurate value of the substation equipment image detection result according to the manual detection result, and construct a detection vector set according to the manual detection result, the substation image detection result, and the importance of the detection information item where {rg 1 ,rg 2 …rg n} are different manual detection results, {mx 1 ,mx 2 …mx n} are different substation image detection results, is the weight of different detection items, and n is the total number of detection information items; calculate the accurate value of the substation equipment image detection result through the formula according to the detection vector set

[0103] The model optimization sub-module judges the detection accuracy of the fast fault recognition model according to the accurate value of the substation equipment fault image monitoring result. If the detection accuracy is low, optimize the fast fault recognition model according to the manual detection result;

[0104] Specifically, the larger the accurate value JZ is, the lower the detection accuracy of the fast fault recognition model. Obtain the detection information items with low detection accuracy according to the accuracy, obtain a large number of substation equipment images as the training set of the fast fault recognition model according to the detection information items, and optimize the training of the model, so that the accuracy of the detection result of the fast fault recognition model is improved after optimization.

[0105] The specific implementation manners described above further elaborate on the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above is only the specific implementation manner of the present invention and is not used to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made on the basis of the technical solution of the present invention shall be included in the protection scope of the present invention.​

Claims

1. A substation equipment image detection method based on rapid fault identification, characterized in that: include: Step S1: The cloud platform collects substation equipment images through an image acquisition device and pre-processes the images to obtain a substation equipment image set; Step S2: The cloud platform extracts substation equipment image fault feature information from the substation equipment image set and classifies and grades the image to obtain a substation equipment image classification and graded fault feature set; Step S3: The cloud platform constructs a fast fault recognition model based on the fault feature set of the substation equipment image classification and grading, and obtains the substation equipment fault image detection result; Step S4: The platform verifies the substation equipment fault image detection results according to the manual detection results, and optimizes the rapid fault identification model according to the verification results.

2. A substation equipment image detection method based on rapid fault identification as claimed in claim 1, characterized in that: The cloud platform collects substation equipment images through image acquisition equipment and preprocesses the images. The sub-steps for obtaining the substation equipment image set are as follows: Step S11: The cloud platform collects substation equipment images through an image acquisition device to obtain substation equipment acquisition images; Step S12: The cloud platform pre-processes the substation equipment collected images to obtain a substation equipment image set.

3. A substation equipment image detection method based on rapid fault identification as claimed in claim 1, characterized in that: The cloud platform extracts the fault feature information of the substation equipment image from the substation equipment image set and classifies and grades it. The sub-steps for obtaining the substation equipment image classification and grading fault feature set are as follows: Step S21: The cloud platform extracts fault features from the substation equipment image set to obtain a substation equipment image fault feature set; Step S22: The cloud platform classifies and grades the substation equipment image fault feature set to obtain the substation equipment image classification and grading fault feature set.

4. A substation equipment image detection method based on rapid fault identification as claimed in claim 1, characterized in that: The cloud platform builds a fast fault recognition model based on the fault feature set of substation equipment image classification and grading. The sub-steps for obtaining the substation equipment fault image detection results are as follows: Step S31, the cloud platform builds a fast fault recognition model based on the fault feature set of substation equipment image classification and grading; Step S32: The platform verifies the substation equipment fault image detection results according to the manual detection results, and optimizes the rapid fault identification model according to the verification results.

5. A substation equipment image detection method based on rapid fault identification as claimed in claim 1, characterized in that: The cloud platform verifies the fault image detection results of substation equipment based on the manual detection results. The sub-steps of optimizing the fast fault identification model based on the verification results are as follows: Step S41: The cloud platform verifies the accuracy of the substation equipment fault image detection result according to the manual detection result, and obtains the accurate value of the substation equipment fault image detection result; Step S42: The cloud platform determines the detection accuracy of the rapid fault identification model according to the precise value of the substation equipment fault image monitoring result. If the detection accuracy is low, the rapid fault identification model is optimized according to the manual detection result.

6. A substation equipment image detection system based on rapid fault identification, characterized in that: include: An image acquisition and processing module acquires substation equipment images through an image acquisition device and pre-processes the images to obtain a substation equipment image set; A classification and grading fault feature set acquisition module extracts fault feature information of substation equipment images from the substation equipment image set and performs classification and grading to obtain a classification and grading fault feature set of substation equipment images; Image detection module, which builds a fast fault recognition model based on the fault feature set of substation equipment image classification and obtains the fault image detection results of substation equipment; The fast fault identification model optimization module verifies the substation equipment fault image detection results based on the manual detection results, and optimizes the fast fault identification model based on the verification results.

7. A substation equipment image detection system based on rapid fault identification as claimed in claim 6, characterized in that: Image acquisition and processing module, specifically including: The image acquisition submodule acquires the substation equipment image through the image acquisition device to obtain the substation equipment acquisition image; The image processing submodule pre-processes the substation equipment collected images to obtain the substation equipment image set.

8. A substation equipment image detection system based on rapid fault identification as claimed in claim 6, characterized in that: The module for acquiring the classification and grading fault feature set includes: The fault feature extraction submodule extracts fault features from the substation equipment image set to obtain the substation equipment image fault feature set; In the classification and grading submodule, the platform classifies and grades the fault feature set of substation equipment images to obtain the classification and grading fault feature set of substation equipment images.

9. A substation equipment image detection system based on rapid fault identification as claimed in claim 6, characterized in that: Image detection module, specifically including: The fast fault identification model building submodule builds a fast fault identification model based on the fault feature set of substation equipment image classification; The image detection result acquisition submodule transmits the real-time collected substation equipment images to be detected to the fast fault identification model to obtain the substation equipment fault image detection results.

10. A substation equipment image detection system based on rapid fault identification as claimed in claim 6, characterized in that: Rapid fault identification model optimization module, including: In the result verification submodule, the cloud platform verifies the accuracy of the substation equipment fault image detection results based on the manual detection results, and obtains the precise value of the substation equipment fault image detection results; The model optimization submodule determines the detection accuracy of the rapid fault identification model according to the precise value of the fault image monitoring results of the substation equipment. If the detection accuracy is low, the rapid fault identification model is optimized according to the manual detection results.

Images