A circuit breaker operating condition classification method, apparatus, device, medium and product

Abstract

The application relates to the technical field of image processing, and provides a circuit breaker working condition classification method, device, equipment, medium and product.The method comprises the following steps: acquiring a circuit breaker image; pre-processing the circuit breaker image to obtain an image of a region of interest, the region of interest comprising a circuit breaker main body, a circuit breaker position indicator and a contact position; the pre-processing at least comprises at least one of region of interest cutting, light self-adaptive correction and angle correction; performing working condition classification on the image of the region of interest based on a pre-trained circuit breaker working condition classification model to obtain a working condition classification result, the working condition classification result comprising a working condition category corresponding to the circuit breaker image and a corresponding confidence, and the working condition category comprising a test position, a working position and an intermediate position; and performing secondary verification on the working condition classification result based on the confidence, and keeping the working condition classification result in the case of successful verification.The technical scheme realizes accurate classification of circuit breaker working conditions.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of image processing technology, and in particular to a method, apparatus, equipment, medium and product for classifying the operating conditions of circuit breakers. Background Technology

[0002] High-voltage circuit breaker condition identification is a core component of ensuring the safe and stable operation of the power grid and a key foundation for intelligent operation and maintenance of power equipment. Accurately determining the real-time operating status of circuit breakers can promptly avoid potential equipment failures, effectively improving the reliability and safety of the power system.

[0003] Traditional circuit breaker condition identification often employs hardware contact detection methods, mainly using mechanical limit switches, electromagnetic sensors, hardware contacts, and other detection components to collect and feedback position signals, and relying on mechanical contact, electromagnetic induction, or contact on / off principles to determine the status.

[0004] Long-term use of testing components can easily lead to aging problems such as wear, loosening, and poor contact, resulting in decreased accuracy, increased probability of misjudgment, and insufficient overall testing stability and reliability. Furthermore, the method of using testing components for operating condition identification can only determine two operating conditions and cannot identify intermediate transitional positions of the circuit breaker, resulting in incomplete operating condition coverage. Summary of the Invention

[0005] This invention provides a method, apparatus, equipment, medium, and product for classifying circuit breaker operating conditions, in order to improve the stability and reliability of circuit breaker operating condition identification.

[0006] According to one aspect of the present invention, a method for classifying circuit breaker operating conditions is provided, comprising:

[0007] Obtain the circuit breaker image;

[0008] The circuit breaker image is preprocessed to obtain a region of interest (ROI) image, which includes the circuit breaker body, the circuit breaker position indicator, and the contact positions. The preprocessing includes at least one of the following: ROI cropping, illumination adaptive correction, and angle correction.

[0009] Based on the pre-trained circuit breaker condition classification model, the region of interest image is classified into conditions to obtain the condition classification results. The condition classification results include the condition category corresponding to the circuit breaker image and the corresponding confidence level. The condition categories include test position, working position and intermediate position.

[0010] The operating condition classification results are validated a second time based on the confidence level. If the validation is successful, the operating condition classification results are maintained.

[0011] According to another aspect of the present invention, a circuit breaker condition classification device is provided, the device comprising:

[0012] Circuit breaker image acquisition module, used to acquire circuit breaker images;

[0013] The circuit breaker image preprocessing module is used to preprocess the circuit breaker image to obtain a region of interest image, which includes the circuit breaker body, the circuit breaker position indicator, and the contact position; the preprocessing includes at least one of the following: region of interest cropping, illumination adaptive correction, and angle correction;

[0014] The working condition classification result acquisition module is used to classify the working condition of the region of interest image based on the pre-trained circuit breaker working condition classification model and obtain the working condition classification result. The working condition classification result includes the working condition category corresponding to the circuit breaker image and the corresponding confidence level. The working condition categories include test position, working position and intermediate position.

[0015] The working condition classification result verification module is used to perform secondary verification of the working condition classification results based on confidence level. If the verification is successful, the working condition classification results are retained.

[0016] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0017] At least one processor; and

[0018] A memory that is communicatively connected to at least one processor; wherein,

[0019] The memory stores a computer program that can be executed by at least one processor, such that the at least one processor is able to perform the circuit breaker condition classification method according to any embodiment of the present invention.

[0020] According to another aspect of the present invention, a computer-readable storage medium is provided, which stores computer instructions for causing a processor to execute and implement the circuit breaker condition classification method of any embodiment of the present invention.

[0021] According to another aspect of the present invention, a computer program product is provided, which, when executed by a processor, implements a circuit breaker condition classification method as described in any of the embodiments of the present invention.

[0022] This invention provides a raw data foundation for operating condition classification by acquiring circuit breaker images. The circuit breaker images are preprocessed to obtain a region of interest (ROI) image, which includes the circuit breaker body, the circuit breaker position indicator, and the contact positions. The preprocessing includes at least one of ROI cropping, adaptive illumination correction, and angle correction. Since the area in the circuit breaker image used to determine the operating condition is relatively small, preprocessing locks the ROI and removes invalid background interference, improving the accuracy of circuit breaker operating condition classification. Based on a pre-trained circuit breaker operating condition classification model, the ROI image is used to classify the operating conditions, resulting in a classification result. The classification result includes the operating condition category corresponding to the circuit breaker image and the corresponding confidence score. The operating condition categories include test position, operating position, and intermediate position. The pre-trained circuit breaker operating condition classification model accurately distinguishes the three operating conditions and outputs confidence scores, achieving stable identification of circuit breaker operating conditions. The operating condition classification results are then validated a second time based on confidence levels. If the validation is successful, the classification results are maintained, and a second verification using confidence levels is performed to avoid misjudgments in the circuit breaker operating condition classification model, thereby improving the accuracy and reliability of the classification results. In summary, the technical solution of this invention preprocesses the circuit breaker image to obtain a region of interest image containing a small target area of ​​the circuit breaker location indicator. Based on the circuit breaker operating condition classification model, the region of interest image is classified, achieving accurate identification of circuit breaker operating conditions. This effectively avoids the problems of easy aging of detection components, insufficient detection reliability, and incomplete operating condition coverage, thus improving the stability and reliability of circuit breaker operating condition identification.

[0023] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a flowchart of a circuit breaker operating condition classification method according to an embodiment of the present invention;

[0026] Figure 2 This is a schematic diagram of the circuit breaker operating condition classification model in an embodiment of the present invention;

[0027] Figure 3 This is a flowchart of the circuit breaker operating condition classification in an embodiment of the present invention;

[0028] Figure 4 This is a schematic diagram of the test position, working position, and intermediate position in an embodiment of the present invention;

[0029] Figure 5 This is a schematic diagram of the structure of a circuit breaker condition classification device according to an embodiment of the present invention;

[0030] Figure 6 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0031] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0032] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0033] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0034] Figure 1 This is a flowchart illustrating a circuit breaker operating condition classification method provided in an embodiment of the present invention. This embodiment is applicable to situations where circuit breaker images are preprocessed and operating conditions are classified using a circuit breaker operating condition classification model. This method can be executed by the circuit breaker operating condition classification device in this embodiment of the invention. This device can be implemented in software and / or hardware, and can be integrated into electronic devices such as computer equipment, servers, mobile terminals, or processors. Figure 1 As shown, the method specifically includes the following steps:

[0035] S110, Obtain circuit breaker image.

[0036] In this embodiment of the invention, the circuit breaker image can be specifically understood as including images of the circuit breaker and its surroundings. The circuit breaker image can intuitively reflect the current mechanical posture of the circuit breaker, the relative positions of its components, and its actual operating appearance, providing a real and effective raw visual data source for subsequent processing.

[0037] Specifically, images of the circuit breakers to be classified are acquired. For example, image acquisition devices deployed at designated monitoring locations can continuously capture real-time or periodic images of the circuit breakers, thereby obtaining clear and valid images. These image acquisition devices include, but are not limited to, industrial cameras, high-definition visual acquisition devices, or embedded camera modules. Deployment locations can include, but are not limited to, unobstructed areas such as the upper part of the high-voltage switchgear, the side wall of the cabinet, or directly in front of the circuit breaker.

[0038] S120. Preprocess the circuit breaker image to obtain a region of interest image, which includes the circuit breaker body, the circuit breaker position indicator, and the contact position; the preprocessing includes at least one of the following: region of interest cropping, illumination adaptive correction, and angle correction.

[0039] In this embodiment of the invention, the region of interest (ROI) image can be specifically understood as an image containing key regions that can be used to determine the circuit breaker's operating condition using a circuit breaker operating condition classification model. The ROI image does not include irrelevant backgrounds or invalid interference content, highlighting the key visual features required for operating condition identification.

[0040] The areas of interest include the circuit breaker body, the circuit breaker position indicator, and the contact positions. The circuit breaker body can be understood as the overall mechanical structure and housing frame of the circuit breaker. The circuit breaker body is the core carrier embodying the overall operating posture and displacement state of the circuit breaker, directly reflecting its working position and mechanical opening / closing state. The circuit breaker position indicator can be understood as an indicator used to mark the physical stopping position of the circuit breaker in real time. The circuit breaker position indicator can change its position or state synchronously with the circuit breaker's operation, serving as a core intuitive marker for determining whether the circuit breaker is in different operating conditions such as the test position, operating position, or intermediate position. The contact positions can be understood as the relative engagement and disengagement positions between the moving and stationary contacts of the circuit breaker. The moving and stationary contacts of the circuit breaker can be completely connected, partially offset, or completely separated, with different positions corresponding to different operating conditions. The contact positions directly reflect the contact engagement, disengagement, and transition states, and are a crucial characteristic for determining the circuit breaker's operating condition.

[0041] Region of Interest (ROI) cropping can be understood as the operation of cropping and removing irrelevant areas from the original circuit breaker image, retaining only the ROI. ROI cropping can highlight small, key components within the ROI image, weaken invalid interference, enhance the visibility of fine features, and improve the accuracy of subsequent operational condition classification. Illumination adaptive correction is the operation of adaptively adjusting image brightness, contrast, and color. Illumination adaptive correction can clearly restore the details of key components even in complex, obstructed, or blurry lighting conditions, ensuring the accuracy of operational condition classification. Angle correction can be understood as correcting angular offsets during image acquisition. Angle correction ensures that key components such as the circuit breaker body, circuit breaker position indicator, and contacts are viewed from a straight angle, avoiding angular deviations that could affect operational condition judgment.

[0042] Specifically, because the key target components in circuit breaker images used for operating condition classification are relatively small, directly using circuit breaker images for classification can easily lead to feature loss, large recognition errors, and inaccurate judgment results. Therefore, preprocessing of the circuit breaker images is necessary to obtain a region of interest (ROI) image. This preprocessing operation should include at least one of the following: ROI cropping, adaptive illumination correction, and angle correction. Through preprocessing, the ROI is accurately focused, enhancing the detailed features of small targets, avoiding the loss of key recognition information, and improving the accuracy and stability of operating condition classification.

[0043] Optionally, the circuit breaker image is preprocessed, including: performing illumination separation on the circuit breaker image to obtain the illumination component and reflection component of the circuit breaker image, and adjusting the reflection component based on the adaptive gain data corresponding to the illumination scene of the circuit breaker image to obtain the circuit breaker image after illumination adaptive correction.

[0044] In this embodiment of the invention, the illumination component can be specifically understood as the brightness component formed on the circuit breaker image by external lighting factors. For example, external lighting factors can include ambient light sources, spatial brightness, occlusion shadows, specular reflections, etc. The illumination component reflects the global illumination intensity changes of the scene and belongs to the environmental interference item of the circuit breaker image. The reflection component can be specifically understood as the inherent reflection component of the object's surface. The reflection component is determined only by the object's own material and structure. The reflection component directly reflects the true characteristics of the target and is not affected by changes in illumination. The lighting scene can be specifically understood as the lighting scene in which the circuit breaker image is acquired. The lighting scene is the basis for generating adaptive gain data that matches the adjustment requirements. The adaptive gain data can be specifically understood as a set of adjustment parameters adaptively generated based on the circuit breaker image. The adaptive gain data adapts to the on-site lighting scene and is used to specifically adjust the pixel performance of the image's reflection component, compensate for dark areas, suppress overly bright areas, and achieve balanced enhancement under different lighting environments.

[0045] Specifically, illumination separation is performed on the circuit breaker image to obtain the illumination component and reflection component. For example, the Retinex (Retina Cortex) algorithm can be used to decompose the circuit breaker image into illumination and reflection components, as shown in the following formula:

[0046]

[0047] in, This represents the pixel value of the circuit breaker image; Indicates the light component; This represents the reflection component. Strong light, backlighting, and shadows in a lighting scene are essentially due to the uneven distribution of the illumination component, affecting the accuracy of subsequent condition recognition. The illumination component is removed, and the reflection component is adjusted based on adaptive gain data corresponding to the lighting scene in the circuit breaker image. Specifically, Gaussian filtering is used to extract the illumination component. Then, the reflection component is adjusted through adaptive gain. The image after illumination correction is obtained. :

[0048]

[0049] Where k is the adaptive gain coefficient, which can be between 1.2 and 1.8; and b is the offset, which can be between 0 and 20. The intensity of the reflection component varies under different lighting conditions. The k and b values ​​can be adaptively adjusted based on the image's grayscale mean. By adjusting the k and b values, weak light compensation and strong light suppression can be achieved, balancing the differences in brightness and darkness in the image, resulting in a more uniform brightness distribution after correction. Illumination-adaptive correction effectively improves the detail recognition of small targets in dim and backlit scenes, reduces the occlusion and weakening of small targets by strong light and shadows, ensures clear and stable small target features of circuit breakers under different lighting conditions, and avoids misjudgments of operating conditions caused by uneven illumination.

[0050] Optionally, the circuit breaker image is preprocessed, including: identifying reference edge lines in the circuit breaker image, determining the image tilt angle based on the reference edge lines, and performing angle correction on the circuit breaker image based on the image tilt angle to obtain an angle-corrected circuit breaker image.

[0051] In this embodiment of the invention, the reference edge line can be specifically understood as a regular, stable boundary line in the circuit breaker image. For example, the reference edge line can be the straight edge of the circuit breaker indicator sign. The reference edge line is regular and has a fixed direction, serving as a reference for angle measurement. The image tilt angle can be specifically understood as the offset angle by which the circuit breaker image deflects relative to the reference edge line. The image tilt angle is the rotational deviation angle of the entire image relative to a standard upright state.

[0052] Specifically, a reference edge line is identified in the circuit breaker image. For example, the reference edge line can be the straight edge of the circuit breaker indicator sign. The image tilt angle is determined based on the reference edge line. This can be achieved by fitting the reference edge line using a Hough transform and calculating the angle between the reference edge line and the horizontal baseline; this angle is the image tilt angle. Angle correction is then performed on the circuit breaker image based on the image tilt angle. A bilinear interpolation algorithm can be used to perform reverse rotation correction on the tilted circuit breaker image according to the image tilt angle, ensuring that the reference edge line is horizontal, eliminating the influence of angle deviation, and obtaining the angle-corrected circuit breaker image. Angle correction can eliminate the deformation and visual deviation caused by angle offset. It also stabilizes the feature representation of small targets, reduces the interference of orientation differences on operating condition classification, and effectively improves classification accuracy.

[0053] Optionally, the circuit breaker image is preprocessed, including: identifying the switchgear structure in the circuit breaker image; identifying the region of interest (ROI) in the circuit breaker image based on the switchgear structure; the ROI includes the areas corresponding to the circuit breaker body, the circuit breaker position indicator, and the contact positions, respectively; and cropping the circuit breaker image based on the ROI to obtain the ROI image. The ROI image includes an overall image of the circuit breaker body, the circuit breaker position indicator, and the contact positions, a magnified first image of the area where the circuit breaker position indicator is located, and a magnified second image of the area where the contact positions are located.

[0054] In this embodiment of the invention, the switchgear structure can be specifically understood as the inherent cabinet outline and hardware layout structure of the switchgear in the circuit breaker image. The switchgear structure serves as a structural reference for distinguishing between redundant background areas and regions of interest. The switchgear is a closed cabinet device that carries the circuit breaker and its associated electrical components, possessing a fixed installation structure and a regular spatial layout, providing a stable structural reference for locating the region of interest. The overall region image can be specifically understood as an image that fully preserves the circuit breaker and its critical components. The overall region image provides complete basic image data for subsequent global feature analysis and component association judgment. The first region magnified image can be specifically understood as an image obtained by magnifying the local target area where the circuit breaker location indicator is located within the overall region image. The first region magnified image focuses on the area where the indicator is located, magnifying and presenting minute markings, scales, and outline information, mitigating the loss of detail caused by long-distance shooting and image compression, and significantly improving the feature recognition of small targets such as the indicator, providing local image support for subsequent operating condition classification. The second region magnified image can be specifically understood as an image obtained by magnifying the local target area where the circuit breaker contacts are located within the overall region image. The second magnified image focuses on the area where the contact is located, specifically magnifying and presenting minute details such as the shape, gap, and contact state of the contact, providing local image support for working condition classification, and ensuring that contact-related features can be accurately extracted and identified.

[0055] Specifically, the switchgear structure in the circuit breaker image is identified. Based on the switchgear structure, the region of interest (ROI) in the circuit breaker image is identified. The circuit breaker image is then cropped based on the ROI to obtain the ROI image. The ROI image includes an overall image of the circuit breaker body, the circuit breaker position indicator, and the contact positions; a first magnified image of the area containing the circuit breaker position indicator; and a second magnified image of the area containing the contact positions. The first and second magnified images can be obtained using an adaptive ROI cropping algorithm. During the ROI cropping process, the areas containing the circuit breaker position indicator and the contact positions are first locally cropped, and then adaptively magnified based on an adaptive magnification factor and offset. The formula is as follows:

[0056]

[0057] in, This is the initial cropped area corresponding to the location of the circuit breaker position indicator or contact. 'a' is the adaptive magnification factor, dynamically adjusted based on the pixel ratio of the cropped target to the entire circuit breaker image. The adaptive magnification factor can be dynamically adjusted according to the percentage of the cropped target pixels; a larger factor value is selected when the percentage is lower, and a relatively smaller factor value is selected when the percentage is higher, thus achieving adaptive magnification for cropped targets of different sizes. The adaptive magnification factor is between 1.2 and 1.5. 'c' is the offset, adaptively determined based on the outer contour size of the cropped target and the initial cropping boundary position. For example, considering the shape and specifications of the circuit breaker indicator or contact location, a fixed pixel margin is reserved outwards around the initial cropped area to dynamically compensate for the cropping boundary.

[0058] Optionally, preprocessing may also include grayscale conversion and brightness equalization, denoising, edge enhancement, size normalization, and contrast enhancement. Grayscale conversion and brightness equalization convert the circuit breaker image to grayscale and use adaptive histogram equalization to unify the image brightness, eliminating the influence of strong light, shadows, and reflections on imaging. Gaussian filtering denoising removes image noise caused by electromagnetic interference in industrial environments through Gaussian smoothing, preserving target edge contours and avoiding noise interference with feature extraction. Edge enhancement strengthens the edges of the region of interest (ROI) of the circuit breaker, highlighting the contour features of small targets, facilitating accurate capture of operational differences by the circuit breaker condition classification model. Size normalization scales the circuit breaker image proportionally to the input size of the circuit breaker condition classification model, ensuring uniform image scale across different distances and angles, and improving recognition stability. Contrast enhancement stretches the grayscale difference between the ROI and the background, amplifying subtle visual differences between the test position, working position, and intermediate position, reducing the classification difficulty of the circuit breaker condition classification model.

[0059] Optionally, after obtaining the region of interest (ROI) image, the method further includes: calculating the total length of effective edge segments and edge continuity of the ROI image. The total length of effective edge segments is used to measure the richness of the contour details of the core structure, and the edge continuity is used to measure the integrity and coherence of the edge lines. If the total length of effective edge segments is greater than or equal to an edge length threshold, and the edge continuity is greater than or equal to an edge continuity threshold, then the ROI image is input into the circuit breaker condition classification model. If the total length of effective edge segments is less than a preset edge length threshold and / or the edge continuity is less than a preset edge continuity threshold, then the image acquisition device is controlled to re-acquire the circuit breaker image.

[0060] In this embodiment of the invention, the total length of effective edge segments can be specifically understood as the sum of the total pixel lengths of the remaining effective edge segments corresponding to the region of interest after filtering and removing noisy edges, fragmented interference segments, and invalid pseudo-edges from the image of the region of interest. The total length of effective edge segments quantifies the contour integrity and detail richness of the region of interest. The larger the total length of edge segments, the more complete the edge information and the more effective image features, which can provide a high-quality input image for subsequent condition classification models. Edge continuity can be specifically understood as a quantitative index of the connectivity integrity of effective edge segments within the region of interest. Edge continuity is used to characterize the coherence and integrity of the edge lines of key components of the circuit breaker. The fewer edge breaks and the smaller the gaps, the higher the edge continuity value, representing a complete target contour without obvious blurring or defects.

[0061] Specifically, the total length of effective edge segments is calculated by performing edge detection on the image of interest, extracting all edge segments, filtering and removing noisy short segments, isolated false edges, and invalid interference edges, and summing the pixel lengths of all remaining effective edge segments. Edge continuity is quantified by statistically analyzing the number of breakpoints and the proportion of connected segments in the effective edges, using the proportion of complete connected edges in the overall effective edges. The total length of effective edge segments and edge continuity are judged based on preset edge length and edge continuity thresholds. This ensures that all images input to the circuit breaker condition classification model possess complete and clear target edge features, improving the accuracy and stability of circuit breaker condition recognition from the source and preventing misclassification and low confidence due to poor image quality. If the total length of effective edge segments is greater than or equal to the edge length threshold, and the edge continuity is greater than or equal to the edge continuity threshold, the image of the region of interest is input into the circuit breaker condition classification model. If the total length of effective edge segments is less than the edge length threshold and the edge continuity is less than the edge continuity threshold, the image acquisition device is controlled to re-acquire the circuit breaker image. If the total length of the effective edge segments is less than the edge length threshold or the edge continuity is less than the edge continuity threshold, the image acquisition device is controlled to re-acquire the circuit breaker image. This forms a closed loop for pre-verification of image quality, automatically triggering a re-acquisition mechanism for unqualified images, adapting to complex on-site shooting environments, and enhancing environmental adaptability and robustness.

[0062] The edge length threshold and edge continuity threshold can be obtained by calculating the average total length of effective edge segments and the average edge continuity, respectively, and their corresponding proportions. For example, 500 images of the circuit breaker are collected on-site, and 200 images are manually selected to meet quality standards. The total length of effective edge segments and the edge continuity of each qualified image are calculated, yielding the average total length of effective edge segments and the average edge continuity. The edge length threshold can be set to 90% of the average total length of effective edge segments. The edge continuity threshold can also be set to 90% of the average edge continuity. These proportions can be adjusted within the range of 70% to 90% depending on the on-site environment. The proportion can be appropriately increased when the ambient lighting is good and appropriately decreased when the lighting is poor.

[0063] S130. Based on the pre-trained circuit breaker condition classification model, the region of interest image is classified according to its condition to obtain the condition classification result. The condition classification result includes the condition category corresponding to the circuit breaker image and the corresponding confidence level. The condition categories include test position, working position and intermediate position.

[0064] In this embodiment of the invention, the circuit breaker operating condition classification model can be specifically understood as a model used to classify operating conditions of an input region of interest image. The circuit breaker operating condition classification model takes the preprocessed region of interest image as input, identifies the operating conditions of the circuit breaker, and automatically outputs the operating condition category and confidence score value corresponding to the test position, operating position, and intermediate position, thereby achieving automated and accurate identification of the circuit breaker operating conditions. The operating condition classification result can be specifically understood as the result output by the circuit breaker operating condition classification model. The operating condition classification result includes the specific operating condition category of the circuit breaker currently in operation, and the confidence score corresponding to the category determination result. The confidence score is the probability value of the circuit breaker operating condition classification model's judgment of the currently determined operating condition category. The higher the confidence score value, the more likely the circuit breaker belongs to the currently determined operating condition category, and the higher the matching degree and credibility of the classification result; the lower the confidence score value, the lower the probability of the currently determined category, and the higher the uncertainty of the identification result.

[0065] Operating conditions are categorized into three types: test position, working position, and intermediate position. The test position is when the circuit breaker is disconnected from the normal energized operating circuit and is used solely for performance testing of opening and closing capabilities and verification of mechanical actions. In this state, the circuit breaker is not under load and only performs mechanical action tests. The working position is when the circuit breaker is fully connected to the power operating circuit, with contacts reliably closing and normally bearing the line load, and is in a normal operating position for long-term stable energized operation. The intermediate position is a transitional position between the working and test positions, neither fully connected to the operating circuit nor fully disconnected from the working position; it is an unstable transitional state.

[0066] Specifically, the region of interest (ROI) image is input into a pre-trained circuit breaker condition classification model to obtain the condition classification result, which includes the condition category corresponding to the circuit breaker image and the corresponding confidence level. By relying on the pre-trained model to identify circuit breaker conditions and combining it with the ROI, the extraction of key features of small targets on the circuit breaker is strengthened, improving the ability to identify the condition of minute components.

[0067] Optionally, the circuit breaker operating condition classification model includes a feature extraction module, a small target enhancement module, and a classification module. The feature extraction module is used to extract multi-scale features from the region of interest image to obtain multi-scale features. The small target enhancement module performs feature fusion and detail enhancement on the multi-scale features in sequence to obtain enhanced features. The classification module is used to locate and classify small targets on the enhanced features to obtain the operating condition classification result.

[0068] In this embodiment of the invention, the feature extraction module can be specifically understood as performing multi-scale feature extraction on the input region of interest image. The feature extraction module comprehensively captures the surface detail features and deep semantic features of the key small targets of the circuit breaker, outputting multi-scale features containing information from different receptive fields, providing basic feature support for operational condition classification. Multi-scale features can be specifically understood as features extracted through different receptive fields. Multi-scale features contain local detail information and global semantic information of the small targets of the circuit breaker, and can adapt to the feature expression needs of small targets of different sizes.

[0069] The small target enhancement module can be understood as a module that fuses multi-scale features and then performs detail enhancement processing. This module strengthens the weak features of small targets on the circuit breaker, compensating for the shortcomings of indistinct features and susceptibility to interference, and outputs more discriminative enhanced features. This provides high-quality feature support for the subsequent classification module's small target localization and operating condition classification. Small targets are key core components in the circuit breaker image that are small in size, have weak features, and are easily affected by background interference. For example, a small target is a circuit breaker location indicator. Enhanced features can be understood as the features output by the small target enhancement module. These enhanced features amplify the weak detail information of the small targets on the circuit breaker and suppress background interference noise. They improve the recognizability and completeness of key component features, making subsequent small target localization and operating condition classification more convenient.

[0070] The classification module can be understood as a module that completes the localization and classification of small targets based on enhanced features. Using enhanced features as input, the module accurately locates the position of key small target areas of the circuit breaker; simultaneously, it performs deep semantic analysis and feature discrimination on the enhanced features, effectively distinguishing subtle structural differences between test positions, operating positions, and intermediate positions. The classification module completes the precise localization of small targets and the refined classification of operating conditions, outputting operating condition classification results that include specific operating condition categories and corresponding confidence levels, ensuring the accuracy and stability of the overall identification.

[0071] Specifically, Figure 2 A schematic diagram of the circuit breaker operating condition classification model is shown. This model includes a feature extraction module, a small target enhancement module, and a classification module. The feature extraction module extracts multi-scale features from the image of interest. The small target enhancement module sequentially performs feature fusion and detail enhancement on the multi-scale features to obtain enhanced features. The classification module outputs the operating condition classification result based on the enhanced features.

[0072] Optionally, the classification module is specifically used for: traversing the enhanced features based on an adaptive sliding window, filtering small target localization features from the enhanced features based on the response feature value of each window, and performing classification processing based on the small target localization features to obtain the working condition classification result.

[0073] In this embodiment of the invention, the adaptive sliding window can be specifically understood as a sliding window whose size can be automatically adjusted according to the actual size of the small target in the enhanced features. Unlike a fixed-size window, the adaptive sliding window can dynamically adjust its size during traversal. The adaptive sliding window accurately captures the feature information of small targets at different scales, avoiding feature omissions or false detections caused by mismatched window sizes, thus improving the robustness and adaptability of small target localization. The response feature value can be specifically understood as a quantified value of the matching between the features within each window and a preset small target feature template. The response feature value reflects the confidence level that the current window contains target features. The higher the value of the response feature value of a window, the stronger the matching degree between the features of that window region and the target features, and the more likely it is to contain the key small target to be located, providing a basis for subsequent screening of effective localization features. The small target localization feature can be specifically understood as the feature corresponding to the window containing the key small target of the circuit breaker. The small target localization feature is obtained after traversing through the adaptive sliding window and filtering by the response feature value.

[0074] Specifically, the classification module iterates through enhanced features using an adaptive sliding window. The adaptive sliding window dynamically adjusts its size based on small targets of different scales in the enhanced feature map. An initial window size and a window adjustment threshold are set. If the feature response intensity within the current window size consistently exceeds the window adjustment threshold, the window size is gradually expanded or reduced to match targets of different scales; otherwise, the window returns to its initial size. Through this adaptive adjustment strategy, the classification module filters windows based on their response feature values ​​and a preset localization threshold. Windows with response values ​​higher than the localization threshold are identified as regions containing small targets, and the corresponding feature vector is extracted from the enhanced feature map as the small target localization feature, providing a basis for subsequent condition classification. For example, when the initial window size is... When the window adjustment threshold is set to 0.5: if the feature response intensity within the current window remains higher than 0.5, the window size is gradually increased or decreased, such as adjusting by 4 pixels each time to match targets of different scales; otherwise, the window is restored to its original size. Initial size. Using an adaptive adjustment strategy, the classification module filters windows based on their response feature values, using a preset localization threshold of 0.7. Windows with response values ​​higher than 0.7 are identified as regions containing small targets, and their corresponding feature vectors are extracted from the enhanced feature map as small target localization features, providing a basis for subsequent condition classification.

[0075] All small target localization features obtained through adaptive sliding window filtering are input into a lightweight classification head. The classification head outputs probability distributions for three types of working conditions: test position confidence, working position confidence, and intermediate position confidence, satisfying that the sum of the test position confidence, working position confidence, and intermediate position confidence is 1. The working condition classification result is output according to the working condition judgment rules: if the test position confidence is greater than a preset first working condition judgment threshold, it is judged as a test position. If the working position confidence is greater than a preset first working condition judgment threshold, it is judged as a working position. A second working condition judgment threshold is set for judging intermediate positions, and the second working condition judgment threshold is lower than the first working condition judgment threshold. If the intermediate position confidence is greater than a preset second working condition judgment threshold, it is judged as an intermediate position. When none of the three probabilities exceed their respective thresholds, it is also judged as an intermediate position, ensuring the completeness of the output results. For example, the first working condition judgment threshold can be set to 0.6, and the second working condition judgment threshold can be set to 0.4. If the reliability of the experimental position is greater than 0.6, it is determined to be the experimental position; if the reliability of the working position is greater than 0.6, it is determined to be the working position; if the reliability of the median position is greater than 0.4, and the reliability of both the experimental and working positions does not exceed 0.6, it is determined to be the median position; when the probabilities of all three do not exceed their respective thresholds, it is also determined to be the median position, ensuring the completeness of the output results. The first and second working condition judgment thresholds can be determined as follows: set multiple sets of first and second working condition judgment thresholds, calculate the classification accuracy under each set of thresholds, and select the threshold pair that results in the highest classification accuracy as the first and second working condition judgment thresholds. The values ​​of the first and second working condition judgment thresholds in this scheme are only examples. Depending on the different requirements for classification accuracy, the first and second working condition judgment thresholds can be appropriately increased or decreased. When higher classification accuracy is required, the first and second working condition judgment thresholds can be appropriately increased; when lower classification accuracy is required, the first and second working condition judgment thresholds can be appropriately decreased.

[0076] It should be noted that the lightweight classification head in the classification model is key to achieving end-to-end inference with non-maximum suppression (NMS). By directly classifying features, the steps of generating redundant candidate boxes and performing NMS filtering in traditional models are eliminated, thus simplifying the inference process.

[0077] Optionally, during the training of the circuit breaker condition classification model, a small-target perceptual label assignment can be adopted to strengthen the focus on the features of small targets, allowing the circuit breaker condition classification model to prioritize learning small target regions and solve the problems of inaccurate recognition under blur, occlusion, and low light conditions. A progressive loss balancing method can be used during the training of the circuit breaker condition classification model to avoid bias towards simpler categories, ensuring balanced convergence of the three types of conditions and significantly improving the accuracy of intermediate position recognition. Alternatively, a hybrid optimizer can be used during the training of the circuit breaker condition classification model to accelerate convergence and enable high-speed and stable operation. For example, the circuit breaker condition classification model can be trained using an adaptive optimization algorithm with a lightweight feature extraction structure, a small-target feature enhancement strategy, and a progressive loss balancing mechanism.

[0078] S140. Perform secondary verification on the working condition classification results based on confidence level. If the verification is successful, maintain the working condition classification results.

[0079] Specifically, the operating condition classification results are validated a second time based on the confidence level. If the validation is successful, the operating condition classification results are maintained. By setting a second validation step, when the confidence level output by the classification model is lower than a preset threshold, the current operating condition classification result is determined to be unreliable, thus improving the reliability of circuit breaker operating condition classification.

[0080] Optionally, the operating condition classification result is validated a second time based on the confidence level. If the validation is successful, the operating condition classification result is retained, including: if the confidence level is greater than or equal to the validation threshold, the operating condition classification result is retained; if the confidence level is less than the validation threshold, the operating condition classification result is determined to be invalid, the circuit breaker image is marked, and the information of the circuit breaker image is recorded, including at least one of the following: image acquisition time, initial classification result, and image source.

[0081] In this embodiment of the invention, the verification threshold can be understood as a preset value used to determine whether the operating condition classification result is reliable. When the confidence level output by the circuit breaker operating condition classification model is greater than or equal to this threshold, the classification result is considered reliable, and the operating condition classification result is maintained; when the confidence level output by the circuit breaker operating condition classification model is lower than this threshold, the classification result is considered unreliable. The specific value of the verification threshold can be set according to the security requirements of the application scenario. For example, in the circuit breaker operating condition identification scenario, the verification threshold can be set to 0.95. The image acquisition time is the shooting time obtained from the circuit breaker image acquisition device. The image acquisition time is used to trace the actual operating condition of the circuit breaker corresponding to the image. The initial classification result is the operating condition classification result directly output by the circuit breaker operating condition classification model. The purpose of recording the initial classification result is to provide maintenance personnel with the original judgment basis of the model, so as to quickly confirm the actual operating condition of the image or mark the wrong sample. The image source can be understood as the device identifier or location information of the circuit breaker image acquisition device. The image source is used to trace the acquisition source of the circuit breaker image. For example, a fixed camera ID or acquisition point name. By recording the source of the images, it is possible to locate whether there are any abnormalities in a specific acquisition device, or to optimize the recognition effect of a specific point.

[0082] Specifically, the circuit breaker condition classification model outputs the circuit breaker condition classification image, along with its corresponding confidence level. If the confidence level is greater than or equal to the verification threshold, the condition classification result is maintained. If the confidence level is less than the verification threshold, the condition classification result is deemed invalid. For example, the verification threshold can be 0.95. When the confidence level is greater than or equal to 0.95, the condition classification result is maintained. If the confidence level is less than 0.95, the condition classification result is deemed invalid. This secondary verification prevents the circuit breaker condition classification model from making low-confidence or even incorrect judgments when features are not significant or image quality is poor, effectively reducing potential operational risks caused by misjudgments of condition conditions. Circuit breaker images with invalid condition classification results are marked, and their information is recorded. A preset exception handling process can be triggered when verification fails, thereby improving the overall security and robustness of circuit breaker condition identification.

[0083] The technical solution of this embodiment provides a raw data foundation for operating condition classification by acquiring circuit breaker images. The circuit breaker images are preprocessed to obtain a region of interest (ROI) image, which includes the circuit breaker body, the circuit breaker position indicator, and the contact positions. Preprocessing includes at least one of ROI cropping, adaptive illumination correction, and angle correction. Since the area in the circuit breaker image used to determine the operating condition is relatively small, preprocessing locks the ROI and removes invalid background interference, improving the accuracy of circuit breaker operating condition classification. Based on a pre-trained circuit breaker operating condition classification model, the ROI image is used to classify the operating conditions, obtaining classification results. These results include the operating condition category corresponding to the circuit breaker image and the corresponding confidence score. Operating condition categories include test position, operating position, and intermediate position. The pre-trained circuit breaker operating condition classification model accurately distinguishes the three operating conditions and outputs confidence scores, achieving stable identification of circuit breaker operating conditions. The operating condition classification results are then validated a second time based on confidence levels. If the validation is successful, the classification results are maintained, and a second verification using confidence levels is performed to avoid misjudgments in the circuit breaker operating condition classification model, thereby improving the accuracy and reliability of the classification results. In summary, the technical solution of this invention preprocesses the circuit breaker image to obtain a region of interest image containing small target areas such as the circuit breaker body, location indicator, and contact positions. Operating condition classification is then performed on the region of interest image based on the circuit breaker operating condition classification model, achieving accurate identification of circuit breaker operating conditions. This effectively avoids the problems of easy aging of detection components, insufficient detection reliability, and incomplete operating condition coverage, thus improving the stability and reliability of circuit breaker operating condition identification.

[0084] Based on the above embodiments, an optional example is provided. This example can be used in scenarios where the area in the circuit breaker image that can be used to determine the operating condition is relatively small, the circuit breaker image is preprocessed to obtain a region of interest image, and the operating condition is classified in the region of interest image using a circuit breaker operating condition classification model.

[0085] First, sample circuit breaker images were acquired, covering the test position, operating position, and intermediate position, with a sample size of 330 images. After enhancement, 660 images were obtained, and the dataset was divided into an 8:1:1 ratio. Based on the initial circuit breaker condition classification model and the sample circuit breaker images, an adaptive optimization algorithm was used for training, employing a lightweight feature extraction structure, a small target feature enhancement strategy, and a progressive loss equalization mechanism. The initial learning rate was 0.001, and the model was trained for 50 epochs to obtain the circuit breaker condition classification model.

[0086] Figure 3The flowchart below illustrates the circuit breaker operating condition classification process. First, the circuit breaker image is acquired. Since the area within the circuit breaker used for operating condition classification is small, preprocessing of the circuit breaker image allows the classification model to focus on key areas, improving feature extraction efficiency and classification accuracy. Preprocessing operations include region of interest (ROI) cropping, adaptive illumination correction, angle correction, grayscale and brightness equalization, denoising, edge enhancement, size normalization, and contrast enhancement to obtain the ROI image. The pre-trained circuit breaker operating condition classification model then performs operating condition classification on the ROI image, yielding the classification results. The classification results include the operating condition category corresponding to the circuit breaker image and its corresponding confidence score. Operating condition categories include test position, operating position, and intermediate position. Figure 4 Schematic diagrams of the test position, working position, and intermediate position are shown respectively. A secondary validation of the confidence level in the operating condition classification results is performed based on a validation threshold. The validation threshold is 0.95. The system checks if the confidence level is greater than or equal to 0.95. If the confidence level is greater than or equal to the validation threshold, the operating condition classification result is maintained and output; if the confidence level is less than the validation threshold, the operating condition classification result is determined to be invalid, the circuit breaker image is reacquired, and the circuit breaker images with invalid operating condition classification results are marked and their information is recorded.

[0087] Figure 5 This is a schematic diagram of a circuit breaker condition classification device provided in an embodiment of the present invention. This embodiment is applicable to situations where circuit breaker images are preprocessed and condition classification is performed using a circuit breaker condition classification model. The device can be implemented in software and / or hardware, and can be integrated into any device that provides circuit breaker condition classification functionality, such as… Figure 5 As shown, the device for classifying circuit breaker operating conditions specifically includes: a circuit breaker image acquisition module 210, a circuit breaker image preprocessing module 220, an operating condition classification result acquisition module 230, and an operating condition classification result verification module 240.

[0088] Circuit breaker image acquisition module 210 is used to acquire circuit breaker images;

[0089] The circuit breaker image preprocessing module 220 is used to preprocess the circuit breaker image to obtain a region of interest image, which includes the circuit breaker body, the circuit breaker position indicator, and the contact position; the preprocessing includes at least one of the following: region of interest cropping, illumination adaptive correction, and angle correction.

[0090] The working condition classification result acquisition module 230 is used to classify the working condition of the region of interest image based on the pre-trained circuit breaker working condition classification model and obtain the working condition classification result. The working condition classification result includes the working condition category corresponding to the circuit breaker image and the corresponding confidence level. The working condition category includes test position, working position and intermediate position.

[0091] The working condition classification result verification module 240 is used to perform secondary verification of the working condition classification result based on confidence level. If the verification is successful, the working condition classification result is maintained.

[0092] The technical solution of this embodiment provides a raw data foundation for operating condition classification by acquiring circuit breaker images. The circuit breaker images are preprocessed to obtain a region of interest (ROI) image, which includes the circuit breaker body, the circuit breaker position indicator, and the contact positions. Preprocessing includes at least one of ROI cropping, adaptive illumination correction, and angle correction. Since the area in the circuit breaker image used to determine the operating condition is relatively small, preprocessing locks the ROI and removes invalid background interference, improving the accuracy of circuit breaker operating condition classification. Based on a pre-trained circuit breaker operating condition classification model, the ROI image is used to classify the operating conditions, obtaining classification results. These results include the operating condition category corresponding to the circuit breaker image and the corresponding confidence score. Operating condition categories include test position, operating position, and intermediate position. The pre-trained circuit breaker operating condition classification model accurately distinguishes the three operating conditions and outputs confidence scores, achieving stable identification of circuit breaker operating conditions. The operating condition classification results are then validated a second time based on confidence levels. If the validation is successful, the classification results are maintained, and a second verification using confidence levels is performed to avoid misjudgments in the circuit breaker operating condition classification model, thereby improving the accuracy and reliability of the classification results. In summary, the technical solution of this invention preprocesses the circuit breaker image to obtain a region of interest image containing small target areas such as the circuit breaker body, location indicator, and contact positions. Operating condition classification is then performed on the region of interest image based on the circuit breaker operating condition classification model, achieving accurate identification of circuit breaker operating conditions. This effectively avoids the problems of easy aging of detection components, insufficient detection reliability, and incomplete operating condition coverage, thus improving the stability and reliability of circuit breaker operating condition identification.

[0093] Based on the above embodiments, optionally, the circuit breaker image preprocessing module 220 is used to: perform illumination separation on the circuit breaker image to obtain the illumination component and reflection component of the circuit breaker image, and adjust the reflection component based on the adaptive gain data corresponding to the illumination scene corresponding to the circuit breaker image to obtain the circuit breaker image after illumination adaptive correction.

[0094] Based on the above embodiments, optionally, the circuit breaker image preprocessing module 220 is used to: identify reference edge lines in the circuit breaker image, determine the image tilt angle based on the reference edge lines, perform angle correction on the circuit breaker image based on the image tilt angle, and obtain the angle-corrected circuit breaker image.

[0095] Based on the above embodiments, optionally, the circuit breaker image preprocessing module 220 is used to: identify the switchgear structure in the circuit breaker image; identify the region of interest in the circuit breaker image based on the switchgear structure; the region of interest includes the areas corresponding to the circuit breaker body, the circuit breaker position indicator, and the contact positions, respectively; and crop the circuit breaker image based on the region of interest to obtain a region of interest image. The region of interest image includes an overall area image of the circuit breaker body, the circuit breaker position indicator, and the contact positions, a first magnified image of the area where the circuit breaker position indicator is located, and a second magnified image of the area where the contact positions are located.

[0096] Based on the above embodiments, optionally, the device further includes a region of interest (ROI) image verification module, used to: calculate the total length of effective edge segments and edge continuity of the ROI image, wherein the total length of effective edge segments is used to measure the richness of the contour details of the core structure, and the edge continuity is used to measure the integrity and coherence of the edge lines; if the total length of effective edge segments is greater than or equal to an edge length threshold, and the edge continuity is greater than or equal to an edge continuity threshold, then the ROI image is input into the circuit breaker condition classification model; if the total length of effective edge segments is less than a preset edge length threshold and / or the edge continuity is less than a preset edge continuity threshold, then the image acquisition device is controlled to re-acquire the circuit breaker image.

[0097] Based on the above embodiments, optionally, the circuit breaker operating condition classification model includes a feature extraction module, a small target enhancement module, and a classification module; wherein, the feature extraction module is used to extract multi-scale features from the region of interest image to obtain multi-scale features; the small target enhancement module sequentially performs feature fusion and detail enhancement on the multi-scale features to obtain enhanced features; the classification module is used to locate and classify small targets on the enhanced features to obtain the operating condition classification result.

[0098] Based on the above embodiments, optionally, the classification module is specifically used for: traversing the enhanced features based on an adaptive sliding window, filtering small target positioning features from the enhanced features based on the response feature value of each window, and performing classification processing based on the small target positioning features to obtain the working condition classification result.

[0099] Based on the above embodiments, optionally, the working condition classification result verification module is used to: maintain the working condition classification result if the confidence level is greater than or equal to the verification threshold; determine that the working condition classification result is invalid if the confidence level is less than the verification threshold, mark the circuit breaker image, and record the information of the circuit breaker image, including at least one of the image acquisition time, initial classification result, and image source.

[0100] The above-described products can perform the methods provided in any embodiment of the present invention, and have the corresponding functional modules and beneficial effects for performing the methods.

[0101] Figure 6 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0102] like Figure 6 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0103] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0104] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as the circuit breaker condition classification method.

[0105] In some embodiments, the circuit breaker condition classification method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the circuit breaker condition classification method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the circuit breaker condition classification method by any other suitable means (e.g., by means of firmware).

[0106] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0107] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0108] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0109] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0110] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0111] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0112] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0113] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements a circuit breaker condition classification method according to any embodiment of the invention.

[0114] In implementing the computer program product, computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0115] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for classifying the operating conditions of circuit breakers, characterized in that, include: Obtain the circuit breaker image; The circuit breaker image is preprocessed to obtain a region of interest (ROI) image, which includes the circuit breaker body, the circuit breaker position indicator, and the contact positions. The preprocessing includes at least one of the following: ROI cropping, illumination adaptive correction, and angle correction. The operating condition classification model of the circuit breaker is used to classify the operating condition of the region of interest image to obtain the operating condition classification result. The operating condition classification result includes the operating condition category corresponding to the circuit breaker image and the corresponding confidence level. The operating condition category includes test position, working position and intermediate position. The test position is the non-working position of the circuit breaker disconnected from the normal power-on operation circuit. The working position is the normal working position of the circuit breaker fully connected to the power operation circuit and operating stably for a long time. The intermediate position is the transition position of the circuit breaker between the working position and the test position. The working condition classification result is verified a second time based on the confidence level. If the verification is successful, the working condition classification result is maintained. The preprocessing of the circuit breaker image includes at least one of the following: The circuit breaker image is subjected to illumination separation to obtain the illumination component and reflection component of the circuit breaker image. The reflection component is adjusted based on the adaptive gain data corresponding to the illumination scene of the circuit breaker image to obtain the circuit breaker image after illumination adaptive correction. Identify reference edge lines in the circuit breaker image, determine the image tilt angle based on the reference edge lines, and perform angle correction on the circuit breaker image based on the image tilt angle to obtain an angle-corrected circuit breaker image. The switchgear structure in the circuit breaker image is identified, and the region of interest (ROI) in the circuit breaker image is identified based on the switchgear structure. The ROI includes the areas corresponding to the circuit breaker body, the circuit breaker position indicator, and the contact positions, respectively. The circuit breaker image is cropped based on the ROI to obtain an ROI image. The ROI image includes an overall image of the circuit breaker body, the circuit breaker position indicator, and the contact positions, a first magnified image of the area where the circuit breaker position indicator is located, and a second magnified image of the area where the contact positions are located. The method further includes, after obtaining the region of interest image: Calculate the total length of effective edge segments and edge continuity of the region of interest image. The total length of effective edge segments is used to measure the richness of the contour details of the core structure, and the edge continuity is used to measure the integrity and coherence of the edge lines. If the total length of the effective edge segments is greater than or equal to the edge length threshold, and the edge continuity is greater than or equal to the edge continuity threshold, then the region of interest image is input into the circuit breaker condition classification model. If the total length of the effective edge segments is less than the preset edge length threshold and / or the edge continuity is less than the preset edge continuity threshold, then the image acquisition device is controlled to re-acquire the circuit breaker image.

2. The method according to claim 1, characterized in that, The circuit breaker operating condition classification model includes a feature extraction module, a small target enhancement module, and a classification module. The feature extraction module extracts multi-scale features from the region of interest image to obtain multi-scale features. The small target enhancement module sequentially fuses and enhances the details of the multi-scale features to obtain enhanced features. The classification module locates and classifies the enhanced features to obtain the operating condition classification result.

3. The method according to claim 2, characterized in that, The classification module is specifically used for: Based on the adaptive sliding window, the enhanced features are traversed, and small target localization features are selected from the enhanced features based on the response feature value of each window. Based on the small target localization features, classification processing is performed to obtain the working condition classification result.

4. The method according to claim 1, characterized in that, The operating condition classification result is then validated a second time based on the confidence level. If the validation is successful, the operating condition classification result is retained, including: If the confidence level is greater than or equal to the verification threshold, the working condition classification result is maintained. If the confidence level is less than the verification threshold, the operating condition classification result is determined to be invalid. The circuit breaker image is then marked, and information about the circuit breaker image is recorded. This information includes at least one of the following: image acquisition time, initial classification result, and image source.

5. A circuit breaker operating condition classification device, characterized in that, include: Circuit breaker image acquisition module, used to acquire circuit breaker images; A circuit breaker image preprocessing module is used to preprocess the circuit breaker image to obtain a region of interest image, wherein the region of interest includes the circuit breaker body, the circuit breaker position indicator, and the contact position; the preprocessing includes at least one of the following: region of interest cropping, illumination adaptive correction, and angle correction. The operating condition classification result acquisition module is used to classify the operating condition of the region of interest image based on a pre-trained circuit breaker operating condition classification model to obtain the operating condition classification result. The operating condition classification result includes the operating condition category corresponding to the circuit breaker image and the corresponding confidence level. The operating condition category includes test position, working position, and intermediate position. The test position is the non-working position of the circuit breaker disconnected from the normal energized operating circuit. The working position is the normal working position of the circuit breaker fully connected to the power operating circuit and operating stably for a long time. The intermediate position is the transition position of the circuit breaker between the working position and the test position. The working condition classification result verification module is used to perform secondary verification on the working condition classification result based on the confidence level, and maintain the working condition classification result if the verification is successful. The circuit breaker image preprocessing module is used to perform illumination separation on the circuit breaker image to obtain the illumination component and reflection component of the circuit breaker image, and adjust the reflection component based on the adaptive gain data corresponding to the illumination scene of the circuit breaker image to obtain the circuit breaker image after illumination adaptive correction. A reference edge line is identified in the circuit breaker image, and the image tilt angle is determined based on the reference edge line. The circuit breaker image is then angle-corrected based on the image tilt angle to obtain an angle-corrected circuit breaker image. The switchgear structure in the circuit breaker image is identified, and a region of interest (ROI) is identified based on the switchgear structure. The ROI includes the areas corresponding to the circuit breaker body, the circuit breaker position indicator, and the contact positions. The circuit breaker image is then cropped based on the ROI to obtain an ROI image. The ROI image includes an overall image of the circuit breaker body, the circuit breaker position indicator, and the contact positions; a first magnified image of the area where the circuit breaker position indicator is located; and a second magnified image of the area where the contact positions are located. The circuit breaker condition classification device includes a region of interest (ROI) image verification module, used to calculate the total length of effective edge segments and edge continuity of the ROI image. The total length of effective edge segments is used to measure the richness of the contour details of the core structure, and the edge continuity is used to measure the integrity and coherence of the edge lines. If the total length of effective edge segments is greater than or equal to an edge length threshold, and the edge continuity is greater than or equal to an edge continuity threshold, then the ROI image is input into the circuit breaker condition classification model. If the total length of effective edge segments is less than a preset edge length threshold and / or the edge continuity is less than a preset edge continuity threshold, then the image acquisition device is controlled to re-acquire the circuit breaker image.

6. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the circuit breaker condition classification method according to any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the circuit breaker condition classification method according to any one of claims 1-4.

8. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the circuit breaker condition classification method according to any one of claims 1-4.

Images