Wind driven generator respirator anomaly detection method and system based on image segmentation

Abstract

The invention discloses a wind driven generator respirator anomaly detection method and system based on image segmentation, and relates to the technical field of respirator anomaly detection.The method comprises the steps that according to image collection equipment, a respirator deployment area is collected, and space-time image frames are determined; returning the space-time image frame, triggering the first image processor, and determining a corrected image frame; pixel scanning is carried out on the corrected image frame, a respirator full life cycle detector is triggered, and a respirator pixel matrix is determined; outputting abnormal detection data; and according to the central control end of the fan management system, executing breathing machine abnormity alarm and replacement instruction issuing. The technical problems that in the prior art, the accuracy of anomaly detection of the respirator of the wind driven generator is insufficient, the respirator is difficult to adapt to complex working conditions, and the safety and stability of large components of a fan are low are solved, and the technical effects that the anomaly of the respirator of the wind driven generator is accurately detected, and the safety and stability of the large components of the fan are improved are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of ventilator anomaly detection, and in particular to a method and system for detecting anomaly of a wind turbine ventilator based on image segmentation. Background Art

[0002] In the operating environment of wind turbine respirators, they often face complex and changing light field conditions such as dim light and backlight. These environmental factors can easily cause image acquisition distortion, affecting the accurate judgment of the respirator status. At the same time, in actual applications, respirators of various specifications vary. Respirators of different specifications differ in structure, internal particle change patterns, etc., and their status exhibits a complex dynamic trend change process throughout their life cycle. At present, there is a lack of a systematic solution for abnormality detection of such respirators. Existing methods are often difficult to adapt to complex light field environments, lack generalization for detection of respirators of multiple specifications, and cannot implement standardized detection processes. It is difficult to meet the needs of efficient and accurate detection of abnormal respirator status in actual engineering.

[0003] The existing technology has technical problems such as insufficient accuracy in detecting abnormalities of wind turbine respirators and difficulty in adapting to complex working conditions, resulting in low safety and stability of large components of the wind turbine. Summary of the Invention

[0004] The present application provides a wind turbine respirator anomaly detection method and system based on image segmentation, which is used to solve the technical problems in the prior art of insufficient accuracy of wind turbine respirator anomaly detection, difficulty in adapting to complex working conditions, and low safety and stability of large components of the wind turbine.

[0005] In view of the above problems, the present application provides a wind turbine respirator anomaly detection method and system based on image segmentation.

[0006] In a first aspect of the present application, a method for detecting abnormalities in a wind turbine respirator based on image segmentation is provided, the method comprising: According to the image acquisition device, the ventilator deployment area is collected to determine the spatiotemporal image frame; the spatiotemporal image frame is transmitted back to trigger the first image processor to perform polarization compensation based on active structured light projection guided by directional demand to determine the corrected image frame; pixel scanning is performed on the corrected image frame to trigger the ventilator full life cycle detector, and a first-order segmentation is performed by performing a first pixel corner point positioning based on the ventilator area to determine the ventilator pixel matrix; for the ventilator pixel matrix, a second segmentation based on the second pixel corner point positioning of the particle color level inside the ventilator and a ratio calculation based on the corner pixel positioning are performed to output abnormality detection data, wherein the calculation target is the ratio of particles of different color levels inside the ventilator; according to the central control terminal of the fan management system, an alarm message based on the abnormality detection data is generated, and a ventilator abnormality alarm and replacement instruction are issued, wherein the multivariate data mode is displayed on the display interface.

[0007] A second aspect of the present application provides a wind turbine respirator anomaly detection system based on image segmentation, the system comprising: The spatiotemporal image frame determination module is used to collect information from the ventilator deployment area according to the image acquisition device and determine the spatiotemporal image frame; the correction image frame determination module is used to transmit the spatiotemporal image frame back, trigger the first image processor, perform polarization compensation based on active structured light projection guided by directional demand, and determine the correction image frame; the first-order detection module is used to perform pixel scanning on the correction image frame, trigger the ventilator full life cycle detector, perform first-order segmentation by performing first pixel corner point positioning based on the ventilator area, and determine the ventilator pixel matrix; the second-order detection module is used to perform second segmentation based on second pixel corner point positioning based on the color level of particles inside the ventilator and ratio calculation based on corner pixel positioning for the ventilator pixel matrix, and output abnormal detection data, wherein the calculation target is the ratio of particles of different color levels inside the ventilator; the abnormal alarm module is used to generate alarm information based on the abnormal detection data according to the central control terminal of the fan management system, execute ventilator abnormality alarm and replacement instruction issuance, wherein the multivariate data mode is displayed on the display interface.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: The image acquisition device captures the ventilator deployment area and determines a spatiotemporal image frame. The spatiotemporal image frame is transmitted back to trigger the first image processor, which performs polarization compensation based on active structured light projection guided by directional demand to determine a corrected image frame. Pixel scanning is performed on the corrected image frame to trigger the ventilator full life cycle detector to determine the ventilator pixel matrix. For the ventilator pixel matrix, a second segmentation based on the second pixel corner point positioning of the ventilator's internal particle color level and a ratio calculation based on the corner pixel positioning are performed to output abnormality detection data. Based on the central control terminal of the wind turbine management system, an alarm message based on the abnormality detection data is generated, and a ventilator abnormality alarm and replacement instruction are issued. This achieves the technical effect of accurately detecting abnormalities in wind turbine ventilators and improving the safety and stability of large wind turbine components. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A schematic flow chart of a method for detecting abnormalities in a wind turbine respirator based on image segmentation according to an embodiment of the present application; Figure 2 Schematic diagram of the structure of the wind turbine respirator anomaly detection system based on image segmentation provided in an embodiment of the present application.

[0011] Description of the accompanying drawings: spatiotemporal image frame determination module 10 , corrected image frame determination module 20 , first-order detection module 30 , second-order detection module 40 , abnormality alarm module 50 . DETAILED DESCRIPTION

[0012] This application provides a wind turbine respirator anomaly detection method and system based on image segmentation, which is used to solve the technical problems in the existing technology that the wind turbine respirator anomaly detection accuracy is insufficient and it is difficult to adapt to complex working conditions, resulting in low safety and stability of large components of the wind turbine.

[0013] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work are within the scope of protection of this application.

[0014] Example 1, as Figure 1As shown, the present application provides a wind turbine respirator anomaly detection method based on image segmentation, the method comprising: Step S100: using an image acquisition device, the ventilator deployment area is acquired to determine a spatiotemporal image frame.

[0015] Specifically, the image acquisition device captures the respirator deployment area to determine spatiotemporal image frames, which is the initial step in the wind turbine respirator anomaly detection process. The captured spatiotemporal image frames are then transmitted back to trigger the first image processor. This processor must first be constructed based on the light field conditions of the respirator assembly area (including at least dark field and backlight field). Through adversarial supervised training for different light field conditions, scalable processing branches are formed, and the input end is set by combining prior nodes of image quality and light field conditions. At the same time, the processed spatiotemporal image frames are used to trigger the respirator full life cycle detector. This detector is constructed based on attention factors of the respirator area and internal particle color levels. Ultimately, it serves to output anomaly detection data and subsequently issue alarms and instructions to the central control terminal. It is the foundation of image data acquisition in the entire detection method, providing raw image information for subsequent steps such as polarization compensation, pixel segmentation, and anomaly judgment.

[0016] Step S200: returning the spatiotemporal image frame, triggering a first image processor to perform polarization compensation based on active structured light projection guided by directional requirements, and determining a corrected image frame.

[0017] Specifically, the spatiotemporal image frame is transmitted back to trigger the first image processor. This processor must first be constructed: by acquiring light field conditions such as dark and backlit fields from the respirator assembly area, adversarially supervised training of an image discriminator-projection generator architecture is performed for each of these light field conditions, forming the first and second processing branches. These parallel and scalable branches together constitute the first image processor. Once constructed, the first prior node based on image quality and the second prior node based on light field conditions are deployed, and the processor inputs are configured accordingly. Upon triggering, the spatiotemporal image frame is first identified, and image quality priors based on the first prior node and light field condition priors based on the second prior node are executed. If image quality meets the requirements, the processor idles. The light field condition prior results then trigger the corresponding processing branch (which includes multiple polarization channels deployed using a joint polarization-depth self-attention scheme). Polarization compensation is then performed using active structured light projection guided by directional demand, ultimately determining the corrected image frame and providing an optimized image foundation for subsequent pixel scanning and segmentation.

[0018] Step S300: performing pixel scanning on the corrected image frame, triggering a ventilator full life cycle detector, performing first-order segmentation by locating a first pixel corner point based on the ventilator area, and determining a ventilator pixel matrix.

[0019] Specifically, pixel scanning is performed on the corrected image frame to trigger the respirator full life cycle detector. The construction of this detector requires the introduction of the first attention factor of the respirator area and the second attention factor of the particle color level inside the respirator, and uses these two factors as constraints and corner pixel positioning as the identification target; after triggering, the detector first scans the corrected image frame to determine the image pixel matrix, and then performs the first corner pixel positioning and corner boundary determination based on the first attention factor on the matrix, wherein the corner positioning is performed with the preset neighborhood pixel difference, and the positioning result is filtered with the first attention factor, and finally the image pixel matrix is ​​segmented according to the determined segmentation pixel boundary to obtain the respirator pixel matrix, which is the basis for the subsequent second segmentation and ratio calculation based on the particle color level inside the respirator. In the whole process, the first image processor and the respirator full life cycle detector are embedded and deployed in the fan management system, and the input end is connected to the image acquisition device to provide support for the continuity of the detection process.

[0020] Step S400: For the ventilator pixel matrix, perform a second segmentation based on second pixel corner location of the particle color scale inside the ventilator and a ratio calculation based on corner pixel location, and output abnormality detection data, wherein the calculation target is the ratio of particles of different color scales inside the ventilator.

[0021] Specifically, for the determined respirator pixel matrix, a second segmentation based on the second pixel corner point positioning of the color level of the particles inside the respirator and a ratio calculation based on the corner pixel positioning need to be performed, and finally the abnormality detection data is output, and the calculation target is clearly the ratio of particles of different color levels inside the respirator. Among them, the second pixel corner point specifically refers to the pixel point where the color scale transition between the first chromaticity and the second chromaticity in the particles inside the respirator occurs. When performing positioning, if the color scale contrast between the first chromaticity and the second chromaticity is less than the preset contrast, pixel contrast enhancement processing must be performed first, and then a ratio calculation based on pixel segmentation is performed based on the color scale pixel boundary obtained by positioning. At the same time, after outputting the abnormality detection data, a first state diagram of the respirator's service status is generated based on the color scale pixel boundary. Combined with the abnormality detection data, the service status curve with the frame timing as the axis is updated to generate a life cycle progress bar. Based on the curve, the lower-level timing is supplemented by situation timing prediction to perform precursor warning management. These abnormality detection data, together with the relevant state diagrams, curves, and progress bars, constitute a multivariate data model, which provides the core basis for the subsequent alarm information generation and interface display of the central control terminal. The entire process relies on a respirator full life cycle detector constructed with the respirator area and internal particle color scale as attention elements.

[0022] Step S500: generating alarm information based on the abnormality detection data according to the central control terminal of the fan management system, executing the abnormality alarm of the ventilator and issuing a replacement instruction, wherein the multivariate data mode is displayed on the display interface.

[0023] Specifically, the central control terminal of the fan management system generates corresponding alarm information based on the aforementioned output abnormality detection data, and then executes the issuance of ventilator abnormality alarms and replacement instructions, while also displaying the information on the display interface in a multivariate data model. This multivariate data model includes the abnormality detection data itself, as well as a first-state diagram of the ventilator's service status generated based on color-scale pixel boundaries, a service status curve with frame timing as the axis and updated with abnormality detection data, a lifecycle progress bar generated based on this curve, and related content of the curve after supplementing the lower-level timing through situation timing prediction. The central control terminal can achieve these functions thanks to the fact that the first image processor and the ventilator full lifecycle detector are both embedded and deployed in the fan management system, and their outputs are connected to the central control terminal, forming a closed loop for the entire abnormality detection process, ensuring the timeliness and accuracy of alarms, instruction issuance, and data display, and providing comprehensive support for staff to understand the ventilator status and take countermeasures.

[0024] In one possible implementation, step S200 further includes: Step S210: obtaining light field conditions of the respirator assembly area, wherein the light field conditions at least include a dark field and a backlight field.

[0025] Step S220: for the light field condition, for the dark field condition, performing adversarial supervised training of the image discriminator-projection generator architecture as the first processing branch.

[0026] Step S230: For the inverse light field condition, perform adversarial supervised training of the image discriminator-projection generator architecture as the second processing branch.

[0027] Step S240: performing the first processing branch and the second processing branch in parallel to determine the first image processor, wherein the processing branch of the first image processor is expandable.

[0028] Specifically, the image acquisition equipment is used to collect light field data from the respirator assembly area at multiple time periods and angles during the initial deployment phase. The built-in light sensor records parameters such as light intensity, polarization direction, and incident angle in different environments. At the same time, combined with auxiliary light field monitoring devices deployed around the assembly area, the feature data of dark fields (such as low light and shadow-covered scenes) and backlight fields (such as direct strong light and scenes with the light source behind the subject) are continuously captured. These data are integrated into a light field condition data set, providing the original light field feature basis for the subsequent construction of the first processing branch and the second processing branch for the dark field and backlight field respectively, ensuring that the first image processor can accurately adapt to the image processing requirements under different light field environments.

[0029] When performing adversarial supervised training of the image discriminator-projection generator architecture for dark field conditions to construct the first processing branch, a generative adversarial network (GAN) is used as the basic framework. The projection generator takes raw dark field image data as input and, through alternating convolutional and deconvolutional layers, generates dark-field-specific active structured light projection parameters and outputs a polarization-compensated predicted image. The image discriminator receives the predicted image output by the generator and a standard calibrated dark field image (as a real sample), extracts features from both, distinguishes authenticity from falsity using a multi-layer perceptron or convolutional neural network, and outputs a probability of recognition. During training, the generator optimizes network parameters using a loss function (such as cross-entropy loss) to minimize the discriminator's accuracy, making the generated compensated image closer to the real sample. The discriminator simultaneously optimizes its parameters to maximize its accuracy, forming an adversarial game between the two. After multiple rounds of iterative training, when the images generated by the generator can make the discrimination accuracy of the discriminator close to the random level (such as 50%), the training converges. At this time, the model composed of the projection generator and the image discriminator serves as the first processing branch, which has the ability to process polarization compensation for dark field conditions.

[0030] When constructing the second processing branch through adversarial supervised training of the image discriminator-projection generator architecture for backlighting conditions, this is achieved based on a generative adversarial network (GAN) framework: the projection generator receives raw image data under backlighting conditions, extracts image features through convolutional layers, and combines this with deconvolutional layers to generate active structured light projection parameters adapted to the backlighting conditions, thereby outputting a polarization-compensated image. The image discriminator then compares the features of the compensated image output by the generator with a standard calibrated image (real sample) under backlighting conditions, and outputs a discrimination probability (determining whether the image is a generated result or a real sample) through a multi-layer perceptron. During training, the generator aims to minimize the discriminator's recognition accuracy and uses a mean squared error loss function to optimize parameters so that the generated image approximates the standard calibrated image. The discriminator aims to maximize recognition accuracy and adjusts its parameters using a cross-entropy loss function, forming an adversarial iteration between the two. After multiple rounds of training, when the images generated by the generator make the discrimination accuracy of the discriminator approach the random level, the training converges. At this time, the model composed of the generator and the discriminator serves as the second processing branch and can be specifically used for polarization compensation processing under inverse light field conditions.

[0031] The first processing branch trained for dark field conditions and the second processing branch trained for inverse light field conditions are set in parallel to determine the first image processor. This means that the two processing branches can independently and simultaneously respond to image processing requirements under different light field conditions. When the input spatiotemporal image frame involves a dark field environment, the first processing branch can be triggered to operate, and if it involves an inverse light field environment, the second processing branch can be activated for processing. At the same time, the processing branches of the first image processor are scalable. In addition to the already constructed dark field and inverse light field processing branches, corresponding processing branches can be added based on other light field conditions that may be encountered in actual applications (such as side light fields, scattered light fields, etc.) through the same image discriminator-projection generator architecture adversarial supervised training method. This allows the first image processor to adapt to more complex and diverse light field environments, providing flexible and comprehensive processing capabilities for subsequent precise polarization compensation of spatiotemporal image frames.

[0032] In one possible implementation, step S240 further includes: Step S241: deploying a first priori node based on image quality and deploying a second priori node based on light field conditions.

[0033] Step S242: setting the input end of the first image processor according to the first a priori node and the second a priori node.

[0034] Specifically, when deploying the first priori node based on image quality, an image clarity assessment algorithm based on the Laplace operator is used. By calculating the second-order derivative change of the image pixel grayscale value, the edge sharpness value of the image is obtained, and it is compared with the preset clarity threshold to determine whether the image quality meets the standard. At the same time, combined with a noise detection algorithm based on a Gaussian noise model, the distribution density of noise pixels in the image is statistically analyzed to assist in verifying the image quality. When deploying the second priori node based on light field conditions, a light field type recognition algorithm based on pixel brightness histogram analysis is used to extract the brightness distribution characteristics of the image. If the brightness mean is lower than the dark field threshold and the distribution is concentrated, it is determined to be a dark field; if the brightness mean is higher than the inverse light field threshold and there is a clear strong light area, it is determined to be an inverse light field. This algorithm can achieve rapid identification of light field conditions, providing an algorithm-level judgment basis for the subsequent triggering of the first image processor processing branch.

[0035] The input end of the first image processor is set according to the first prior node and the second prior node, and is implemented by establishing a linkage mechanism with the spatiotemporal image frame recognition. The input end will first identify the spatiotemporal image frame, and then execute the image quality prior based on the first prior node to determine the first prior result, and at the same time execute the light field condition prior based on the second prior node to determine the second prior result; if the first prior result meets the standard, the input end will control the first image processor to idle, and then trigger the corresponding target processing branch (first processing branch or second processing branch) according to the second prior result. Since the processing branch has built-in multi-polarization channels based on the joint self-attention deployment of polarization-depth, the setting of the input end needs to ensure that these channels can operate efficiently with the triggering of the target processing branch, so that the first image processor can perform polarization compensation in a targeted manner to ensure the accuracy of the corrected image frame.

[0036] In one possible implementation, step S242 further includes: Step S2421: Identify the spatiotemporal image frame, perform image quality prior based on the first prior node, and determine a first prior result.

[0037] Step S2422: executing the light field condition prior based on the second prior node to determine a second prior result.

[0038] If the first a priori result meets the criteria, the first image processor is idled, and the target processing branch is triggered based on the second a priori result, where the target processing branch is the first processing branch or the second processing branch. The processing branch has built-in multi-polarization channels based on a joint self-attention deployment of polarization-depth.

[0039] Specifically, the returned spatiotemporal image frames are first identified. This identification process mainly extracts the basic features of the image as the basis for subsequent processing. Subsequently, the image quality prior is executed based on the deployed first prior node, that is, the quality of the identified spatiotemporal image frames is evaluated based on the preset standards related to image quality in the first prior node (such as clarity, noise level and other indicators). By comparing the evaluation result with the preset threshold, the first prior result (meets the standard or does not meet the standard) is finally determined, providing a basis for judging the image quality for the subsequent triggering of the processing branch of the first image processor.

[0040] Execute the light field condition prior based on the second prior node, that is, according to the preset standards related to the light field conditions in the deployed second prior node, analyze and judge the light field environment of the identified spatiotemporal image frame, and focus on identifying whether it belongs to a defined light field type such as a dark field or a backlight field; by comparing the analysis result with the preset light field condition characteristics, the second prior result is finally determined, providing a judgment basis for the light field environment for the subsequent triggering of the target processing branch of the first image processor (the first processing branch or the second processing branch).

[0041] When the first prior result meets the target, the first image processor enters an idle state, performing no additional preprocessing and directly triggering the corresponding target processing branch based on the second prior result. If the light field condition is dark, the first processing branch is triggered; if the light field condition is inverted, the second processing branch is triggered. The multi-path polarization channel built into the processing branch is constructed using a joint self-attention deployment of polarization and depth. First, independent polarization-sensing sub-channels are set up for different polarization directions (such as 0°, 45°, 90°, and 135°), each equipped with a dedicated polarization filter module to capture polarization information in the corresponding direction. Simultaneously, a depth estimation sub-channel is integrated to obtain scene depth data through phase difference calculation using structured light projection. Based on this, a joint self-attention mechanism is constructed to perform cross-channel correlation analysis between the features of each polarization sub-channel and the features of the depth sub-channel, focusing on the coordinated changes in polarization characteristics and depth information in the respirator region. Weighted weighting is then used to enhance the feature response in key areas, ultimately forming a multi-path coordinated polarization compensation processing channel to accurately address polarization interference under different light field conditions and improve the quality of the corrected image frames.

[0042] In one possible implementation, step S300 further includes: Step S310: introducing a first attention element, wherein the first attention element is a respirator area.

[0043] Step S320: introducing a second attention factor, wherein the second attention factor is the color scale of particles inside the respirator.

[0044] Step S330: constructing the respirator full life cycle detector based on the first attention factor and the second attention factor as constraints and the positioning of corner pixels as the recognition target.

[0045] Specifically, the first attention factor is introduced, with the respirator region explicitly defined as its specific content. This means that when building a full-lifecycle respirator detector, attention is prioritized to the respirator region in the image. By defining the characteristic range of the respirator region (such as shape, position, and other attributes), the detector can focus specifically on this region during subsequent image processing, filtering out interference from irrelevant background information. This lays the foundation for accurately locating the pixel corners of the respirator region, ensuring that subsequent segmentation operations can accurately focus on the respirator region.

[0046] A second attention factor is introduced, specifically specifying the color gradation of particles inside the respirator. This means that when building a full-lifecycle respirator detector, in addition to focusing on the overall respirator area, the color gradation of particles inside the respirator will also be specifically examined. By defining the characteristic ranges of different color gradations (such as brightness and saturation), the detector can accurately identify and distinguish the color gradation differences of these particles in subsequent processing, providing a targeted focus for color gradation-based segmentation and ratio calculations, thereby more accurately capturing the state changes within the respirator.

[0047] The specific implementation method of constructing a full-life cycle detector for respirators is as follows: first, a detection model framework based on deep learning is built, and the first attention factor (respirator area) and the second attention factor (color level of particles inside the respirator) are converted into constraint parameters for model training. By masking the respirator area and classifying and labeling the internal particle color levels in the training data, a constrained training sample set is formed; then, a corner detection layer is embedded in the model, and the improved Harris corner detection algorithm is used as the basic operator. Combined with the attention mechanism, the weight parameters of the first and second attention factors are integrated into the corner response function, so that the model gives priority to the respirator area and the area with significant internal color level changes during feature extraction, thereby enhancing the positioning accuracy of corner pixels; the model parameters are optimized through multiple rounds of iterative training, so that the detector can accurately locate the corner pixels that meet the constraints in the input image, and finally a full-life cycle detector for respirators is constructed with corner pixel positioning as the core recognition target and constrained by dual attention factors.

[0048] In one possible implementation, step S300 further includes: Step S340: Scanning the calibrated image frame according to the respirator full life cycle detector to determine an image pixel matrix.

[0049] Step S350: For the image pixel matrix, perform first corner pixel positioning and corner boundary based on the first attention element to determine the segmentation pixel boundary, wherein the first corner pixel positioning is performed with a preset neighborhood pixel difference, and the positioning result is screened with the first attention element.

[0050] Step S360: Segment the image pixel matrix according to the segmented pixel boundaries to determine a ventilator pixel matrix.

[0051] Specifically, based on the established full-lifecycle detector for respirators, a comprehensive pixel-level scan of the polarization-compensated corrected image frame is performed. During this scan, the detector reads information about each pixel in the corrected image frame, row by row and column by column, including its coordinate position, grayscale value, and color parameters. This information is systematically organized into an ordered matrix to determine the image pixel matrix. This matrix fully represents the pixel distribution of the corrected image frame, providing the underlying data for subsequent segmentation operations based on corner pixel location.

[0052] For the determined image pixel matrix, the first corner pixel location and corner boundary determination are performed using the first attention factor (the respirator region) as a guide. Corner pixels are identified by using a preset neighborhood pixel difference method. This method calculates the grayscale value difference between each pixel and the pixels within a preset range of its surroundings. When the difference reaches a set threshold, the pixel is marked as a potential corner. Subsequently, these potential corners are filtered based on the first attention factor, retaining the corner pixels belonging to the respirator region. The corner boundary is then formed by connecting these filtered corner pixels, and the segmentation pixel boundary is finally determined, providing an accurate boundary basis for the subsequent segmentation of the respirator pixel matrix.

[0053] Based on the determined segmentation pixel boundaries, a segmentation operation is performed on the image pixel matrix. By defining the range enclosed by the segmentation pixel boundaries, the pixels in the image pixel matrix within this range are screened out. These pixels together constitute the pixel set of the respirator area. At the same time, other pixels outside the boundaries are excluded, and finally a pixel matrix containing only the respirator area is formed, that is, the respirator pixel matrix is ​​determined, providing an accurate pixel-level data foundation for subsequent further detection of particles inside the respirator.

[0054] In one possible implementation, step S400 further includes: Step S410: For the respirator pixel matrix, perform second corner pixel positioning based on the second attention factor to determine the color scale pixel boundary, wherein the second corner pixel is a pixel point where the color scale of the first chromaticity and the second chromaticity in the particles inside the respirator changes gradually, wherein if the color scale contrast between the first chromaticity and the second chromaticity is less than the preset contrast, perform pixel contrast enhancement processing.

[0055] Step S420: performing pixel segmentation-based ratio calculation according to the color scale pixel boundary to generate abnormality detection data.

[0056] Specifically, within the respirator pixel matrix, a second corner pixel location is performed, focusing on the second attention factor (the color scale of the particles within the respirator) to determine the color scale pixel boundaries. The second corner pixels here specifically refer to the pixels where the color scale transitions between the first and second chromaticities within the particles within the respirator. The color difference between normal and abnormal particles is predefined, for example, blue is set as the first chromaticity (representing a normal color) and white is set as the second chromaticity (representing an abnormal color). Then, using a pixel corner location algorithm, the boundary between the blue and white particle regions is automatically identified by analyzing the pixel-level distribution characteristics of the two color particles. The pixels on these boundaries are the corner pixels. Based on this, the ratio of the number of pixels in the blue and white particle regions can be directly calculated based on the demarcated blue and white particle ranges. This ratio is used to determine whether the respirator status is abnormal. Furthermore, if the color scale contrast between the first chromaticity (blue) and the second chromaticity (white) falls below a preset contrast threshold, pixel contrast enhancement is performed to improve the distinction between the two colors, ensuring the accuracy of boundary identification and subsequent calculations.

[0057] Based on the determined color-scale pixel boundaries, the particle regions of different color scales within the respirator's pixel matrix are segmented at the pixel level, defining the pixel ranges occupied by particles of the first chromaticity (e.g., the normal color blue) and particles of the second chromaticity (e.g., the abnormal color white). Subsequently, the number of pixels within these two regions is counted, and the ratio between them is calculated (e.g., the ratio of blue particle pixels to white particle pixels, or the ratio of white particle pixels to the total particle pixels). By comparing this ratio with a preset normal range, detection data is generated that reflects whether the particle status within the respirator is abnormal, i.e., abnormal detection data, providing a basis for subsequent warnings and command issuance.

[0058] In one possible implementation, step S420 further includes: Step S421: generating a first state diagram based on the service state of the respirator according to the color scale pixel boundary.

[0059] Step S422: updating the service status curve of the ventilator according to the abnormality detection data, wherein the service status curve is axially oriented with the frame time sequence.

[0060] Step S423: Generate a life cycle progress bar of the respirator according to the service status curve.

[0061] Step S424: Based on the service status curve, taking the ventilator specifications and operating mechanism as the benchmark, by performing situation time series prediction, the service status curve is supplemented with lower time series according to the predicted trend, wherein ventilator precursor warning management is performed according to the predicted trend.

[0062] Step S425: wherein the abnormality detection data, the first state diagram and the service state curve, and the life cycle progress bar are used as a multivariate data model and displayed and updated on a visual interface.

[0063] Specifically, the specific implementation method for generating the first state diagram based on the service status of the respirator is as follows: first, the different color scale particle areas defined by the color scale pixel boundaries are analyzed, and the features of the blue particle area representing the normal state and the white particle area representing the abnormal state are extracted respectively; then, a simplified atlas form is used to present the blue particle area as a continuous blue filled block, and a specific code "N" (representing normal) is assigned, and the white particle area is presented as an interval white grid block, and a specific code "A" (representing abnormal) is assigned; at the same time, the pixel ratio data of each area is marked on the edge of the atlas. By converting the physical acquisition image into an intuitive map containing color blocks, codes and quantitative data, a first state diagram is formed, which clearly reflects the service status of the respirator.

[0064] The respirator's service status curve is updated based on the generated anomaly detection data. This curve, with frame timing as the axis, records the ratio data of different color-level particles determined by each test from the time the respirator is installed to the current state. This data is chronologically constructed into a graph with a time-series frame rate. By mapping newly acquired anomaly detection data to the corresponding frame timing position, the curve is supplemented or corrected, allowing the service status curve to dynamically and continuously reflect the changes in the respirator's status at different time points, intuitively presenting its status trends during service.

[0065] A lifecycle progress bar is generated based on the updated respirator service status curve. This bar uses the service status curve to show the changing state of the respirator from installation to the current state, and combines this with the respirator's expected service life to divide the entire lifecycle into several stages. By calculating the current state within the lifecycle, this bar is presented as an intuitive progress bar. For example, different colors are used to distinguish between normal, warning, and abnormal stages. The progress bar clearly displays the percentage of time the respirator has been used and the remaining lifespan, helping users quickly understand the stage of the lifecycle.

[0066] The long short-term memory network (LSTM) algorithm is used for situation time series prediction: first, the abnormal detection data (particle ratios of different color levels) arranged in frame time series in the service status curve is used as the input sequence, and a multi-dimensional feature matrix is ​​constructed by combining the respirator specification parameters (such as rated service life, particle capacity threshold) and operation mechanism data (such as average daily operating time, load fluctuation coefficient); the long-term dependencies in the data sequence are captured through the gating mechanism of the LSTM network, and the model is trained to learn the change pattern under normal conditions; the trained model is used to predict the particle ratio change trend in the future period to obtain a predicted sequence; this sequence is supplemented into the service status curve as a lower-level time series to form a complete trend curve; at the same time, an early warning threshold is set (such as the abnormal particle ratio reaches 30%). If the value of a certain time series point in the predicted trend touches or exceeds the threshold, the early warning mechanism is triggered and a precursor warning signal is generated.

[0067] Abnormal detection data, the first state diagram, the service status curve, and the life cycle progress bar are integrated into a multivariate data model and displayed and updated in a visual interface. Abnormal detection data presents the ratio of particles of different color levels inside the respirator in numerical form, intuitively reflecting the current degree of abnormality; the first state diagram displays the spatial characteristics of particle distribution through simplified diagrams or coding forms; the service status curve uses frame time series as the axis to dynamically present historical and predicted state change trends; and the life cycle progress bar indicates the life cycle stage of the respirator in the form of progress percentage. These data are arranged in different regions in the interface. Through the linkage and interactive function, users can click on any data item to view detailed related information, realizing the synchronous update and comprehensive display of multivariate data, and providing a comprehensive and intuitive reference for the respirator status.

[0068] In one possible implementation, step S500 further includes: Step S510: The first image processor and the ventilator life cycle detector are embedded and deployed in the fan management system, with the input end connected to the image acquisition device and the output end connected to the central control end of the fan management system.

[0069] Specifically, the first image processor and the ventilator lifecycle detector are embedded within the fan management system, forming part of the system. The inputs of these two components are connected to the image acquisition device to receive spatiotemporal image frames of the ventilator deployment area captured by the image acquisition device. Their outputs are connected to the fan management system's central control terminal to transmit processed corrected image frames, anomaly detection data, and other information to the central control terminal, providing data support for operations such as generating alarm information, executing alarms, and issuing replacement instructions, thereby achieving the integrated operation of the entire anomaly detection process within the fan management system.

[0070] Embodiment 2 is based on the same inventive concept as the wind turbine respirator abnormality detection method based on image segmentation in the above embodiment. Figure 2 As shown, the present application provides a wind turbine respirator anomaly detection system based on image segmentation. The system and method embodiments in the present application are based on the same inventive concept. The system includes: The spatiotemporal image frame determination module 10 is used to collect images of the ventilator deployment area using an image acquisition device and determine spatiotemporal image frames.

[0071] The corrected image frame determination module 20 is configured to transmit back the spatiotemporal image frame, trigger the first image processor, perform polarization compensation based on active structured light projection guided by directional requirements, and determine the corrected image frame.

[0072] The first-order detection module 30 is configured to perform pixel scanning on the corrected image frame, trigger a ventilator full life cycle detector, perform first-order segmentation by locating first pixel corners based on the ventilator region, and determine a ventilator pixel matrix.

[0073] The second-order detection module 40 is used to perform a second segmentation based on the second pixel corner point positioning of the particle color scale inside the respirator and a ratio calculation based on the corner pixel positioning for the respirator pixel matrix, and output anomaly detection data, wherein the calculation target is the ratio of particles of different color scales inside the respirator.

[0074] The abnormality alarm module 50 is used to generate alarm information based on the abnormality detection data according to the central control terminal of the fan management system, execute the abnormality alarm of the ventilator and issue a replacement instruction, wherein the multivariate data mode is displayed on the display interface.

[0075] Furthermore, the system is also used to implement the following functions: Obtain light field conditions of a respirator assembly area, wherein the light field conditions include at least a dark field and an inverse light field; for the light field conditions, perform architectural adversarial supervised training of an image discriminator-projection generator for the dark field conditions as a first processing branch; for the inverse light field conditions, perform architectural adversarial supervised training of an image discriminator-projection generator as a second processing branch; and perform the first processing branch and the second processing branch in parallel to determine the first image processor, wherein the processing branch of the first image processor is expandable.

[0076] Furthermore, the system is also used to implement the following functions: A first a priori node is deployed based on image quality, and a second a priori node is deployed based on light field conditions; and an input end of the first image processor is set according to the first a priori node and the second a priori node.

[0077] Furthermore, the system is also used to implement the following functions: Identify the spatiotemporal image frame, execute the image quality prior based on the first prior node, and determine a first prior result; execute the light field condition prior based on the second prior node, and determine a second prior result; wherein, if the first prior result is up to standard, idle the first image processor, and trigger the target processing branch according to the second prior result, wherein the target processing branch is the first processing branch or the second processing branch; wherein the processing branch has a built-in multi-polarization channel based on the joint self-attention deployment of polarization-depth.

[0078] Furthermore, the system is also used to implement the following functions: A first attention factor is introduced, wherein the first attention factor is the respirator area; a second attention factor is introduced, wherein the second attention factor is the color level of particles inside the respirator; based on the first and second attention factors as constraints and the positioning of corner pixels as the identification target, the respirator full life cycle detector is constructed.

[0079] Furthermore, the system is also used to implement the following functions: According to the respirator full life cycle detector, the corrected image frame is scanned to determine the image pixel matrix; for the image pixel matrix, first corner pixel positioning and corner boundary based on the first attention factor are performed to determine the segmentation pixel boundary, wherein the first corner pixel positioning is performed with a preset neighborhood pixel difference, and the positioning result is screened with the first attention factor; according to the segmentation pixel boundary, the image pixel matrix is ​​segmented to determine the respirator pixel matrix.

[0080] Furthermore, the system is also used to implement the following functions: For the respirator pixel matrix, second corner pixel positioning based on the second attention factor is performed to determine the color scale pixel boundary, wherein the second corner pixel is a pixel point where the color scale of the first chromaticity and the second chromaticity in the particles inside the respirator changes gradually, wherein if the color scale contrast between the first chromaticity and the second chromaticity is less than a preset contrast, pixel contrast enhancement processing is performed; according to the color scale pixel boundary, a ratio calculation based on pixel segmentation is performed to generate abnormality detection data.

[0081] Furthermore, the system is also used to implement the following functions: According to the color scale pixel boundary, a first state diagram based on the service status of the ventilator is generated; according to the abnormality detection data, the service status curve of the ventilator is updated, wherein the service status curve takes the frame timing as the axis; according to the service status curve, a life cycle progress bar of the ventilator is generated; according to the service status curve, with the ventilator specifications and operation mechanism as the benchmark, by performing situation timing prediction, the service status curve is supplemented with lower-level timing according to the prediction trend, wherein the ventilator precursor warning management is executed according to the prediction trend; wherein the abnormality detection data, the first state diagram and the service status curve, and the life cycle progress bar are used as multivariate data models and displayed and updated on a visual interface.

[0082] Furthermore, the system is also used to implement the following functions: The first image processor and the respirator full life cycle detector are embedded and deployed in the fan management system, with the input end connected to the image acquisition device and the output end connected to the central control end of the fan management system.

[0083] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0084] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

[0085] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. A wind turbine respirator anomaly detection method based on image segmentation, characterized in that: The method comprises: According to the image acquisition equipment, the ventilator deployment area is collected to determine the spatiotemporal image frame; Transmitting back the spatiotemporal image frame, triggering a first image processor to perform polarization compensation based on active structured light projection guided by directional requirements, and determining a corrected image frame; Performing pixel scanning on the corrected image frame, triggering a full life cycle detector for the ventilator, performing first-order segmentation by locating a first pixel corner point based on the ventilator region, and determining a ventilator pixel matrix; For the ventilator pixel matrix, performing a second segmentation based on second pixel corner location of the particle color scale inside the ventilator and a ratio calculation based on the corner pixel location, and outputting abnormality detection data, wherein the calculation target is the ratio of particles of different color scales inside the ventilator; According to the central control terminal of the fan management system, alarm information based on the abnormal detection data is generated, and the abnormal alarm and replacement instructions of the ventilator are executed, wherein the multivariate data mode is displayed on the display interface.

2. The method according to claim 1, wherein Before triggering the first image processor, the construction of the first image processor includes: Acquiring light field conditions of a respirator assembly area, wherein the light field conditions at least include a dark field and a backlight field; For the light field condition, for the dark field condition, performing adversarial supervised training of the image discriminator-projection generator architecture as a first processing branch; For adverse light field conditions, perform adversarial supervised training of the image discriminator-projection generator architecture as the second processing branch; The first processing branch and the second processing branch are performed in parallel to determine the first image processor, wherein the processing branch of the first image processor is scalable.

3. The method according to claim 2, wherein After determining the first image processor, the method includes: Deploy the first prior node based on image quality and the second prior node based on light field conditions; An input terminal of the first image processor is set according to the first a priori node and the second a priori node.

4. The method according to claim 3, wherein Triggering the first image processor, including: Identifying the spatiotemporal image frame, performing an image quality prior based on the first prior node, and determining a first prior result; executing a light field condition prior based on the second prior node to determine a second prior result; If the first a priori result is up to standard, the first image processor is idle, and a target processing branch is triggered according to the second a priori result, wherein the target processing branch is the first processing branch or the second processing branch; Among them, the processing branch has built-in multi-polarization channels based on polarization-depth joint self-attention deployment.

5. The method according to claim 1, wherein Before triggering the respirator full life cycle detector, the construction of the respirator full life cycle detector includes: Introducing a first element of attention, wherein the first element of attention is a respirator area; Introducing a second attention factor, wherein the second attention factor is the color scale of particles inside the respirator; The respirator full life cycle detector is constructed based on the first and second attention factors as constraints and the positioning of corner pixels as the recognition target.

6. The method according to claim 1, wherein Perform first-order segmentation based on first pixel corner location of the respiratory region, including: Scanning the calibration image frame according to the respirator full life cycle detector to determine an image pixel matrix; For the image pixel matrix, performing first corner pixel positioning and corner boundary based on the first attention factor to determine the segmentation pixel boundary, wherein the first corner pixel positioning is performed using a preset neighborhood pixel difference, and the positioning results are screened using the first attention factor; The image pixel matrix is ​​segmented according to the segmented pixel boundaries to determine a respirator pixel matrix.

7. The method for detecting abnormality of a wind turbine respirator based on image segmentation according to claim 6, wherein: Performing a second segmentation based on second pixel corner location of particle color levels inside the respirator and a ratio calculation based on corner pixel location, including: For the respirator pixel matrix, performing second corner pixel positioning based on the second attention factor to determine a color scale pixel boundary, wherein the second corner pixel is a pixel point where the color scale of the first chromaticity and the second chromaticity in the particles inside the respirator gradually changes, and if the color scale contrast between the first chromaticity and the second chromaticity is less than a preset contrast, performing pixel contrast enhancement processing; According to the color scale pixel boundary, a ratio calculation based on pixel segmentation is performed to generate abnormality detection data.

8. The method according to claim 7, wherein After generating anomaly detection data, including: generating a first state map based on the service state of the respirator according to the color scale pixel boundary; updating a service status curve of the respirator according to the abnormality detection data, wherein the service status curve is oriented with a frame time sequence as an axis; generating a life cycle progress bar for the respirator according to the service status curve; Based on the service status curve, taking the specifications and operation mechanism of the respirator as a benchmark, by performing a situation time series prediction, the service status curve is supplemented with a lower time series according to the prediction trend, wherein the respirator precursor warning management is performed according to the prediction trend; The abnormality detection data, the first state diagram and the service state curve, and the life cycle progress bar are used as a multivariate data model and updated on a visual interface.

9. The wind turbine respirator anomaly detection method based on image segmentation according to claim 1, characterized in that: The first image processor and the respirator full life cycle detector are embedded and deployed in the fan management system, with the input end connected to the image acquisition device and the output end connected to the central control end of the fan management system.

10. The wind turbine respirator anomaly detection system based on image segmentation is characterized by: The system is used to implement the wind turbine respirator anomaly detection method based on image segmentation according to any one of claims 1 to 9, and the system comprises: a spatiotemporal image frame determination module, configured to collect images of the ventilator deployment area according to an image acquisition device and determine spatiotemporal image frames; a correction image frame determination module, configured to transmit back the spatiotemporal image frame, trigger the first image processor, perform polarization compensation based on active structured light projection guided by directional requirements, and determine the correction image frame; a first-order detection module configured to perform pixel scanning on the corrected image frame, trigger a full-lifecycle detector for the respirator, and perform first-order segmentation by locating a first pixel corner point based on the respirator region to determine a respirator pixel matrix; A second-order detection module is configured to perform, for the respirator pixel matrix, a second segmentation based on second pixel corner location of particle color levels inside the respirator and a ratio calculation based on corner pixel location, and output abnormality detection data, wherein the calculation target is the ratio of particles of different color levels inside the respirator; The abnormal alarm module is used to generate alarm information based on the abnormal detection data according to the central control terminal of the fan management system, execute the abnormal alarm and replacement instruction of the ventilator, and display the multivariate data mode on the display interface.

Images