Shell defect detection method and system for wind power control cabinet

Abstract

The invention provides a shell defect detection method and system for a wind power control cabinet, and belongs to the technical field of wind power equipment detection.The method comprises the steps that target data corresponding to a shell of the wind power control cabinet are determined based on detection requirements, and the target data comprise image data and point cloud data; processing the image data to obtain a target shell image, and processing the point cloud data to obtain target point cloud data; and obtaining a shell defect detection result of the wind power control cabinet according to the target shell image and the target point cloud data. According to the shell defect detection method and system of the wind power control cabinet, clearer target data is obtained through image data processing and point cloud data processing, various defects of the shell of the wind power control cabinet can be comprehensively and accurately detected through the target data, and the detection efficiency and accuracy are improved.

Description

Technical Field

[0001] The present disclosure belongs to the technical field of wind power equipment detection, and more particularly, relates to a method and system for detecting shell defects of a wind power control cabinet. Background Art

[0002] As an important part of a wind power generation system, the integrity of the shell of a wind power control cabinet is crucial for ensuring the normal operation of internal electrical equipment. Currently, there are certain limitations in the method for detecting shell defects of a wind power control cabinet; a single data source is difficult to accurately reflect the defect condition of the shell. Summary of the Invention

[0003] The purpose of the present disclosure is to provide a method and system for detecting shell defects of a wind power control cabinet, so as to improve the speed and accuracy of detecting shell defects of a wind power control cabinet.

[0004] In the first aspect of the embodiments of the present disclosure, a method for detecting shell defects of a wind power control cabinet is provided, including: Determining target data corresponding to the shell of the wind power control cabinet based on detection requirements, where the target data includes image data and point cloud data; Processing the image data to obtain a target shell image, and processing the point cloud data to obtain target point cloud data; Obtaining a detection result of shell defects of the wind power control cabinet according to the target shell image and the target point cloud data.

[0005] In the second aspect of the embodiments of the present disclosure, a system for detecting shell defects of a wind power control cabinet is provided, including: An image data processing module for processing the image data of the shell of the wind power control cabinet to obtain a target shell image; A point cloud data processing module for processing the point cloud data of the shell of the wind power control cabinet to obtain target point cloud data; A monitoring result output module for inputting the target shell image and the target point cloud data into a target classification model to obtain a detection result of shell defects of the wind power control cabinet.

[0006] The beneficial effects of the method and system for detecting shell defects of a wind power control cabinet provided by the embodiments of the present disclosure are as follows: By processing the image data and the point cloud data, clearer target data is obtained, and the target data can comprehensively and accurately detect various defects of the shell of the wind power control cabinet, improving the detection efficiency and accuracy. Description of the Drawings

[0007] To more clearly illustrate the technical solutions in the embodiments of the present disclosure, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings in the following description are only some embodiments of the present disclosure. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0008] Figure 1 It is a schematic flowchart of a method for detecting the shell defects of a wind power control cabinet provided by an embodiment of the present disclosure; Figure 2 It is a structural block diagram of a system for detecting the shell defects of a wind power control cabinet provided by an embodiment of the present disclosure; Figure 3 It is a schematic block diagram of an electronic device provided by an embodiment of the present disclosure. Detailed implementation manners

[0009] In the following description, for the purpose of illustration rather than limitation, specific details such as specific system structures and technologies are presented in order to thoroughly understand the embodiments of the present disclosure. However, those skilled in the art should understand that the present disclosure can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid unnecessary details from interfering with the description of the present disclosure.

[0010] To make the purpose, technical solutions, and advantages of the present disclosure clearer, the following will be described through specific embodiments in conjunction with the drawings.

[0011] Please refer to Figure 1 , Figure 1 It is a schematic flowchart of a method for detecting the shell defects of a wind power control cabinet provided by an embodiment of the present disclosure. The method includes: S101: Determine the target data corresponding to the shell of the wind power control cabinet based on the detection requirements; the target data includes image data and point cloud data.

[0012] In this embodiment, the detection requirements include: appearance defect detection requirements and structural integrity detection requirements; among them, the appearance defect detection requirements include: detecting fine cracks, depressions, and coating integrity on the shell surface, etc.; the structural integrity detection requirements include: the overall structure, welds, deformations, depressions, and hole position deviations of the control cabinet shell, etc. The detection requirements also include the user's requirement to detect multiple or one shell of the wind power control cabinet; Deploy a multi-sensor array based on detection requirements. The multi-sensor array includes: a high-resolution industrial camera array, a 3D lidar, and a structured light scanner. In this embodiment, image data and point cloud data are used as target data. The image data of the wind power control cabinet housing to be detected is obtained through the high-resolution industrial camera array. The point cloud data of the wind power control cabinet housing to be detected is obtained through the 3D lidar and the structured light scanner.

[0013] S102: Process the image data to obtain a target housing image, and process the point cloud data to obtain target point cloud data.

[0014] In this embodiment, for the image data, it is necessary to adjust the clarity of the image data to obtain target image data. The image data is easily affected by factors such as illumination, environment, and shooting angle, which affect the image quality. Therefore, it is necessary to process the image, such as adjusting the clarity of the image through methods such as image enhancement and image weakening, so as to obtain target image data. The surface texture and visible defect information of the housing are collected through industrial cameras; for example, surface defects (scratches, rust), texture anomalies (paint peeling), and micro-size changes (weld width fluctuations).

[0015] In this embodiment, the point cloud data of the housing is obtained through a 3D laser scanner to characterize deformation, depression, and hole position deviation. Since there may be noise points and outliers in the process of collecting point cloud data, it is necessary to perform filtering processing on the collected point cloud data. Common filtering methods include statistical filtering, radius filtering, bilateral filtering, etc. Through filtering processing, the noise points and outliers in the point cloud data can be effectively removed, and the quality of the point cloud data can be improved.

[0016] If multiple scanners are used or scanning is performed from different angles, it is necessary to register multiple point cloud data to obtain the complete point cloud data of the housing. Point cloud registration methods include the iterative closest point algorithm, feature-based registration algorithms, etc. Through point cloud registration, point cloud data from different sources can be aligned to the same coordinate system.

[0017] It is also necessary to segment different parts of the point cloud data. For example, segment the housing surface from the surrounding environment and mark the areas where defects may exist. A clustering-based segmentation method, such as the density-based spatial clustering of applications with noise (DBSCAN) algorithm, can be used to segment the point cloud data. Through point cloud segmentation, the target area can be separated from the complex point cloud data, facilitating subsequent geometric feature extraction.

[0018] S103: Obtain the detection result of the housing defects of the wind power control cabinet according to the target housing image and the target point cloud data.

[0019] In this embodiment, the image features of the target housing image and the geometric features of the target point cloud data are fused to obtain more comprehensive and accurate feature information. The feature-level fusion method can be used to splice the image features and geometric features together to form a new feature vector; or the decision-level fusion method can be used to perform defect judgment based on the image features of the target housing image and the geometric features of the target point cloud data respectively, and then fuse the judgment results. Machine learning or deep learning models are used to classify and identify the fused features to determine whether there are defects in the wind power control cabinet housing and the types of defects. For example, the target housing image and the target point cloud data are spatially and temporally aligned to generate a fused feature matrix; the fused feature matrix is input into a pre-trained defect classification model (for example, a random forest model, a support vector machine model) to output the types, positions, and severity levels of the housing defects.

[0020] As can be seen from the above, the housing defect detection method of the wind power control cabinet of the present disclosure processes the image data and the point cloud data to obtain clearer target data, and the target data can comprehensively and accurately detect various defects of the wind power control cabinet housing, improving the detection efficiency and accuracy.

[0021] In an embodiment of the present disclosure, processing the image data of the wind power control cabinet housing to obtain a target housing image includes: Performing grayscale processing on the image data based on an adaptive grayscale algorithm to obtain a first grayscale image; Performing denoising processing on the first grayscale image based on a filtering algorithm to obtain a target housing image.

[0022] In this embodiment, the adaptive grayscale takes into account the color distribution, contrast, and other characteristics of different regions of the first plastic pipe image. The image can be divided into multiple sub-regions, and the grayscale parameters are calculated for each sub-region, so that the converted grayscale image can better retain details and features in each local region. The grayscale processing is the process of converting a color image into a grayscale image. The first grayscale image is the grayscale image obtained after the image data is processed by the adaptive grayscale algorithm, which converts the color information in the original color image into grayscale values and retains the details of different regions by using the adaptive grayscale algorithm.

[0023] The filtering algorithm can reduce or eliminate the influence of noise by processing the pixel values in the image. The filtering algorithm can adopt mean filtering, Gaussian filtering, median filtering, etc. Different filtering algorithms are applicable to different types of noise. Denoising processing is to remove the noise introduced in the processes of image acquisition, transmission, etc., and improve the image quality. Noise can be manifested as random bright spots, dark spots or salt-and-pepper-like interferences in the image, affecting the analysis and understanding of the image content. The target housing image is the image obtained after denoising the first grayscale image by the filtering algorithm. After denoising, the noise interference in the image is reduced, and the true features of the housing are clearer, providing cleaner image data for subsequent processing.

[0024] In an embodiment of the present disclosure, when processing the image data of the wind power control cabinet housing to obtain the target housing image, it further includes: graying the original image in the image data based on the brightness partition weighting algorithm to generate a first grayscale image; Specifically, divide the RGB channels of the original image into a high-brightness area , a medium-brightness area and a low-brightness area ; dynamically adjust the gray weight coefficient according to the brightness partition.

[0025] The gray calculation formula is:

[0026] where is the first dynamic weight coefficient, is the second dynamic weight coefficient, is the third dynamic weight coefficient, .

[0027] For example, the second brightness threshold is greater than the first brightness threshold; the gray value of the low-brightness area is less than the first brightness threshold; the gray value of the medium-brightness area is greater than or equal to the first brightness threshold and less than the second brightness threshold; the gray value of the high-brightness area is greater than or equal to the second brightness threshold; the first brightness threshold and the second brightness threshold for dividing the high-brightness area, the medium-brightness area and the low-brightness area can be determined according to the gray value of each pixel of the original image; the first brightness threshold and the second brightness threshold can also be determined by the three-quantile method or an adaptive method (such as Otsu multi-threshold segmentation).

[0028] In this embodiment, based on the contrast (standard deviation) of the R, G, and B channels in the brightness area, calculate the weight of each brightness partition: The weight calculation formula is: w F =(σ R +σ G +σ B ) / σ F where F ∈ {R, G, B}, σ R is the contrast of the high - brightness area, σ G is the contrast of the medium - brightness area, and σ B is the contrast of the low - brightness area, and w R + w G + w B = 1.

[0029] In this embodiment, by enhancing the contrast of the highlight area and the details in the dark part, more image information can be retained compared with the traditional grayscale method.

[0030] In an embodiment of the present disclosure, denoising processing is performed on the first grayscale image based on a filtering algorithm to obtain a target housing image, including: Determining a first Gaussian kernel based on the resolution of the first grayscale image; Determining a target standard deviation based on the noise intensity and noise suppression coefficient of the first grayscale image; Adjusting the first Gaussian kernel based on the target standard deviation to obtain a target Gaussian kernel; Performing denoising processing on the first grayscale image based on the target Gaussian kernel to obtain a target housing image.

[0031] In this embodiment, during the image acquisition process, due to factors such as sensor noise and illumination changes, the grayscale - converted first grayscale image may contain noise. Noise will interfere with subsequent feature extraction and defect detection, so denoising processing is required. The present disclosure adopts an adaptive filtering method based on a Gaussian kernel, and dynamically adjusts the Gaussian kernel parameters according to the resolution and noise intensity of the image. The image resolution reflects the fineness of the image. Images with different resolutions require Gaussian kernels of different sizes for denoising. Images with high resolution contain more details and can use larger Gaussian kernels to smooth the image; images with low resolution use smaller Gaussian kernels to avoid over - blurring details.

[0032] Specifically, according to the resolution of the first grayscale image calculate the initial Gaussian kernel size ; The calculation formula is:

[0033] where W and H are the width and height of the image respectively, and is an odd number; to construct a first Gaussian kernel, with its initial standard deviation where R is a preset value, such as 0.3.

[0034] Calculate the noise intensity of the first grayscale image by the local variance method, and the calculation formula is:

[0035] Among them, is the number of non-overlapping local regions in the image, is the i th region's variance; According to the noise intensity and a preset noise suppression coefficient , calculate the target standard deviation , and the calculation formula is: ; The target Gaussian kernel is

[0036] Among them, x and y are the pixel coordinates within the Gaussian kernel.

[0037] In this embodiment, the adjusted target Gaussian kernel is used to perform a convolution operation on the first grayscale image, replacing the value of each pixel in the image with the weighted average of the pixels in its neighborhood, and the weights are determined by the Gaussian kernel. It can effectively smooth the image, remove noise, and retain the details of the image as much as possible.

[0038] In an embodiment of the present disclosure, based on a filtering algorithm, the first grayscale image is denoised to obtain a target housing image, and it further includes: In response to the noise intensity of the first grayscale image being greater than or equal to the first noise intensity, increase the noise suppression coefficient by the first step size; In response to the noise intensity of the first grayscale image being less than the first noise intensity, decrease the noise suppression coefficient by the second step size.

[0039] In this embodiment, the noise intensity of the first grayscale image is greater than or equal to the first noise intensity, and the noise suppression coefficient is increased by the first step size. When the noise intensity is large, increasing the noise suppression coefficient can enhance the denoising effect and make the image smoother.

[0040] The noise intensity of the first grayscale image is less than the first noise intensity, and the noise suppression coefficient is decreased by the second step size. When the noise intensity is small, decreasing the noise suppression coefficient can reduce the impact on the image details and avoid over-smoothing.

[0041] In this embodiment, calculate the first step size according to the first step size base value, the noise intensity, the first noise intensity, and a preset maximum noise intensity; Calculate the second step size according to the second step size base value, the noise intensity, and the first noise intensity; The calculation formula for the first step size is:

[0042] Among them, is the first step size base value, is the preset maximum noise intensity, is the first step size, is the first noise intensity; The calculation formula for the second step size is:

[0043] wherein, is the second step size base value, is the second step size.

[0044] Specifically, the first step size and the second step size adaptively change with the difference between the noise intensity and the threshold value, rapidly increasing the suppression intensity in strong noise and finely adjusting in weak noise; it can effectively smooth the image, remove noise, and at the same time retain the details of the image.

[0045] This embodiment further includes: adjusting the first step size and the second step size according to the mean gray difference of adjacent pixels (local contrast) in the first grayscale image; wherein, if the mean gray difference of adjacent pixels is greater than the preset value, the first step size and the second step size are calculated according to the attenuation coefficient; the attenuation coefficient can be obtained based on experience. In order to avoid the edge blurring problem of traditional Gaussian filtering and improve the defect detection accuracy, the present disclosure reduces the step size adjustment amplitude in high-contrast regions (which may be the defect edges) to prevent feature loss caused by excessive smoothing, achieving a balance between noise suppression and retaining image features.

[0046] In an embodiment of the present disclosure, processing the point cloud data of the wind power control cabinet housing to obtain target point cloud data includes: Preprocessing the point cloud data to obtain first point cloud data; Calculating the curvature information of the first point cloud data and adding the curvature information to the first point cloud data to obtain second point cloud data; Registering the second point cloud data by using the normal distribution transformation algorithm to obtain the target point cloud data.

[0047] In this embodiment, there are situations where the point cloud data is unevenly distributed, such as noise points and outliers. Preprocessing improves the quality of the point cloud data, providing high-quality point cloud data for subsequent curvature calculation.

[0048] Curvature is an important feature describing the local geometry of the point cloud surface, which can reflect the concavity and convexity of the point cloud surface; in the defect detection of the wind power control cabinet housing, the curvature information can help us better identify surface defects, such as cracks and holes.

[0049] The calculated curvature information is added as a new attribute to the first point cloud data to obtain the second point cloud data containing the curvature information; when the second point cloud data is registered, the voxels can be optimized through the curvature distribution dynamics; the registration efficiency of the normal distribution transformation algorithm for the second point cloud data and the alignment accuracy of the defect area are improved.

[0050] In one embodiment of the present disclosure, a normal distribution transformation algorithm is used to register the second point cloud data to obtain point cloud data, including: Determine the voxel size according to the point cloud density of the target point cloud data; determining a resizing factor for the voxel size based on the curvature information; Adjust the voxel size based on the resizing factor to obtain a target voxel size; The second point cloud data is registered based on the normal distribution transformation algorithm and the target voxel size to obtain the target point cloud data.

[0051] In this example, different densities of point cloud data require different voxel sizes for effective processing. When the point cloud density is high, using a smaller voxel size can retain more detail information; when the point cloud density is low, a larger voxel size can improve computational efficiency and avoid inaccurate statistical information caused by too few points in a voxel.

[0052] In areas with larger curvature, the point cloud surface changes more dramatically and may contain more detailed information; while in areas with smaller curvature, the point cloud surface is relatively smooth. Therefore, in order to better preserve the details of areas with larger curvature during the registration process, the voxel size needs to be adjusted according to the curvature information.

[0053] In one embodiment of the present disclosure, in response to a curvature value in the curvature information being less than a first curvature threshold, increasing the size adjustment coefficient by a third step size; In response to a curvature value in the curvature information being greater than a second curvature threshold, the resizing factor is decreased by a fourth step size.

[0054] In this embodiment, the point cloud density is determined by calculating the average number of points in a certain area of ​​the point cloud. The appropriate voxel size is determined based on a pre-set density-voxel size mapping relationship. The curvature value of each point in the point cloud is calculated, and then the size adjustment coefficient is determined based on the curvature distribution.

[0055] When the curvature value is small, it means that the surface of the area is relatively smooth, such as the large flat surface of the wind turbine control cabinet housing. Increasing the size adjustment factor at this time can enable subsequent processing (such as point cloud meshing, filtering, etc.) to be performed in a larger range, which helps to capture the subtle defects or feature changes that may exist in the area; When the curvature value is large, it indicates that the surface of this area is highly curved, and there may be features such as edges, corners, or defects. For example, around the mounting holes or at the welds of the wind power control cabinet housing. Reducing the size adjustment coefficient can make the processing more precise and focus on these local features.

[0056] Through the intelligent dynamic optimization of the size adjustment coefficient in the detection of the housing defects of the wind power control cabinet, the present disclosure achieves the balance between global features and local details, and improves the accuracy and efficiency of defect detection.

[0057] In an embodiment of the present disclosure, the method for detecting housing defects of the wind power control cabinet further includes: For each point in the second point cloud data, calculate its curvature gradient , which is defined as the curvature change rate between this point and the adjacent points in the neighborhood:

[0058] where N is the number of points in the neighborhood, is a point in the second point cloud data; is the curvature value of the th point; According to the curvature value and the curvature gradient , classify the point cloud area as: ≥ and ≥ ), with high curvature and drastic change, which can be the defect edge; Gentle transition area ( and ), with low curvature and gentle change, which can be a normal plane; Noise disturbance area ( and ), with low curvature and high gradient, which can be noise or microstructures; is the curvature gradient threshold, is the first curvature threshold.

[0059] In the smooth mutation area, in response to the curvature value in the curvature information being greater than the first curvature threshold, increase the size adjustment coefficient by the fifth step size; In the gentle transition area, in response to the curvature value in the curvature information being less than the second curvature value threshold, reduce the size adjustment coefficient by the sixth step size.

[0060] The calculation formula for the fifth step size is:

[0061] is the fifth step size, is the fifth step size base value.

[0062] The calculation formula for the sixth step size is as follows:

[0063] Wherein, is the sixth step size, is the base value of the sixth step size, is the threshold value of the second curvature value.

[0064] The present disclosure introduces the curvature gradient as the basis for region classification to distinguish the real defect edge (high curvature + high gradient) from the noise perturbation (low curvature + high gradient), avoiding the misjudgment problem of the traditional single curvature threshold, suppressing the noise interference while retaining the defect features, improving the registration robustness, and at the same time improving the registration accuracy of the defect region.

[0065] In an embodiment of the present disclosure, according to the target housing image and the target point cloud data, the housing defect detection result of the wind power control cabinet is obtained, including: Construct a fusion matrix according to the target housing image and the target point cloud data, and the matrix dimensions of the fusion matrix include: texture abnormality degree, curvature mutation coefficient, and three-dimensional deformation parameter; Input the fusion matrix into the target classification model to obtain the housing defect detection result of the wind power control cabinet; Wherein, the target classification model is obtained by training a random forest model based on the first training data; the first training data is the fusion matrix data of the historical defect samples.

[0066] In this embodiment, a fusion matrix is constructed according to the target image and the target point cloud data. The fusion matrix includes information in multiple dimensions, and the information in the dimensions can comprehensively and accurately describe the characteristics of the housing of the target wind power control cabinet. Among them, the matrix dimensions of the fusion matrix include: texture abnormality degree, curvature mutation coefficient, and three-dimensional deformation parameter; the texture abnormality degree is determined by performing a detailed texture analysis on the target image. Image processing algorithms are used to extract the texture features of the target image. The calculation of the curvature mutation coefficient depends on the target point cloud data. The point cloud data is an accurate record of the three-dimensional spatial information on the surface of the target object. By processing and analyzing these data, the curvature information of the object surface can be obtained; the curvature mutation coefficient is calculated for the defect candidate region in the target point cloud data. The three-dimensional deformation parameter is used to describe the shape change of the housing of the target wind power control cabinet in the three-dimensional space. The registered target point cloud data is aligned with the standard model of the housing of the target wind power control cabinet, and the deformation distance of each point is calculated to obtain the three-dimensional deformation situation of the housing of the target wind power control cabinet.

[0067] The target classification model is constructed based on the random forest algorithm; random forest is an ensemble learning method that combines multiple decision trees for classification and regression tasks, with high accuracy and robustness. The constructed fusion matrix is input into the trained target classification model, and the model analyzes and judges the fusion matrix according to the learned feature patterns; the defect detection results of the target wind power control cabinet shell are output, which can be specific defect types, such as scratches, dents, deformations, etc., or the severity level of the defects. For example, if the model determines that there are obvious texture abnormalities and curvature mutations in the wind power control cabinet shell corresponding to the fusion matrix, it can output the defect detection results of surface scratches and local dents, and give the corresponding severity level according to the specific situation of the defects.

[0068] In an embodiment of the present disclosure, obtaining the defect detection result of the wind power control cabinet shell according to the target shell image and the target point cloud data further includes: Calculating a first ratio according to the matrix dimension number and ratio formula of the fusion matrix, and determining a compensation value according to the first ratio, where the compensation value is used to compensate the basic depth of the decision tree in the target classification model; Performing weighted calculation on the compensation value and the basic depth to obtain the target depth of the decision tree in the target classification model.

[0069] In this embodiment, calculating a first ratio according to the number of the maximum feature categories and the minimum feature categories of the matrix dimensions in the fusion matrix and the ratio formula; when the first ratio is greater than the ratio threshold, calculating a compensation value according to the preset compensation value calculation formula; performing weighted calculation on the compensation value and the basic depth to obtain the target depth of the decision tree in the target classification model; constraining the target classification model according to the target depth; Among them, the features include: scratches, dents, deformations, etc.

[0070] In this embodiment, the compensation value calculation formula is:

[0071] Among them, is the compensation value, is the first ratio, is an intermediate parameter used to suppress the sudden increase in the compensation amount at high ; is the number of the minimum feature categories.

[0072] This embodiment can dynamically adjust the depth of the decision tree in the target classification model according to the feature information in the fusion matrix, thereby improving the performance of the model for defect detection of the target wind power control cabinet shell and better adapting to different data features and actual application scenarios.

[0073] Corresponding to the above-mentioned embodiment of the method for defect detection of the wind power control cabinet shellFigure 2 The structural block diagram of the outer shell defect detection device for a wind power control cabinet provided by an embodiment of the present disclosure. For the sake of convenience, only the parts related to the embodiments of the present disclosure are shown. Refer to Figure 2 , the outer shell defect detection device 20 of the wind power control cabinet includes: a target data determination module 21, a target data processing module 22, and a monitoring result output module 23.

[0074] Among them, the target data determination module 21 is used to determine the target data corresponding to the outer shell of the wind power control cabinet based on the detection requirements; the target data includes image data and point cloud data; The target data processing module 22 is used to process the image data to obtain a target outer shell image; it is also used to process the point cloud data to obtain target point cloud data; The monitoring result output module 23 is used to obtain the outer shell defect detection result of the wind power control cabinet according to the target outer shell image and the target point cloud data.

[0075] In an embodiment of the present disclosure, the target data processing module 22 is specifically used for: Based on the adaptive grayscale algorithm, perform grayscale processing on the image data to obtain a first grayscale image; Based on the filtering algorithm, perform denoising processing on the first grayscale image to obtain a target outer shell image.

[0076] In an embodiment of the present disclosure, the target data processing module 22 is specifically used for: Determine the first Gaussian kernel based on the resolution of the first grayscale image; Determine the target standard deviation based on the noise intensity and the noise suppression coefficient of the first grayscale image; Adjust the first Gaussian kernel based on the target standard deviation to obtain a target Gaussian kernel; Based on the target Gaussian kernel, perform denoising processing on the first grayscale image to obtain a target outer shell image.

[0077] In an embodiment of the present disclosure, the target data processing module 22 is specifically used for: In response to the noise intensity of the first grayscale image being greater than or equal to the first noise intensity, increase the noise suppression coefficient by the first step size; In response to the noise intensity of the first grayscale image being less than the first noise intensity, decrease the noise suppression coefficient by the second step size.

[0078] In an embodiment of the present disclosure, the target data processing module 22 is specifically used for: Perform preprocessing on the point cloud data to obtain first point cloud data; Calculate the curvature information of the first point cloud data, and add the curvature information to the first point cloud data to obtain second point cloud data; The second point cloud data is registered using the normal distribution transformation algorithm to obtain the target point cloud data.

[0079] In an embodiment of the present disclosure, the target data processing module 22 is specifically configured to: Determine the voxel size according to the point cloud density of the target point cloud data; Determine the size adjustment coefficient of the voxel size according to the curvature information; Adjust the voxel size based on the size adjustment coefficient to obtain the target voxel size; Register the second point cloud data based on the normal distribution transformation algorithm and the target voxel size to obtain the target point cloud data.

[0080] In an embodiment of the present disclosure, the target data processing module 22 is specifically configured to: In response to the curvature value in the curvature information being less than the first curvature threshold, increase the size adjustment coefficient by the third step length; In response to the curvature value in the curvature information being greater than the second curvature threshold, decrease the size adjustment coefficient by the fourth step length.

[0081] In an embodiment of the present disclosure, the monitoring result output module 23 is specifically configured to: Construct a fusion matrix based on the target shell image and the target point cloud data. The matrix dimensions of the fusion matrix include: texture anomaly degree, curvature mutation coefficient, and three-dimensional deformation parameter; Input the fusion matrix into the target classification model to obtain the shell defect detection result of the wind power control cabinet; Among them, the target classification model is trained based on the first training data for the random forest model; the first training data is the fusion matrix data of historical defect samples.

[0082] In an embodiment of the present disclosure, the monitoring result output module 23 is specifically configured to: Determine the basic depth of the decision tree in the target classification model based on the number of features of the fusion matrix; Calculate the first ratio according to the number of matrix dimensions of the fusion matrix and the ratio formula, and determine the compensation value according to the first ratio. The compensation value is used to compensate the basic depth of the decision tree in the target classification model; Perform weighted calculation on the compensation value and the basic depth to obtain the target depth of the decision tree in the target classification model.

[0083] The shell defect detection system of the wind power control cabinet provided by the embodiments of the present disclosure can comprehensively and accurately detect various defects of the shell of the wind power control cabinet by integrating image features and geometric features and combining advanced image processing and point cloud data processing technologies, improving the efficiency and accuracy of detection.

[0084] See Figure 3 ,Figure 3 A schematic block diagram of an electronic device provided by an embodiment of the present disclosure. As Figure 3 shown, the electronic device 300 in this embodiment may include: one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The above-mentioned processors 301, input devices 302, output devices 303, and memories 304 communicate with each other through a communication bus 305. The memory 304 is used to store computer programs, and the computer programs include program instructions. The processor 301 is used to execute the program instructions stored in the memory 304. Among them, the processor 301 is configured to call the program instructions to execute the functions of each module in the above device embodiments, for example Figure 2 shown, the functions of the target data determination module 22, the target data processing module 22, and the monitoring result output module 23.

[0085] It should be understood that in the embodiments of the present disclosure, the so-called processor 301 may be a central processing unit (CPU), and this processor may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or this processor may also be any conventional processor, etc.

[0086] The input device 302 may include a touchpad, a fingerprint acquisition sensor (for acquiring the fingerprint information and the direction information of the fingerprint of the user), a microphone, etc., and the output device 303 may include a display (such as an LCD), a speaker, etc.

[0087] The memory 304 may include a read-only memory and a random access memory, and provide instructions and data to the processor 301. A part of the memory 304 may also include a non-volatile random access memory. For example, the memory 304 may also store information about the device type.

[0088] In specific implementation, the processors 301, input devices 302, and output devices 303 described in the embodiments of the present disclosure may execute the implementation manners described in the first embodiment and the second embodiment of the method for detecting the shell defects of the wind power control cabinet provided by the embodiments of the present disclosure, and may also execute the implementation manner of the electronic device described in the embodiments of the present disclosure, which will not be elaborated here.

[0089] In another embodiment of the present disclosure, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program, and the computer program includes program instructions. When the program instructions are executed by a processor, all or part of the processes in the method of the above embodiment are implemented. It can also be completed by instructing related hardware through the computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal, and software distribution medium, etc.

[0090] The computer-readable storage medium can be an internal storage unit of the electronic device in any of the foregoing embodiments, such as the hard disk or memory of the electronic device. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device. Further, the computer-readable storage medium can also include both the internal storage unit and the external storage device of the electronic device. The computer-readable storage medium is used to store the computer program and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store the data that has been output or will be output.

[0091] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been generally described according to their functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present disclosure.

[0092] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the above-described electronic devices and units can refer to the corresponding processes in the foregoing method embodiments, and will not be described herein again.

[0093] In several embodiments provided by this application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed coupling or direct coupling or communication connection to each other can be an indirect coupling or communication connection through some interfaces or units, or can also be an electrical, mechanical or other form of connection.

[0094] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place, or can also be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiments of the present disclosure.

[0095] In addition, each functional unit in various embodiments of the present disclosure can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0096] The above is only the specific implementation manner of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present disclosure can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should be covered within the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be subject to the protection scope of the claims.

Claims

1. A method for detecting shell defects of a wind power control cabinet, characterized in that: include: Determine target data corresponding to the wind power control cabinet housing based on the detection requirements, wherein the target data includes image data and point cloud data; Processing the image data to obtain a target shell image, and processing the point cloud data to obtain target point cloud data; According to the target shell image and the target point cloud data, a shell defect detection result of the wind power control cabinet is obtained.

2. The method for detecting shell defects of a wind power control cabinet according to claim 1, characterized in that: The step of processing the image data to obtain a target shell image includes: Performing grayscale processing on the image data based on an adaptive grayscale algorithm to obtain a first grayscale image; The first grayscale image is denoised based on a filtering algorithm to obtain the target shell image.

3. The method for detecting shell defects of a wind power control cabinet according to claim 2, characterized in that: Performing denoising on the first grayscale image based on a filtering algorithm to obtain the target shell image includes: determining a first Gaussian kernel based on a resolution of the first grayscale image; determining a target standard deviation based on the noise intensity and the noise suppression coefficient of the first grayscale image; Adjusting the first Gaussian kernel based on the target standard deviation to obtain a target Gaussian kernel; The first grayscale image is denoised based on the target Gaussian kernel to obtain the target shell image.

4. The method for detecting shell defects of a wind power control cabinet according to claim 3, characterized in that: Also includes: In response to the noise intensity of the first grayscale image being greater than or equal to a first noise intensity, increasing the noise suppression coefficient by a first step; In response to the noise intensity of the first grayscale image being less than the first noise intensity, the noise suppression coefficient is reduced with a second step size.

5. The method for detecting shell defects of a wind power control cabinet according to claim 1, characterized in that: The step of processing the point cloud data to obtain target point cloud data includes: Preprocessing the point cloud data to obtain first point cloud data; Calculating curvature information of the first point cloud data, and adding the curvature information to the first point cloud data to obtain second point cloud data; The second point cloud data is registered using a normal distribution transformation algorithm to obtain target point cloud data.

6. The method for detecting shell defects of a wind power control cabinet according to claim 5, characterized in that: The second point cloud data is registered using a normal distribution transformation algorithm to obtain target point cloud data, including: Determining a voxel size according to the point cloud density of the target point cloud data; determining a size adjustment coefficient of the voxel size according to the curvature information; Adjusting the voxel size based on the size adjustment coefficient to obtain a target voxel size; The second point cloud data is registered based on a normal distribution transformation algorithm and the target voxel size to obtain target point cloud data.

7. The method for detecting shell defects of a wind power control cabinet according to claim 6, characterized in that: Also includes: In response to a curvature value in the curvature information being less than a first curvature threshold, increasing the size adjustment coefficient by a third step size; In response to a curvature value in the curvature information being greater than a second curvature threshold, the resizing coefficient is reduced by a fourth step size.

8. The method for detecting shell defects of a wind power control cabinet according to claim 1, characterized in that: The step of obtaining the shell defect detection result of the wind power control cabinet according to the target shell image and the target point cloud data includes: Constructing a fusion matrix according to the target shell image and the target point cloud data, wherein the matrix dimensions of the fusion matrix include: texture abnormality, curvature mutation coefficient and three-dimensional deformation parameter; Inputting the fusion matrix into a target classification model to obtain a shell defect detection result of the wind power control cabinet; Among them, the target classification model is obtained by training the random forest model based on the first training data; the first training data is the fusion matrix data of the historical defect samples.

9. The method for detecting shell defects of a wind power control cabinet according to claim 8, characterized in that: Also includes: Calculating a first ratio according to the number of matrix dimensions of the fusion matrix and a ratio formula, and determining a compensation value according to the first ratio, wherein the compensation value is used to compensate for a basic depth of a decision tree in a target classification model; The compensation value and the basic depth are weightedly calculated to obtain a target depth of the decision tree in the target classification model.

10. A shell defect detection system for a wind power control cabinet, characterized in that: include: A target data determination module is used to determine target data corresponding to the wind power control cabinet housing based on detection requirements; the target data includes image data and point cloud data; A target data processing module is used to process the image data to obtain a target shell image; and is also used to process the point cloud data to obtain target point cloud data; The monitoring result output module is used to obtain the shell defect detection result of the wind power control cabinet according to the target shell image and the target point cloud data.

Images