Substation equipment defect detection method based on MWD algorithm
Through the MWD algorithm-based substation equipment defect detection method, the damaged contour is extracted using image matching and GM algorithm, which solves the problems of high hardware requirements and slow prediction speed in the prior art, and achieves efficient and accurate equipment defect detection.
Patent Information
- Application Number
- CN202411338713.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-25
- Publication Date
- 2024-12-20
AI Technical Summary
In the prior art, the defect detection method of substation equipment has high requirements for hardware equipment, and the prediction target calculation volume is large, resulting in a long prediction response time and a slow prediction speed.
The defect detection method of substation equipment based on MWD algorithm is adopted, and the image matching, redundant information removal and image data output are used to match feature points using SIFT algorithm and FLANN matcher, and the median filtering and binarization processing are performed in combination with GM algorithm to extract Broken contour images, reducing hardware costs and data volume requirements.
Improve detection efficiency and accuracy, reduce hardware and algorithm deployment costs, and achieve fast and accurate identification of damaged areas.
Smart Images
Figure CN119168995A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of equipment defect detection, in particular to a substation equipment defect detection method based on an MWD algorithm. Background Art
[0002] With the vigorous development of my country's power industry, the scale of the power network continues to grow. Maintaining a stable power supply is crucial to ensuring social and economic development and people's living needs. High-quality power services can not only improve people's quality of life, but also provide reliable power support for corporate production. Timely inspection and maintenance of power facilities can ensure the safe and stable operation of the power grid, avoid power failures and power outages, and thus ensure the normal operation of society and the safety of people's lives and property. However, in the field of power inspection, the traditional manual inspection method is limited by the quality of personnel and geographical environment, the inspection efficiency is low, and there are problems such as incomplete data records and easy loss. Although drone inspection technology has developed rapidly in recent years, it still faces challenges in terms of stability, endurance and data transmission security in complex environments. In addition, the operation and maintenance costs of drones are relatively high, and the technology and management system need to be further improved.
[0003] Therefore, how to combine the advantages of manual inspection and drone inspection to improve the efficiency and safety of power inspection is an important issue that the current power industry needs to solve. At present, most of the power station defect detection algorithms are based on deep learning networks. The deep learning network model requires a large amount of high-quality image data and accurate data annotation during the initial training. For each detection target, a large amount of data is required as model training support. The number of parameters in the neural network is often in the millions. In actual prediction, the hardware equipment requirements are high and the prediction target calculation volume is large, resulting in a long actual prediction response time and slow prediction speed. Summary of the invention
[0004] The purpose of the present invention is to provide a substation equipment defect detection method based on the MWD algorithm to solve the problems of the existing defect detection method having high requirements on hardware equipment, large amount of prediction target calculation, long prediction response time and slow prediction speed.
[0005] The purpose of the present invention is achieved through the following technical solutions:
[0006] A method for detecting defects in substation equipment based on an MWD algorithm comprises the following steps:
[0007] S1. Collect image data of substation equipment and transmit the image data to the inspection host, which transmits the data to the data analysis host for image data processing, wherein the image data includes standard images and detection images;
[0008] S2, the data analysis host performs image preprocessing on the received standard image and test image, removes noise, and converts the image into a grayscale image;
[0009] S3, executing the MWD algorithm on the grayscale of step S2, wherein the MWD algorithm performs image matching between the detection image and the standard image to obtain a valid detection image of the same size as the standard image;
[0010] S4, the MWD algorithm detects the standard image and the effective detection image, cuts the invalid part of the effective detection image, retains the corresponding area data in the standard image, and obtains the difference image through absolute value difference processing;
[0011] S5, the difference image is input into the GM algorithm, the median filter operation is performed on the two images to remove noise, the gradient threshold is calculated, and then the damaged contour image of the effective detection image is obtained after binarization processing and morphological operation;
[0012] S6. The image output part of the MWD algorithm extracts the data of the damaged contour image, obtains detailed contour data, draws the contour on the standard map, outputs the detection result, and feeds it back to the patrol host.
[0013] Furthermore, the image matching part of step S3 uses the SIFT algorithm to obtain feature points of the detection image and the standard image, obtains matching points after matching by the FLANN matcher, creates a transformation matrix based on the matching points, and obtains a valid detection image of the same size as the standard image after perspective transformation of the detection image.
[0014] Furthermore, in the step S4, the invalid part of the valid detection image is cut out, and the corresponding area data in the standard image is retained. Specifically, the image is traversed from the top, bottom, left and right directions until valid non-zero data exists, and the black boundary is cut out to zero to obtain a valid image of the same size as the standard image.
[0015] Furthermore, the calculated gradient threshold in step S5 is to calculate the gradient threshold in the x and y directions of the image.
[0016] Furthermore, in the step S4, a histogram equalization process is performed on the standard image and the effective detection image before the absolute value difference process.
[0017] The present invention has the following advantages:
[0018] 1. The MWD algorithm is used for equipment defect detection. First, the images are matched, then the redundant image information is extracted, and finally the information is output. The standard image is used as a benchmark, and the image taken with the actual inspection target is used. By comparing and analyzing the differences between the two, the damaged area can be accurately detected. The detection efficiency and accuracy are high. When the algorithm is deployed, there is no need to train the model like the deep learning algorithm, which eliminates the data volume requirement of the algorithm, reduces hardware costs, and reduces algorithm deployment costs.
[0019] 2. The GM algorithm structure designed in the present invention uses gradient calculation to detect the position of the damaged area and calculates the pixel mean of the damaged area as the binarization threshold. A more accurate binarization threshold is obtained through the GM algorithm structure, and the damaged contour after the binarization operation is more accurate. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 It is a schematic diagram of the process of the present invention.
[0021] Figure 2 It is a flowchart of the image matching part in the MWD algorithm of the present invention.
[0022] Figure 3 It is a schematic diagram of the process after image matching is completed in the MWD algorithm of the present invention.
[0023] Figure 4 It is a schematic diagram of the flow of the GM algorithm of the present invention.
[0024] Figure 5 This is a comparison chart of the effects of the GM algorithm of the present invention. DETAILED DESCRIPTION
[0025] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.
[0026] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0027] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.
[0028] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.
[0029] In the description of the present invention, it should be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, or the positions or positional relationships in which the invented product is usually placed when in use, or the positions or positional relationships commonly understood by those skilled in the art, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.
[0030] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0031] refer to Figure 1-5 As shown, one embodiment of the present invention is:
[0032] A method for detecting defects in substation equipment based on an MWD algorithm comprises the following steps:
[0033] S1. Collect image data of substation equipment and transmit the image data to the inspection host, which transmits the data to the data analysis host for image data processing, wherein the image data includes standard images and detection images;
[0034] S2, the data analysis host performs image preprocessing on the received standard image and test image, removes noise, and converts the image into a grayscale image;
[0035] S3, executing the MWD algorithm on the grayscale of step S2, wherein the MWD algorithm performs image matching between the detection image and the standard image to obtain a valid detection image of the same size as the standard image;
[0036] S4, the MWD algorithm detects the standard image and the effective detection image, cuts the invalid part of the effective detection image, retains the corresponding area data in the standard image, and obtains the difference image through absolute value difference processing;
[0037] S5, the difference image is input into the GM algorithm, the median filter operation is performed on the two images to remove noise, the gradient threshold is calculated, and then the damaged contour image of the effective detection image is obtained after binarization processing and morphological operation;
[0038] S6. The image output part of the MWD algorithm extracts the data of the damaged contour image, obtains detailed contour data, draws the contour on the standard map, outputs the detection result, and feeds it back to the patrol host.
[0039] The MWD algorithm is used for equipment defect detection. When the algorithm is deployed, there is no need to perform model training like the deep learning algorithm, which eliminates the algorithm's data volume requirements, reduces hardware costs, and reduces algorithm deployment costs. The core requirements of the algorithm of the present invention are only a standard image of the detection target as a comparison benchmark and a photographic image of the target during an actual inspection. By comparing and analyzing the differences between the two, the damaged area can be quickly and accurately identified, and the fit of the output contour can be improved. This design not only reduces the difficulty of data collection, but also greatly improves the efficiency and accuracy of detection.
[0040] The MWD algorithm designed in the present invention mainly consists of three parts, namely image matching, redundant information removal and image data output. The MWD algorithm structure is used to achieve image matching, improve the quality of the detected image, meet the preconditions for image difference calculation, and ensure the efficient operation of the algorithm. Corresponding to the steps of this method, the SIFT algorithm is used in the image matching part of the S3 step to obtain the feature points of the detection image and the standard image, and the matching points are obtained after matching by the FLANN matcher. A transformation matrix is created based on the matching points, and after the perspective transformation operation of the detection image, an effective detection image with the same size as the standard image is obtained.
[0041] Furthermore, since the image after perspective transformation may have black borders due to different image sizes, image angles, etc., it is necessary to remove redundant information of the image. Specifically, in step S4, the invalid part of the valid detection image is cut off, and the corresponding area data in the standard image is retained. Specifically, the image is traversed from the top, bottom, left, and right directions until there is valid non-zero data, and the black border is cut off to zero to obtain a valid image of the same size as the standard image, and the standard image and the valid detection image are subjected to histogram equalization before the absolute value difference processing.
[0042] When detecting breakage, the absolute value difference function is used to detect whether there is damage. If there is damage, the damaged outline will be highlighted in the value returned by the function, and then the gradient search is used to calculate the mean value of the contour pixels. After binarization and morphological operations, the precise damaged outline of the detection device is obtained. Specifically, the gradient threshold is calculated to calculate the gradient threshold in the x and y directions of the image.
[0043] For the specific GM algorithm process, refer to Figure 4As shown, after the filtering is terminated, the gradient in the x and y directions is calculated, and the gradient in the x and y directions is specifically limited, and then converted to uint8 type, and the corrosion operation and the expansion operation are performed in sequence, and then the contour of the broken area is detected, and the pixel value in the broken contour is calculated. In this embodiment, the Sobel operator is specifically used to calculate the gradient in the difference image, and the morphological operation is used to obtain the gradient contour image of the damaged range. According to the contour position coordinates in the gradient image, the corresponding pixel value is obtained, and the mean value of the damaged area image is calculated as the subsequent binarization operation threshold. Through the GM algorithm structure, a more accurate binarization threshold is obtained to better realize the contour range extraction, improve the degree of fitting between the defect contour and the actual defect, and make the defect range more accurate.
[0044] This application uses the GM algorithm to calculate the image threshold, and obtains a better effect of the defect range binarization operation. Figure 5 The GM algorithm effect comparison chart is presented. In the first row, the four detection images are resized after image preprocessing, converted to grayscale and blurred by mean filter. The second row of images are binarized using the GM algorithm gradient threshold after the absolute difference of the MWD algorithm. The third row is the binarized image with a threshold of 30.
[0045] from Figure 5 It can be seen that for the complete connected defect, that is, the leftmost image, there is little difference between the effects of the 30 threshold and the gradient threshold, and the binarized image is only a problem of the size of the defect contour; for the disconnected large defect contours, that is, the defect contours in the second column, it can be seen that the difference between the two is that the gradient threshold is more sensitive to the contour range, and the binarized image with a fixed threshold of 30 has a poor effect, showing only a small part of the defect; when detecting images with smaller defects in the third column, the gradient threshold shows better robustness, and will not fail to detect due to the small defect range. The fixed threshold has a poor effect and cannot detect the defects of the equipment; the fourth column of images is two defect contours, and it is obvious that the gradient threshold has a better effect on the extraction of the contour range, and the contour range is more accurate.
[0046] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A substation equipment defect detection method based on MWD algorithm, characterized in that: The following steps are involved: S1. Collect image data of substation equipment and transmit the image data to the inspection host, which transmits the data to the data analysis host for image data processing, wherein the image data includes standard images and detection images; S2, the data analysis host performs image preprocessing on the received standard image and test image, removes noise, and converts the image into a grayscale image; S3, executing the MWD algorithm on the grayscale of step S2, wherein the MWD algorithm performs image matching between the detection image and the standard image to obtain a valid detection image of the same size as the standard image; S4, the MWD algorithm detects the standard image and the effective detection image, cuts the invalid part of the effective detection image, retains the corresponding area data in the standard image, and obtains the difference image through absolute value difference processing; S5, the difference image is input into the GM algorithm, the median filter operation is performed on the two images to remove noise, the gradient threshold is calculated, and then the damaged contour image of the effective detection image is obtained after binarization processing and morphological operation; S6. The image output part of the MWD algorithm extracts the data of the damaged contour image, obtains detailed contour data, draws the contour on the standard map, outputs the detection result, and feeds it back to the patrol host.
2. The method for detecting substation equipment defects based on the MWD algorithm according to claim 1, characterized in that: The image matching part of step S3 uses the SIFT algorithm to obtain feature points of the detection image and the standard image, obtains matching points after matching by the FLANN matcher, creates a transformation matrix based on the matching points, and obtains a valid detection image of the same size as the standard image after perspective transformation of the detection image.
3. The method for detecting substation equipment defects based on the MWD algorithm according to claim 1, characterized in that: In the step S4, the invalid part of the effective detection image is cut off, and the corresponding area data in the standard image is retained. Specifically, the image is traversed from the top, bottom, left and right directions until valid non-zero data exists, and the black border is cut off to zero to obtain a valid image with the same size as the standard image.
4. The method for detecting substation equipment defects based on the MWD algorithm according to claim 1, characterized in that: The calculated gradient threshold in step S5 is to calculate the gradient threshold in the x and y directions of the image.
5. The method for detecting substation equipment defects based on the MWD algorithm according to claim 1, characterized in that: In the step S4, a histogram equalization process is performed on the standard image and the effective detection image before the absolute value difference process.