Mathematical morphology-based image recognition method for icing thickness of isolator switch

Abstract

A mathematical morphology-based image recognition method for the icing thickness of an isolator switch, aiming to solve the problem of poor accuracy of existing edge detection methods in recognizing an image of the icing thickness of an isolator switch. The method comprises: acquiring an iced isolator switch image; performing image preprocessing on the iced isolator switch image to generate a grayscale enhanced image; using mathematical morphology to perform edge extraction on the grayscale enhanced image to obtain an edge image; performing edge selection on the edge image to obtain an extracted image of the iced isolator switch; and using acquired original edge data of an un-iced isolator switch and the extracted image of the iced isolator switch to perform comparison calculation to calculate the icing thickness of the isolator switch. The beneficial effect is that the icing thickness of the isolator switch can be accurately obtained.

Description

A mathematical morphology-based image recognition method for isolating switches with icing thickness Technical Field

[0001] This invention relates to a method for image recognition of ice thickness on disconnect switches. Background Technology

[0002] In the process of image processing and analysis of disconnector switches covered with ice, edge detection methods are currently widely used. These methods employ first-order edge detection operators (Sobel or Prewitt operators) and second-order edge detection operators (Canny operators) for edge detection. First-order edge detection operators have low sensitivity and can avoid interference from the background environment, but their output edge detection results are intermittent and have severe graininess. Second-order edge detection operators have higher sensitivity, but require high-quality original images. If the original image is poor, the output line edges will show obvious distortion and breakage. Therefore, existing edge detection methods have poor accuracy in identifying disconnector switch ice thickness images. Summary of the Invention

[0003] The purpose of this invention is to solve the problem of poor accuracy in the existing edge detection method for recognizing the ice thickness image of disconnect switches, and to propose a method for recognizing the ice thickness image of disconnect switches based on mathematical morphology.

[0004] The present invention provides a method for image recognition of icing thickness of disconnect switches based on mathematical morphology, comprising the following steps:

[0005] The method for recognizing the ice thickness of the disconnect switch includes the following steps:

[0006] Step 1: Obtain the raw edge data of the un-iced isolating switch through the monitoring system; the raw edge data includes the diameter d1 of the isolating switch when it is not iced and the number of pixels x1 of the isolating switch when it is not iced.

[0007] Step 2: Obtain images of the ice-covered isolating switches through the monitoring system;

[0008] Step 3: Perform image preprocessing on the ice-covered isolating switch image obtained in Step 2 to generate a grayscale enhanced image;

[0009] Step 4: Use mathematical morphology to extract edges from the grayscale enhanced image generated in Step 3 to obtain the edge image;

[0010] Step 5: Perform edge selection on the edge image obtained in Step 4 to obtain the extracted image of the ice-covered isolation switch;

[0011] Step 6: Compare the original edge data of the un-iced isolating switch obtained in Step 5 with the extracted image of the iced isolating switch obtained in Step 4 to calculate the icing thickness of the isolating switch.

[0012] The specific formula for the comparison calculation is as follows:

[0013] Where H is the icing thickness of the disconnecting switch, d1 is the diameter of the disconnecting switch when it is not iced, d2 is the diameter of the disconnecting switch when it is iced, x1 is the number of pixels when the disconnecting switch is not iced, and x2 is the number of pixels when the disconnecting switch is iced.

[0014] Preferably, the image preprocessing in step three includes image rotation, image grayscale conversion, and grayscale image enhancement;

[0015] Image rotation refers to rotating the length direction of the image of the ice-covered isolating switch to the horizontal direction.

[0016] Preferably, the image grayscale conversion process is as follows: a floating-point algorithm is used to convert the rotated image of the ice-covered isolation switch into a grayscale image.

[0017] Preferably, the grayscale image enhancement process is as follows: the grayscale image is enhanced using histogram equalization to obtain a grayscale enhanced image.

[0018] Preferably, the specific steps for obtaining the edge image in step four are as follows:

[0019] Mathematical morphology is used to detect the edges of disconnect switches. Eight sets of structures are defined, the weights of each set of structures are calculated, and the structures are inserted into the grayscale enhanced image. The grayscale enhanced image is then thresholded using an iterative optimal thresholding method. The thresholded image is then inverted, and finally, the edge image is output using an erosion algorithm.

[0020] Preferably, the specific steps for obtaining the image of the ice-covered isolating switch in step five are as follows:

[0021] The position information of the isolating switch is extracted by a straight line detection method, and the image of the ice-covered isolating switch is obtained by fitting the angle bisector.

[0022] Compared with the prior art, the present invention has the following advantages:

[0023] This invention proposes an image recognition method for ice thickness of disconnect switches based on mathematical morphology. For the target iced disconnect switch image captured by the monitoring system, the boundary extracted by mathematical morphology fusion is used to obtain the number of pixels when the disconnect switch is iced. By importing the diameter of the un-iced disconnect switch and the number of pixels of the un-iced disconnect switch for comparison and calculation, the ice thickness of the disconnect switch can be accurately determined, thereby improving the accuracy of ice thickness image detection. Attached Figure Description

[0024] Figure 1 is a flowchart of an image recognition method for ice thickness of disconnect switches based on mathematical morphology, as described in Specific Implementation 1.

[0025] Figure 2 shows the grayscale enhancement image generated in Specific Implementation Method 4;

[0026] Figure 3 shows the edge image effect obtained in Specific Implementation Method 5. Detailed Implementation

[0027] Specific Implementation Method 1: Referring to Figure 1, this implementation method describes an image recognition method for ice thickness of disconnecting switches based on mathematical morphology. The method includes the following steps:

[0028] Step 1: Obtain the raw edge data of the un-iced isolating switch through the monitoring system; the raw edge data includes the diameter d1 of the isolating switch when it is not iced and the number of pixels x1 of the isolating switch when it is not iced.

[0029] Step 2: Obtain images of the ice-covered isolating switches through the monitoring system;

[0030] Step 3: Perform image preprocessing on the ice-covered isolating switch image obtained in Step 2 to generate a grayscale enhanced image;

[0031] Step 4: Use mathematical morphology to extract edges from the grayscale enhanced image generated in Step 3 to obtain the edge image;

[0032] Step 5: Perform edge selection on the edge image obtained in Step 4 to obtain the extracted image of the ice-covered isolation switch;

[0033] Step 6: Compare the original edge data of the un-iced isolating switch obtained in Step 5 with the extracted image of the iced isolating switch obtained in Step 4 to calculate the icing thickness of the isolating switch.

[0034] The specific formula for the comparison calculation is as follows:

[0035] Where H is the icing thickness of the disconnecting switch, d1 is the diameter of the disconnecting switch when it is not iced, d2 is the diameter of the disconnecting switch when it is iced, x1 is the number of pixels when the disconnecting switch is not iced, and x2 is the number of pixels when the disconnecting switch is iced.

[0036] In this embodiment, the image of the iced isolating switch obtained in step one is captured by the monitoring system; the extracted image of the iced isolating switch in step four is the number of pixels when the isolating switch is iced; the original edge data of the uniced isolating switch in step five includes the diameter of the isolating switch when it is not iced and the number of pixels when the isolating switch is not iced; after obtaining the boundary of the iced isolating switch by pixel-level fusion of the isolating switch boundary, the coordinates of the line edge pixels are obtained, and then the relevant data of the isolating switch when it is not iced are read. The ice thickness of the isolating switch is calculated according to the proportional relationship of the corresponding pixels to determine the current ice status of the isolating switch.

[0037] Specific Implementation Method 2: This implementation method further defines the image recognition method for ice thickness of disconnecting switches based on mathematical morphology described in Specific Implementation Method 1. In this implementation method, the image preprocessing in step 2 includes image rotation, image grayscale conversion, and grayscale image enhancement.

[0038] Image rotation refers to rotating the length direction of the image of the ice-covered isolating switch to the horizontal direction.

[0039] In this embodiment, the disconnect switch is a cylindrical structure, and the length direction of the ice-covered disconnect switch image is the central axis direction of the ice-covered disconnect switch. After the ice-covered disconnect switch image obtained in step three is successively rotated, grayscaled, and enhanced, the overall grayscale distribution of the processed grayscale enhanced image is more uniform compared with the original image, thereby achieving the purpose of image enhancement and improving the accuracy of ice thickness image detection.

[0040] Specific Implementation Method 3: This implementation method further defines the image recognition method for ice thickness of isolating switches based on mathematical morphology described in Specific Implementation Method 2. In this implementation method, the image grayscale conversion process is as follows: a floating-point algorithm is used to convert the rotated image of the ice-covered isolating switch into a grayscale image.

[0041] Specific Implementation Method Four: This implementation method is described in conjunction with Figure 2. This implementation method further defines the image recognition method for ice thickness of disconnecting switches based on mathematical morphology described in Specific Implementation Method Three. In this implementation method, the grayscale image enhancement process is as follows: the grayscale image is enhanced by histogram equalization to obtain a grayscale enhanced image.

[0042] In this embodiment, histogram equalization is used to increase the global contrast of the icing isolator image, especially when the contour contrast of the icing isolator image is quite similar. This method allows for better distribution of brightness across the histogram, thus enhancing local contrast without affecting the overall contrast. Histogram equalization achieves grayscale enhancement by effectively expanding commonly used brightness levels. The image is preprocessed using Python programming, and the resulting grayscale-enhanced image of the icing isolator is shown in Figure 2.

[0043] Specific Implementation Method Five: This implementation method is further defined in conjunction with Figure 3. It further refines the image recognition method for ice thickness of disconnecting switches based on mathematical morphology described in Specific Implementation Method One. In this implementation method, the specific steps for obtaining the edge image in step four are as follows:

[0044] Mathematical morphology is used to detect the edges of disconnect switches. Eight sets of structures are defined, the weights of each set of structures are calculated, and the structures are inserted into the grayscale enhanced image. The grayscale enhanced image is then thresholded using an iterative optimal thresholding method. The thresholded image is then inverted, and finally, the edge image is output using an erosion algorithm.

[0045] In this embodiment, mathematical morphology is first used to detect the edges of the disconnecting switch, defining eight sets of structures. These eight sets of structures include two lengths and four phase angles; the two lengths are 3mm and 5mm; and the four phase angles are 0°, 45°, 90°, and 135°. Then, the weights corresponding to each set of structures are calculated, and the structures are inserted into the grayscale enhancement image. The grayscale enhancement image is then thresholded using an iterative optimal thresholding method. The thresholded image is then inverted, and finally, an edge image is output using an erosion algorithm. Edge detection is performed on the preprocessed image using mathematical morphology, and the output result is shown in Figure 3.

[0046] Specific Implementation Method Six: This implementation method further defines the image recognition method for ice thickness of isolating switches based on mathematical morphology described in Specific Implementation Method One. In this implementation method, the specific steps for obtaining the extracted image of the ice-covered isolating switch in step five are as follows:

[0047] The position information of the isolating switch is extracted by a straight line detection method, and the image of the ice-covered isolating switch is obtained by fitting the angle bisector.

[0048] In this embodiment, the method extracts the outline of the isolating switch from its environmental background by edge selection before image fusion. Since the edge detection result of the isolating switch is an edge contour line composed of straight lines, while the edge detection result of the interfering background in its environment is usually a curve, the position information of the isolating switch is extracted by the straight line detection method, and the ice-covered isolating switch is extracted by fitting the angle bisector.

[0049] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for image recognition of icing thickness of disconnect switches based on mathematical morphology, characterized in that, The method for recognizing the ice thickness of the disconnector switch includes the following steps: Step 1: Obtain the raw edge data of the un-iced isolating switch through the monitoring system; the raw edge data includes the diameter d1 of the isolating switch when it is not iced and the number of pixels x1 of the isolating switch when it is not iced. Step 2: Obtain images of the ice-covered isolating switches through the monitoring system; Step 3: Perform image preprocessing on the ice-covered isolating switch image obtained in Step 2 to generate a grayscale enhanced image; Step 4: Use mathematical morphology to extract edges from the grayscale enhanced image generated in Step 3 to obtain the edge image; Step 5: Perform edge selection on the edge image obtained in Step 4 to obtain the extracted image of the ice-covered isolation switch; Step 6: Compare the original edge data of the un-iced isolating switch obtained in Step 5 with the extracted image of the iced isolating switch obtained in Step 4 to calculate the icing thickness of the isolating switch. The specific formula for the comparison calculation is as follows: Where H is the icing thickness of the disconnecting switch, d1 is the diameter of the disconnecting switch when it is not iced, d2 is the diameter of the disconnecting switch when it is iced, x1 is the number of pixels when the disconnecting switch is not iced, and x2 is the number of pixels when the disconnecting switch is iced.

2. The method for image recognition of icing thickness of disconnect switches based on mathematical morphology according to claim 1, characterized in that, Step 3, image preprocessing, includes image rotation, image grayscale conversion, and grayscale image enhancement. Image rotation refers to rotating the length direction of the image of the ice-covered isolating switch to the horizontal direction.

3. The method for image recognition of icing thickness of disconnect switches based on mathematical morphology according to claim 2, characterized in that, The process of converting the image to grayscale is as follows: a floating-point algorithm is used to convert the rotated image of the ice-covered isolation switch into a grayscale image.

4. The method for image recognition of icing thickness of disconnecting switches based on mathematical morphology according to claim 3, characterized in that, The grayscale image enhancement process is as follows: the grayscale image is enhanced using histogram equalization to obtain a grayscale enhanced image.

5. The method for image recognition of icing thickness of disconnect switches based on mathematical morphology according to claim 1, characterized in that, The specific steps for obtaining the edge image in step four are as follows: Mathematical morphology is used to detect the edges of disconnect switches. Eight sets of structures are defined, the weights of each set of structures are calculated, and the structures are inserted into the grayscale enhanced image. The grayscale enhanced image is then thresholded using an iterative optimal thresholding method. The thresholded image is then inverted, and finally, the edge image is output using an erosion algorithm.

6. The method for image recognition of icing thickness of disconnect switches based on mathematical morphology according to claim 1, characterized in that, The specific steps for obtaining the image of the ice-covered isolating switch in step five are as follows: The position information of the isolating switch is extracted by a straight line detection method, and the image of the ice-covered isolating switch is obtained by fitting the angle bisector.

Images