A foreign body detection method for transmission lines based on white point rate method

Abstract

The present invention discloses a method for detecting foreign objects in power transmission lines based on a white point rate method, comprising the following steps: 1. Obtaining images of power transmission lines in a target area through an unmanned aerial vehicle inspection; 2. Obtaining a preliminary image for foreign object identification by image preprocessing, and performing improved Otsu threshold segmentation and morphological closing operations; 3. Obtaining a horizontal image of the power transmission lines through image tilt correction; 4. Screening out the power transmission lines by performing Hough transform line detection on the corrected image area; 5. Determining the contour type information based on the vertical coordinate position of the power transmission line contour, performing internal white point statistics on the contour, and outputting the corresponding foreign object contour type for contours that meet the white point rate. The present invention can distinguish between foreign objects attached to power transmission lines and those hanging on them, thereby reducing missed detections and false detections caused by foreign objects attached during inspections, and providing effective support for fault analysis during power transmission line inspections.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and in particular to a method for detecting foreign matter in a power transmission line based on a white point rate method. Background Art

[0002] Transmission lines are the primary carrier of power transmission and play a vital role in the stable and safe operation of power transmission systems. Transmission lines are exposed to the elements and are subject to various influences. Foreign objects on transmission lines can significantly affect their normal operation. In particular, foreign objects hanging from transmission lines, especially those close to the ground, are more likely to cause power outages or accidents. Therefore, foreign object identification on transmission lines is extremely important.

[0003] The existing method for detecting foreign objects is primarily manual line inspection. However, with the development of high-voltage, high-power, and long-distance transmission lines, the geographical environments traversed by transmission networks are becoming increasingly complex, making manual inspections increasingly difficult. To reduce workload and improve efficiency, recent years have seen the emergence of drones equipped with optical equipment to capture images of transmission lines. Computers are then used to intelligently process the large amounts of image data collected during inspections to determine whether foreign objects are present.

[0004] At present, traditional image processing algorithms have the problem of high recognition error rate when judging foreign objects on power lines, and the problem of poor detection of foreign objects hanging on power lines in image foreign object recognition. Summary of the Invention

[0005] In order to overcome the shortcomings of the above-mentioned prior art, the present invention provides a transmission line foreign body detection method based on the white point rate method, so as to effectively identify foreign bodies attached to the transmission line and hanging foreign bodies, reduce missed detections and false detections caused by the attachment of foreign bodies during inspections, and thus provide effective support for inspection fault analysis of transmission lines.

[0006] In order to achieve the above-mentioned object, the present invention adopts the following technical solutions:

[0007] The invention provides a method for detecting foreign matter in a power transmission line based on a white point rate method, which comprises the following steps:

[0008] Step 1: Use a drone equipped with a camera and data transmission equipment to inspect the power transmission line and obtain images of the power transmission line in the target area, and then transmit them to the foreign object detection system on the ground.

[0009] Step 2: The foreign object detection system performs image preprocessing on the received transmission line image of the target area to obtain a foreign object recognition preliminary image, wherein the preprocessing operation includes weighted grayscale processing and median filtering;

[0010] Step 3: Perform improved Otsu threshold segmentation on the early image of foreign body recognition to obtain a binary image; then perform morphological closing operation on the binary image to obtain a denoised binary image;

[0011] Step 4: Perform Hough transform detection on the denoised binary image to obtain each transmission line, and perform tilt correction on the denoised binary image based on the rotation direction of the longest transmission line to obtain a corrected image of the transmission line tending to be horizontal;

[0012] Step 5: After performing Hough transform on the corrected image, all straight line segments in the corrected image are obtained and the straight lines are screened. The straight line screening operation includes: the length and slope of the straight line;

[0013] Step 6: Statistically calculate the vertical coordinates of the filtered straight line segments, and define the area between the minimum and maximum vertical coordinates as the transmission line area;

[0014] Step 7: After selecting the minimum circumscribed rectangular outline of the connected area in the rectified image, the selected outline is filtered. The contour screening operation includes: the area and the aspect ratio of the contour;

[0015] Step 8: Compare the vertical coordinates of the endpoints of the filtered contour with the vertical coordinates of the transmission line area. If any of the vertical coordinates of the four endpoints of the contour is within the range of the vertical coordinates of the transmission line area, it indicates that the type of the corresponding contour is an attached contour; otherwise, it indicates that the type of the corresponding contour is a hanging contour.

[0016] Step 9: Count the white dots inside the minimum circumscribed rectangle and use it to set the number of white dots S inside the outline. Fault and the contour area S cont The ratio between them is used as the white point rate, different white point rates are set for the two divided contour types, and the contours that meet the corresponding white point rates are output as the corresponding foreign body types.

[0017] The method for detecting foreign objects in a power transmission line based on the white rate method according to the present invention is also characterized in that the improved Otsu threshold segmentation process in step 3 includes the following steps:

[0018] Step A1: Divide the grayscale values of the image in the early stage of foreign object recognition into two types: the foreground part C1 with grayscale levels [0, k] and the background part C2 with grayscale levels [k+1, L-1]; k represents the set threshold and L represents the grayscale level of the image;

[0019] Let p A and p B They represent the proportion of C1 and C2 to all pixels in the early stage of foreign object recognition image, and p A +p B =1; where p i represents the number of pixels with gray level i in C1 or C2;

[0020] Calculate the grayscale mean of C1 and the grayscale mean of C2

[0021] Step A2: Calculate the intra-class variance of the foreground part C1 and the intra-class variance of the background part C2 Obtaining the intra-class variance of the early image for foreign object recognition m is the grayscale mean of the image in the early stage of foreign body recognition;

[0022] Step A3: Calculate the average grayscale value of the foreground part C1 after adding the intra-class variance and the average gray value of the background part C2 Calculate the average gray value (m * ) 2 =(m 2 +θ) / 2;

[0023] Step A4: Use formula (1) to calculate the optimal threshold t for segmenting the image in the early stage of foreign body recognition and obtain a binary image:

[0024] t=ArgMax[P A ×(m A * -m * ) 2 +P B ×(m B * -m * ) 2 ] (1)

[0025] In formula (1), ArgMax represents the maximum value operation.

[0026] The screening in step 5 is performed as follows:

[0027] For any j-th straight line segment l in the rectified image j The slope k i , when k i ∈[-δ,δ], retain the jth straight line segment l j , otherwise, remove the jth straight line segment l j ; Where δ represents the slope screening threshold;

[0028] Calculate the jth line segment l j Length Lj , and use the screening conditions shown in formula (2) to screen. If formula (2) is satisfied, the jth straight line segment l is retained. j , otherwise, remove the jth straight line segment l j ;

[0029]

[0030] In formula (2), n represents the number of straight line segments obtained.

[0031] The electronic device of the present invention includes a memory and a processor, wherein the memory is used to store a program that supports the processor to execute the power transmission line foreign object detection method, and the processor is configured to execute the program stored in the memory.

[0032] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program executes the steps of the power transmission line foreign object detection method when the computer program is executed by a processor.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] 1. The method proposed in the present invention can accurately identify horizontal transmission lines, providing a new approach for the location of transmission lines and foreign object fault detection during power inspections.

[0035] 2. The present invention realizes the identification of attached foreign objects and hanging foreign objects under complex backgrounds, proposes a new solution for the simultaneous identification of two types of foreign objects on transmission lines, improves the intelligent inspection of transmission lines, and promotes the development of foreign object detection and identification in transmission lines. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a flow chart of the method for identifying foreign objects in power transmission lines according to the present invention. DETAILED DESCRIPTION

[0037] In this embodiment, reference Figure 1 A method for detecting foreign objects in a power transmission line based on a white point rate method comprises the following steps:

[0038] Step A: Use a drone equipped with a camera and data transmission equipment to inspect the power transmission lines and obtain images of the power transmission lines in the target area, which are then transmitted to the foreign object detection system on the ground.

[0039] Step B: The foreign object detection system performs image preprocessing on the received transmission line image of the target area to obtain a preliminary image for foreign object recognition, wherein the preprocessing operation includes: weighted grayscale processing and median filtering;

[0040] In this embodiment, the weighted grayscale processing is to convert the image from an RGB image to a grayscale image, as shown in formula (3):

[0041] I(x,y)=0.299R(x,y)+0.578G(x,y)+0.144B(x,y) (3)

[0042] In formula (3), I(x,y) represents the grayscale value, R(x,y), G(x,y), and B(x,y) represent the values of the red, green, and blue channels of the pixel, respectively, and their values range from 0 to 255.

[0043] The method of median filter image processing is as follows:

[0044] A fixed-size window kernel is selected. In general, the window size of the median filter is an odd number, such as a 3×3 or 5×5 square. The pixels within the window kernel are sorted, and the middle value is taken as the grayscale value of the pixel.

[0045] Step C: Perform improved Otsu threshold segmentation on the early image of foreign body recognition to obtain a binary image; then perform morphological closing operation on the binary image to obtain a denoised binary image;

[0046] Step C1: Divide the grayscale values of the image in the early stage of foreign object recognition into two types: the foreground part C1 with grayscale levels [0, k] and the background part C2 with grayscale levels [k+1, L-1]; k represents the set threshold and L represents the grayscale level of the image;

[0047] Let p A and p B They represent the proportion of C1 and C2 to all pixels in the early stage of foreign object recognition image, and p A +p B =1; where p i represents the number of pixels with gray level i in C1 or C2;

[0048] Calculate the grayscale mean of C1 and the grayscale mean of C2

[0049] Step C2: Calculate the intra-class variance of the foreground part C1 and the intra-class variance of the background part C2 Obtaining the intra-class variance of the early image for foreign object recognition m is the grayscale mean of the image in the early stage of foreign body recognition;

[0050] Step C3: Calculate the average grayscale value of the foreground part C1 after adding the intra-class variance and the average gray value of the background part C2 Calculate the average gray value (m * ) 2 =(m 2 +θ) / 2;

[0051] Step C4: Use formula (1) to calculate the optimal threshold t for segmenting the image in the early stage of foreign body recognition and obtain a binary image:

[0052] t=ArgMax[P A ×(m A * -m * ) 2 +P B ×(m B * -m * ) 2 ] (1)

[0053] In formula (1), ArgMax represents the maximum value operation.

[0054] The morphological closing operation is shown in formula (4):

[0055]

[0056] In formula (4), A represents the image before processing, and B represents the processed structure element.

[0057] In a specific implementation, the original image is subjected to a dilation operation of a structural element B and then an erosion operation of the same structural element B is performed. The structural element B is a flat disk structure with a radius of 3.

[0058] Step D: Perform Hough transform detection on the denoised binary image to obtain the individual transmission lines, and perform tilt correction on the denoised binary image based on the rotation direction of the longest transmission line to obtain a corrected image of the transmission lines tending to be horizontal;

[0059] The specific method of tilt correction in step D includes the following steps:

[0060] Step D1: Perform Canny edge detection on the binary image to obtain the edge map of the image;

[0061] Step D2: Detecting power lines in the image using Hough transform;

[0062] Step D3: Using the rotation direction of the longest transmission line as a standard, perform tilt correction on the denoised binary image.

[0063] Step E: After performing Hough transform on the rectified image, all straight line segments in the rectified image are obtained and the straight lines are screened. The operations of screening the straight line segments include: the length and slope of the straight line;

[0064] The straight line segment screening is performed as follows:

[0065] E1: For any j-th straight line segment l in the rectified image j The slope k i , when k i ∈[-δ,δ], retain the jth straight line segment l j , otherwise, remove the jth straight line segment l j ; Wherein, δ represents the slope screening threshold; in this embodiment, δ = 0.2;

[0066] E2: Calculate the jth straight line segment l j Length L j , and use the screening conditions shown in formula (2) to screen. If formula (2) is satisfied, the jth straight line segment l is retained. j , otherwise, remove the jth straight line segment l j ;

[0067]

[0068] In formula (2), n represents the number of straight line segments obtained.

[0069] Step F: Statistical analysis of the vertical coordinates of the filtered straight line segments is performed, and the area between the minimum and maximum vertical coordinates is defined as the transmission line area. The extraction of the vertical coordinates of the straight line segments includes the following steps:

[0070] F1: The two-dimensional image of size M*N is regarded as a rectangular coordinate system with the origin at the upper left corner;

[0071] F2: Extract each transmission line straight segment l i Endpoint coordinates (x i ,y i )'s y-coordinate i ;

[0072] F3: Change the vertical coordinate y of the power line in the image i The range between the minimum and maximum values of is defined as the transmission line area.

[0073] Step G: After selecting the minimum circumscribed rectangle outline of the connected area in the rectified image, the selected outline is filtered. The contour screening operation includes: the area and aspect ratio of the contour;

[0074] The steps for screening the circumscribed rectangle outline in step G are as follows:

[0075] G1: calculate the area S of each minimum circumscribed rectangle contour cont Calculation, and according to the screening conditions of 1500≤S cont ≤35000 for screening, and only the minimum circumscribed rectangular outline that meets the conditions is retained;

[0076] G2: Then calculate the width-to-length ratio w / h of each minimum circumscribed rectangle outline and filter it according to the width-to-length ratio screening condition of 0≤w / h≤7.

[0077] Step H: Set the vertical coordinate y of the endpoint of the filtered contour cont Perform statistics and compare the ordinate value with the ordinate of the transmission line area. If any of the ordinates of the four endpoints of the contour is within the range of the ordinate of the transmission line area, it means that the type of the corresponding contour is an attached contour. Otherwise, it means that the type of the corresponding contour is a hanging contour.

[0078] Step I: After counting the white points inside the minimum circumscribed rectangle, the statistical value S of the white pixels inside the outline is obtained. Fault ; Set the number of white pixels within the outline S Fault and the contour area S cont The ratio S Fault / S cont And as the white point rate WhiteRate, different white point rates are set for the two types of contours after division, among which the white point rate judgment condition for hanging foreign matter is set to WhiteRate≥0.35, and the white point rate judgment condition for attached foreign matter is set to WhiteRate≥0.15; the contour that meets the corresponding white point rate is output as the corresponding foreign matter type, otherwise, the corresponding contour is deleted.

[0079] In this embodiment, an electronic device includes a memory and a processor, wherein the memory is used to store a program that supports the processor to execute the above method, and the processor is configured to execute the program stored in the memory.

[0080] In this embodiment, a computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method are executed.

Claims

1. A method for detecting foreign objects in a power transmission line based on a white point rate method, characterized in that: The following steps are involved: Step 1: Use a drone equipped with a camera and data transmission equipment to inspect the power transmission line and obtain images of the power transmission line in the target area, and then transmit them to the foreign object detection system on the ground. Step 2: The foreign object detection system performs image preprocessing on the received transmission line image of the target area to obtain a foreign object recognition preliminary image, wherein the preprocessing operation includes weighted grayscale processing and median filtering; Step 3: Perform improved Otsu threshold segmentation on the early image of foreign body recognition to obtain a binary image; then perform morphological closing operation on the binary image to obtain a denoised binary image; Step A1: Divide the grayscale values of the image in the early stage of foreign body recognition into two types, namely grayscale The foreground part and the gray level are The background part of the image; k represents the set threshold, L represents the gray level of the image; make and Respectively and The ratio of all pixels in the image before foreign object recognition; Step A2: Calculate the intra-class variance of the foreground part and the intra-class variance of the background part, and calculate the intra-class variance of the image in the early stage of foreign object recognition; Step A3: Calculate the average grayscale value of the foreground after adding the intra-class variance and background The average gray value , calculate the average gray value of the image in the early stage of foreign body recognition after adding the intra-class variance ; Step A4: Use formula (1) to calculate the optimal threshold t for segmenting the image in the early stage of foreign body recognition and obtain a binary image: (1) In formula (1), Indicates the maximum value operation; Step 4: Perform Hough transform detection on the denoised binary image to obtain each transmission line, and perform tilt correction on the denoised binary image based on the rotation direction of the longest transmission line to obtain a corrected image of the transmission line tending to be horizontal; Step 5: After performing Hough transform on the corrected image, all straight line segments in the corrected image are obtained and the straight lines are screened. The straight line screening operation includes: the length and slope of the straight line; Step 6: Statistically calculate the vertical coordinates of the filtered straight line segments, and define the area between the minimum and maximum vertical coordinates as the transmission line area; Step 7: After selecting the minimum circumscribed rectangular outline of the connected area in the rectified image, the selected outline is filtered. The contour screening operation includes: the area and the aspect ratio of the contour; Step 8: Compare the vertical coordinates of the endpoints of the filtered contour with the vertical coordinates of the transmission line area. If any of the vertical coordinates of the four endpoints of the contour is within the range of the vertical coordinates of the transmission line area, it indicates that the type of the corresponding contour is an attached contour; otherwise, it indicates that the type of the corresponding contour is a hanging contour. Step 9: Count the white dots inside the minimum circumscribed rectangle and use it to set the number of white dots inside the outline. and the area of the contour The ratio between them is used as the white point rate, different white point rates are set for the two divided contour types, and the contours that meet the corresponding white point rates are output as the corresponding foreign body types.

2. The method for detecting foreign matter in a power transmission line based on the white point rate method according to claim 1, characterized in that: The screening in step 5 is performed as follows: For any j-th straight line segment in the rectified image The slope ,when When , retain the jth straight line segment , otherwise, remove the jth straight line segment ;in, represents the slope screening threshold; Calculate the jth straight line segment Length , and use the screening conditions shown in formula (2) to screen. If formula (2) is satisfied, the jth straight line segment is retained. , otherwise, remove the jth straight line segment ; (2) In formula (2), Indicates getting the number of straight line segments.

3. An electronic device comprising a memory and a processor, characterized in that: The memory is used to store a program that supports the processor to execute the power transmission line foreign object detection method according to claim 1 or 2, and the processor is configured to execute the program stored in the memory.

4. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for detecting foreign objects in a power transmission line according to claim 1 or 2 are executed.

Images