A method, system and drone for defect detection of distribution lines

Abstract

The present invention relates to the technical field of defect detection of distribution lines, and specifically discloses a defect detection method, system and drone for distribution lines, including: based on the current inspection video of the distribution line, obtaining all processed video frames of the current inspection video of the distribution line and the comparison video frames of each processed video frame; based on all processed video frames of the current inspection video of the distribution line and the comparison video frames of all processed video frames, obtaining all video frame control groups; based on the processed video frames in each video frame control group, obtaining the restored video frames of the processed video frames in each video frame control group; based on the restored video frames of the processed video frames in all video frame control groups and a preset defect detection model, obtaining the defect detection result of the distribution line. The present invention realizes defect detection of distribution lines in rainy and snowy weather, and improves the efficiency of defect detection.

Description

Technical Field

[0001] The present invention relates to the technical field of power distribution line defect detection, and in particular to a power distribution line defect detection method, system and drone. Background Art

[0002] At present, in the power system, the inspection and maintenance of distribution lines are the key links to ensure the safe and stable operation of the power grid. The most primitive power inspection is through manual inspection, but this traditional manual inspection is time-consuming, slow in detection and has certain safety hazards. However, with the advancement of science and technology, today's power inspection has been replaced by drone inspection instead of the original manual inspection. However, in the actual inspection process, weather conditions often become an important factor affecting the efficiency and accuracy of drone inspections. Especially in rainy and snowy weather, it is difficult to ensure the accuracy of the inspection results due to the obstruction of rain or snowflakes. Therefore, how to realize defect detection of distribution lines in rainy and snowy weather has become an urgent issue for drone inspections.

[0003] However, the existing defect detection methods, systems and drones for distribution lines only screen positive and negative samples through weighted loss, improve the distribution speed of positive and negative samples, shorten the training time and avoid additional parameters that need to be optimized, but do not consider how to realize defect detection of distribution lines in rainy and snowy weather. For example, the patent with the publication number "CN116485731A" and the patent name "Defect Detection Method, System, Identification Terminal and Drone for Distribution Lines" includes the following steps: the inspection image of the distribution line is detected by a defect detection model based on the YOLO network; the YOLO network determines the number of positive samples by intersection and union ratio, and screens positive and negative samples by weighted loss, improves the distribution speed of positive and negative samples, shortens the training time, and avoids additional parameters that need to be optimized. However, the patent only screens positive and negative samples by weighted loss, improves the distribution speed of positive and negative samples, shortens the training time and avoids additional parameters that need to be optimized, but does not consider how to realize defect detection of distribution lines in rainy and snowy weather.

[0004] Therefore, the present invention proposes a method, system and drone for defect detection of distribution lines. Summary of the invention

[0005] The present invention provides a method, system and drone for defect detection of distribution lines, which are used to accurately obtain the comparison video frame of each processed video frame of the current inspection video of the distribution line according to the current inspection video of the distribution line, and then obtain all video frame comparison groups according to all processed video frames of the current inspection video of the distribution line and the comparison video frames of all processed video frames, so as to facilitate the screening of subsequent video frame control groups, obtain all video frame control groups according to all video frame comparison groups, accurately obtain the grid standard deviation of the processed video frames in each video frame control group according to the processed video frames in each video frame control group, and realize the uniformity of the pixel value distribution of all pixel points of the processed video frames in each video frame control group. The data are quantized, and then according to the processed video frames in each video frame control group and the grid standard deviation of the corresponding processed video frames, the initial restored video frames of the processed video frames in each video frame control group are accurately obtained. According to the initial restored video frames and the compared video frames of the processed video frames in each video frame control group, the restored video frames of the processed video frames in each video frame control group are obtained, which effectively weakens the influence of rain and snow on the image clarity of the processed video frames in each video frame control group. Finally, according to the restored video frames of the processed video frames in all video frame control groups and the preset defect detection model, the defect detection results of the distribution lines are accurately obtained, so as to realize defect detection of distribution lines in rainy and snowy weather and improve the efficiency of defect detection.

[0006] The present invention provides a method for detecting defects in a power distribution line, comprising:

[0007] S1: based on the current inspection video of the distribution line, obtaining all processed video frames of the current inspection video of the distribution line, and based on all processed video frames of the current inspection video of the distribution line, obtaining a comparison video frame of each processed video frame of the current inspection video of the distribution line;

[0008] S2: based on all processed video frames of the current inspection video of the distribution line and the comparison video frames of all processed video frames, obtain all video frame comparison groups, and based on all video frame comparison groups, obtain all video frame control groups;

[0009] S3: based on the processed video frames in each video frame control group, obtaining the grid standard deviation of the processed video frames in each video frame control group, based on the processed video frames in each video frame control group and the grid standard deviation of the corresponding processed video frames, obtaining the initial restored video frames of the processed video frames in each video frame control group, based on the initial restored video frames of the processed video frames in each video frame control group and the comparison video frames, obtaining the restored video frames of the processed video frames in each video frame control group;

[0010] S4: Obtain defect detection results of the distribution lines based on the restored video frames of the processed video frames in all video frame control groups and a preset defect detection model.

[0011] Preferably, a defect detection method for a distribution line, S1: based on a current inspection video of the distribution line, obtaining all processed video frames of the current inspection video of the distribution line, and based on all processed video frames of the current inspection video of the distribution line, obtaining a comparison video frame of each processed video frame of the current inspection video of the distribution line, including:

[0012] Obtain all video frames of the current inspection video of the distribution line, and obtain the shooting position of each video frame of the current inspection video of the distribution line;

[0013] If the distance between each preset shooting position and the shooting position of each video frame of the current inspection video of the power distribution line is less than the preset interval distance, the corresponding video frame of the current inspection video of the power distribution line is regarded as the processing video frame corresponding to the preset shooting position;

[0014] All processed video frames of all preset shooting positions are used as processed video frames of the current inspection video of the distribution line;

[0015] Based on all processed video frames of the current inspection video of the distribution line, a comparison video frame of each processed video frame of the current inspection video of the distribution line is obtained.

[0016] Preferably, the defect detection method of the distribution line obtains a comparison video frame of each processed video frame of the current inspection video of the distribution line based on all processed video frames of the current inspection video of the distribution line, including:

[0017] Obtain all video frames of a preset standard inspection video of a power distribution line, and obtain a shooting position of each video frame of the preset standard inspection video of a power distribution line;

[0018] The video frame of the preset standard inspection video that is closest to the shooting position of all video frames of the preset standard inspection video of the distribution line and the shooting position of each processed video frame of the current inspection video of the distribution line is used as the comparison video frame of the corresponding processed video frame of the current inspection video of the distribution line.

[0019] Preferably, the defect detection method for the distribution line, S2: based on all processed video frames of the current inspection video of the distribution line and the comparison video frames of all processed video frames, obtaining all video frame comparison groups, and based on all video frame comparison groups, obtaining all video frame control groups, including:

[0020] Each processed video frame of the current inspection video of the distribution line and the comparison video frame of the corresponding processed video frame are used as a video frame comparison group;

[0021] Obtaining all similar video frame groups based on the comparison video frames of the processed video frames in all video frame comparison groups and the preset first image recognition model;

[0022] Based on each similar video frame group and a preset second image recognition model, obtaining the number of line pixel areas of the comparison video frame of each processed video frame in each similar video frame group;

[0023] The comparison video frame of the processed video frame with the largest number of line pixel areas in each similar video frame group and the processed video frame in the video frame comparison group corresponding to the corresponding comparison video frame are used as a video frame control group.

[0024] Preferably, the defect detection method for distribution lines obtains the grid standard deviation of the processed video frames in each video frame control group based on the processed video frames in each video frame control group, including:

[0025] Divide the processed video frames in each video frame control group into a preset number of grids of equal size, and obtain all grids of the processed video frames in each video frame control group;

[0026] The pixel extraction value is set to 0, and the number of pixel points in each grid of the processed video frame in each video frame control group whose pixel value is the same as the pixel extraction value is regarded as the extraction number of each grid of the processed video frame in each video frame control group;

[0027] The standard deviation of the extraction numbers of all grids of the processed video frames in each video frame control group is regarded as the standard deviation of the processed video frames in each video frame control group;

[0028] The pixel extraction value is set to 1, and the number of pixel points in each grid of the processed video frame in each video frame control group whose pixel value is the same as the pixel extraction value is regarded as the extraction number of each grid of the processed video frame in each video frame control group;

[0029] The standard deviation of the number of extractions of all grids of the processed video frames in each video frame control group is also regarded as the standard deviation of the processed video frames in each video frame control group;

[0030] Continue to traverse the pixel extraction value incrementally by unit value until the pixel extraction value reaches 255, and obtain all standard deviations of the processed video frames in each video frame control group;

[0031] The standard deviation with the smallest value among all standard deviations of the processed video frames in each video frame control group is taken as the grid standard deviation of the processed video frames in each video frame control group.

[0032] Preferably, the defect detection method of the distribution line obtains the initial restored video frame of the processed video frame in each video frame control group based on the processed video frame in each video frame control group and the grid standard deviation of the corresponding processed video frame, including:

[0033] The pixel extraction value corresponding to the grid standard deviation of the processed video frame in each video frame control group is regarded as the rain and snow pixel value of the processed video frame in each video frame control group;

[0034] The pixel points whose pixel values ​​are the same as the rain and snow pixel values ​​of the processed video frames in the corresponding video frame control group among all the pixel points of the processed video frames in each video frame control group are regarded as the processed pixel points of the processed video frames in each video frame control group, and all the pixel points except the processed pixel points among all the pixel points of the processed video frames in each video frame control group are regarded as the non-processed pixel points of the processed video frames in each video frame control group;

[0035] The area consisting of all adjacent processed pixel points of the processed video frame in each video frame control group is regarded as the processed area of ​​the processed video frame in each video frame control group, and the area consisting of all non-processed pixel points of the processed video frame in each video frame control group is regarded as the non-processed area of ​​the processed video frame in each video frame control group;

[0036] If the number of pixel points in each processing area of ​​the processed video frame in each video frame control group is less than the preset judgment number, the pixel values ​​of all pixel points in the corresponding processing area of ​​the processed video frame in the corresponding video frame control group are adjusted to be the same as the adjacent boundary pixel values ​​of the corresponding processing area of ​​the processed video frame in the corresponding video frame control group, so as to obtain the initial restored video frame of the processed video frame in each video frame control group.

[0037] Preferably, the defect detection method of the distribution line obtains the restored video frame of the processed video frame in each video frame control group based on the initial restored video frame and the comparison video frame of the processed video frame in each video frame control group, comprising:

[0038] A combined range of all processing regions of the processed video frames in each video frame control group, in which the number of pixels is not less than a preset judgment number, is used as a restoration region of an initial restoration video frame of the processed video frames in each video frame control group, and a combined range of all processing regions of the processed video frames in each video frame control group, in which the number of pixels is less than a preset judgment number, is used as a first non-restoration region of an initial restoration video frame of the processed video frames in each video frame control group;

[0039] Using the non-processed region of the processed video frame in each video frame control group as the second non-restored region of the initial restored video frame of the processed video frame in each video frame control group;

[0040] Based on the comparison video frame of the processed video frame in each video frame control group and the restored area, the first non-restored area and the second non-restored area of ​​the initial restored video frame, the replacement value of each pixel in the restored area of ​​the initial restored video frame of the processed video frame in each video frame control group is obtained, that is:

[0041]

[0042] Wherein, 6 is the replacement value of the currently calculated pixel in the restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, ε1 is the pixel value of the currently calculated pixel in the restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, β1 is the average value of the pixel values ​​of all pixels in the first non-restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, τ1 is the average value of the pixel values ​​of all pixels in the second non-restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, ε2 is the average value of the pixel values ​​of all pixels in the comparison video frame of the processed video frame in the currently calculated video frame control group, β2 is the minimum value of the pixel values ​​of all pixels in the first non-restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, τ2 is the minimum value of the pixel values ​​of all pixels in the second non-restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, ln is the natural logarithm, and the value of the natural constant e is 2.718;

[0043] Based on the replacement value of each pixel point in the restoration area of ​​the initial restoration video frame of the processed video frame in each video frame control group, the restored video frame of the processed video frame in each video frame control group is obtained.

[0044] Preferably, the defect detection method of the distribution line obtains the restored video frame of the processed video frame in each video frame control group based on the replacement value of each pixel point in the restored area of ​​the initial restored video frame of the processed video frame in each video frame control group, comprising:

[0045] If the replacement value of each pixel point in the restoration area of ​​the initial restored video frame of the processed video frame in each video frame control group is greater than the preset replacement value threshold, the pixel value of the corresponding pixel point in the restoration area of ​​the initial restored video frame of the processed video frame in the corresponding video frame control group is adjusted to be the same as the average value of the pixel values ​​of all pixels in the comparison video frame of the processed video frame in the corresponding video frame control group, so as to obtain the restored video frame of the processed video frame in each video frame control group.

[0046] The present invention provides a distribution line defect detection system, which is used to execute any one of the distribution line defect detection methods in embodiments 1 to 9, including:

[0047] A first processing module is used to obtain all processed video frames of the current inspection video of the distribution line based on the current inspection video of the distribution line, and obtain a comparison video frame of each processed video frame of the current inspection video of the distribution line based on all processed video frames of the current inspection video of the distribution line;

[0048] A second processing module is used to obtain all video frame comparison groups based on all processed video frames of the current inspection video of the distribution line and the comparison video frames of all processed video frames, and obtain all video frame control groups based on all video frame comparison groups;

[0049] A restoration module, for obtaining a grid standard deviation of the processed video frames in each video frame control group based on the processed video frames in each video frame control group, obtaining an initial restored video frame of the processed video frames in each video frame control group based on the processed video frames in each video frame control group and the grid standard deviation of the corresponding processed video frames, and obtaining a restored video frame of the processed video frames in each video frame control group based on the initial restored video frame and the comparison video frame of the processed video frames in each video frame control group;

[0050] The detection module is used to obtain the defect detection result of the distribution line based on the restored video frames of the processed video frames in all video frame control groups and the preset defect detection model.

[0051] The present invention provides a drone, comprising: an integrated controller, a flight platform and a wireless transmission device; the integrated controller is used to obtain a fixed waypoint preset by an operator, and interact with the flight platform to send the fixed waypoint to the flight platform;

[0052] The flight platform is used to convert the fixed waypoints into corresponding flight paths and execute them;

[0053] The wireless transmission device is used to wirelessly transmit the current inspection video of the distribution line between the mobile control terminal and the drone.

[0054] The beneficial effects of the present invention compared with the prior art are as follows: according to the current inspection video of the distribution line, the comparison video frame of each processed video frame of the current inspection video of the distribution line is accurately obtained, and then according to all the processed video frames of the current inspection video of the distribution line and the comparison video frames of all the processed video frames, all the video frame comparison groups are obtained, which is convenient for screening out the subsequent video frame control groups, and according to all the video frame comparison groups, all the video frame control groups are obtained, and according to the processed video frames in each video frame control group, the grid standard deviation of the processed video frames in each video frame control group is accurately obtained, so as to realize the quantification of the uniformity of the pixel value distribution of all pixel points of the processed video frames in each video frame control group. Then, according to the processed video frames in each video frame control group and the grid standard deviation of the corresponding processed video frames, the initial restored video frames of the processed video frames in each video frame control group are accurately obtained, and according to the initial restored video frames and the compared video frames of the processed video frames in each video frame control group, the restored video frames of the processed video frames in each video frame control group are obtained, which effectively weakens the influence of rain and snow on the image clarity of the processed video frames in each video frame control group. Finally, according to the restored video frames of the processed video frames in all video frame control groups and the preset defect detection model, the defect detection results of the distribution lines are accurately obtained, so as to realize defect detection of distribution lines in rainy and snowy weather and improve the efficiency of defect detection.

[0055] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structure specifically pointed out in the written application document.

[0056] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0058] Figure 1 This is a flow chart of a method for detecting defects in a power distribution line according to an embodiment of the present invention;

[0059] Figure 2 Schematic diagram of a defect detection system for a power distribution line in an embodiment of the present invention. DETAILED DESCRIPTION

[0060] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0061] Embodiment 1:

[0062] The present invention provides a method for detecting defects in a power distribution line. Figure 1 ,include:

[0063] S1: based on the current inspection video of the distribution line, obtaining all processed video frames of the current inspection video of the distribution line, and based on all processed video frames of the current inspection video of the distribution line, obtaining a comparison video frame of each processed video frame of the current inspection video of the distribution line;

[0064] S2: based on all processed video frames of the current inspection video of the distribution line and the comparison video frames of all processed video frames, obtain all video frame comparison groups, and based on all video frame comparison groups, obtain all video frame control groups;

[0065] S3: based on the processed video frames in each video frame control group, obtaining the grid standard deviation of the processed video frames in each video frame control group, based on the processed video frames in each video frame control group and the grid standard deviation of the corresponding processed video frames, obtaining the initial restored video frames of the processed video frames in each video frame control group, based on the initial restored video frames of the processed video frames in each video frame control group and the comparison video frames, obtaining the restored video frames of the processed video frames in each video frame control group;

[0066] S4: Obtain defect detection results of the distribution lines based on the restored video frames of the processed video frames in all video frame control groups and a preset defect detection model.

[0067] In this embodiment, the current inspection video of the distribution line is the inspection video of the distribution line currently captured by the drone.

[0068] In this embodiment, all processed video frames of the current inspection video of the power distribution line are partial video frames screened out after processing all processed video frames of the current inspection video of the power distribution line.

[0069] In this embodiment, the comparison video frames of the processed video frames are video frames in all video frames of the preset standard inspection video of the distribution line, which can be compared with each processed video frame of the current inspection video of the distribution line.

[0070] In this embodiment, the video frame comparison group is a combination of each processed video frame of the current inspection video of the distribution line and a comparison video frame corresponding to the processed video frame.

[0071] In this embodiment, the video frame control group is a partial video frame comparison group that is screened out from all video frame comparison groups and has the best analysis effect on the shooting location of a single distribution line (the shooting location includes the location, such as the insulator on the power tower 500 meters in front of the transformer station).

[0072] In this embodiment, the grid standard deviation is a value obtained based on the processed video frames in each video frame control group and can reflect the uniformity of the pixel value distribution of all pixels of the processed video frames in each video frame control group.

[0073] In this embodiment, the initially restored video frame is a video frame obtained by initially restoring the processed video frame in each video frame control group based on the processed video frame in each video frame control group and the grid standard deviation of the corresponding processed video frame.

[0074] In this embodiment, the restored video frame is a video frame obtained by secondary restoration of the initial restored video frame of the processed video frame in each video frame control group based on the initial restored video frame and the comparison video frame of the processed video frame in each video frame control group.

[0075] In this embodiment, a preset defect detection model uses restored video frames of processed video frames in a large number of pre-collected video frame control groups as model input, and manually annotated judgment results of whether there are defects in the distribution lines in the restored video frames of the processed video frames in the corresponding video frame control groups as model outputs. The trained restored video frames that can be input into the processed video frames in the video frame control group can output the judgment results of whether there are defects in the distribution lines in the restored video frames of the processed video frames in the corresponding video frame control group.

[0076] In this embodiment, the defect detection result of the power distribution line is a summary of the judgment results of whether there are defects in the power distribution line in the restored video frames of the processed video frames in all the video frame control groups.

[0077] The beneficial effects of the above technology are as follows: according to the current inspection video of the distribution line, the comparison video frame of each processed video frame of the current inspection video of the distribution line is accurately obtained, and then according to all the processed video frames of the current inspection video of the distribution line and the comparison video frames of all the processed video frames, all video frame comparison groups are obtained, which is convenient for the subsequent screening of video frame control groups, and according to all the video frame comparison groups, all video frame control groups are obtained, and according to the processed video frames in each video frame control group, the grid standard deviation of the processed video frames in each video frame control group is accurately obtained, which realizes the quantification of the uniformity of the pixel value distribution of all pixel points of the processed video frames in each video frame control group, and then according to The processed video frames in each video frame control group and the grid standard deviation of the corresponding processed video frames are used to accurately obtain the initial restored video frames of the processed video frames in each video frame control group. According to the initial restored video frames and the comparison video frames of the processed video frames in each video frame control group, the restored video frames of the processed video frames in each video frame control group are obtained, which effectively weakens the influence of rain and snow on the image clarity of the processed video frames in each video frame control group. Finally, according to the restored video frames of the processed video frames in all video frame control groups and the preset defect detection model, the defect detection results of the distribution lines are accurately obtained, so as to realize defect detection of distribution lines in rainy and snowy weather and improve the efficiency of defect detection.

[0078] Embodiment 2:

[0079] Based on Example 1, a defect detection method for a distribution line, S1: based on a current inspection video of the distribution line, obtaining all processed video frames of the current inspection video of the distribution line, and based on all processed video frames of the current inspection video of the distribution line, obtaining a comparison video frame of each processed video frame of the current inspection video of the distribution line, including:

[0080] Obtain all video frames of the current inspection video of the distribution line, and obtain the shooting position of each video frame of the current inspection video of the distribution line;

[0081] If the distance between each preset shooting position and the shooting position of each video frame of the current inspection video of the power distribution line is less than the preset interval distance, the corresponding video frame of the current inspection video of the power distribution line is regarded as the processing video frame corresponding to the preset shooting position;

[0082] All processed video frames of all preset shooting positions are used as processed video frames of the current inspection video of the distribution line;

[0083] Based on all processed video frames of the current inspection video of the distribution line, a comparison video frame of each processed video frame of the current inspection video of the distribution line is obtained.

[0084] In this embodiment, the shooting position of each video frame of the current inspection video of the distribution line is the shooting position of each video frame of the current inspection video of the distribution line acquired by the drone (the real-time spatial position of the drone when shooting).

[0085] In this embodiment, the preset shooting position is a shooting position preset by the drone operator.

[0086] In this embodiment, the preset interval distance is a preset interval distance used to obtain all processed video frames of each preset shooting position, for example, 3 meters.

[0087] The beneficial effects of the above technology are: based on the current inspection video of the distribution line, the preset shooting position and the preset interval distance, all processed video frames of the current inspection video of the distribution line are accurately obtained, which is convenient for the subsequent acquisition of the comparison video frame of each processed video frame of the current inspection video of the distribution line, and based on all the processed video frames of the current inspection video of the distribution line, the comparison video frame of each processed video frame of the current inspection video of the distribution line is obtained.

[0088] Embodiment 3:

[0089] On the basis of Example 2, the defect detection method of the distribution line obtains a comparison video frame of each processed video frame of the current inspection video of the distribution line based on all processed video frames of the current inspection video of the distribution line, including:

[0090] Obtain all video frames of a preset standard inspection video of a power distribution line, and obtain a shooting position of each video frame of the preset standard inspection video of a power distribution line;

[0091] The video frame of the preset standard inspection video that is closest to the shooting position of all video frames of the preset standard inspection video of the distribution line and the shooting position of each processed video frame of the current inspection video of the distribution line is used as the comparison video frame of the corresponding processed video frame of the current inspection video of the distribution line.

[0092] In this embodiment, the preset standard inspection video is a pre-selected inspection video that has been used to detect defects in the power distribution line and has no abnormalities in the detection results.

[0093] The beneficial effect of the above technology is: based on the preset standard inspection video of the distribution line and all the processed video frames of the current inspection video of the distribution line, the comparison video frame of each processed video frame of the current inspection video of the distribution line is accurately obtained. This embodiment gives in detail a specific method for obtaining the comparison video frame of each processed video frame of the current inspection video of the distribution line based on all the processed video frames of the current inspection video of the distribution line.

[0094] Embodiment 4:

[0095] Based on Example 1, a defect detection method for a distribution line, S2: based on all processed video frames of the current inspection video of the distribution line and the comparison video frames of all processed video frames, obtaining all video frame comparison groups, and based on all video frame comparison groups, obtaining all video frame control groups, including:

[0096] Each processed video frame of the current inspection video of the distribution line and the comparison video frame of the corresponding processed video frame are used as a video frame comparison group;

[0097] Obtaining all similar video frame groups based on the comparison video frames of the processed video frames in all video frame comparison groups and the preset first image recognition model;

[0098] Based on each similar video frame group and a preset second image recognition model, obtaining the number of line pixel areas of the comparison video frame of each processed video frame in each similar video frame group;

[0099] The comparison video frame of the processed video frame with the largest number of line pixel areas in each similar video frame group and the processed video frame in the video frame comparison group corresponding to the corresponding comparison video frame are used as a video frame control group.

[0100] In this embodiment, the first image recognition model is preset to use comparison video frames of processed video frames in a large number of pre-collected video frame comparison groups as model input, and the results of manually dividing the comparison video frames of the processed video frames in all video frame comparison groups according to the locations of the photographed distribution lines are used as model output. The trained model that can input the comparison video frames of the processed video frames in all video frame comparison groups can output the results of dividing the comparison video frames of the processed video frames in all video frame comparison groups.

[0101] In this embodiment, the similar video frame group is a combination of partial comparison video frames whose shooting locations of the distribution lines are consistent (the shooting location includes the location, such as the insulators on the power tower 500 meters in front of the transformer station) screened out from the comparison video frames of the processed video frames in all video frame comparison groups.

[0102] In this embodiment, the second image recognition model is preset to use a large number of similar video frame groups collected in advance as model input, and the number of line pixel areas of the comparison video frame of each processed video frame in each similar video frame group that has been manually annotated as model output. The trained model can input similar video frame groups and output the number of line pixel areas of the comparison video frame of each processed video frame in the corresponding similar video frame group.

[0103] In this embodiment, the number of line pixel regions is the number of pixel points in the region occupied by the shooting location of the power distribution line in the comparison video frame of each processed video frame in each similar video frame group.

[0104] The beneficial effects of the above technology are as follows: based on all processed video frames of the current inspection video of the distribution line and the comparison video frames of all processed video frames, all video frame comparison groups are obtained, and based on all video frame comparison groups, the preset first image recognition model and the preset second image recognition model, all video frame control groups are accurately obtained, thereby realizing the selection of some video frame comparison groups from all video frame comparison groups that have the best analysis effect on the shooting location of a single distribution line.

[0105] Embodiment 5:

[0106] On the basis of Example 1, the defect detection method for distribution lines obtains the grid standard deviation of the processed video frames in each video frame control group based on the processed video frames in each video frame control group, including:

[0107] Divide the processed video frames in each video frame control group into a preset number of grids of equal size, and obtain all grids of the processed video frames in each video frame control group;

[0108] The pixel extraction value is set to 0, and the number of pixel points in each grid of the processed video frame in each video frame control group whose pixel value is the same as the pixel extraction value is regarded as the extraction number of each grid of the processed video frame in each video frame control group;

[0109] The standard deviation of the extraction numbers of all grids of the processed video frames in each video frame control group is regarded as the standard deviation of the processed video frames in each video frame control group;

[0110] The pixel extraction value is set to 1, and the number of pixel points in each grid of the processed video frame in each video frame control group whose pixel value is the same as the pixel extraction value is regarded as the extraction number of each grid of the processed video frame in each video frame control group;

[0111] The standard deviation of the number of extractions of all grids of the processed video frames in each video frame control group is also regarded as the standard deviation of the processed video frames in each video frame control group;

[0112] Continue to traverse the pixel extraction value incrementally by unit value until the pixel extraction value reaches 255, and obtain all standard deviations of the processed video frames in each video frame control group;

[0113] The standard deviation with the smallest value among all standard deviations of the processed video frames in each video frame control group is taken as the grid standard deviation of the processed video frames in each video frame control group.

[0114] In this embodiment, the preset number is a preset number of grids divided into grids of equal size, for example, 10.

[0115] In this embodiment, the incremental traversal is to increment the pixel extraction value by unit value (1) to obtain a new standard deviation of the processed video frames in each video frame control group, until the pixel extraction value increases to 255 and stops, to obtain all standard deviations (255) of the processed video frames in each video frame control group.

[0116] The beneficial effects of the above technology are: based on the processed video frames in each video frame control group, the grid standard deviation of the processed video frames in each video frame control group is obtained, which is convenient for subsequently obtaining the initial restored video frames of the processed video frames in each video frame control group. This embodiment gives in detail a specific method for obtaining the grid standard deviation of the processed video frames in each video frame control group based on the processed video frames in each video frame control group.

[0117] Embodiment 6:

[0118] On the basis of Example 5, the defect detection method for distribution lines obtains the initial restored video frame of the processed video frame in each video frame control group based on the processed video frame in each video frame control group and the grid standard deviation of the corresponding processed video frame, including:

[0119] The pixel extraction value corresponding to the grid standard deviation of the processed video frame in each video frame control group is regarded as the rain and snow pixel value of the processed video frame in each video frame control group;

[0120] The pixel points whose pixel values ​​are the same as the rain and snow pixel values ​​of the processed video frames in the corresponding video frame control group among all the pixel points of the processed video frames in each video frame control group are regarded as the processed pixel points of the processed video frames in each video frame control group, and all the pixel points except the processed pixel points among all the pixel points of the processed video frames in each video frame control group are regarded as the non-processed pixel points of the processed video frames in each video frame control group;

[0121] The area consisting of all adjacent processed pixel points of the processed video frame in each video frame control group is regarded as the processed area of ​​the processed video frame in each video frame control group, and the area consisting of all non-processed pixel points of the processed video frame in each video frame control group is regarded as the non-processed area of ​​the processed video frame in each video frame control group;

[0122] If the number of pixel points in each processing area of ​​the processed video frame in each video frame control group is less than the preset judgment number, the pixel values ​​of all pixel points in the corresponding processing area of ​​the processed video frame in the corresponding video frame control group are adjusted to be the same as the adjacent boundary pixel values ​​of the corresponding processing area of ​​the processed video frame in the corresponding video frame control group, so as to obtain the initial restored video frame of the processed video frame in each video frame control group.

[0123] In this embodiment, the preset determination number is a preset number of pixel points for determining whether to adjust the pixel values ​​of all pixel points in each processing area of ​​the processed video frame in the video frame control group.

[0124] In this embodiment, the adjacent boundary pixel value of each processing area of ​​the processed video frame in each video frame control group is the average of the pixel values ​​of all pixels adjacent to the edge of each processing area of ​​the processed video frame in the corresponding video frame control group in the processed video frame in each video frame control group.

[0125] The beneficial effects of the above technology are: according to the processed video frames in each video frame control group and the grid standard deviation of the corresponding processed video frames, the rain and snow pixel values ​​of the processed video frames in each video frame control group are obtained, and then according to the rain and snow pixel values ​​of the processed video frames in each video frame control group, all the processed areas of the processed video frames in each video frame control group are obtained, and according to all the processed areas of the processed video frames in each video frame control group, the initial restored video frames of the processed video frames in each video frame control group are obtained. This embodiment gives in detail a specific method for obtaining the initial restored video frames of the processed video frames in each video frame control group based on the processed video frames in each video frame control group and the grid standard deviation of the corresponding processed video frames.

[0126] Embodiment 7:

[0127] On the basis of Example 6, the defect detection method for distribution lines obtains the restored video frame of the processed video frame in each video frame control group based on the initial restored video frame and the comparison video frame of the processed video frame in each video frame control group, including:

[0128] A combined range of all processing regions of the processed video frames in each video frame control group, in which the number of pixels is not less than a preset judgment number, is used as a restoration region of an initial restoration video frame of the processed video frames in each video frame control group, and a combined range of all processing regions of the processed video frames in each video frame control group, in which the number of pixels is less than a preset judgment number, is used as a first non-restoration region of an initial restoration video frame of the processed video frames in each video frame control group;

[0129] Using the non-processed region of the processed video frame in each video frame control group as the second non-restored region of the initial restored video frame of the processed video frame in each video frame control group;

[0130] Based on the comparison video frame of the processed video frame in each video frame control group and the restored area, the first non-restored area and the second non-restored area of ​​the initial restored video frame, the replacement value of each pixel in the restored area of ​​the initial restored video frame of the processed video frame in each video frame control group is obtained, that is:

[0131]

[0132] Wherein, 6 is the replacement value of the currently calculated pixel in the restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, ε1 is the pixel value of the currently calculated pixel in the restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, β1 is the average value of the pixel values ​​of all pixels in the first non-restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, τ1 is the average value of the pixel values ​​of all pixels in the second non-restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, ε2 is the average value of the pixel values ​​of all pixels in the comparison video frame of the processed video frame in the currently calculated video frame control group, β2 is the minimum value of the pixel values ​​of all pixels in the first non-restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, τ2 is the minimum value of the pixel values ​​of all pixels in the second non-restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, ln is the natural logarithm, and the value of the natural constant e is 2.718;

[0133] Based on the replacement value of each pixel point in the restoration area of ​​the initial restoration video frame of the processed video frame in each video frame control group, the restored video frame of the processed video frame in each video frame control group is obtained.

[0134] In this embodiment, the replacement value is a numerical value obtained based on the comparison video frame of the processed video frame in each video frame control group and the restored area, the first non-restored area and the second non-restored area of ​​the initial restored video frame, which can characterize whether the pixel value of each pixel point in the restored area of ​​the initial restored video frame of the processed video frame in each video frame control group needs to be replaced.

[0135] The beneficial effects of the above technology are: according to the initial restored video frame and the comparison video frame of the processed video frame in each video frame control group, the replacement value of each pixel point in the restoration area of ​​the initial restored video frame of the processed video frame in each video frame control group is accurately obtained, and then according to the replacement value of each pixel point in the restoration area of ​​the initial restored video frame of the processed video frame in each video frame control group, the restored video frame of the processed video frame in each video frame control group is obtained, which effectively weakens the influence of rain and snow on the image clarity of the processed video frame in each video frame control group.

[0136] Embodiment 8:

[0137] On the basis of Example 7, the defect detection method for a distribution line obtains a restored video frame of a processed video frame in each video frame control group based on the replacement value of each pixel point in the restored area of ​​the initial restored video frame of the processed video frame in each video frame control group, including:

[0138] If the replacement value of each pixel point in the restoration area of ​​the initial restored video frame of the processed video frame in each video frame control group is greater than the preset replacement value threshold, the pixel value of the corresponding pixel point in the restoration area of ​​the initial restored video frame of the processed video frame in the corresponding video frame control group is adjusted to be the same as the average value of the pixel values ​​of all pixels in the comparison video frame of the processed video frame in the corresponding video frame control group, so as to obtain the restored video frame of the processed video frame in each video frame control group.

[0139] In this embodiment, the preset replacement value threshold is a replacement value threshold that is preset to determine whether to adjust the pixel value of each pixel point in the restoration area of ​​the initial restoration video frame of the processed video frame in each video frame control group.

[0140] The beneficial effect of the above technology is: according to the replacement value of each pixel point in the restoration area of ​​the initial restored video frame of the processed video frame in each video frame control group and the preset replacement value threshold, the restored video frame of the processed video frame in each video frame control group is obtained. This embodiment gives in detail a specific method for obtaining the restored video frame of the processed video frame in each video frame control group based on the replacement value of each pixel point in the restoration area of ​​the initial restored video frame of the processed video frame in each video frame control group.

[0141] Embodiment 9:

[0142] The present invention provides a distribution line defect detection system, which is used to perform any one of the distribution line defect detection methods in embodiments 1 to 9, referring to Figure 2 ,include:

[0143] A first processing module is used to obtain all processed video frames of the current inspection video of the distribution line based on the current inspection video of the distribution line, and obtain a comparison video frame of each processed video frame of the current inspection video of the distribution line based on all processed video frames of the current inspection video of the distribution line;

[0144] A second processing module is used to obtain all video frame comparison groups based on all processed video frames of the current inspection video of the distribution line and the comparison video frames of all processed video frames, and obtain all video frame control groups based on all video frame comparison groups;

[0145] A restoration module, for obtaining a grid standard deviation of the processed video frames in each video frame control group based on the processed video frames in each video frame control group, obtaining an initial restored video frame of the processed video frames in each video frame control group based on the processed video frames in each video frame control group and the grid standard deviation of the corresponding processed video frames, and obtaining a restored video frame of the processed video frames in each video frame control group based on the initial restored video frame and the comparison video frame of the processed video frames in each video frame control group;

[0146] The detection module is used to obtain the defect detection result of the distribution line based on the restored video frames of the processed video frames in all video frame control groups and the preset defect detection model.

[0147] The beneficial effects of the above technology are as follows: according to the current inspection video of the distribution line, the comparison video frame of each processed video frame of the current inspection video of the distribution line is accurately obtained, and then according to all the processed video frames of the current inspection video of the distribution line and the comparison video frames of all the processed video frames, all video frame comparison groups are obtained, which is convenient for the subsequent screening of video frame control groups, and according to all the video frame comparison groups, all video frame control groups are obtained, and according to the processed video frames in each video frame control group, the grid standard deviation of the processed video frames in each video frame control group is accurately obtained, which realizes the quantification of the uniformity of the pixel value distribution of all pixel points of the processed video frames in each video frame control group, and then according to The processed video frames in each video frame control group and the grid standard deviation of the corresponding processed video frames are used to accurately obtain the initial restored video frames of the processed video frames in each video frame control group. According to the initial restored video frames and the comparison video frames of the processed video frames in each video frame control group, the restored video frames of the processed video frames in each video frame control group are obtained, which effectively weakens the influence of rain and snow on the image clarity of the processed video frames in each video frame control group. Finally, according to the restored video frames of the processed video frames in all video frame control groups and the preset defect detection model, the defect detection results of the distribution lines are accurately obtained, so as to realize defect detection of distribution lines in rainy and snowy weather and improve the efficiency of defect detection.

[0148] Embodiment 10:

[0149] The present invention provides a drone, comprising: an integrated controller, a flight platform and a wireless transmission device; the integrated controller is used to obtain a fixed waypoint preset by an operator, and interact with the flight platform to send the fixed waypoint to the flight platform;

[0150] The flight platform is used to convert the fixed waypoints into corresponding flight paths and execute them;

[0151] The wireless transmission device is used to wirelessly transmit the current inspection video of the distribution line between the mobile control terminal and the drone.

[0152] In this embodiment, the fixed waypoint is a spatial position that is pre-set by the operator and that the drone needs to approximately pass through on its flight path (the closest distance between the drone and the fixed waypoint is less than 2 meters).

[0153] The beneficial effect of the above technology is: providing a drone that can perform efficient and high-accuracy detection of defects in distribution lines based on image analysis.

[0154] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention, and the present invention is also intended to include these changes and modifications.

Claims

1. A method for detecting defects in a power distribution line, characterized in that: include: S1: based on the current inspection video of the distribution line, obtaining all processed video frames of the current inspection video of the distribution line, and based on all processed video frames of the current inspection video of the distribution line, obtaining a comparison video frame of each processed video frame of the current inspection video of the distribution line; S2: based on all processed video frames of the current inspection video of the distribution line and the comparison video frames of all processed video frames, obtain all video frame comparison groups, and based on all video frame comparison groups, obtain all video frame control groups; S3: based on the processed video frames in each video frame control group, obtaining the grid standard deviation of the processed video frames in each video frame control group, based on the processed video frames in each video frame control group and the grid standard deviation of the corresponding processed video frames, obtaining the initial restored video frames of the processed video frames in each video frame control group, based on the initial restored video frames of the processed video frames in each video frame control group and the comparison video frames, obtaining the restored video frames of the processed video frames in each video frame control group; S4: obtaining a defect detection result of the distribution line based on the restored video frames of the processed video frames in all video frame control groups and a preset defect detection model; Wherein, based on the processed video frames in each video frame control group, obtaining the grid standard deviation of the processed video frames in each video frame control group includes: Divide the processed video frames in each video frame control group into a preset number of grids of equal size, and obtain all grids of the processed video frames in each video frame control group; The pixel extraction value is set to 0, and the number of pixel points in each grid of the processed video frame in each video frame control group whose pixel value is the same as the pixel extraction value is regarded as the extraction number of each grid of the processed video frame in each video frame control group; The standard deviation of the extraction numbers of all grids of the processed video frames in each video frame control group is regarded as the standard deviation of the processed video frames in each video frame control group; The pixel extraction value is set to 1, and the number of pixel points in each grid of the processed video frame in each video frame control group whose pixel value is the same as the pixel extraction value is regarded as the extraction number of each grid of the processed video frame in each video frame control group; The standard deviation of the number of extractions of all grids of the processed video frames in each video frame control group is also regarded as the standard deviation of the processed video frames in each video frame control group; Continue to traverse the pixel extraction value incrementally by unit value until the pixel extraction value reaches 255, and obtain all standard deviations of the processed video frames in each video frame control group; The standard deviation with the smallest value among all the standard deviations of the processed video frames in each video frame control group is regarded as the grid standard deviation of the processed video frames in each video frame control group; Wherein, based on the processed video frames in each video frame control group and the grid standard deviation of the corresponding processed video frames, obtaining the initial restored video frames of the processed video frames in each video frame control group includes: The pixel extraction value corresponding to the grid standard deviation of the processed video frame in each video frame control group is regarded as the rain and snow pixel value of the processed video frame in each video frame control group; The pixel points whose pixel values ​​are the same as the rain and snow pixel values ​​of the processed video frames in the corresponding video frame control group among all the pixel points of the processed video frames in each video frame control group are regarded as the processed pixel points of the processed video frames in each video frame control group, and all the pixel points except the processed pixel points among all the pixel points of the processed video frames in each video frame control group are regarded as the non-processed pixel points of the processed video frames in each video frame control group; The area consisting of all adjacent processed pixel points of the processed video frame in each video frame control group is regarded as the processed area of ​​the processed video frame in each video frame control group, and the area consisting of all non-processed pixel points of the processed video frame in each video frame control group is regarded as the non-processed area of ​​the processed video frame in each video frame control group; If the number of pixel points in each processing area of ​​the processed video frame in each video frame control group is less than the preset judgment number, the pixel values ​​of all pixel points in the corresponding processing area of ​​the processed video frame in the corresponding video frame control group are adjusted to be the same as the adjacent boundary pixel values ​​of the corresponding processing area of ​​the processed video frame in the corresponding video frame control group, so as to obtain the initial restored video frame of the processed video frame in each video frame control group; Wherein, based on the initial restored video frame and the comparison video frame of the processed video frame in each video frame control group, obtaining the restored video frame of the processed video frame in each video frame control group includes: A combined range of all processing regions of the processed video frames in each video frame control group, in which the number of pixels is not less than a preset judgment number, is used as a restoration region of an initial restoration video frame of the processed video frames in each video frame control group, and a combined range of all processing regions of the processed video frames in each video frame control group, in which the number of pixels is less than a preset judgment number, is used as a first non-restoration region of an initial restoration video frame of the processed video frames in each video frame control group; Using the non-processed region of the processed video frame in each video frame control group as the second non-restored region of the initial restored video frame of the processed video frame in each video frame control group; Based on the comparison video frame of the processed video frame in each video frame control group and the restored area, the first non-restored area and the second non-restored area of ​​the initial restored video frame, the replacement value of each pixel in the restored area of ​​the initial restored video frame of the processed video frame in each video frame control group is obtained, that is: Wherein, δ is the replacement value of the currently calculated pixel in the restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, ε1 is the pixel value of the currently calculated pixel in the restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, β1 is the mean value of the pixel values ​​of all pixels in the first non-restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, τ1 is the mean value of the pixel values ​​of all pixels in the second non-restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, ε2 is the mean value of the pixel values ​​of all pixels in the comparison video frame of the processed video frame in the currently calculated video frame control group, β2 is the minimum value of the pixel values ​​of all pixels in the first non-restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, τ2 is the minimum value of the pixel values ​​of all pixels in the second non-restored area of ​​the initial restored video frame of the processed video frame in the currently calculated video frame control group, ln is the natural logarithm, and the value of the natural constant e is 2.718; Obtaining a restored video frame of the processed video frame in each video frame control group based on a replacement value of each pixel point in a restored area of ​​the initial restored video frame of the processed video frame in each video frame control group; Wherein, based on the replacement value of each pixel point in the restoration area of ​​the initial restoration video frame of the processed video frame in each video frame control group, obtaining the restored video frame of the processed video frame in each video frame control group includes: If the replacement value of each pixel point in the restoration area of ​​the initial restored video frame of the processed video frame in each video frame control group is greater than the preset replacement value threshold, the pixel value of the corresponding pixel point in the restoration area of ​​the initial restored video frame of the processed video frame in the corresponding video frame control group is adjusted to be the same as the average value of the pixel values ​​of all pixels in the comparison video frame of the processed video frame in the corresponding video frame control group, so as to obtain the restored video frame of the processed video frame in each video frame control group.

2. The method for detecting defects in a power distribution line according to claim 1, characterized in that: S1: Based on the current inspection video of the distribution line, all processed video frames of the current inspection video of the distribution line are obtained, and based on all processed video frames of the current inspection video of the distribution line, a comparison video frame of each processed video frame of the current inspection video of the distribution line is obtained, including: Obtain all video frames of the current inspection video of the distribution line, and obtain the shooting position of each video frame of the current inspection video of the distribution line; If the distance between each preset shooting position and the shooting position of each video frame of the current inspection video of the power distribution line is less than the preset interval distance, the corresponding video frame of the current inspection video of the power distribution line is regarded as the processing video frame corresponding to the preset shooting position; All processed video frames of all preset shooting positions are used as processed video frames of the current inspection video of the distribution line; Based on all processed video frames of the current inspection video of the distribution line, a comparison video frame of each processed video frame of the current inspection video of the distribution line is obtained.

3. The method for detecting defects in a power distribution line according to claim 2, characterized in that: Based on all processed video frames of the current inspection video of the distribution line, a comparison video frame of each processed video frame of the current inspection video of the distribution line is obtained, including: Obtain all video frames of a preset standard inspection video of a power distribution line, and obtain a shooting position of each video frame of the preset standard inspection video of a power distribution line; The video frame of the preset standard inspection video that is closest to the shooting position of all video frames of the preset standard inspection video of the distribution line and the shooting position of each processed video frame of the current inspection video of the distribution line is used as the comparison video frame of the corresponding processed video frame of the current inspection video of the distribution line.

4. The method for detecting defects in a power distribution line according to claim 1, characterized in that: S2: Based on all processed video frames of the current inspection video of the distribution line and the comparison video frames of all processed video frames, all video frame comparison groups are obtained, and based on all video frame comparison groups, all video frame control groups are obtained, including: Each processed video frame of the current inspection video of the distribution line and the comparison video frame of the corresponding processed video frame are used as a video frame comparison group; Obtaining all similar video frame groups based on the comparison video frames of the processed video frames in all video frame comparison groups and the preset first image recognition model; Based on each similar video frame group and a preset second image recognition model, obtaining the number of line pixel areas of the comparison video frame of each processed video frame in each similar video frame group; The comparison video frame of the processed video frame with the largest number of line pixel areas in each similar video frame group and the processed video frame in the video frame comparison group corresponding to the corresponding comparison video frame are used as a video frame control group.

5. A defect detection system for a power distribution line, characterized in that: A method for detecting defects in a power distribution line according to any one of claims 1 to 4, comprising: A first processing module is used for S1: obtaining all processed video frames of the current inspection video of the distribution line based on the current inspection video of the distribution line, and obtaining a comparison video frame of each processed video frame of the current inspection video of the distribution line based on all processed video frames of the current inspection video of the distribution line; A second processing module is used to obtain all video frame comparison groups based on all processed video frames of the current inspection video of the distribution line and the comparison video frames of all processed video frames, and obtain all video frame control groups based on all video frame comparison groups; A restoration module, for obtaining a grid standard deviation of the processed video frames in each video frame control group based on the processed video frames in each video frame control group, obtaining an initial restored video frame of the processed video frames in each video frame control group based on the processed video frames in each video frame control group and the grid standard deviation of the corresponding processed video frames, and obtaining a restored video frame of the processed video frames in each video frame control group based on the initial restored video frame and the comparison video frame of the processed video frames in each video frame control group; The detection module is used to obtain the defect detection result of the distribution line based on the restored video frames of the processed video frames in all video frame control groups and the preset defect detection model.

6. A drone comprising: An integrated controller, a flying platform and a wireless transmission device; characterized in that the integrated controller is used to obtain fixed waypoints pre-set by the operator, interact with the flying platform, and send the fixed waypoints to the flying platform; The flight platform is used to convert the fixed waypoints into corresponding flight paths and execute them; The wireless transmission device is used to wirelessly transmit the current inspection video of the distribution line between the mobile control terminal and the drone.

Images