Distribution line defect identification early warning method and system based on unmanned aerial vehicle

Abstract

The invention discloses a distribution line defect identification early warning method and system based on an unmanned aerial vehicle, and solves the problems that whether a distribution line has defects or not cannot be detected and omnibearing image acquisition cannot be performed on the defective distribution line in the prior art, and the method comprises the steps that the unmanned aerial vehicle acquires images and videos along the distribution line; the processing module processes the image and transmits the processed image to the identification module; the recognition module adopts a deep learning recognition model to judge whether the image has defects, if yes, the communication module transmits a recognition result to a second system, the unmanned aerial vehicle executes an inspection action, and the communication module transmits the image and video collected by the inspection action to a second target address; if not, the communication module transmits the collected image and video to a first target address; the display module marks the number and position of the unmanned aerial vehicle acquiring the defect image, and the early warning module transmits early warning information. According to the invention, whether the distribution line has defects can be detected, and omnibearing image acquisition can be carried out on the defective distribution line.

Description

Technical Field

[0001] The present invention relates to the technical field of line inspection, and in particular to a method and system for identifying and warning defects in power distribution lines based on unmanned aerial vehicles. Background Art

[0002] Manual inspections will plan the power plant inspection route map based on the equipment distribution location and power plant equipment inspection points. Regardless of whether it is a simplified inspection, a comprehensive inspection, or a special inspection, the operation and maintenance personnel need to complete the inspection work of each equipment in sequence according to the planned inspection route map.

[0003] In the prior art, inspection robots are often used to conduct line inspections by walking along the ground wire on the distribution line. There are detection blind spots in the distribution line, and there is a problem that it is impossible to detect whether there are defects in the distribution line and to perform all-round image acquisition of the defective distribution line.

[0004] For example, the Chinese invention patent with announcement number CN107706871A, named Overhead Distribution Line, has two composite crossarms fixed on the same horizontal plane of the tension pole, which can meet the longitudinal tension required for tension resistance. Furthermore, by utilizing the composite nature of the composite crossarm body, the dry arc distance can be greatly increased, the electrical lightning resistance level can be improved, and accidents such as lightning strikes, line breaks, and flashovers can be avoided. The composite crossarm is maintenance-free and does not require regular inspection and maintenance, greatly reducing labor costs. The use of composite crossarms can also omit tension insulators and simplify the line structure. However, there is a problem of being unable to detect whether there are defects in the distribution line and to perform all-round image acquisition of the defective distribution line. Summary of the invention

[0005] In view of the deficiency that the existing technology cannot detect whether there are defects in distribution lines and perform all-round image acquisition of defective distribution lines, the present invention proposes a distribution line defect identification and early warning method and system based on drone, which can detect whether there are defects in distribution lines and perform all-round image acquisition of defective distribution lines.

[0006] The following is a technical solution of the present invention, a distribution line defect identification and early warning method based on a drone, the drone is connected to the second system, the drone includes a processing module, an identification module and a communication module, the second system includes a display module and an early warning module, and the method includes the following steps: S1: UAV collects images and videos along the power distribution line; S2: The processing module processes the image and transmits the processed image to the recognition module; S3: The recognition module uses a deep learning recognition model to determine whether there are defects in the image. If so, the communication module transmits the recognition result to the second system, the drone performs an inspection action, and the communication module transmits the images and videos collected by the inspection action to the second target address of the second system; if not, the communication module transmits the collected images and videos to the first target address of the second system; S4: The display module marks the number and location of the drone that obtained the defect image, and the warning module transmits the warning information.

[0007] In this solution, the drone collects images along the distribution line, the processing module processes the images, and transmits the processed images to the recognition module. The recognition module uses a deep learning recognition model to determine whether the image has defects, which can detect whether the distribution line has defects and improve the accuracy of detection. If there are defects, the drone performs an inspection action, and the communication module transmits the images and videos collected by the inspection action to the second target address. The inspection action collects images and videos in all directions, and classifies the defective images and videos by changing the storage address of the images and videos, which is convenient for checking the defective images and videos; if there are no defects, the communication module transmits the collected images and videos to the first target address. The display module marks the number and location of the drone that obtains the defective image, and the early warning module transmits early warning information, which can promptly remind the operation and maintenance personnel that there are defects in the distribution line.

[0008] Preferably, in S1, the UAV twin model is matched according to the UAV information, the distribution line twin model is matched according to the distribution line information, and the position of the UAV twin model in the distribution line twin model is updated according to the real-time position information of the UAV.

[0009] In this solution, digital twin technology is used to match the distribution line twin model according to the distribution line information, set the distribution line as the background, match the drone twin model according to the drone information, and update the position of the drone twin model in the distribution line twin model according to the real-time position information of the drone. The relative position relationship between the drone and the distribution line can be intuitively viewed.

[0010] Preferably, in S2, the processing module performs denoising, enhancement and segmentation processing on the image.

[0011] In this scheme, the image is preprocessed before deep learning to improve the accuracy of deep learning.

[0012] Preferably, the denoising process comprises the following steps: selecting a sliding window of a specified size, moving the sliding window pixel by pixel on the image, sorting the grayscale values ​​of the corresponding pixels by the sliding window and finding the middle value, and using the middle value as the grayscale value of the pixel where the sliding window is located; The enhancement process includes the following steps: using the Laplace operator to process the original image to generate a grayscale mutation image, superimposing the grayscale mutation image and the original image to generate a sharpened image, and performing a linear transformation process on the sharpened image to obtain an enhanced image; The segmentation process includes the following steps: an adaptive algorithm calculates the threshold of each pixel in the image, and the obtained threshold matrix is ​​used to realize binarization.

[0013] In this scheme, the image is denoised, enhanced and segmented before deep learning to improve the accuracy of deep learning.

[0014] Preferably, in S3, the drone performs an inspection action, including the following steps: S31: The UAV uses the first vertical distance to inspect the distribution line; S32: When there are defects in the image, the drone circles the transmission line at a first vertical distance and a first angle for several times to collect images and videos; S33: the drone ascends and descends to a second vertical distance, and circles the transmission line at a second angle for several times at the second vertical distance to collect images and videos; S34: The UAV is lifted to a first vertical distance and inspects along the distribution line.

[0015] In this solution, all-round image acquisition of the target detection points of the distribution line is carried out at different heights and angles, so that the defects of the distribution line can be intuitively checked.

[0016] Preferably, the first angle is between 30° and 60°, the first angle and the second angle are complementary to each other, and the second vertical height is greater than the first vertical height.

[0017] In this solution, all-round image acquisition of the target detection points of the distribution line is carried out at different heights and angles, so that the defects of the distribution line can be intuitively checked.

[0018] Preferably, drones in different states are rendered differently: the green drone twin model corresponds to a drone that is currently patrolling and has not yet performed an inspection action; the red drone twin model corresponds to a drone that is currently patrolling and is performing an inspection action; the orange drone twin model corresponds to a drone that is currently patrolling and has performed an inspection action; the gray drone twin model corresponds to a drone that has not been patrolled.

[0019] In this scheme, in the display module, the drone twin models are classified according to whether the drone has conducted an inspection and whether defects in the distribution lines have been found, so that the mission status of all drone twin models can be intuitively viewed.

[0020] A distribution line defect identification and early warning system based on an unmanned aerial vehicle comprises: an unmanned aerial vehicle, a first system and a second system, wherein the second system connects the unmanned aerial vehicle and the first system.

[0021] In this solution, it is possible to detect whether there are defects in the distribution line and to perform all-round image acquisition of the defective distribution line.

[0022] Preferably, the drone comprises: Main control module, used to control driving module, energy supply module, position module, lighting module, intercom module and communication module; A power supply module, used to supply power to the drone; Driving module, used to control the flight direction, flight altitude and flight trajectory of the drone; Position module, used to locate the real-time position information of the drone; Light module, used to provide lighting for the drone; Intercom module for real-time conversation; A camera module, used for collecting images of power distribution lines; Video recording module, used to collect distribution line video; A processing module, used for processing the images collected by the camera module; The recognition module is used to compare the image processed by the processing module with the standard image to determine the defects of the distribution line; The communication module is used to transmit the image captured by the camera module to the processing module, transmit the recognition result of the recognition module to the second system, and transmit the image captured by the camera module and the video captured by the video recording module to the second system.

[0023] In this solution, it is possible to detect whether there are defects in the distribution line and to perform all-round image acquisition of the defective distribution line.

[0024] Preferably, the first system includes a charging module, a point-changing module and a detection module. The detection module is signal-connected to the battery-changing module, and the battery-changing module is signal-connected to the charging module. In this solution, the first system is provided with a detection module, a charging module and a battery-changing module. The detection module is signal-connected to the battery-changing module, and the battery-changing module is signal-connected to the charging module. The detection module detects whether the drone is docked in place through a sensor. The battery-changing module is used to take out the energy supply module of the target drone, put the energy supply module into the charging module for charging, and take out another charged energy supply module and put it into the drone. It is convenient for the drone to automatically change the battery.

[0025] The second system includes a setting module, a rendering module, a display module and an early warning module. The setting module is connected to the rendering module, and the rendering module is connected to the display module. The setting module is used to set the inspection route and the collection trajectory. The inspection route of the drone is set according to the distribution line. Each drone inspects a specific distribution line, and each distribution line specifies several drones for inspection. The setting module is wirelessly connected to the drone through the communication module of the second system, and sends the inspection route and the collection trajectory to the main control module. The main control module controls the driving module to fly according to the inspection route and the collection trajectory. The setting module sets the drone information according to the drone model, and sets the distribution line information. The setting module sends the drone information and the distribution line information to the rendering module. The rendering module can be a cloud rendering server, which is used to match the drone twin model according to the drone information, match the distribution line twin model according to the distribution line information, and update the position of the drone twin model in the distribution line twin model according to the drone location information transmitted by the drone location module. The display module is used to display the drone twin model, the distribution line twin model and the video. The display module is connected to the drone via the communication module of the second system to receive images and videos transmitted from the drone.

[0026] The beneficial effect of the present invention is that by using a drone to detect whether there are defects in the distribution line and performing all-round image acquisition of the defective distribution line, the defect location can be obtained in real time, and an early warning can be issued in time to remind the operation and maintenance personnel of the defects in the distribution line. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 A schematic structural diagram of a distribution line defect identification and early warning system based on an unmanned aerial vehicle in the present invention.

[0028] Figure 2 The present invention is a flow chart of a distribution line defect identification and early warning system based on unmanned aerial vehicle.

[0029] In the figure, 1 is a drone; 2 is the first system; 3 is the second system; 101 is a main control module; 102 is a power supply module; 103 is a driving module; 104 is a position module; 105 is a lighting module; 106 is an intercom module; 107 is a camera module; 108 is a video recording module; 109 is a processing module; 110 is an identification module; 111 is a communication module; 201 is a charging module; 202 is a battery replacement module; 203 is a detection module; 301 is a setting module; 302 is a rendering module; 303 is a display module; 304 is an early warning module. DETAILED DESCRIPTION

[0030] The technical solution of the present invention is further specifically described below through embodiments and in conjunction with the accompanying drawings.

[0031] Example: This embodiment is a distribution line defect recognition and early warning system based on drones, such as Figure 1 As shown, it includes: a drone 1, a first system 2 and a second system 3, and the second system 3 is connected to the drone 1 and the first system 2 respectively.

[0032] The drone 1 is provided with a main control module 101, a power supply module 102, a driving module 103, a position module 104, a lighting module 105, an intercom module 106, a camera module 107, a video module 108, a processing module 109, an identification module 110 and a communication module 111. The driving module 103, the function module 102, the position module 104, the lighting module 105, the intercom module 106 and the communication module 111 are respectively connected to the main control module 101, the camera module 107 is connected to the processing module 109 through the communication module, and the camera module 107 and the video module 108 are connected to the second system through the communication module; the identification module 110 is connected to the processing module 109, and the identification module 110 is connected to the second system through the communication module. The main control module 101 is used to control the driving module 103, the energy supply module 102, the position module 104, the lighting module 105 and the communication module 111; the energy supply module 102 is used to supply power to the UAV 1; the driving module 103 controls the flight direction, flight altitude and flight trajectory of the UAV 1, and the driving module 103 is provided with a four-wing propeller for the launch and flight of the UAV 1, and the four-wing propeller is respectively arranged in the front, back, left and right directions of the UAV 1 to facilitate the stability of the UAV 1; the position module 104 is used to locate the real-time position information of the UAV 1 and transmit it to the second system 3; the lighting module 105 is used to provide lighting for the UAV 1, and the lighting module 105 is located directly above the camera module 107 and the video recording module 108; the intercom module 106 is used for real-time calls, and voice warnings and voice collection near the UAV 1 are performed through the intercom module 106; the camera module 107 can be a camera, which is used to collect images of the power distribution line and the surrounding environment; the video recording module 108 It can be a monitor or a camera for monitoring the distribution line and the surrounding environment; the processing module 109 is used to process the image collected by the camera module 107; the recognition module 110 is used to compare the image processed by the processing module 109 with the standard image, determine whether the collected image has defects and abnormalities, make a distribution line defect judgment, and obtain a recognition result, and the recognition result will also be added with the corresponding drone 1 number; the communication module 111 is used to transmit the image collected by the camera module 107 to the processing module 109, transmit the recognition result of the recognition module 110 to the second system 3, and transmit the image collected by the camera module 107 and the video collected by the video module 108 to the first target address in the second system 3, wherein, when the recognition result is abnormal, the image and video collected by the drone 1 during the inspection action are transmitted to the second target address in the second system 3.

[0033] The processing module 109 obtains the image captured by the camera module 107 and performs denoising on the image, including the following steps: selecting a sliding window of a specified size; moving the sliding window pixel by pixel on the image; sorting the grayscale values ​​of the corresponding pixels by the sliding window and finding the middle value; using the middle value as the grayscale value of the pixel where the sliding window is located. Performing image enhancement processing on the image includes the following steps: processing the original image with the Laplace operator to generate a grayscale mutation image; superimposing the grayscale mutation image and the original image to generate a sharpened image; performing linear transformation processing on the sharpened image to obtain an enhanced image. The expression of linear transformation is as follows: f ( i , j )= m '+(( n '- m ') / ( n - m ))( h ( i , j )- m ) In the formula, h ( i , j ) is an image, [ m , n ] is the grayscale range of the image, f ( i , j ) is the transformed image, [ m ', n '] is the grayscale range of the transformed image.

[0034] The enhanced image is segmented, including the following steps: an adaptive algorithm is used to calculate the threshold of each pixel in the image; and the obtained threshold matrix is ​​used to realize binarization. The processing module 109 processes the image and transmits the processed image to the recognition module 110 .

[0035] The recognition module 110 uses a deep learning recognition model for defect detection. The input layer of the deep learning recognition model is connected to the first convolution layer, the first convolution layer is connected to the first BN layer, the first BN layer is connected to the pooling layer, the pooling layer is connected to the second convolution layer, the second convolution layer is connected to the second BN layer, the second BN layer is connected to the second pooling layer, and the third convolution layer is obtained by Resnet and add processing. The first fully connected layer outputs 4096 neurons, and the neural network unit is temporarily discarded from the network with a probability of 0.5. The second fully connected layer outputs 1024 neurons, and the neural network unit is temporarily discarded from the network with a probability of 0.5. The final output category is calculated by the softmax function. The recognition result of the recognition module 110 is transmitted to the second system 3 through the communication module 111. The input layer size is 64×64×3, the first convolutional layer size is 64×3×3, the second convolutional layer size is 128×128×3, the third convolutional layer size is 128×128×3, the pooling layer size is 2×2, and the stride is 2. The training steps of each BN layer are as follows: solve the mean and variance of each training batch of data; use the mean and variance to standardize the batch of data, Get a 0-1 distribution; adjust the value by the scale factor, and then add an offset by the translation factor to get the output value.

[0036] The first system 2 is provided with a detection module 203, a charging module 201 and a battery exchange module 202. The detection module 203 is connected to the battery exchange module 202 by signal, and the battery exchange module 202 is connected to the charging module 201 by signal. The detection module 203 detects whether the drone 1 is docked in place through a sensor. The battery exchange module 202 is used to take out the energy supply module 102 of the target drone 1, put the energy supply module 102 into the charging module 201 for charging, and take out another charged energy supply module 102 and put it into the drone 1.

[0037] The second system 3 includes a setting module 301, a rendering module 302, a display module 303 and an early warning module 304. The setting module 301 is connected to the rendering module 302, and the rendering module 302 is connected to the display module 303.

[0038] The setting module 301 is used to set the inspection route and collection trajectory. The inspection route of the drone 1 is set according to the distribution line. Each drone 1 inspects a specific distribution line, and each distribution line specifies several drones 1 for inspection. The setting module 301 is wirelessly connected to the drone 1 through the communication module of the second system, and sends the inspection route and collection trajectory to the main control module 101. The main control module 101 controls the driving module to fly according to the inspection route and collection trajectory. The setting module 301 sets the drone information according to the drone model and the distribution line information. The setting module 301 sends the drone information and the distribution line information to the rendering module 302.

[0039] The rendering module 302 may be a cloud rendering server. The rendering module 302 is connected to the drone 1 through the communication module of the second system to obtain the location information of the drone 1 located by the location module 104. The rendering module 302 is used to match the twin model of the drone 1 according to the information of the drone 1, match the twin model of the distribution line according to the information of the distribution line, and update the position of the twin model of the drone 1 in the twin model of the distribution line according to the location information of the drone 1 transmitted by the location module 104 of the drone 1. That is, the position of the drone 1 in the distribution line can be intuitively checked through the visualized twin model.

[0040] The display module 303 is used to display the twin model of drone 1, the twin model of the power distribution line, and the video. The display module 303 is connected to the drone 1 through the communication module of the second system, and receives the images and videos transmitted by the drone 1. The display module 303 displays the positions of all drones 1 on the power distribution line through the twin model of drone 1 and the twin model of the power distribution line. The video is the video collected by the video recording module 108 of the specified drone 1, and the real-time inspection of the drone 1 is displayed through the video. Drones 1 in different states are rendered differently. The green twin model of drone 1 indicates that drone 1 is inspecting and has not performed the inspection action; the red twin model of drone 1 indicates that drone 1 is inspecting and performing the inspection action; the orange twin model of drone 1 indicates that drone 1 is inspecting and has performed the inspection action; the gray twin model of drone 1 indicates that drone 1 has not inspected. The display module 303 gives priority to displaying the video collected by the red twin model of drone 1 through the inspection action. Select a specific twin model of drone 1 to view the video collected by the drone 1 through the inspection action. The display module 303 includes a normal module and an abnormal module. The normal module stores images and videos obtained by the drone during normal inspection, i.e., without performing inspection actions. The abnormal module stores images and videos transmitted when the drone 1 is recognized as abnormal, and the display module highlights the number and location of the drone 1 in combination with the drone information and the drone location information. The second system 3 receives the usage of the first system 2, and the display module 303 displays it accordingly in a form.

[0041] The inspection action is the collection action performed by the drone 1 when the recognition result is abnormal. During normal inspection, the drone 1 inspects the distribution line at the first vertical distance. When the recognition result is abnormal, the drone 1 collects images and videos at the first vertical distance and at the first angle with the transmission line. After two weeks of circling, the drone 1 rises and falls to the second vertical distance, and collects images and videos at the second vertical distance and at the second angle with the transmission line. After two weeks of circling, the drone 1 rises and falls to the first vertical distance to inspect along the distribution line. The first angle is 30° to 60°, the first angle and the second angle are complementary, and the second vertical height is greater than the first vertical height.

[0042] The early warning module 304 is connected to the drone through the communication module of the second system, receives the recognition result sent by the drone 1 recognition module 110, and sends the drone 1 number and drone 1 position transmitted by the drone 1 due to the abnormal recognition result to the mobile phone or email of the operation and maintenance personnel. The early warning module 304 uses a horn or warning light for direct reminder.

[0043] Drone 1, first system 2, and second system 3 build a stable and reliable data return link to achieve real-time transmission of the video shot by drone 1 to second system 3. Through fast reconnection technology, transmission optimization is achieved and video transmission is quickly restored.

[0044] like Figure 2 As shown, a distribution line defect identification and early warning method based on drones includes the following steps: S1: UAV 1 collects images and videos along the power distribution line; S2: the processing module 109 processes the image and transmits the processed image to the recognition module 110; S3: The recognition module 110 uses a deep learning recognition model to determine whether there are defects in the image. If so, the communication module 111 transmits the recognition result to the second system 3, the drone 1 performs an inspection action, and the communication module 111 transmits the images and videos collected by the inspection action to the second target address of the second system; if not, the communication module 111 transmits the collected images and videos to the first target address of the second system; S4: The display module 303 marks the number and position of the drone 1 that obtains the defect image, and the warning module 304 transmits the warning information.

[0045] This method uses drones to detect whether there are defects in distribution lines and collects all-round images of defective distribution lines. It can obtain the defect location in real time and issue early warnings in time to remind operation and maintenance personnel of defects in distribution lines.

[0046] As a preferred solution of this embodiment, step S1 specifically includes: The setting module of the second system inputs the drone information and the distribution line information, sets the drone inspection route and collection trajectory according to the drone information and the distribution line information, and sends the inspection route and collection trajectory to drone 1 through the wireless network. The drone control driving module flies according to the inspection route and collection trajectory, and collects images and videos of the distribution line along the distribution line.

[0047] As a preferred solution of this embodiment, step S2 specifically includes: The image is subjected to denoising, enhancement and segmentation processing, and the processed image is subjected to recognition processing.

[0048] The denoising process includes: A sliding window of a specified size is selected, and the sliding window moves pixel by pixel on the image. The sliding window sorts the grayscale values ​​of the corresponding pixels and finds the middle value, and the middle value is used as the grayscale value of the pixel where the sliding window is located; Enhanced processing includes: The original image is processed with Laplace operator to generate a grayscale mutation image, the grayscale mutation image and the original image are superimposed to generate a sharpened image, and the sharpened image is subjected to linear transformation to obtain an enhanced image; The segmentation process includes: The adaptive algorithm calculates the threshold of each pixel in the image and uses the obtained threshold matrix to achieve binarization.

[0049] The processed image is sent to the recognition module through the communication module, and the recognition module determines the defects of the distribution line based on the image.

[0050] As a preferred solution of this embodiment, step S3 specifically includes: A deep learning recognition model is used to determine whether there are defects in the distribution line. The recognition process includes: The input layer of the deep learning recognition model is connected to the first convolutional layer, the first convolutional layer is connected to the first BN layer, the first BN layer is connected to the pooling layer, the pooling layer is connected to the second convolutional layer, the second convolutional layer is connected to the second BN layer, the second BN layer is connected to the second pooling layer, and the third convolutional layer is obtained by Resnet and add processing. The first fully connected layer outputs 4096 neurons, and the neural network unit is temporarily discarded from the network with a probability of 0.5. The second fully connected layer outputs 1024 neurons, and the neural network unit is temporarily discarded from the network with a probability of 0.5. The final output category is calculated by the softmax function.

[0051] The input layer size is 64×64×3, the first convolutional layer size is 64×3×3, the second convolutional layer size is 128×128×3, the third convolutional layer size is 128×128×3, the pooling layer size is 2×2, and the stride is 2. The training steps of each BN layer are as follows: solve the mean and variance of each training batch of data; use the mean and variance to standardize the batch of data, Get a 0-1 distribution; adjust the value by the scale factor, and then add an offset by the translation factor to get the output value.

[0052] After obtaining the recognition result, the drone transmits the recognition result to the second system. When a defect is identified, the drone is controlled to perform an inspection action, and the collected images and videos are transmitted and saved separately according to the defect situation. The drone performs the inspection action, which specifically includes the following steps: S31: The UAV uses the first vertical distance to inspect the distribution line; S32: When there are defects in the image, the drone circles the transmission line at a first vertical distance and a first angle for several times to collect images and videos; S33: the drone ascends and descends to a second vertical distance, and circles the transmission line at a second angle for several times at the second vertical distance to collect images and videos; S34: The UAV is lifted to a first vertical distance and inspects along the distribution line.

[0053] In the inspection action, the specific first angle is 30° to 60°, the first angle and the second angle are complementary to each other, assuming that the first angle is 30°, the second angle is 60°. In addition, the second vertical height is set to be greater than the first vertical height.

[0054] As a preferred solution of this embodiment, step S4 specifically includes: The second system receives the recognition result and obtains the corresponding drone number and location information. If a defect is identified, the drone number and location of the defect image are marked on the display module of the second system.

[0055] In addition, the display module of the second system uses the twin model of drones and the twin model of distribution lines, as well as the recognition information of drones on the captured images to display the positions of all drones on the distribution lines and the status of drones. Different renderings are performed for drones in different states. Specifically, the green twin model of drone 1 indicates that drone 1 is inspecting and has not performed inspection actions; the red twin model of drone 1 indicates that drone 1 is inspecting and performing inspection actions; the orange twin model of drone 1 indicates that drone 1 is inspecting and has performed inspection actions; the gray twin model of drone 1 indicates that drone 1 has not inspected. And the display module gives priority to displaying the video collected by the red twin model of drone through inspection actions.

[0056] The early warning module receives the recognition results sent by the drone. If it is judged that there are defects in the image, the early warning module sends the drone number and drone location transmitted by the drone due to the abnormal recognition result to the operation and maintenance personnel's mobile phone or email, so as to promptly remind the operation and maintenance personnel of the defects in the power distribution line. At the same time, the early warning module uses a horn or warning light for direct early warning reminders.

[0057] The specific embodiments described herein are merely examples of the spirit of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in similar ways, but they will not deviate from the spirit of the present invention or exceed the scope defined by the appended claims.

[0058] Although the terms such as drone, first system, second system, main control module, power supply module, etc. are frequently used in this article, the possibility of using other terms is not excluded. These terms are used only to more conveniently describe and explain the essence of the present invention; interpreting them as any additional restrictions is contrary to the spirit of the present invention.

Claims

1. A distribution line defect identification and early warning method based on drones, characterized in that: The following steps are involved: S1: UAV collects images and videos along the power distribution line; S2: The processing module processes the image and transmits the processed image to the recognition module; S3: The recognition module uses a deep learning recognition model to determine whether the image has defects. If so, the communication module transmits the recognition result to the second system, the drone performs an inspection action, and the communication module transmits the images and videos collected by the inspection action to the second target address of the second system; if not, the communication module transmits the collected images and videos to the first target address of the second system; S4: The display module marks the number and location of the drone that obtained the defect image, and the warning module transmits the warning information.

2. According to the method for identifying and warning power distribution line defects based on drones in claim 1, it is characterized in that: In step S1, the UAV twin model is matched according to the UAV information, the distribution line twin model is matched according to the distribution line information, and the position of the UAV twin model in the distribution line twin model is updated according to the real-time position information of the UAV.

3. The method for identifying and warning power line defects based on drones according to claim 1 is characterized in that: In step S2, the processing module performs denoising, enhancement and segmentation processing on the image.

4. The method for identifying and warning power line defects based on drones according to claim 3 is characterized in that: The denoising process includes the following steps: selecting a sliding window of a specified size, moving the sliding window pixel by pixel on the image, sorting the grayscale values ​​of the corresponding pixels and finding the middle value, and using the middle value as the grayscale value of the pixel where the sliding window is located; The enhancement process includes the following steps: using the Laplace operator to process the original image to generate a grayscale mutation image, superimposing the grayscale mutation image and the original image to generate a sharpened image, and performing a linear transformation process on the sharpened image to obtain an enhanced image; The segmentation process includes the following steps: an adaptive algorithm calculates the threshold of each pixel in the image, and the obtained threshold matrix is ​​used to realize binarization.

5. The method for identifying and warning power line defects based on drones according to claim 1 is characterized in that: In step S3, the drone performs an inspection operation, including the following steps: S31: The UAV uses the first vertical distance to inspect the distribution line; S32: When there are defects in the image, the drone circles the transmission line at a first vertical distance and a first angle for several times to collect images and videos; S33: the drone ascends and descends to a second vertical distance, and circles the transmission line at a second angle for several times at the second vertical distance to collect images and videos; S34: The UAV is lifted to a first vertical distance and inspects along the distribution line.

6. The method for identifying and warning power line defects based on drones according to claim 5 is characterized in that: The first angle is 30° to 60°, the first angle and the second angle are complementary to each other, and the second vertical height is greater than the first vertical height.

7. A distribution line defect identification and early warning method based on drone according to claim 1 or 2, characterized in that: Drones in different states are rendered differently: the green drone twin model corresponds to the drone being inspected but not performing inspection actions; the red drone twin model corresponds to the drone being inspected and performing inspection actions; The orange drone twin model corresponds to the drone that is currently inspecting and has performed inspection actions; The orange drone twin model corresponds to the drone that has not been inspected.

8. A distribution line defect identification and early warning system based on drones, implementing a distribution line defect identification and early warning method based on drones as claimed in any one of claims 1 to 7, characterized in that: include: The drone, the first system and the second system, wherein the second system is connected to the drone and the first system respectively.

9. The UAV-based distribution line defect identification and early warning system according to claim 8 is characterized in that: Drones include: Main control module, driving control module, energy supply module, position module, lighting module, intercom module and communication module; Energy supply module, which provides power for the drone; Driving module, controls the flight direction, flight altitude and flight trajectory of the drone; Position module, locates the real-time position information of the drone; Light module, providing lighting for the drone; Intercom module for real-time conversation; A camera module to collect images of power distribution lines; Video recording module, collecting distribution line video; A processing module, for processing the images collected by the camera module; The recognition module compares the image processed by the processing module with the standard image to determine the defects of the distribution line; The communication module transmits the image captured by the camera module to the processing module, transmits the recognition result of the recognition module to the second system, and transmits the image captured by the camera module and the video captured by the video recording module to the second system.

10. The UAV-based distribution line defect identification and early warning system according to claim 8 is characterized in that: The first system includes a charging module, a switching module and a detection module, the detection module is connected to the battery switching module by signal, and the battery switching module is connected to the charging module by signal; The second system includes a setting module, a rendering module, a display module and an early warning module. The setting module is connected to the rendering module, and the rendering module is connected to the display module.

Images