A visual orientation method for UAV distribution network tower inspection

Abstract

The present invention discloses a visual orientation method for unmanned aerial vehicle distribution network pole tower inspection, including a visual positioning method for the current pole tower orientation, and a direction correction method for the lower base pole tower to autonomously find the tower; the position of the pole tower is determined by adding an identification frame outside the insulator on the distribution network pole tower column photographed by the unmanned aerial vehicle; the current pole tower orientation is judged according to the direction of the connecting vertical line; an identification frame is added to the current pole tower, cross arm, and overhead wire image; the direction of automatic tower search is corrected during the adaptive inspection process; the position of the pole tower identification frame and the overhead wire identification frame are combined to assist the unmanned aerial vehicle to accurately confirm the position of the lower base pole tower and fly along the wire to the top of the lower base pole tower. The above method is used in combination, firstly the current pole tower orientation is located according to the insulator of the current pole tower, and then this is used as a reference to provide azimuth calibration for the lower base pole tower in combination with the cross arm and wire direction of the current pole tower, effectively improving the route accuracy of the unmanned aerial vehicle autonomous pole tower inspection.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle inspection of electric power systems, and in particular to a visual orientation method for unmanned aerial vehicle inspection of distribution network towers. Background Art

[0002] With the development of drone technology and the deepening of its business integration with the power industry, drones have shown advantages in power grid inspections, such as flexibility, comprehensive perspectives, and high efficiency, compared with manual inspections and manned helicopter inspections. In recent years, drones have been widely used in power grid inspections. The traditional drone distribution network inspection method is still similar to the transmission inspection type, using a fixed route inspection mode, which requires a large cost for route planning. However, with the continuous development of artificial intelligence technologies such as machine vision, new possibilities have been provided for the automation of distribution network inspections. At present, some drone-related units have begun to tackle the adaptive drone inspection technology based on machine vision.

[0003] However, there are many insurmountable obstacles in the process of UAV distribution network adaptive inspection. 1. It is difficult to orient the tower, and it is difficult to form an inspection route for key inspection equipment such as insulators; 2. The distribution network environment is complex, with many obstacles such as tree barriers and cross-lines, which pose a great safety hazard to the flight safety of drones; 3. The autonomous tower search technology is difficult, and there are many interference factors such as street lights, which places high requirements on the tower recognition model.

[0004] Therefore, how to solve the above problems, eliminate interference factors, improve the drone's positioning of the current tower direction, and correct the orientation of the next tower based on the current tower has become a technical problem that needs to be solved urgently by people in this field. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a visual orientation method for drone distribution network tower inspection. First, the current tower orientation is located according to the insulator of the current tower, and then this is used as a reference to provide azimuth calibration for the lower tower in conjunction with the crossarm and conductor direction of the current tower, thereby effectively improving the route accuracy of the drone's autonomous tower inspection.

[0006] A visual orientation method for UAV distribution network tower inspection, including a visual positioning method for the current tower orientation and a direction correction method for autonomous tower search of the lower tower;

[0007] In the visual positioning method of the current pole tower orientation, the position of the pole tower is determined by adding an identification frame outside the insulator on the distribution network tower column photographed by the drone; then the centers of multiple identification frames are connected, and invalid connections are removed based on the connection situation and the identification frame of the insulator; finally, the cross arm direction of the current pole tower is determined by the direction of the remaining insulator connection lines, and the current pole tower orientation is determined according to the direction of the vertical line of the connection lines;

[0008] In the direction correction method for autonomous tower search of the lower base tower, an identification frame is added to the image of the current tower, crossarm and overhead wire based on the recognition results of the crossarm, overhead wire and current tower orientation of the distribution network tower photographed by the drone; the position and size of the overhead wire identification frame and the crossarm identification frame are analyzed to correct the direction of automatic tower search during the adaptive inspection process; the position of the tower identification frame and the overhead wire identification frame are combined to assist the drone in accurately confirming the position of the lower base tower and flying along the wire to the top of the lower base tower.

[0009] Preferably, the visual positioning method of the current tower orientation comprises the following steps:

[0010] A1 controls the drone to hover above the tower and adjusts the gimbal angle to -90°;

[0011] A2 identifies all the insulators in the picture, obtains the insulator identification frame, and calculates the center point of the identification frame, that is, the average value of the coordinates of the upper left corner and the lower right corner;

[0012] A3 calculates the vectors of the lines connecting the center points of all insulators, and connects the lines from the center point of the upper leftmost insulator to other insulators in sequence. Insulators that have been connected by vectors will not be connected by reverse vectors.

[0013] A4 builds an insulator identification model on a distribution network pole, and traverses the relationship between all center point connection segments and other insulator identification frames one by one, so as to remove the connection lines in the insulator connection lines that do not intersect with other insulator identification frames;

[0014] The direction of the remaining line of A5 represents the direction of the tower cross arm, and the direction of the vertical line is the current tower direction.

[0015] Preferably, step A4 includes the following sub-steps:

[0016] A4.1 Calculate the center point A'(xA, yA) of the identification frame of insulator A, where xA = (xA1 + xA2) / 2, yA = (yA1 + yA2) / 2, and the center point C'(xC, yC) of the identification frame of insulator C, where xC = (xC1 + xC2) / 2, yC = (yC1 + yC2) / 2;

[0017] A4.2 Calculate the center point vector of insulator A and insulator C:

[0018] A4.3 Calculate the vectors of the two diagonals of the identification box of insulator B:

[0019]

[0020] A4.4 Calculate the vectors connecting the line segment A'C' and the four vertices of the identification box of insulator B

[0021]

[0022] A4.5 Calculate the cross product of the vector of line segment A'C' and the four vertices of the identification box of insulator B with the vector of line segment A'C';

[0023] The calculation results of A'C' and diagonal B1B2 are:

[0024]

[0025] The calculation results of C'A' and diagonal B3B4 are:

[0026]

[0027] A4.6 Calculate the intersection of the two diagonals of the identification frame of insulator B and line segment A'C' respectively;

[0028] If the calculation result of A'C' and diagonal line B1B2 and the calculation result of C'A' and diagonal line B3B4 are both less than 0, then the two diagonals of the identification frame of insulator B intersect with line segment A'C';

[0029] If the calculation result of A'C' and diagonal line B1B2 and the calculation result of C'A' and diagonal line B3B4 are not less than 0, the two diagonals of the identification frame of insulator B do not intersect with line segment A'C';

[0030] A4.7 Remove the connecting lines of the insulator that do not intersect with other insulator identification frames;

[0031] A4.8 Repeat selecting other connection lines and insulator identification boxes and executing steps A4.1-A4.7 until all insulators and connection lines have been traversed.

[0032] Preferably, during step A4, the lower left corner of the overall image of the current pole tower photographed by the drone is taken as the coordinate zero point, and the coordinates of the upper left corner and the lower right corner of multiple insulator identification frames are obtained based on the coordinate zero point.

[0033] Preferably, the method for correcting the direction of the lower base tower autonomously finding the tower comprises the following steps:

[0034] B1: Assume that the width of the image is 2a and the height is 2b; after obtaining the current tower orientation by the visual positioning method of the current tower orientation, adjust the angle of the drone to be consistent with the direction of the lower tower;

[0035] B2 identifies the current tower cross arm D, obtains the tower identification frame (xD1, yD1), (xD2, yD2); identifies the overhead wire L, obtains the wire identification frame (xL1, yL1), (xL2, yL2);

[0036] B3 calculates the width of the tower cross arm W = xD2-xD1;

[0037] The width of the B4 conductor is similar to the width of the tower crossarm. Calculate the distance Wx from the horizontal center of the conductor to the horizontal center of the screen;

[0038] If the center point of the identification frame of wire L is on the left side of the screen, that is, (xL1+xL2) / 2≤a, then Wx=xL1-a+W / 2;

[0039] If the center point of the identification frame of the wire L is on the right side of the screen, that is, (xL1+xL2) / 2>a, then Wx=xL2-aW / 2;

[0040] B5 calculates the angle between the horizontal center of the wire and the horizontal center line of the screen α = arctan (Wx / b);

[0041] B6 adjusts the drone angle according to the angle α. When α is positive, the drone is adjusted clockwise, and when α is negative, the drone is adjusted counterclockwise.

[0042] B7 To ensure the calibration result, repeat steps 1-6 until (xL1+xL2) / 2=a;

[0043] After the B8 direction adjustment is completed, adjust the gimbal angle to -30°, identify the lower base tower T, and obtain the tower identification frame (xT1, yT1), (xT2, yT2); when the overhead wire identification frame intersects with the lower base tower identification frame, that is, xL1≥xT1≥xT2≥xL2 is satisfied, or any of xL1≥xT1≥xL2 and xL1≥xT2≥xL2 is satisfied; it means that the current tower is the lower base tower, and then the drone can continue to fly in the current direction to the lower base tower position; if none of them are satisfied, it means that the correct lower base tower position is not identified in the current picture.

[0044] Preferably, in the direction correction method for autonomous tower search of the lower base tower, the lower left corner of the overall picture of the current tower taken by the drone is used as the coordinate zero point, and the coordinates of the upper left corner and lower right corner of the identification frame of the crossarm on the distribution network column, the overhead wire and the current tower are obtained based on the coordinate zero point.

[0045] Preferably, the drone is a multi-rotor drone equipped with a photographing device and an RTK differential positioning system.

[0046] The advantages and technical effects of the present invention are:

[0047] (1) The visual positioning method of the current pole tower orientation of the present invention can calculate the orientation of the pole tower and provide basic data for automatically generating inspection routes for distribution network pole towers. It can ensure that the drone can capture the distribution network pole tower and its key components at the best angle, improving the image quality and recognition accuracy. At the same time, it can ensure that the image data of each inspection is consistent and comparable, facilitating subsequent data analysis and identification of potential faults. This can reduce inspection time, improve overall inspection efficiency, and reduce inspection costs.

[0048] (2) The direction correction method for autonomous tower search of the lower base tower of the present invention can guide the UAV to complete the line-following flight. On the one hand, it can assist the UAV to complete the autonomous inspection. On the other hand, the line-following flight process can ensure that the UAV always flies above the line, avoiding the risk of interference or damage to the UAV caused by the conductor. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a flow chart of the visual positioning method of the current tower orientation of the present invention;

[0050] Figure 2 The present invention is a flow chart of the method for correcting the direction of the lower base tower autonomously finding the tower. DETAILED DESCRIPTION

[0051] In order to further understand the content, features and effects of the present invention, the following embodiments are given as examples and described in detail with reference to the accompanying drawings. It should be noted that the embodiments are illustrative, not restrictive, and the protection scope of the present invention cannot be limited thereby.

[0052] A visual orientation method for UAV distribution network tower inspection, including a visual positioning method for the current tower orientation and a direction correction method for autonomous tower search of the lower tower;

[0053] In the visual positioning method of the current pole tower orientation, the position of the pole tower is determined by adding an identification frame outside the insulator on the distribution network tower column photographed by the drone; then the centers of multiple identification frames are connected, and invalid connections are removed based on the connection situation and the identification frame of the insulator; finally, the cross arm direction of the current pole tower is determined by the direction of the remaining insulator connection lines, and the current pole tower orientation is determined according to the direction of the vertical line of the connection lines;

[0054] In the direction correction method for autonomous tower search of the lower base tower, an identification frame is added to the image of the current tower, crossarm and overhead wire based on the recognition results of the crossarm, overhead wire and current tower orientation of the distribution network tower photographed by the drone; the position and size of the overhead wire identification frame and the crossarm identification frame are analyzed to correct the direction of automatic tower search during the adaptive inspection process; the position of the tower identification frame and the overhead wire identification frame are combined to assist the drone in accurately confirming the position of the lower base tower and flying along the wire to the top of the lower base tower.

[0055] Preferably, the visual positioning method of the current tower orientation comprises the following steps:

[0056] A1 controls the drone to hover above the tower and adjusts the gimbal angle to -90°;

[0057] A2 identifies all the insulators in the picture, obtains the insulator identification frame, and calculates the center point of the identification frame, that is, the average value of the coordinates of the upper left corner and the lower right corner;

[0058] A3 calculates the vectors of the lines connecting the center points of all insulators, and connects the lines from the center point of the upper leftmost insulator to other insulators in sequence. Insulators that have been connected by vectors will not be connected by reverse vectors.

[0059] A4 builds an insulator identification model on a distribution network pole, and traverses the relationship between all center point connection segments and other insulator identification frames one by one, so as to remove the connection lines in the insulator connection lines that do not intersect with other insulator identification frames;

[0060] The direction of the remaining line of A5 represents the direction of the tower cross arm, and the direction of the vertical line is the current tower direction.

[0061] Preferably, step A4 includes the following sub-steps:

[0062] A4.1 Calculate the center point A'(xA, yA) of the identification frame of insulator A, where xA = (xA1 + xA2) / 2, yA = (yA1 + yA2) / 2, and the center point C'(xC, yC) of the identification frame of insulator C, where xC = (xC1 + xC2) / 2, yC = (yC1 + yC2) / 2;

[0063] A4.2 Calculate the center point vector of insulator A and insulator C:

[0064] A4.3 Calculate the vectors of the two diagonals of the identification box of insulator B:

[0065]

[0066] A4.4 Calculate the vectors connecting the line segment A'C' and the four vertices of the identification box of insulator B

[0067]

[0068] A4.5 Calculate the cross product of the vector of line segment A'C' and the four vertices of the identification box of insulator B with the vector of line segment A'C';

[0069] The calculation results of A'C' and diagonal B1B2 are:

[0070]

[0071] The calculation results of C'A' and diagonal B3B4 are:

[0072]

[0073] A4.6 Calculate the intersection of the two diagonals of the identification frame of insulator B and line segment A'C' respectively;

[0074] If the calculation result of A'C' and diagonal line B1B2 and the calculation result of C'A' and diagonal line B3B4 are both less than 0, then the two diagonals of the identification frame of insulator B intersect with line segment A'C';

[0075] If the calculation result of A'C' and diagonal line B1B2 and the calculation result of C'A' and diagonal line B3B4 are not less than 0, the two diagonals of the identification frame of insulator B do not intersect with line segment A'C';

[0076] A4.7 Remove the connecting lines of the insulator that do not intersect with other insulator identification frames;

[0077] A4.8 Repeat selecting other connection lines and insulator identification boxes and executing steps A4.1-A4.7 until all insulators and connection lines have been traversed.

[0078] Preferably, during step A4, the lower left corner of the overall image of the current pole tower photographed by the drone is taken as the coordinate zero point, and the coordinates of the upper left corner and the lower right corner of multiple insulator identification frames are obtained based on the coordinate zero point.

[0079] Preferably, the method for correcting the direction of the lower base tower autonomously finding the tower comprises the following steps:

[0080] B1: Assume that the width of the image is 2a and the height is 2b; after obtaining the current tower orientation by the visual positioning method of the current tower orientation, adjust the angle of the drone to be consistent with the direction of the lower tower;

[0081] B2 identifies the current tower cross arm D, obtains the tower identification frame (xD1, yD1), (xD2, yD2); identifies the overhead wire L, obtains the wire identification frame (xL1, yL1), (xL2, yL2);

[0082] B3 calculates the width of the tower cross arm W = xD2-xD1;

[0083] The width of the B4 conductor is similar to the width of the tower crossarm. Calculate the distance Wx from the horizontal center of the conductor to the horizontal center of the screen;

[0084] If the center point of the identification frame of wire L is on the left side of the screen, that is, (xL1+xL2) / 2≤a, then Wx=xL1-a+W / 2;

[0085] If the center point of the identification frame of the wire L is on the right side of the screen, that is, (xL1+xL2) / 2>a, then Wx=xL2-aW / 2;

[0086] B5 calculates the angle between the horizontal center of the wire and the horizontal center line of the screen α = arctan (Wx / b);

[0087] B6 adjusts the drone angle according to the angle α. When α is positive, the drone is adjusted clockwise, and when α is negative, the drone is adjusted counterclockwise.

[0088] B7 To ensure the calibration result, repeat steps 1-6 until (xL1+xL2) / 2=a;

[0089] After the B8 direction adjustment is completed, adjust the gimbal angle to -30°, identify the lower base tower T, and obtain the tower identification frame (xT1, yT1), (xT2, yT2); when the overhead wire identification frame intersects with the lower base tower identification frame, that is, xL1≥xT1≥xT2≥xL2 is satisfied, or any of xL1≥xT1≥xL2 and xL1≥xT2≥xL2 is satisfied; it means that the current tower is the lower base tower, and then the drone can continue to fly in the current direction to the lower base tower position; if none of them are satisfied, it means that the correct lower base tower position is not identified in the current picture.

[0090] Preferably, in the direction correction method for autonomous tower search of the lower base tower, the lower left corner of the overall picture of the current tower taken by the drone is used as the coordinate zero point, and the coordinates of the upper left corner and lower right corner of the identification frame of the crossarm on the distribution network column, the overhead wire and the current tower are obtained based on the coordinate zero point.

[0091] Preferably, the drone is a multi-rotor drone equipped with a photographing device and an RTK differential positioning system.

[0092] In addition, the present invention is preferred, and the first purpose of the present invention is to solve the problem of distribution network tower orientation during the UAV distribution network adaptive inspection process. This method can determine the distribution network tower orientation through the arrangement pattern of the distribution network tower insulators, thereby adjusting the tower inspection route, so that the planned route can accurately and effectively inspect the distribution network tower. At the same time, the direction of the next tower can be determined according to the tower orientation, providing a data support for the realization of autonomous tower search by UAVs in adaptive inspection. Figure 1 As shown, when the drone identifies the current direction of the tower, it identifies the insulators in the picture through the image transmission picture of the drone hovering directly above the tower, and determines whether the center of the identification frame intersects with the identification frame based on the arrangement pattern of the insulators to determine the direction of the tower.

[0093] In addition, the second purpose of the present invention is to solve the problem of unclear target identification and incorrect flight direction adjustment during the autonomous tower search of the drone. This method can correct the tower direction calculated in the first purpose of the invention through the identification results of the crossarms and wires of the distribution network tower, and support the drone to accurately confirm the position of the lower base tower and fly above the lower base tower. Figure 2 As shown, when the UAV identifies the position of the lower base pole tower, it determines whether the direction of the overhead wires is consistent with the UAV by identifying the pole tower crossarms, overhead wires and the pole tower body, thereby calibrating the direction of the lower base pole tower in the pole tower invention one, and assisting in determining the lower base pole tower from a complex identification environment, guiding the UAV to fly accurately to the lower base pole tower.

[0094] Finally, all the parts not described in the present invention adopt mature products and mature technical means in the prior art.

[0095] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all these improvements and changes should fall within the scope of protection of the appended claims of the present invention.

Claims

1. A visual orientation method for unmanned aerial vehicle distribution network tower inspection, characterized by: Including the visual positioning method of the current tower orientation, and the direction correction method of the lower tower autonomously finding the tower; In the visual positioning method of the current pole tower orientation, the position of the pole tower is determined by adding an identification frame outside the insulator on the distribution network tower column photographed by the drone; then the centers of multiple identification frames are connected, and invalid connections are removed based on the connection situation and the identification frame of the insulator; finally, the cross arm direction of the current pole tower is determined by the direction of the remaining insulator connection lines, and the current pole tower orientation is determined according to the direction of the vertical line of the connection lines; In the method for correcting the direction of the lower base tower autonomously finding the tower, an identification frame is added to the image of the current tower, cross arm and overhead wire according to the recognition result of the cross arm, overhead wire and the current tower direction of the distribution network tower photographed by the drone; Analyze the position and size of the overhead wire identification frame and the cross arm identification frame, and correct the direction of automatic tower search during the adaptive inspection process; combine the position of the tower identification frame and the overhead wire identification frame to assist the drone to accurately confirm the position of the lower base tower and fly along the wire to the top of the lower base tower; The visual positioning method of the current tower orientation comprises the following steps: A1 controls the drone to hover above the tower and adjusts the gimbal angle to -90°; A2 identifies all the insulators in the picture, obtains the insulator identification frame, and calculates the center point of the identification frame, that is, the average value of the coordinates of the upper left corner and the lower right corner; A3 calculates the vectors of the lines connecting the center points of all insulators, and connects the lines from the center point of the upper leftmost insulator to other insulators in sequence. Insulators that have been connected by vectors will not be connected by reverse vectors. A4 builds an insulator identification model on a distribution network pole, and traverses the relationship between all center point connection segments and other insulator identification frames one by one, so as to remove the connection lines in the insulator connection lines that do not intersect with other insulator identification frames; The direction of the remaining line of A5 represents the direction of the tower cross arm, and the direction of the vertical line of the line is the current tower direction; The step A4 comprises the following sub-steps: A4.1 Calculate the center point A'(xA, yA) of the identification frame of insulator A, where xA = (xA1 + xA2) / 2, yA = (yA1 + yA2) / 2, and the center point C'(xC, yC) of the identification frame of insulator C, where xC = (xC1 + xC2) / 2, yC = (yC1 + yC2) / 2; A4.2 Calculate the center point vector of insulator A and insulator C: A4.3 Calculate the vectors of the two diagonals of the identification box of insulator B: A4.4 Calculate the vectors connecting the line segment A'C' and the four vertices of the identification box of insulator B A4.5 Calculate the cross product of the vector of line segment A'C' and the four vertices of the identification box of insulator B with the vector of line segment A'C'; The calculation results of A'C' and diagonal B1B2 are: The calculation results of C'A' and diagonal B3B4 are: A4.6 Calculate the intersection of the two diagonals of the identification frame of insulator B and line segment A'C' respectively; If the calculation result of A'C' and diagonal line B1B2 and the calculation result of C'A' and diagonal line B3B4 are both less than 0, then the two diagonals of the identification frame of insulator B intersect with line segment A'C'; If the calculation result of A'C' and diagonal line B1B2 and the calculation result of C'A' and diagonal line B3B4 are not less than 0, the two diagonals of the identification frame of insulator B do not intersect with line segment A'C'; A4.7 Remove the connecting lines of the insulator that do not intersect with other insulator identification frames; A4.8 Repeat selecting other connection lines and insulator identification boxes and executing steps A4.1-A4.7 until all insulators and connection lines have been traversed.

2. The visual orientation method for unmanned aerial vehicle distribution network tower inspection according to claim 1 is characterized by: When performing step A4, the lower left corner of the overall image of the current pole tower photographed by the drone is used as the coordinate zero point, and the coordinates of the upper left corner and the lower right corner of multiple insulator identification frames are obtained based on the coordinate zero point.

3. The visual orientation method for unmanned aerial vehicle distribution network tower inspection according to claim 1 is characterized in that: The method for correcting the direction of the lower base tower autonomously finding the tower comprises the following steps: B1: Assume that the width of the image is 2a and the height is 2b; after obtaining the current tower orientation by the visual positioning method of the current tower orientation, adjust the angle of the drone to be consistent with the direction of the lower tower; B2 identifies the current tower cross arm D, obtains the tower identification frame (xD1, yD1), (xD2, yD2); identifies the overhead wire L, obtains the wire identification frame (xL1, yL1), (xL2, yL2); B3 calculates the width of the tower cross arm W = xD2-xD1; The width of the B4 conductor is similar to the width of the tower crossarm. Calculate the distance Wx from the horizontal center of the conductor to the horizontal center of the screen; If the center point of the identification frame of wire L is on the left side of the screen, that is, (xL1+xL2) / 2≤a, then Wx=xL1-a+W / 2; If the center point of the identification frame of the wire L is on the right side of the screen, that is, (xL1+xL2) / 2>a, then Wx=xL2-aW / 2; B5 calculates the angle between the horizontal center of the wire and the horizontal center line of the screen α = arctan (Wx / b); B6 adjusts the drone angle according to the angle α. When α is positive, the drone is adjusted clockwise, and when α is negative, the drone is adjusted counterclockwise. B7 To ensure the calibration result, repeat steps 1-6 until (xL1+xL2) / 2=a; After the B8 direction adjustment is completed, adjust the gimbal angle to -30°, identify the lower base tower T, and obtain the tower identification frame (xT1, yT1), (xT2, yT2); when the overhead wire identification frame intersects with the lower base tower identification frame, that is, xL1≥xT1≥xT2≥xL2 is satisfied, or any of xL1≥xT1≥xL2 and xL1≥xT2≥xL2 is satisfied; it means that the current tower is the lower base tower, and then the drone can continue to fly in the current direction to the lower base tower position; if none of them are satisfied, it means that the correct lower base tower position is not identified in the current picture.

4. The visual orientation method for unmanned aerial vehicle distribution network tower inspection according to claim 3 is characterized by: In the direction correction method for autonomous tower search of the lower base tower, the lower left corner of the overall picture of the current tower photographed by the drone is used as the coordinate zero point, and the coordinates of the upper left corner and lower right corner of the identification frame of the cross arm on the distribution network column, the overhead wire and the current tower are obtained based on the coordinate zero point.

5. The visual orientation method for unmanned aerial vehicle distribution network tower inspection according to claim 1 is characterized by: The drone is a multi-rotor drone equipped with a camera and an RTK differential positioning system.

Images