Power grid inspection method and system based on unmanned aerial vehicle

Abstract

The invention relates to the technical field of unmanned aerial vehicles, in particular to a power grid inspection method and system based on an unmanned aerial vehicle, and the method comprises the steps: obtaining a target tower and judging the type of the tower in combination with real-time visual detection and radar; locking the target pole tower and positioning in real time to determine a target position; if the target position is determined, entering a visual guidance flight following mode, otherwise, entering a radar line simulation flight mode until the target position is determined; before following of the current target pole tower is finished, different next pole tower searching strategies are executed according to the pole tower type, if the pole tower is a forked pole tower, a branch is manually selected, and if the pole tower is not the forked pole tower, locking of the current target pole tower is relieved when routing inspection crosses a preset position; and repeating the steps to continuously fly. According to the invention, the remote sensing advantage of vision and the short-distance detection advantage of radar can be deeply combined, the flight mode is adaptively switched, and the line patrol efficiency, accuracy and intelligent degree are improved.

Description

Technical Field

[0001] The present application relates to the technical field of unmanned aerial vehicles, and particularly to a power grid inspection method and system based on unmanned aerial vehicles. Background Art

[0002] Electricity is an important infrastructure for ensuring the development of the national economy and the normal operation of people's lives, and regular inspections are required to ensure the safety and stability of the power system. The transmission lines of the power grid can generally be roughly divided into main network lines composed of transmission lines of 110 kV and above and distribution network lines mainly composed of transmission lines of 35 kV and below. At present, for the main network lines, three-dimensional point cloud modeling of the channels has been basically realized nationwide, and the ability to carry out autonomous inspection by unmanned aerial vehicles has been achieved. However, the number of distribution network lines is huge, the amount of its transmission lines is more than 5 times that of the main network lines, and the line distribution form is more complex than that of the main network lines, and the line change frequency is also higher. Therefore, it is difficult for the distribution network lines to carry out three-dimensional point cloud modeling centrally through services like the main network lines, and the inspection difficulty is further increased. Compared with the main network lines, the distribution network lines are more directly related to the electricity consumption of the masses in production and life, directly affect the reputation of the power grid service, have high operation and maintenance requirements, and greater operation and maintenance pressure. Especially in mountainous areas, the distribution network has low poles, many tree obstacles, and complex environments, and the per capita operation and maintenance mileage is 3 times that of the main network lines, and the operation and maintenance force is even weaker.

[0003] To sum up, the distribution network lines have complex environments and a large number, which are difficult to cover by manual inspection. Relying on point cloud detection of poles has huge requirements for computing power and equipment costs, while relying solely on visual detection is prone to errors. Especially in the complex distribution network line environment, there are high operation risks, and at the same time, its technical difficulty is high, safety is low, and it is impossible to establish a ledger.

[0004] Therefore, how to achieve efficient, highly accurate and safe power grid inspection in complex environments has become an urgent problem to be solved. Summary of the Invention

[0005] The technical problem to be solved by the present application is how to achieve efficient, highly accurate and safe power grid inspection in complex environments.

[0006] One of the technical solutions adopted by the present application to solve its technical problems is: a power grid inspection method based on an unmanned aerial vehicle, characterized in that the method includes the following steps:

[0007] S1. Combine real-time visual detection and radar to obtain the target pole and judge the pole type;

[0008] S2. Determine the current target pole and the target position. If the current target pole is determined, the target position of the current target pole is located in real time and enter the visual guidance following flight mode, otherwise enter the radar line following flight mode until the current target pole and the target position are determined;

[0009] S3. Before the follow - up flight of the current target tower ends, different next - tower search strategies are executed according to the tower type;

[0010] If it is a bifurcated tower, reference data and / or operation instructions are obtained, and the next target tower is determined based on the obtained reference data and / or operation instructions;

[0011] If it is not a bifurcated tower, when the visual - guidance follow - up flight mode patrols across a preset position, the locking of the current target tower is released, and then the next target tower is searched in combination with radar wire information.

[0012] Furthermore, in step S1, obtaining the target tower by combining real - time visual detection and radar and judging the tower type specifically includes:

[0013] The tower to be detected is obtained through real - time visual detection and a tower detection frame is formed;

[0014] Towers with the number of pixels in the tower detection frame lower than the threshold t1 are excluded;

[0015] For each remaining tower after exclusion, the deviation angle Δθ between the yaw angle relative to the nose orientation and the current wire direction is obtained;

[0016] The target tower is obtained by combining the deviation angle Δθ and the coordinates of the tower detection frame in a preset order.

[0017] Furthermore, the obtaining the target tower by combining the deviation angle Δθ and the coordinates of the tower detection frame in a preset order specifically includes:

[0018] The Δθ of each tower is sorted in ascending order of magnitude, and the number Ntower of towers with the absolute value less than the threshold t2 is counted;

[0019] If Ntower = 1, the target tower is output;

[0020] If Ntower > 1, for the Ntower towers with the absolute value less than the threshold t2, they are sorted in ascending order of the ordinate of the center point of the tower detection frame, and the tower with the largest ordinate value is output as the target tower;

[0021] If Ntower = 0, the threshold t2 is enlarged, but not exceeding the maximum threshold t3, and Ntower is re - counted.

[0022] Furthermore, the deviation angle Δθ is obtained through the following relational expression:

[0023] where x is the horizontal - direction pixel deviation from the tower tip of the tower to the principal point of the image, f xis the horizontal focal length of the camera, θ g is the current yaw angle of the camera, θ pl is the angle between the wire extracted by the radar and the nose direction of the aircraft.

[0024] Furthermore, in step S1, combining real-time visual detection and radar to obtain the target tower and judge the tower type specifically further includes:

[0025] Obtain the current target tower of visual detection, and combine the tower wire direction segmented by the radar to judge the tower type;

[0026] If the current target tower of visual detection is a bifurcated tower, or the number of tower wire branches segmented by the radar is greater than 2, then the current target tower is a bifurcated tower;

[0027] Otherwise, the current target tower is not a bifurcated tower.

[0028] Furthermore, in step S3, before the end of following the current target tower, different next tower search strategies are executed according to the tower type, which specifically includes the following steps:

[0029] If it is a bifurcated tower, the telemetry shows the bifurcated branches, and an operator selects any one branch to obtain the reference direction data; based on the obtained reference direction data, control the pan-tilt to search for the next target tower.

[0030] If it is not a bifurcated tower, then when the visual guidance following flight mode patrols beyond the preset position, unlock the current target tower; then combine the radar wire information to control the pan-tilt camera to turn to the radar wire direction and search for the next target tower.

[0031] Furthermore, when the visual guidance following flight mode patrols beyond the preset position and unlocks the target tower, it specifically includes:

[0032] Obtain the camera attitude angle during the visual guidance following flight mode patrol, and the camera attitude angle includes at least one of the camera pitch angle, the camera yaw angle, and the camera roll angle;

[0033] When patrolling to the preset position, any one of the camera attitude angles reaches the specified angle, so that the target position of the target tower is located at the boundary of the camera field of view;

[0034] When patrolling beyond the preset position, the target tower is out of the camera field of view range to unlock the target tower.

[0035] Furthermore, when patrolling to the preset position, any one of the camera attitude angles reaches the specified angle, specifically includes:

[0036] Obtain the current pitch angle of the camera as p cThe target pitch angle at the avoidance preset position is p i The camera focal length is f;

[0037] When patrolling below the preset position;

[0038] Then the current pitch angle of the camera is p c Satisfy: Among them, is the upper half vertical field of view angle, and α is the vertical direction angle formed by the preset position and the camera optical axis in the camera coordinate system;

[0039] When patrolling above the preset position;

[0040] Then the current pitch angle of the camera is p c Satisfy: Among them, is the lower half vertical field of view angle, and α is the vertical direction angle formed by the preset position and the camera optical axis in the camera coordinate system.

[0041] The beneficial effect of the power grid inspection method based on the unmanned aerial vehicle in this application is as follows: This method deeply combines the long-distance perception advantage of vision and the close-range thin wire detection advantage of radar, adaptively switches the flight mode according to different scenarios, and controls the pan-tilt camera to intelligently lock, unlock, and search for the pole tower, improving the efficiency, accuracy, and intelligence level of line inspection. By combining real-time vision detection and radar technology, refined inspection can be achieved, and the status of equipment in complex power grids can be monitored and evaluated more carefully. At the same time, accurate data collection can be realized, providing reliable data support for subsequent analysis and decision-making. And it can adapt to complex distribution network environments, including mountainous areas and other areas with complex terrain, improving the inspection coverage. Using intelligent technology to switch the flight mode according to the determination of the target position improves the automation level during the inspection process, has low skill requirements for operators, and reduces the operation threshold. It can reduce the burden of manual operation, improve work efficiency, and at the same time reduce potential risks and improve the safety of operations. In addition, the equipment cost is low, which can reduce the investment cost.

[0042] One of the technical solutions adopted by this application to solve its technical problems is: It includes an unmanned aerial vehicle and a vision module, a radar module, a positioning module, and a control module arranged on the unmanned aerial vehicle,

[0043] The vision module and the radar module are used to combine real-time vision detection and radar to obtain the target pole tower and judge the type of the pole tower;

[0044] The positioning module is used to lock the target pole tower and perform real-time positioning to determine the target position;

[0045] The control module is used to switch the flight mode according to whether the positioning is successful or not according to the target position. If the target position is determined, the flight mode is entered into the visual guidance flight mode, otherwise the radar line-following flight mode is entered until the target position is determined.

[0046] The control module is also used to execute different next tower search strategies according to the tower type;

[0047] If it is a bifurcated pole tower, obtaining reference data and / or operation instructions, and determining the next target pole tower based on the obtained reference data and / or operation instructions;

[0048] If it is not a bifurcated pole tower, when the inspection passes the preset position in the vision-guided follow-up mode, the vision module releases the lock on the target pole tower; and then searches for the next target pole tower in combination with the radar wire information.

[0049] Furthermore, the vision module includes a camera and a vision sensor, and the vision sensor is used to obtain a camera attitude angle, and the camera attitude angle includes at least one of a camera pitch angle, a camera yaw angle, and a camera roll angle.

[0050] Furthermore, the camera attitude angle of the camera satisfies the following conditions:

[0051] When inspecting to a preset position, any one of the camera attitude angles reaches a specified angle, so that the target position of the target tower is located at the boundary of the camera field of view;

[0052] When the inspection passes over the preset position, the target pole tower exceeds the field of view of the camera so as to release the lock on the target pole tower.

[0053] Furthermore, the camera pitch angle satisfies the following conditions:

[0054] When the inspection reaches below the preset position;

[0055] The current pitch angle of the camera is p c satisfy: Among them, p c is the current pitch angle of the camera, p i is the pitch angle of the target to avoid the preset position, f is the focal length of the camera, is the upper half vertical field of view angle, α is the vertical angle between the preset position and the camera optical axis in the camera coordinate system;

[0056] When the inspection reaches above the preset position;

[0057] The current pitch angle of the camera is p c satisfy: Among them, p c is the current pitch angle of the camera, p iTo avoid the target pitch angle at the preset position, f is the camera focal length, is the lower half vertical field of view angle, and α is the vertical direction angle formed by the preset position and the camera optical axis in the camera coordinate system.

[0058] The beneficial effects of the power grid inspection system based on UAVs in this application are as follows: It can deeply combine the long-distance perception advantage of vision and the close-range thin wire detection advantage of radar, adaptively switch the flight mode according to different scenarios, and control the pan-tilt camera to intelligently lock, unlock, and search for the pole towers, improving the efficiency, accuracy, and intelligence level of line inspection. Brief Description of the Drawings

[0059] The following further describes this application with reference to the drawings and embodiments.

[0060] Figure 1 is the flowchart of the power grid inspection method based on UAVs in the embodiment of this application;

[0061] Figure 2 is the flowchart of the power grid inspection system based on UAVs in the embodiment of this application executing the power grid inspection method;

[0062] Figure 3 is the system framework diagram of the power grid inspection system based on UAVs in the embodiment of this application executing the power grid inspection method;

[0063] Figure 4 is the schematic diagram of pole tower detection in the embodiment of this application;

[0064] Figure 5 is the schematic diagram of controlling the pitch angle to avoid any point in the field of view when the pan-tilt camera is below the preset position in the embodiment of this application;

[0065] Figure 6 is the schematic diagram of controlling the pitch angle to avoid any point in the field of view when the pan-tilt camera is above the preset position in the embodiment of this application;

[0066] Figure 7 is the schematic diagram of the pan-tilt camera locking and unlocking the pole tower in the embodiment of this application;

[0067] Figure 8 is the schematic diagram of the vertical direction angle formed by the midpoint at the top of the pole tower and the camera optical axis in the embodiment of this application;

[0068] Figure 9 is the schematic diagram of a non-bifurcated pole tower in the embodiment of this application;

[0069] Figure 10 is the schematic diagram of a bifurcated pole tower in the embodiment of this application. Detailed Embodiments

[0070] In order to make the technical means, creative features, achieved purposes and effects realized by this application easy to understand, the following further elaborates this application in combination with specific implementation manners.

[0071] The technical solutions in the embodiments of this application will be clearly and completely described below. Obviously, the described embodiments are only one of the embodiments of this application, rather than all embodiments. Based on the embodiments in this application, other embodiments popularized by those of ordinary skill in the art all fall within the protection scope of this application.

[0072] Please refer to Figure 1 As shown, this application discloses a power grid inspection method based on an unmanned aerial vehicle. The method includes the following steps:

[0073] S1. Combine real-time visual detection and radar to obtain the target tower and judge the tower type;

[0074] S2. Determine the current target tower and the target position. If the current target tower is determined, the target position of the current target tower is located in real time and the visual guidance following flight mode is entered. Otherwise, the radar wire following flight mode is entered until the current target tower and the target position are determined;

[0075] S3. Before the end of following the current target tower, different next tower search strategies are executed according to the tower type;

[0076] If it is a bifurcated tower, reference data and / or operation instructions are obtained, and the next target tower is determined based on the obtained reference data and / or operation instructions. Among them, the reference data can specifically be the bifurcated branch shown by telemetry, and the operation instructions can be manual operation instructions or operation information preset by a computer. For example, by manually selecting any branch, reference direction data is obtained, and the next target tower is searched based on the obtained reference direction data;

[0077] If it is not a bifurcated tower, when the inspection reaches a preset position, the locking of the current target tower is released, and then the next target tower is searched in combination with the radar wire information. Among them, the preset position can specifically be an unlocking point or an avoidance point;

[0078] S4. Repeat steps S1 to S3, and the unmanned aerial vehicle flies continuously.

[0079] In step S1, as Figure 4As shown in the figure, the pole tower can be detected in real time through the vision module. The detected pole towers can be classified into target pole towers, side pole towers, and distant pole towers according to different preset conditions. Then, by combining the yaw angle of the pole tower in the camera coordinate system with the yaw angle of the wire segmented by the radar module, the target pole tower, that is, the pole tower that needs to be followed currently, is screened out from the detected numerous pole towers, and then the target pole tower is tracked. Among them, the vision module can include a pole tower detection and tracking module, and the radar module can include a radar wire extraction module. In addition, the drone includes a flight controller, and the vision module can also include a pan-tilt camera module.

[0080] Specifically, the target pole tower can be the pole tower that is screened out as the target pole tower by being the one with the most consistent direction with the radar wire and the closest to the drone (the ordinate value of the center point of the pole tower detection frame is the largest).

[0081] As Figure 2 and Figure 3 As shown in the figure, the power grid inspection system based on the drone disclosed in the present application can be used to execute the power grid inspection method based on the drone as described above. The power grid inspection system based on the drone includes a drone and a vision module, a radar module, a positioning module, and a control module arranged on the drone.

[0082] The vision module and the radar module are used to combine real-time vision detection and radar to obtain the target pole tower and / or the pole tower type.

[0083] The positioning module is used to lock the target pole tower and perform real-time positioning to determine the target position.

[0084] The control module is used to switch the flight mode according to whether the positioning is successful based on the target position. If the target position is determined, it enters the vision-guided following flight mode; otherwise, it enters the radar wire-following flight mode until the target position is determined.

[0085] The control module is also used to execute different next pole tower search strategies according to the pole tower type.

[0086] If it is a bifurcated pole tower, reference data and / or operation instructions are obtained, and the next target pole tower is determined based on the obtained reference data and / or operation instructions.

[0087] If it is not a bifurcated pole tower, when the inspection in the vision-guided following flight mode crosses the preset position, the vision module releases the lock on the target pole tower; then the next target pole tower is searched in combination with the radar wire information.

[0088] When the drone executes the drone-based power grid inspection method as described above, pre-flight preparations are required, such as setting the HOME point, that is, the landing point after the task is completed. Subsequently, the function is enabled. In the semi-automatic mode, manually control the drone to hover a few meters directly above the wire, with the nose facing the first target tower. Then switch to the automatic mode, enable the intelligent line inspection function, and the pan-tilt camera enters the default posture.

[0089] In some further embodiments, in step S1, real-time visual detection and radar are combined to obtain the target tower and judge the tower type, such as Figure 4 As shown, the tower can be detected in real time by the vision module. The detected towers can be classified into target towers, side towers, and distant towers according to different preset conditions. The tower detection and tracking module detects the towers in real time, and then combines the yaw angle of the tower in the camera coordinate system with the yaw angle of the wire segmented by the radar wire extraction module to screen out the target tower from the numerous detected towers, that is, the tower that needs to be followed currently, and then tracks the target tower. Specifically, it includes:

[0090] Obtain the towers to be detected through real-time visual detection and form tower detection frames;

[0091] Exclude the towers with the number of pixels in the tower detection frame lower than the threshold t1 to eliminate too far towers or possible misdetection results;

[0092] Obtain the deviation angle Δθ between the yaw angle of each remaining tower relative to the nose direction and the current wire direction after exclusion; where the nose direction refers to the direction pointed by the drone's nose, and the wire direction is the extension direction of the current wire;

[0093] Combine the deviation angle Δθ and the coordinates of the tower detection frame to obtain the target tower in a preset order.

[0094] In some further embodiments, exclude the side towers and screen out the towers located in the main channel direction: all towers are sorted in ascending order according to Δθ. The smaller the value, the smaller the deviation between the tower and the wire direction extracted by the radar wire extraction module, that is, the greater the possibility that the tower is connected to the wire, and the greater the probability that the corresponding tower is the target tower. Subsequently, exclude the distant towers located in the main channel direction and screen the target tower. The nearest tower can be screened out from the towers located in the main channel direction according to the ordinate value of the center point of the tower detection frame.

[0095] In some further embodiments, the combining the deviation angle Δθ and the coordinates of the tower detection frame to obtain the target tower in a preset order specifically includes:

[0096] Sort the Δθ of each tower in ascending order of magnitude, and count the number Ntower of towers with the absolute value less than the threshold t2;

[0097] If Ntower = 1, output the target tower.

[0098] If Ntower > 1, then for the Ntower towers whose absolute value is less than the threshold t2, sort them in ascending order according to the vertical coordinate value of the center point of the tower detection box, and output the tower with the first pixel count as the target tower.

[0099] If Ntower = 0, then increase the threshold t2, but not exceeding the maximum threshold t3, and re - count Ntower.

[0100] Among them, in some embodiments, the determination of the threshold t1 can be: count the number of area pixels of the tower detection box manually marked, and obtain the minimum value of the number of area pixels of the tower detection box manually marked, and use this minimum value as the threshold t1. In some specific embodiments, taking an image with 1920*1080 image pixels as an example, the minimum value of the number of area pixels of the tower detection box manually marked is 2500, and this minimum value is set as the threshold t1 to be 2500, which is used to filter out the detection boxes with an area less than 2500.

[0101] Among them, in some embodiments, the threshold t2 can be set to the half horizontal field of view angle of the UAV camera / 3, and the threshold t2 can be set to the half horizontal field of view angle of the UAV camera / 2.

[0102] In some further embodiments, the deviation angle Δθ is obtained through the following relational expression:

[0103] Among them, x is the horizontal pixel deviation from the tip of the tower to the principal point of the image, f x is the horizontal focal length of the camera, θ g is the current yaw angle of the camera, θ pl is the angle between the wire extracted by the radar and the nose direction.

[0104] In step S1, combining real - time visual detection and radar to obtain the target tower and judge the tower type, specifically further includes:

[0105] Obtain the current target tower of visual detection, and combine the tower wire direction segmented by the radar to judge the tower type;

[0106] If the current target tower type detected visually is a bifurcated tower, or the number of tower wire branches segmented by the radar is greater than 2, then the current target tower is a bifurcated tower; for example, as Figure 10 shown, when the UAV is following the target tower 2, if it is detected visually as bifurcated, or the number of its wire branches segmented by the radar is 3, then the second tower is a bifurcated tower.

[0107] Otherwise, the current target pole tower is not a bifurcated pole tower.

[0108] In step S2, after the pole tower detection and tracking module filters out the target pole tower, the pixel coordinates of the midpoint at the top of the pole tower are calculated through the detection frame of the target pole tower. Then, the pan-tilt camera is controlled to lock the midpoint at the top of the target pole tower, and at the same time, the target pole tower is sent to the pole tower positioning module for target pole tower positioning and corresponding target waypoint sending.

[0109] In step S2, if the current target pole tower and the target position are determined, enter the visual guidance following flight mode, and at the same time, the yaw angle and altitude of the UAV can be corrected through the radar module. In the visual guidance following flight mode, after the UAV flight controller receives the target waypoint information, it controls the UAV to autonomously follow the target pole tower. At the same time, the flight controller receives the yaw and altitude deviation correction amounts from the radar wire extraction module, and real-time corrects the flight yaw angle and flight altitude of the UAV relative to the wire, and controls the UAV to fly within the specified horizontal and vertical ranges of the wire during the process of following the target pole tower.

[0110] If the determination of the target position fails, enter the radar wire imitation flight mode until the target position is determined to obtain the target waypoint.

[0111] For example, when encountering situations such as two pole towers being far apart beyond the visual perception range, or the pole tower being blocked by houses or trees, resulting in positioning failure, enter the radar wire imitation flight mode; when the wire is blocked and other reasons cause the radar wire imitation flight mode to not be entered normally, switch to the visual guidance following flight mode.

[0112] In some further embodiments, before the end of following the current target pole tower in step S3, different next pole tower search strategies are executed according to the pole tower type, specifically including:

[0113] If it is a bifurcated pole tower, the telemetry shows the bifurcated branch, and an operator selects any branch to obtain the reference direction data; based on the obtained reference direction data, the pan-tilt is controlled to search for the next target pole tower.

[0114] For example, if encountering non-bifurcated pole towers such as Figure 9 shown straight lines, broken lines, etc., after the pan-tilt camera is unlocked, control the pan-tilt camera to turn to the wire yaw direction extracted by the radar wire extraction module, search for the next pole tower on the same distribution network line, and then repeat the above steps 3-6. The pole tower detection and tracking module detects, tracks and filters the next target pole tower in real time, and the positioning module locates the position of the target pole tower. After the UAV finishes following the current target pole tower, the UAV automatically follows the next target pole tower. By using this method, the position of the next target pole tower can be located before the UAV finishes following the current pole tower. After the UAV finishes following the current target pole tower, the UAV can fly to the next target pole tower without decelerating.

[0115] For example, as Figure 10 shown, when the UAV is following the second tower, if visual detection shows it is a bifurcation, or the radar divides the number of wire branches into 3, then the second tower is a bifurcated tower. Telemetry shows the bifurcated branch, and any branch is manually selected to obtain reference direction data. Then the pan-tilt camera automatically looks up in the direction of the next tower to be followed for search. At the same time, the tower detection and tracking module continuously detects and tracks the next target tower, and the positioning module locates the position of the target tower. After the UAV finishes following the current target tower, the UAV automatically follows the next target tower in the direction of the bifurcated channel.

[0116] If it is not a bifurcated tower, when patrolling and passing the preset position in the visual guidance following flight mode, the locking of the current target tower is released; then, combining the radar wire information, the pan-tilt camera is controlled to turn to the radar wire direction to search for the next target tower.

[0117] In some further embodiments, when in the visual guidance following flight mode and passing the preset position, if it is not a bifurcated tower, releasing the locking of the target tower specifically includes:

[0118] Obtaining the camera attitude angle during the visual guidance following flight mode patrol, where the camera attitude angle includes at least one of the camera pitch angle, camera yaw angle, and camera roll angle;

[0119] When patrolling to the preset position, any one of the camera attitude angles reaches a specified angle, so that the target position of the target tower is located at the boundary of the camera field of view; where the target position can specifically be the midpoint of the tower top, or any other preset point position on the tower that can be obtained.

[0120] When patrolling beyond the preset position, the target tower is out of the camera field of view range to release the locking of the target tower.

[0121] In some further embodiments, when patrolling to the preset position, any one of the camera attitude angles reaching a specified angle specifically includes:

[0122] Obtaining the current camera pitch angle as p c , the target pitch angle for avoiding the preset position as p i , and the camera focal length as f;

[0123] As Figure 5 shown, when patrolling below the preset position;

[0124] then the current camera pitch angle (pitch) is p c satisfies: where is the upper half vertical field of view angle, and α is the vertical direction angle formed by the preset position and the camera optical axis in the camera coordinate system;

[0125] As shown Figure 6 when patrolling above the preset position;

[0126] the current pitch angle of the camera is p c satisfying: wherein, is the lower half vertical field of view angle, and α is the vertical direction angle formed by the preset position and the optical axis of the camera in the camera coordinate system.

[0127] In step S3, as the distance between the UAV and the pole tower shortens, at a certain critical moment, at the specified pitch angle, the midpoint of the top of the pole tower is exactly located at the boundary of the lower half vertical field of view of the pan-tilt camera and then goes out of the field of view. In the case of non-bifurcated pole towers, unlocking too early will result in still being able to see the current pole tower after unlocking, which is likely to cause changes in the tracking number of the same pole tower, problems such as repeated positioning and affecting positioning accuracy, and is not conducive to subsequent following flight. Unlocking too late will result in late detection, tracking, positioning, and following flight of the next pole tower, which may cause the UAV to decelerate.

[0128] In order to control the unlocking of the pan-tilt camera when it just can't see the currently following pole tower, according to the above method of controlling the rotation of the pan-tilt camera to avoid any point in the field of view, when the pan-tilt angle p of the pan-tilt camera c satisfies the following relational expression, the pan-tilt camera unlocks and looks up to the specified pitch angle p i and no longer locks the pole tower:

[0129]

[0130] This method is not affected by factors such as the flight altitude of the UAV, uphill and downhill of the distribution network channel, etc., and avoids premature or late unlocking of the pan-tilt camera.

[0131] In some further embodiments, in step S4, steps S1 to S3 are repeated. After the UAV finishes following the current target pole tower, the UAV automatically follows the next target pole tower.

[0132] Using this method, the position of the next target pole tower can be located before the UAV finishes following the current pole tower. After the UAV finishes following the current target pole tower, the UAV can fly to the next target pole tower without decelerating.

[0133] Repeat the above steps. The UAV conducts continuous line patrol flight. Before the UAV reaches the current target tower being followed, control the camera to search for the next target tower, and then repeat the above steps to detect, track, and locate the next target tower. If the positioning is successful, after the UAV finishes following the current target tower, enter the visual guidance following mode, and at the same time, the radar corrects the yaw angle and altitude of the UAV. If the positioning fails, after the UAV finishes following the current target tower, enter the radar line-following flight mode until the target waypoint is successfully located, and then switch to the visual guidance following mode. Realize the continuous line patrol flight of the UAV.

[0134] The ground control system displays the telemetry data in real time for ground monitoring and remote control. During the intelligent line patrol flight of the UAV, the ground control software displays videos, tower types, etc. in real time for ground monitoring and remote control. For example, when encountering a bifurcated tower, an operator selects the direction of the bifurcated channel to be inspected through the ground control software.

[0135] When the mission ends, turn off the function and return automatically. After completing intelligent line patrol tasks such as automatic collection of power distribution point clouds, turn off the intelligent line patrol function switch, and click to return to the HOME point and then land automatically.

[0136] This application also provides a power grid inspection system based on a UAV, including a UAV and a vision module, a radar module, a positioning module, and a control module provided on the UAV;

[0137] The vision module and the radar module are used to combine real-time vision detection and radar to obtain the target tower and determine the type of the current target tower;

[0138] The positioning module is used to lock the target tower and perform real-time positioning to determine the target position. Specifically, the positioning module can locate the position of the target tower according to information such as pixel deviation and camera parameters;

[0139] The control module is used to switch the flight mode according to whether the target position is determined. If the target position is determined, enter the visual guidance following mode; if the determination of the target position fails, enter the radar line-following flight mode until the target position is determined; wherein, the visual guidance following mode includes the UAV flying following the waypoint, and at the same time, using the radar to correct the flight yaw angle and flight altitude of the UAV relative to the wire in real time; the radar line-following flight mode includes using the radar to control the flight yaw angle and flight altitude of the UAV relative to the wire in real time;

[0140] The control module is also used to execute different search strategies for the next tower according to the tower type;

[0141] If it is a bifurcated tower, the telemetry shows the bifurcated branches, and an operator selects any branch to obtain the reference direction data; based on the obtained reference direction data, control the pan-tilt to search for the next target tower.

[0142] If it is not a forked pole tower, when the visually guided follow-up mode is used to inspect over the preset position, the lock on the current target pole tower is released; then, combined with the radar wire information, the gimbal camera is controlled to turn to the direction of the radar wire to search for the next target pole tower.

[0143] In some further embodiments, the vision module includes a camera and a vision sensor, the vision sensor is used to obtain a camera attitude angle, and the camera attitude angle includes at least one of a camera pitch angle, a camera yaw angle, and a camera roll angle.

[0144] In some further embodiments, the camera attitude angle of the camera satisfies the following conditions:

[0145] When inspecting to a preset position, any one of the camera attitude angles reaches a specified angle, so that the target position of the target tower is located at the boundary of the camera field of view;

[0146] When the inspection passes over the preset position, the target pole tower exceeds the field of view of the camera so as to release the lock on the target pole tower.

[0147] In some further embodiments, the camera pitch angle satisfies the following conditions:

[0148] When the inspection reaches below the preset position;

[0149] The current pitch angle of the camera is p c satisfy: Among them, p c is the current pitch angle of the camera, p i is the pitch angle of the target to avoid the preset position, f is the focal length of the camera, is the upper half vertical field of view angle, α is the vertical angle between the preset position and the camera optical axis in the camera coordinate system;

[0150] When the inspection reaches above the preset position;

[0151] The current pitch angle of the camera is p c satisfy: Among them, p c is the current pitch angle of the camera, p i is the pitch angle of the target to avoid the preset position, f is the focal length of the camera, is the lower half vertical field of view angle, and α is the vertical angle formed by the preset position in the camera coordinate system and the camera optical axis.

[0152] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium provided in this application and used in the embodiments can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0153] It should be noted that in this document, the terms "include", "comprise", or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that includes a series of elements includes not only those elements but also other elements not expressly listed, or elements that are inherent to such process, apparatus, article, or method. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, apparatus, article, or method that includes such element.

[0154] The foregoing has shown and described the basic principles, main features, and advantages of the present application. Those skilled in the art should understand that the present application is not limited by the above embodiments. The above embodiments and the descriptions in the specification only illustrate the principles of the present application. Without departing from the spirit and scope of the present application, the present application will have various changes and improvements, and all such changes and improvements fall within the scope claimed by the present application. The scope claimed by the present application is defined by the appended claims and their equivalents.

Claims

1. A power grid inspection method based on drones, characterized in that: The method comprises: S1. Combine real-time visual detection and radar to obtain the target tower and determine the tower type; S2, judging whether the current target pole tower is determined, if the current target pole tower is determined, locating the target position of the current target pole tower in real time and entering the visual guidance following flight mode, otherwise entering the radar line-following flight mode until the current target pole tower and the target position are determined; S3. Before the current target tower is finished following the flight, different next tower search strategies are executed according to the tower type: If it is a bifurcated pole tower, obtaining reference data and / or operation instructions, and determining the next target pole tower based on the obtained reference data and / or operation instructions; If it is not a bifurcated pole tower, when the preset position is passed in the visually guided follow-up mode inspection, the lock on the current target pole tower is released, and then the next target pole tower is searched for in combination with the radar wire information.

2. The method for power grid inspection based on drones according to claim 1, characterized in that: In step S1, the target tower is acquired and the tower type is determined by combining real-time visual detection and radar. Specifically, the following steps are performed: Acquire the pole tower to be inspected through real-time visual inspection and form a pole tower detection frame; Exclude the towers whose detection frame pixel count is lower than the threshold t1; Obtain the deviation angle Δθ between the yaw angle of each tower remaining after exclusion relative to the nose direction and the current wire direction; The target tower is acquired in a preset order by combining the deviation angle Δθ with the coordinates of the tower detection frame.

3. The method for power grid inspection based on drone according to claim 2, characterized in that: The step of combining the deviation angle Δθ with the coordinates of the tower detection frame to obtain the target tower in a preset order specifically includes: Sort the Δθ of each tower in order of size, and count the number of towers N whose absolute value of Δθ is less than the threshold t2 tower ; If N tower =1, output the target tower; If N tower >1, then the absolute value of Δθ is less than the threshold value t2 for N tower The towers are sorted in order of size according to the vertical coordinates of the center points of the tower detection frames, and the tower with the largest vertical coordinate value is output as the target tower; If N tower = 0, then enlarge the threshold t2, but not exceeding the maximum threshold t3, and re-count N tower .

4. The method for power grid inspection based on drone according to claim 2 or 3, characterized in that: The deviation angle Δθ is obtained by the following relationship: Where x is the horizontal pixel deviation from the top of the tower to the image principal point, f x is the horizontal focal length of the camera, θ g is the current yaw angle of the camera, θ pl It is the angle between the wires and the direction of the nose as detected by the radar.

5. The method for power grid inspection based on drones according to claim 1 is characterized in that: The step S1 combines real-time visual detection and radar to obtain the target tower and determine the tower type, and specifically includes: Get the current target tower detected by visual inspection, and determine the tower type based on the direction of the tower wires segmented by radar; If the current target pole tower detected by visual inspection is a forked pole tower and / or the number of pole tower wire branches segmented by radar is greater than 2, then the current target pole tower is a forked pole tower; Otherwise, the current target tower is not a bifurcated tower.

6. The method for power grid inspection based on drones according to claim 1, characterized in that: In step S3, before the following flight of the current target tower ends, different next tower search strategies are executed according to the tower type, specifically including: If it is a bifurcated pole tower, the telemetry shows a bifurcated branch, and any branch is manually selected to obtain reference direction data; based on the obtained reference direction data, the PTZ is controlled to search for the next target pole tower; If it is not a forked pole tower, when the visually guided follow-up mode is used to inspect over the preset position, the lock on the current target pole tower is released; then, combined with the radar wire information, the gimbal camera is controlled to turn to the direction of the radar wire to search for the next target pole tower.

7. The method for power grid inspection based on drones according to claim 1, characterized in that: When the inspection in the visually guided follow-up mode passes over a preset position, unlocking the target pole tower specifically includes: Acquire a camera attitude angle during the inspection process of the vision-guided follow-up mode, wherein the camera attitude angle includes at least one of a camera pitch angle, a camera yaw angle, and a camera roll angle; When inspecting to a preset position, any one of the camera attitude angles reaches a specified angle, so that the target position of the target tower is located at the boundary of the camera field of view; When the inspection passes over the preset position, the target pole tower exceeds the field of view of the camera so as to release the lock on the target pole tower.

8. The method for power grid inspection based on drones according to claim 7, characterized in that: When inspecting to a preset position, any one of the camera posture angles reaches a specified angle, specifically including: Get the current pitch angle of the camera as p c , the pitch angle of the target to avoid the preset position is p i , the focal length of the camera is f; When the inspection reaches below the preset position; The current pitch angle of the camera is p c satisfy: in, is the upper half vertical field of view angle, α is the vertical angle between the preset position and the camera optical axis in the camera coordinate system; When the inspection reaches above the preset position; The current pitch angle of the camera is p c satisfy: in, is the lower half vertical field of view angle, and α is the vertical angle formed by the preset position in the camera coordinate system and the camera optical axis.

9. The power grid inspection system based on drones is characterized by: It includes a drone and a vision module, a radar module, a positioning module and a control module arranged on the drone. The visual module and radar module are used to combine real-time visual detection and radar to acquire the target tower and determine the tower type; The positioning module is used to lock the target tower and locate it in real time to determine the target position; The control module is used to switch the flight mode according to whether the positioning is successful or not according to the target position. If the target position is determined, the flight mode is entered into the visual guidance flight mode, otherwise the radar line-following flight mode is entered until the target position is determined. The control module is also used to execute different next tower search strategies according to the tower type; If it is a bifurcated pole tower, obtaining reference data and / or operation instructions, and determining the next target pole tower based on the obtained reference data and / or operation instructions; If it is not a bifurcated pole tower, when the inspection passes over the preset position in the visual guided follow-up mode, the visual module releases the lock on the target pole tower; Then combine the radar wire information to search for the next target tower.

10. The power grid inspection system based on drone according to claim 9, characterized in that: The vision module includes a camera and a vision sensor, wherein the vision sensor is used to obtain a camera attitude angle, and the camera attitude angle includes at least one of a camera pitch angle, a camera yaw angle, and a camera roll angle.

11. The power grid inspection system based on drone according to claim 10, characterized in that: The camera attitude angle of the camera satisfies the following conditions: When inspecting to a preset position, any one of the camera attitude angles reaches a specified angle, so that the target position of the target tower is located at the boundary of the camera field of view; When the inspection passes over the preset position, the target pole tower exceeds the field of view of the camera so as to release the lock on the target pole tower.

12. The power grid inspection system based on drone according to claim 10 or 11, characterized in that: The camera pitch angle meets the following conditions: When the inspection reaches below the preset position; The current pitch angle of the camera is p c satisfy: Among them, p c is the current pitch angle of the camera, p i is the pitch angle of the target to avoid the preset position, f is the focal length of the camera, is the upper half vertical field of view angle, α is the vertical angle between the preset position and the camera optical axis in the camera coordinate system; When the inspection reaches above the preset position; The current pitch angle of the camera is p c satisfy: Among them, p c is the current pitch angle of the camera, p i is the pitch angle of the target to avoid the preset position, f is the focal length of the camera, is the lower half vertical field of view angle, and α is the vertical angle formed by the preset position in the camera coordinate system and the camera optical axis.

Images