Wind power blade detection method, device and equipment based on unmanned aerial vehicle tracking and medium

Abstract

The invention relates to a wind power blade detection method, device and equipment based on unmanned aerial vehicle tracking and a medium, and belongs to the field of unmanned aerial vehicle detection. The method comprises the following steps: presetting wind power blade detection initial information; an unmanned aerial vehicle working point is preset, and the unmanned aerial vehicle moves to the unmanned aerial vehicle working point based on the wind power blade detection initial information and obtains the working distance of the unmanned aerial vehicle; the unmanned aerial vehicle carries out wind power blade target identification on the wind power blade through the wind power blade detection device and obtains a target wind power blade image, and the wind power blade target identification is realized through a YOLOv8 target detection algorithm; generating an unmanned aerial vehicle tracking path through an improved path generation method based on the target wind power blade image; and the unmanned aerial vehicle performs wind power blade defect detection according to the unmanned aerial vehicle tracking path based on the PPO algorithm, so that the wind power blade defects are comprehensively detected in real time without shutdown, the labor cost is reduced, the detection efficiency is improved, and the misjudgment risk is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicle detection, and in particular relates to a wind turbine blade detection method, device, equipment and medium based on unmanned aerial vehicle tracking. Background Art

[0002] With the increasing concern about climate change and environmental pollution, wind energy has become a popular form of clean and environmentally friendly energy. Converting wind energy into electricity through wind turbines is one of the common ways of energy conversion. Wind energy technology has made great progress in the past few decades. The continuous innovation and development of wind energy technology has led to a continuous decline in the cost of wind power generation, making it more competitive than traditional energy sources. The cost reduction is mainly attributed to the technical improvements of wind turbines, the improvement of manufacturing efficiency and the advantages of large-scale production. In order to make full use of wind power resources, more and more wind turbines have been built.

[0003] For wind power generation, wind turbine blades are the core components of the wind power generation system, which directly affects the efficiency and reliability of wind power generation. Therefore, it is extremely important to ensure the normal operation of wind turbine blades during wind power generation. Due to the special nature of their work, wind turbine blades need to be exposed to the outside for a long time and are easily corroded and damaged. If damaged blades are not found and repaired in time, it is difficult to ensure the safe operation of the wind power generation system. Therefore, the inspection of wind turbine blades is an important part of the wind power generation process. At present, the existing inspection method of wind turbine blades is mainly manual. It is necessary to shut down the wind turbine set first and then conduct manual inspection. This method will undoubtedly greatly reduce the efficiency of wind power generation, and cannot ensure comprehensive inspection of wind turbine blades. In addition, this method cannot detect the status of wind turbine blades in real time. In addition, frequent on and off operations will also affect the life of the wind turbine generator set. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, the present invention provides a wind turbine blade detection method, device, equipment and medium based on drone tracking.

[0005] The purpose of the present invention can be achieved through the following technical solutions:

[0006] A wind turbine blade detection method based on drone tracking, the implementation of the wind turbine blade detection method comprising the following steps:

[0007] S1: Preset initial information for wind turbine blade detection, where the initial information for wind turbine blade detection includes initial information for wind turbine blades and initial information for drones;

[0008] S2: Preset a UAV working point, and the UAV moves to the UAV working point based on the initial wind turbine blade detection information and obtains the UAV working distance;

[0009] S3: The UAV uses a wind turbine blade detection device to identify the wind turbine blade target and obtains the target wind turbine blade image, wherein the wind turbine blade target identification is achieved by using the YOLOv8 target detection algorithm;

[0010] S4: Generate a UAV tracking path based on the target wind turbine blade image by using an improved path generation method;

[0011] S5: The UAV performs wind turbine blade defect detection based on the PPO algorithm according to the UAV tracking path.

[0012] Preferably, the initial information of the wind turbine blades specifically includes the wind turbine blade rotation speed and the wind turbine blade rotation center position, and the initial information of the drone includes the drone initial position and the detection initial distance, the drone initial position is directly in front of the wind turbine blade rotation center position, and the detection initial distance is the distance between the wind turbine blade rotation center position and the drone initial position.

[0013] Preferably, the UAV working distance is the distance between the UAV working point and the rotation center position of the wind turbine blade. The UAV is equipped with the wind turbine blade detection device, and the wind turbine blade detection device includes a laser radar and an industrial camera. The laser radar is used to obtain the UAV working distance.

[0014] Preferably, the step S4 specifically includes:

[0015] S401: Detecting the wind turbine blade edge of the target wind turbine blade image by using a Canny edge detection algorithm and obtaining a target wind turbine blade boundary map;

[0016] S402: Generate the UAV tracking path based on the target wind turbine blade boundary map.

[0017] Preferably, the step S402 specifically includes:

[0018] S402-1: extracting special boundary points of the target wind turbine blade boundary map, wherein the special boundary points are top points of a single wind turbine blade, and the special boundary points include a first special boundary point, a second special boundary point, and a third special boundary point;

[0019] S402-2: Connect the special boundary points to obtain a special boundary point polygon, randomly select two sides of the special boundary point polygon as the midline to obtain the polygon midline, and obtain the centroid of the target wind turbine blade boundary map through the intersection of the polygon midlines;

[0020] S402-3: Obtain a set of wind turbine blade detection paths, wherein the set of wind turbine blade detection paths includes a first wind turbine blade detection path, a second wind turbine blade detection path, a third wind turbine blade detection path, and a fourth wind turbine blade detection path. The first wind turbine blade detection path is a directional path connecting the centroid of the target wind turbine blade boundary map to the first special boundary point. The second wind turbine blade detection path is a directional path connecting the first special boundary point to the second special boundary point. The third wind turbine blade detection path is a directional path connecting the second special boundary point to the centroid of the target wind turbine blade boundary map. The fourth wind turbine blade detection path is a directional path connecting the centroid of the target wind turbine blade boundary map to the third special boundary point. The UAV tracking path is obtained through the wind turbine blade detection path set.

[0021] Preferably, the step S5 specifically includes:

[0022] S501: discretizing the drone tracking path to obtain N drone tracking path discrete points with equal distances;

[0023] S502: The UAV uses the centroid of the target wind turbine blade boundary map as a detection starting point and traverses the discrete points of the UAV tracking path according to the PPO algorithm;

[0024] S503: The UAV performs defect detection on the wind turbine blade based on the Simple Net defect detection algorithm.

[0025] Preferably, the step S502 specifically includes:

[0026] S502-1: Preset a shooting cycle, and acquire a wind turbine blade image set by the industrial camera according to the shooting cycle, wherein the wind turbine blade image set is a collection of wind turbine blade images;

[0027] S502-2: Obtain the center point of the wind turbine blade image of the wind turbine blade image, preset a discrete point distance threshold and obtain a target distance, the target distance being the distance between the center point of the wind turbine blade image and the discrete point of the UAV tracking path, and when the target distance is less than or equal to the discrete point distance threshold, set the discrete point of the UAV tracking path as the next target point, and the next target point is the discrete point of the UAV tracking path that the UAV will reach next;

[0028] S502-3: Based on the PPO algorithm, the discrete points of the UAV tracking path are traversed to obtain the minimum discrete distance, which is the distance between the UAV and the nearest discrete point of the UAV tracking path, and the algorithm input state is obtained. The algorithm input state is the input state of the PPO algorithm at time t. The algorithm input state includes the working distance difference, the target distance, the discrete point orientation, the discrete point distance, and the wind turbine blade speed. The working distance difference is the difference between the UAV working distance and the ideal UAV working distance. The discrete point orientation is the orientation of the wind turbine blade image center point of the UAV tracking path discrete point located in the wind turbine blade image. The discrete point distance is the distance between the UAV tracking path discrete point and the target wind turbine blade boundary map center point. The algorithm input state is expressed as:

[0029] s t =(Dis t ,dis t ,θ t , Δ t ,ω)

[0030] Dis t =min(Radar t )-ΔD

[0031]

[0032]

[0033] Among them, s t Input state for the algorithm, Dis t is the working distance difference, dis t is the target distance, θ t is the discrete point orientation, Δ t is the distance between discrete points, ω is the speed of wind turbine blades, min(Radar t ) is the minimum discrete distance, ΔD is the UAV working distance, x j is the horizontal coordinate of the special boundary point, y j is the ordinate of the special boundary point, x i is the horizontal coordinate of the discrete points on the UAV tracking path, y i is the ordinate of the discrete points of the UAV tracking path, x0 is the abscissa of the center of gravity of the target wind turbine blade boundary map, and y0 is the ordinate of the center of gravity of the target wind turbine blade boundary map;

[0034] S502-4: Obtain the reward value of the PPO algorithm, and output the drone tracking path adjustment information based on the reward value, wherein the drone tracking path adjustment information includes the drone tracking path no need to adjust information, the drone tracking path abnormal information, and the drone tracking path need to adjust information. The expression of the reward value is:

[0035]

[0036] Among them, rt is the reward value, Dis t is the working distance difference, ε is the discrete point distance threshold, Tanh is the hyperbolic tangent, and dis t is the target distance, θ t is the discrete point orientation, min(Radar t ) is the minimum discrete distance, ΔD is the UAV working distance;

[0037] Determine the reward value, and when the reward value is 100, feedback information indicating that no adjustment is required for the tracking path of the drone;

[0038] When the reward value is in the open interval (-100, 100), the abnormal information of the drone tracking path is fed back to a terminal;

[0039] When the reward value is -100, information indicating that the drone tracking path needs to be adjusted is fed back, and the drone automatically adjusts the drone tracking path.

[0040] A wind turbine blade detection device based on drone tracking, used to execute the wind turbine blade detection method described above, characterized in that it includes an initialization module, a target recognition module, a path generation module, and a defect detection module;

[0041] The initialization module is used to preset the initial information of wind turbine blade detection, which includes the initial information of wind turbine blades and the initial information of drones, preset the working point of the drone, and the drone moves to the working point of the drone based on the initial information of wind turbine blade detection and obtains the working distance of the drone;

[0042] The target recognition module is used for the UAV to perform wind turbine blade target recognition on the wind turbine blade through the wind turbine blade detection device and obtain the target wind turbine blade image, and the wind turbine blade target recognition is realized by the YOLOv8 target detection algorithm;

[0043] The path generation module is used to generate a UAV tracking path based on the target wind turbine blade image by using an improved path generation method;

[0044] The defect detection module is used by the drone to perform wind turbine blade defect detection according to the drone tracking path based on the PPO algorithm.

[0045] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned wind turbine blade detection method when executing the program.

[0046] A storage medium containing computer executable instructions, wherein the computer executable instructions are used to execute the above-mentioned wind turbine blade detection method when executed by a computer processor.

[0047] The beneficial effects of the present invention are:

[0048] (1) By using drones to inspect wind turbine blade defects, there is no need to shut down the wind turbine, which significantly improves the efficiency of wind power generation and reduces the impact of frequent on / off operations on the life of the wind turbine. It also realizes intelligent inspection to ensure comprehensive and real-time inspection of wind turbine blades, saves manpower, improves inspection efficiency, and significantly reduces the risk of misjudgment caused by insufficient work experience or carelessness of inspectors.

[0049] (2) The UAV tracking path is generated based on the boundary map of the target wind turbine blade, and the tracking route of the UAV is fully designed to ensure defect detection in all areas of the wind turbine blade;

[0050] (3) Based on the PPO algorithm, wind turbine blade defect detection is performed according to the UAV tracking path, without the need to establish a complex mathematical model, and has better generalization ability and robustness;

[0051] (4) Based on the reward value, the tracking path adjustment information of the UAV is output, and the detection path of the UAV is adjusted in time to avoid the serious consequences of detection errors caused by the UAV yaw or even damage to the wind turbine blades;

[0052] (5) Set the working distance of the drone to ensure that the industrial camera carried by the drone will not cause unclear images due to being too far away from the wind turbine blades, thereby affecting the success rate of wind turbine blade defect detection, nor will it cause image loss due to being too close to the wind turbine blades. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0054] Figure 1 The present invention is a flowchart of the steps of a wind turbine blade detection method based on drone tracking. DETAILED DESCRIPTION

[0055] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0056] The working principle and use process of the present invention:

[0057] See also Figure 1 , a wind turbine blade detection method based on drone tracking, comprising:

[0058] S1: Preset initial information for wind turbine blade detection, where the initial information for wind turbine blade detection includes initial information for wind turbine blades and initial information for drones;

[0059] S2: Preset a UAV working point, and the UAV moves to the UAV working point based on the initial wind turbine blade detection information and obtains the UAV working distance;

[0060] S3: The UAV uses a wind turbine blade detection device to identify the wind turbine blade target and obtains the target wind turbine blade image, wherein the wind turbine blade target identification is achieved by using the YOLOv8 target detection algorithm;

[0061] S4: Generate a UAV tracking path based on the target wind turbine blade image by using an improved path generation method;

[0062] S5: The UAV performs wind turbine blade defect detection based on the PPO algorithm according to the UAV tracking path.

[0063] In this embodiment, initial information for wind turbine blade detection is preset, and the initial information for wind turbine blade detection includes initial information for wind turbine blades and initial information for drones. The initial information for wind turbine blades specifically includes the rotation speed of wind turbine blades and the rotation center position of wind turbine blades. The initial information for drones includes an initial position of drones and an initial distance for detection. The initial position of drones is directly in front of the rotation center position of wind turbine blades, and the initial distance for detection is the distance between the rotation center position of wind turbine blades and the initial position of drones.

[0064] In this embodiment, a UAV working point is preset, and the UAV moves to the UAV working point based on the initial information of the wind turbine blade detection and obtains the UAV working distance, where the UAV working distance is the distance between the UAV working point and the rotation center of the wind turbine blade. The UAV is equipped with the wind turbine blade detection device, and the wind turbine blade detection device includes a laser radar and an industrial camera. The laser radar is used to obtain the UAV working distance to ensure that the industrial camera carried by the UAV will not cause unclear images due to being too far away from the wind turbine blades, thereby affecting the success rate of wind turbine blade defect detection, and will not cause image loss due to being too close to the wind turbine blades.

[0065] In this embodiment, a UAV tracking path is generated based on the target wind turbine blade image by an improved path generation method, and a comprehensive design is made for the tracking route of the UAV to ensure defect detection in all areas of the wind turbine blade. The specific implementation can be performed through the following steps:

[0066] S401: Detecting the wind turbine blade edge of the target wind turbine blade image by using a Canny edge detection algorithm and obtaining a target wind turbine blade boundary map;

[0067] S402: generating the UAV tracking path based on the target wind turbine blade boundary map;

[0068] In this embodiment, the UAV tracking path is generated based on the target wind turbine blade boundary map, which can be implemented by the following steps:

[0069] S402-1: extracting special boundary points of the target wind turbine blade boundary map, wherein the special boundary points are top points of a single wind turbine blade, and the special boundary points include a first special boundary point, a second special boundary point, and a third special boundary point;

[0070] S402-2: Connect the special boundary points to obtain a special boundary point polygon, randomly select two sides of the special boundary point polygon as the midline to obtain the polygon midline, and obtain the centroid of the target wind turbine blade boundary map through the intersection of the polygon midlines;

[0071] S402-3: Obtain a set of wind turbine blade detection paths, wherein the set of wind turbine blade detection paths includes a first wind turbine blade detection path, a second wind turbine blade detection path, a third wind turbine blade detection path, and a fourth wind turbine blade detection path. The first wind turbine blade detection path is a directional path connecting the centroid of the target wind turbine blade boundary map to the first special boundary point. The second wind turbine blade detection path is a directional path connecting the first special boundary point to the second special boundary point. The third wind turbine blade detection path is a directional path connecting the second special boundary point to the centroid of the target wind turbine blade boundary map. The fourth wind turbine blade detection path is a directional path connecting the centroid of the target wind turbine blade boundary map to the third special boundary point. The UAV tracking path is obtained through the wind turbine blade detection path set.

[0072] In this embodiment, the drone performs wind turbine blade defect detection according to the drone tracking path based on the PPO algorithm, without shutting down the wind turbine generator set, significantly improving the efficiency of wind power generation, reducing the impact of frequent power on and off operations on the life of the wind turbine generator set, and realizing intelligent detection, ensuring comprehensive and real-time detection of wind turbine blades, saving manpower, improving detection work efficiency, and significantly reducing the risk of misjudgment caused by insufficient work experience or carelessness of detection personnel. It can be implemented specifically through the following steps:

[0073] S501: discretizing the drone tracking path to obtain N drone tracking path discrete points with equal distances;

[0074] S502: The UAV takes the centroid of the target wind turbine blade boundary map as the detection starting point, and traverses the discrete points of the UAV tracking path according to the PPO algorithm, without the need to establish a complex mathematical model, and has better generalization ability and robustness;

[0075] S503: The UAV performs defect detection on the wind turbine blade based on the Simple Net defect detection algorithm.

[0076] In this embodiment, the drone uses the centroid of the target wind turbine blade boundary map as the detection starting point and traverses the discrete points of the drone tracking path according to the PPO algorithm, which can be specifically implemented by the following steps:

[0077] S502-1: Preset a shooting cycle, and acquire a wind turbine blade image set by the industrial camera according to the shooting cycle, wherein the wind turbine blade image set is a collection of wind turbine blade images;

[0078] S502-2: Obtain the center point of the wind turbine blade image of the wind turbine blade image, preset a discrete point distance threshold and obtain a target distance, the target distance being the distance between the center point of the wind turbine blade image and the discrete point of the UAV tracking path, and when the target distance is less than or equal to the discrete point distance threshold, set the discrete point of the UAV tracking path as the next target point, and the next target point is the discrete point of the UAV tracking path that the UAV will reach next;

[0079] S502-3: Based on the PPO algorithm, the discrete points of the UAV tracking path are traversed to obtain the minimum discrete distance, which is the distance between the UAV and the nearest discrete point of the UAV tracking path, and the algorithm input state is obtained. The algorithm input state is the input state of the PPO algorithm at time t. The algorithm input state includes the working distance difference, the target distance, the discrete point orientation, the discrete point distance, and the wind turbine blade speed. The working distance difference is the difference between the UAV working distance and the ideal UAV working distance. The discrete point orientation is the orientation of the wind turbine blade image center point of the UAV tracking path discrete point located in the wind turbine blade image. The discrete point distance is the distance between the UAV tracking path discrete point and the target wind turbine blade boundary map center point. The algorithm input state is expressed as:

[0080] s t =(Dis t ,dis t ,θ t , Δ t ,ω)

[0081] Dis t =min(Radar t )-ΔD,

[0082]

[0083] Among them, s t Input state for the algorithm, Dist is the working distance difference, dis t is the target distance, θ t is the discrete point orientation, Δ t is the distance between discrete points, ω is the speed of wind turbine blades, min(Radar t ) is the minimum discrete distance, ΔD is the UAV working distance, x j is the horizontal coordinate of the special boundary point, y j is the ordinate of the special boundary point, x i is the horizontal coordinate of the discrete points on the UAV tracking path, y i is the ordinate of the discrete points of the UAV tracking path, x0 is the abscissa of the center of gravity of the target wind turbine blade boundary map, and y0 is the ordinate of the center of gravity of the target wind turbine blade boundary map;

[0084] S502-4: Obtain the reward value of the PPO algorithm, output the UAV tracking path adjustment information based on the reward value, and adjust the detection path of the UAV in time to avoid the serious consequences of detection errors caused by the yaw of the UAV and even damage to the wind turbine blades. The UAV tracking path adjustment information includes the information that the UAV tracking path does not need to be adjusted, the UAV tracking path abnormal information, and the UAV tracking path needs to be adjusted. The expression of the reward value is:

[0085]

[0086] Among them, rt is the reward value, Dis t is the working distance difference, ε is the discrete point distance threshold, Tanh is the hyperbolic tangent, and dis t is the target distance, θ t is the discrete point orientation, min(Radar t ) is the minimum discrete distance, ΔD is the UAV working distance;

[0087] Determine the reward value, and when the reward value is 100, feedback information indicating that no adjustment is required for the tracking path of the drone;

[0088] When the reward value is in the open interval (-100, 100), the abnormal information of the drone tracking path is fed back to a terminal;

[0089] When the reward value is -100, information that the drone tracking path needs to be adjusted is fed back, and the drone automatically adjusts the drone tracking path;

[0090] A wind turbine blade detection device based on drone tracking, comprising an initialization module, a target recognition module, a path generation module, and a defect detection module;

[0091] The initialization module is used to preset the initial information of wind turbine blade detection, which includes the initial information of wind turbine blades and the initial information of drones, preset the working point of the drone, and the drone moves to the working point of the drone based on the initial information of wind turbine blade detection and obtains the working distance of the drone;

[0092] The target recognition module is used for the UAV to perform wind turbine blade target recognition on the wind turbine blade through the wind turbine blade detection device and obtain the target wind turbine blade image, and the wind turbine blade target recognition is realized by the YOLOv8 target detection algorithm;

[0093] The path generation module is used to generate a UAV tracking path based on the target wind turbine blade image by using an improved path generation method;

[0094] The defect detection module is used by the drone to perform wind turbine blade defect detection according to the drone tracking path based on the PPO algorithm.

[0095] The computer storage medium of the embodiment of the present invention may adopt any combination of one or more computer-readable media. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device.

[0096] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0097] The program code included on the computer readable medium can be transmitted with any appropriate medium, including but not limited to wireless, electric wire, optical cable, RF, etc., or any suitable combination of the above. The computer program code for performing the operation of the present invention can be written in one or more programming languages ​​or their combinations, and the programming language includes object-oriented programming languages-such as Java, Smalltalk, C++, and also includes conventional procedural programming languages-such as "C" language or similar programming languages. The program code can be executed completely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on the remote computer, or completely on the remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0098] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A wind turbine blade detection method based on drone tracking, characterized in that: The implementation of the wind turbine blade detection method comprises the following steps: S1: Preset the initial information for wind turbine blade detection, the initial information for wind turbine blade detection includes the initial information for wind turbine blades and the initial information for drones, the initial information for wind turbine blades specifically includes the wind turbine blade rotation speed and the wind turbine blade rotation center position, the initial information for drones includes the initial position of drones and the initial distance for detection, the initial position of drones is directly in front of the rotation center position of the wind turbine blades, and the initial distance for detection is the distance between the rotation center position of the wind turbine blades and the initial position of drones; S2: Preset a UAV working point. The UAV moves to the UAV working point based on the initial wind turbine blade detection information and obtains the UAV working distance. The UAV working distance is the distance between the UAV working point and the rotation center of the wind turbine blade. The UAV is equipped with the wind turbine blade detection device. The wind turbine blade detection device includes a laser radar and an industrial camera. The laser radar is used to obtain the UAV working distance. S3: The UAV uses a wind turbine blade detection device to identify the wind turbine blade target and obtains the target wind turbine blade image, wherein the wind turbine blade target identification is achieved by using the YOLOv8 target detection algorithm; S4: Generate a UAV tracking path based on the target wind turbine blade image by using an improved path generation method; S5: The UAV performs wind turbine blade defect detection based on the PPO algorithm according to the UAV tracking path.

2. The wind turbine blade detection method according to claim 1, characterized in that: The step S4 specifically includes: S401: Detecting the wind turbine blade edge of the target wind turbine blade image by using a Canny edge detection algorithm and obtaining a target wind turbine blade boundary map; S402: Generate the UAV tracking path based on the target wind turbine blade boundary map.

3. The wind turbine blade detection method according to claim 2, characterized in that: The step S402 specifically includes: S402-1: extracting special boundary points of the target wind turbine blade boundary map, wherein the special boundary points are top points of a single wind turbine blade, and the special boundary points include a first special boundary point, a second special boundary point, and a third special boundary point; S402-2: Connect the special boundary points to obtain a special boundary point polygon, randomly select two sides of the special boundary point polygon as the midline to obtain the polygon midline, and obtain the centroid of the target wind turbine blade boundary map through the intersection of the polygon midlines; S402-3: Obtain a set of wind turbine blade detection paths, and obtain the UAV tracking path based on the set of wind turbine blade detection paths.

4. The wind turbine blade detection method according to claim 3, characterized in that: The step S402-3 is specifically as follows: The wind turbine blade detection path set includes a first wind turbine blade detection path, a second wind turbine blade detection path, a third wind turbine blade detection path, and a fourth wind turbine blade detection path. The first wind turbine blade detection path is a directional path connecting the centroid of the target wind turbine blade boundary map to the first special boundary point. The second wind turbine blade detection path is a directional path connecting the first special boundary point to the second special boundary point. The third wind turbine blade detection path is a directional path connecting the second special boundary point to the centroid of the target wind turbine blade boundary map. The fourth wind turbine blade detection path is a directional path connecting the centroid of the target wind turbine blade boundary map to the third special boundary point. The UAV tracking path is obtained through the wind turbine blade detection path set.

5. The wind turbine blade detection method according to claim 1, characterized in that: The step S5 specifically includes: S501: discretizing the drone tracking path to obtain N drone tracking path discrete points with equal distances; S502: The UAV uses the centroid of the target wind turbine blade boundary map as a detection starting point and traverses the discrete points of the UAV tracking path according to the PPO algorithm; S503: The UAV performs defect detection on the wind turbine blade based on the Simple Net defect detection algorithm.

6. The wind turbine blade detection method according to claim 5, characterized in that: The step S502 specifically includes: S502-1: Preset a shooting cycle, and acquire a wind turbine blade image set by the industrial camera according to the shooting cycle, wherein the wind turbine blade image set is a collection of wind turbine blade images; S502-2: Obtain the center point of the wind turbine blade image of the wind turbine blade image, preset a discrete point distance threshold and obtain a target distance, the target distance being the distance between the center point of the wind turbine blade image and the discrete point of the UAV tracking path, and when the target distance is less than or equal to the discrete point distance threshold, set the discrete point of the UAV tracking path as the next target point, and the next target point is the discrete point of the UAV tracking path that the UAV will reach next; S502-3: Based on the PPO algorithm, the discrete points of the UAV tracking path are traversed to obtain the minimum discrete distance, which is the distance between the UAV and the nearest discrete point of the UAV tracking path, and the algorithm input state is obtained. The algorithm input state is the input state of the PPO algorithm at time t. The algorithm input state includes the working distance difference, the target distance, the discrete point orientation, the discrete point distance, and the wind turbine blade speed. The working distance difference is the difference between the UAV working distance and the ideal UAV working distance. The discrete point orientation is the orientation of the wind turbine blade image center point of the UAV tracking path discrete point located in the wind turbine blade image. The discrete point distance is the distance between the UAV tracking path discrete point and the target wind turbine blade boundary map center point. The algorithm input state is expressed as: s t =(Dis t ,dis t ,the t ,D t ,oh), Haze t =my(Radar t )-ΔD, Among them, s t Input state for the algorithm, Dis t is the working distance difference, dis t is the target distance, θ t is the discrete point orientation, Δ t is the distance between discrete points, ω is the speed of wind turbine blades, min(Radar t ) is the minimum discrete distance, ΔD is the UAV working distance, x j is the horizontal coordinate of the special boundary point, y j is the ordinate of the special boundary point, x i is the horizontal coordinate of the discrete points on the UAV tracking path, y i is the ordinate of the discrete points of the UAV tracking path, x0 is the abscissa of the center of gravity of the target wind turbine blade boundary map, and y0 is the ordinate of the center of gravity of the target wind turbine blade boundary map; S502-4: Obtain a reward value of the PPO algorithm, and output UAV tracking path adjustment information based on the reward value, wherein the UAV tracking path adjustment information includes information that the UAV tracking path does not need to be adjusted, information that the UAV tracking path is abnormal, and information that the UAV tracking path needs to be adjusted.

7. The wind turbine blade detection method according to claim 6, characterized in that: The step S502-4 specifically includes: The expression of the reward value is: Among them, rt is the reward value, Dis t is the working distance difference, ε is the discrete point distance threshold, Tanh is the hyperbolic tangent, and dis t is the target distance, θ t is the discrete point orientation, min(Radar t ) is the minimum discrete distance, ΔD is the UAV working distance; Determine the reward value, and when the reward value is 100, feedback information indicating that no adjustment is required for the tracking path of the drone; When the reward value is in the open interval (-100, 100), the abnormal information of the drone tracking path is fed back to a terminal; When the reward value is -100, information indicating that the drone tracking path needs to be adjusted is fed back, and the drone automatically adjusts the drone tracking path.

8. A wind turbine blade detection device based on drone tracking, characterized in that: It includes initialization module, target recognition module, path generation module and defect detection module; The initialization module is used to preset the initial information of wind turbine blade detection, which includes the initial information of wind turbine blades and the initial information of drones, preset the working point of the drone, and the drone moves to the working point of the drone based on the initial information of wind turbine blade detection and obtains the working distance of the drone; The target recognition module is used for the UAV to perform wind turbine blade target recognition on the wind turbine blade through the wind turbine blade detection device and obtain the target wind turbine blade image, and the wind turbine blade target recognition is realized by the YOLOv8 target detection algorithm; The path generation module is used to generate a UAV tracking path based on the target wind turbine blade image by using an improved path generation method; The defect detection module is used by the drone to perform wind turbine blade defect detection according to the drone tracking path based on the PPO algorithm.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the wind turbine blade detection method as described in any one of claims 1-7 is implemented.

10. A storage medium containing computer executable instructions, characterized in that: The computer executable instructions are used to execute the wind turbine blade detection method as described in any one of claims 1 to 7 when executed by a computer processor.

Images