Method for Locating Faulty Photovoltaic Modules in a Photovoltaic Field

Abstract

The present invention discloses a method for locating a faulty photovoltaic module in a photovoltaic field, comprising the following steps: obtaining a panoramic view of the photovoltaic field, and obtaining the corresponding geographical coordinate range of the photovoltaic field according to the first position information of the first acquisition position and the panoramic view; obtaining a fault image, where the fault image is an image including at least the faulty photovoltaic module, and obtaining the position coordinates of the faulty photovoltaic module according to the second position information of the second acquisition position and the fault image; and calibrating the faulty photovoltaic module in the panoramic view according to the position coordinates and the geographical coordinate range. The present invention improves the accuracy of locating faulty modules in a photovoltaic field, and by first determining the longitude and latitude range of the photovoltaic field and then determining its position in the photovoltaic field according to the faulty longitude and latitude, reduces the workload of fault location. By using the method of character recognition of image watermarks to obtain parameter data during flight, the dependence on flight data is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of fault detection of photovoltaic modules, and particularly relates to a method for locating faulty photovoltaic modules in a photovoltaic field. Background Art

[0002] Related research shows that with the continuous development of solar energy technology, more and more photovoltaic power stations will be put into use. However, during the use of solar panels, their operation may be abnormal due to reasons such as foreign object occlusion or hot spots. In severe cases, it may even directly lead to the scrapping of the entire panel and then the shutdown of the photovoltaic string. Therefore, conducting daily inspections of photovoltaic power stations, detecting the operation status of photovoltaic modules in real time, and promptly eliminating faults are important tasks in the operation of large-scale photovoltaic power stations.

[0003] Currently, the inspection method of photovoltaic power stations has gradually developed in the technical direction from time-consuming and laborious manual inspection to automated inspection. Using drones with cameras for inspection is one of the most typical application scenarios. However, due to the generally large area of photovoltaic fields, the number of images captured by drones is large and the content is highly homogeneous, resulting in a low accuracy rate for locating fault positions. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to overcome the defect of low accuracy in the position of faulty photovoltaic modules in a photovoltaic power station in the prior art, and provide a method for locating faulty photovoltaic modules in a photovoltaic field.

[0005] The present invention solves the above technical problem through the following technical solutions:

[0006] The present invention provides a method for locating faulty photovoltaic modules in a photovoltaic field, including the following steps:

[0007] Obtain a panoramic view of the photovoltaic field, and obtain the corresponding geographical coordinate range of the photovoltaic field based on the first position information of the first acquisition position, the first camera field of view angle of the drone, and the panoramic view. The first acquisition position is the position where the first acquisition device is located when obtaining the panoramic view; wherein, the first acquisition device includes the drone; the first camera field of view angle is the camera field of view angle of the drone when obtaining the panoramic view of the photovoltaic field; obtain a fault image, which is an image including at least a faulty photovoltaic module. Obtain the position coordinates of the faulty photovoltaic module based on the second position information of the second acquisition position, the second camera field of view angle of the drone, and the fault image. The second acquisition position is the position where the second acquisition device is located when obtaining the fault image; wherein, the second acquisition device includes the drone; the second camera field of view angle is the camera field of view angle of the drone when obtaining the fault image;

[0008] Calibrate the faulty photovoltaic module in the panoramic view according to the position coordinates and the geographical coordinate range.

[0009] Preferably, the first acquisition device includes a drone for obtaining a panoramic view of the photovoltaic field, including:

[0010] Obtaining the panoramic view based on the drone;

[0011] Obtaining the corresponding geographical coordinate range of the photovoltaic field according to the first position information of the first acquisition position and the panoramic view, including:

[0012] Obtaining the coordinate range according to the size of the panoramic view, the first camera field of view angle of the drone, and the first position information, where the first position information includes the longitude and latitude, flight altitude of the drone during shooting, and the clockwise azimuth angle between the drone flight direction and the preset reference direction

[0013] Preferably, obtaining the coordinate range according to the size of the panoramic view, the first camera field of view angle of the drone, and the first position information includes:

[0014] Constructing a first Cartesian coordinate system, and obtaining the corresponding first coordinate range of the panoramic view in the first Cartesian coordinate system according to the size of the panoramic view, the first camera field of view angle, and the flight altitude;

[0015] Obtaining the corresponding geographical coordinate range of the photovoltaic field according to the first coordinate range, the longitude and latitude of the drone during shooting, and the clockwise azimuth angle.

[0016] Preferably, the first Cartesian coordinate system takes the projection point of the drone in the panoramic view as the origin, the horizontal direction from left to right along the panoramic view as the positive x-axis direction, and the vertical direction from bottom to top along the panoramic view as the positive y-axis direction. Then, the coordinates (X′, Y′) corresponding to the boundary corner points of the panoramic view in the first Cartesian coordinate system are represented as:

[0017]

[0018] where ∠FOV represents the first camera field of view angle, h represents the flight altitude, l represents the length of the panoramic view, and w represents the width of the panoramic view.

[0019] Preferably, the geographical coordinates (X1, Y1) of the boundary corner points are represented as:

[0020] X1 = X′cos(θ) - Y′sin(θ) + X,

[0021] Y1 = Y′cos(θ) + X′sin(θ) + Y,

[0022] where θ represents the clockwise azimuth angle; (X, Y) represents the longitude and latitude of the drone during shooting.

[0023] Preferably, obtaining the fault image includes:

[0024] Obtain a component image, which includes at least one photovoltaic component in a photovoltaic field;

[0025] Identify the component image to determine whether the photovoltaic component in the component image is a faulty photovoltaic component. If so, determine the component image as a faulty image.

[0026] Preferably, the second acquisition device includes a drone. Obtaining the component image includes:

[0027] Based on the drone's inspection and shooting of the photovoltaic field, obtain the component image and mark watermark information in the component image. The watermark information includes second position information, and the second position information includes the longitude, latitude, and altitude of the second acquisition position.

[0028] Preferably, obtaining the component image includes:

[0029] Based on the drone's inspection and shooting of the photovoltaic field to traverse the photovoltaic field, obtain the video data of the photovoltaic field, and mark watermark information in each frame of the video data;

[0030] Perform frame extraction on the video data to obtain the component image.

[0031] Preferably, performing frame extraction on the video data to obtain the component image includes:

[0032] Based on the ground workstation receiving the video data from the drone, perform frame extraction on the video data based on the ground workstation to obtain the component image.

[0033] Preferably, obtaining the position coordinates of the faulty photovoltaic component according to the second position information of the second acquisition position and the faulty image includes:

[0034] Extract the first longitude and latitude information from the watermark information of the faulty image, and extract the second longitude and latitude information from the watermark information of the auxiliary image. The auxiliary image is the next frame image adjacent to the faulty image in the video data;

[0035] Obtain the first azimuth angle according to the first longitude and latitude information and the second longitude and latitude information. The first azimuth angle is the angle between the flight direction of the drone at the second acquisition position and the first preset reference direction;

[0036] Wherein, the first azimuth angle The first longitude and latitude information is (X2, Y2), and the second longitude and latitude information is (X3, Y3).

[0037] Preferably, obtaining the position coordinates of the faulty photovoltaic component according to the second position information of the second acquisition position and the faulty image further includes:

[0038] Extract the first altitude information from the watermark information of the faulty image;

[0039] Obtain the position coordinates of the faulty photovoltaic module based on the fault image, the first longitude and latitude information, the first azimuth angle, the first altitude information, and the second camera field of view angle.

[0040] Preferably, obtaining the position coordinates of the faulty photovoltaic module based on the fault image, the first longitude and latitude information, the first azimuth angle, the first altitude information, and the second camera field of view angle includes:

[0041] Construct a second Cartesian coordinate system, and obtain the fault coordinates corresponding to the faulty photovoltaic module in the second Cartesian coordinate system according to the size of the fault image, the second camera field of view angle of the unmanned aerial vehicle, and the first altitude information;

[0042] Obtain the position coordinates of the faulty photovoltaic module based on the fault coordinates, the first longitude and latitude information, the first azimuth angle, and the first altitude information.

[0043] Preferably, the positioning method further includes:

[0044] Output the position coordinates of the faulty photovoltaic module.

[0045] Preferably, identifying the component image to determine whether the photovoltaic module in the component image is the faulty photovoltaic module includes:

[0046] Identify the component image based on the OCR (Optical Character Recognition) character recognition model to determine whether the photovoltaic module in the component image is the faulty photovoltaic module.

[0047] Preferably, the OCR character recognition model is generated based on the YOLO V5 (an algorithm) algorithm.

[0048] The positive and progressive effects of the present invention are as follows: The present invention improves the accuracy of locating faulty components in a photovoltaic field, and by first determining the longitude and latitude range of the photovoltaic field and then determining its position within the photovoltaic field based on the fault longitude and latitude, reduces the workload of fault location. The method of obtaining parameter data during flight by recognizing image watermarks with characters reduces the dependence on flight data. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a flowchart of a method for locating a faulty photovoltaic module in a photovoltaic field according to a preferred embodiment of the present invention.

[0050] Figure 2 It is a schematic diagram of an unmanned aerial vehicle obtaining a panoramic view of a photovoltaic field in a method for locating a faulty photovoltaic module in a photovoltaic field according to a preferred embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0051] The present invention will be further described below by way of a preferred embodiment, but the present invention is not limited to the scope of the described embodiment.

[0052] This embodiment provides a method for locating a faulty photovoltaic module in a photovoltaic field. Referring to Figure 1 , the method for locating a faulty photovoltaic module in the photovoltaic field includes the following steps:

[0053] Step S1: Obtain a panoramic view of the photovoltaic field, and obtain the geographical coordinate range corresponding to the photovoltaic field according to the first position information of the first acquisition position and the panoramic view. The first acquisition position is the position where the first acquisition device obtains the panoramic view.

[0054] Step S2: Obtain a fault image, where the fault image is an image including at least the faulty photovoltaic module, and obtain the position coordinates of the faulty photovoltaic module according to the second position information of the second acquisition position and the fault image. The second acquisition position is the position where the second acquisition device obtains the fault image.

[0055] Step S3: Calibrate the faulty photovoltaic module in the panoramic view according to the position coordinates and the geographical coordinate range.

[0056] In specific implementation, first, in step S1, a panoramic view of the photovoltaic field is obtained.

[0057] As an optional implementation manner, referring to Figure 2 , in step S1, a panoramic view of the photovoltaic field is obtained based on the unmanned aerial vehicle 101. Among them, the unmanned aerial vehicle 101 takes a picture of the photovoltaic field 102 from above the photovoltaic field, and the camera of the unmanned aerial vehicle is vertically downward. Among them, the flight height of the unmanned aerial vehicle 101 relative to the ground of the photovoltaic field 102 is represented by h, and the field of view angle of the camera of the unmanned aerial vehicle 101 is represented by ∠FOV. Let the length of the picture of the panoramic view of the photovoltaic field obtained by shooting be represented by l, and the width of the picture of the panoramic view of the photovoltaic field be represented by w. Then, a first Cartesian coordinate system is constructed. Among them, with the panoramic view of the photovoltaic field as the reference, the first Cartesian coordinate system takes the projection point O corresponding to the unmanned aerial vehicle 101 as the origin, the direction from left to right as the positive direction of the x-axis, and the direction from bottom to top as the positive direction of the y-axis. According to the flight height h, the camera field of view angle ∠FOV, the length l of the picture of the panoramic view of the photovoltaic field, and the width w of the picture of the panoramic view of the photovoltaic field, the real coordinates (X′, Y′) corresponding to the boundary corner points of the panoramic view of the photovoltaic field in the first Cartesian coordinate system can be obtained, and its calculation method is as follows:

[0058]

[0059] Among them, the flight direction of the drone, the nose direction, and the upward direction of the width direction of the image captured by the camera (i.e., the bottom-to-top direction of the panoramic view of the photovoltaic field) are in the same direction (i.e., the positive direction of the y-axis of the first Cartesian coordinate system). Taking the geographical due north direction as the preset reference direction, the clockwise azimuth angle between the flight direction of the drone and the north is represented by θ, and the longitude and latitude of the drone when taking pictures are (X, Y). Combining with the coordinate rotation transformation formula, the longitude and latitude coordinates (X1, Y1) (i.e., geographical coordinates) of the boundary corner points of the panoramic view can be calculated. The operation formula is as follows:

[0060] X1 = X'cos(θ) - Y'sin(θ) + X (3)

[0061] Y1 = Y'cos(θ) + X'sin(θ) + Y (4)

[0062] Correspondingly, the longitude and latitude range (i.e., geographical coordinate range) corresponding to the coverage range of the panoramic view of the photovoltaic field can be obtained, that is, the longitude and latitude coordinates (geographical coordinates) of each location in the panoramic view of the photovoltaic field can be obtained.

[0063] Then, in step S2, component images are acquired based on the drone. The component images include at least one photovoltaic component in the photovoltaic field, and the component images also carry watermark information, which at least includes the longitude and latitude information and flight altitude information of the drone when acquiring the component images. In specific implementation, the drone is used to inspect and photograph the photovoltaic components in an automatic constant-speed cruising manner, with the camera facing vertically downward, collecting high-definition videos of the photovoltaic components with longitude, latitude, and flight altitude watermarks, and saving the video data in the camera. After the inspection and photographing, the video data covers each photovoltaic component in the photovoltaic field. Therefore, the component images corresponding to each photovoltaic component can be obtained from the video data.

[0064] In specific implementation, the drone transmits the above video data stored in the camera to the ground workstation. The ground workstation performs frame extraction processing on the video data to form an inspection image database, and the inspection image database includes the component images corresponding to each photovoltaic component.

[0065] Then, fault images are identified from the component images. The fault images are the component images corresponding to the faulty photovoltaic components. As an optional implementation manner, a photovoltaic component surface fault recognition model trained by an artificial intelligence algorithm is used to analyze and identify the component images. If it is determined that the photovoltaic component corresponding to the component image is faulty, then the component image is determined as a fault image.

[0066] Next, determine the position of the faulty photovoltaic module in the panoramic view of the photovoltaic field based on the fault image. In specific implementation, an OCR character recognition model is used to recognize the fault image to extract the first longitude and latitude information (X2, Y2) and the first altitude information. Among them, the first longitude and latitude information is the longitude and latitude information contained in the watermark information of the fault image, and the first altitude information is the flight altitude information contained in the watermark information of the fault image. The OCR character recognition model is also used to recognize the auxiliary image to extract the second longitude and latitude information (X3, Y3) and the second altitude information. Among them, the auxiliary image is the next frame image adjacent to the fault image in the inspection image database. The second longitude and latitude information is the longitude and latitude information contained in the watermark information of the auxiliary image, and the second altitude information is the flight altitude information contained in the watermark information of the auxiliary image. Since the fault image and the auxiliary image are two adjacent frame images, the first altitude information is the same as the second altitude information, which is represented by h1.

[0067] Then, determine the longitude and latitude coordinates of the faulty photovoltaic module according to the first longitude and latitude information, the second longitude and latitude information, and the shooting altitude information.

[0068] In specific implementation, the first azimuth angle is obtained according to the first longitude and latitude information and the second longitude and latitude information. Among them, the first azimuth angle is the clockwise azimuth angle between the flight direction of the drone when obtaining the fault image and the north, which is represented by θ1. Then:

[0069]

[0070] Assume that the installation height of the photovoltaic panel of the photovoltaic module is h2. Then the distance between the drone and the photovoltaic panel of the photovoltaic module is (h1 - h2). Substitute this distance into formulas (1) and (2) to obtain the actual scene size corresponding to the fault image. Then, map the actual scene size corresponding to the fault image and the position of the faulty photovoltaic component in the fault image to obtain the real coordinates (X″, Y″) of this position in the coordinate system with the center point of the image as the origin, the right side of the length direction of the image as the positive x-axis direction, and the upper side of the width direction as the positive y-axis direction. By substituting this coordinate and the flight azimuth angle θ1 of the drone into formulas (3) and (4), the longitude and latitude coordinates of the faulty photovoltaic module can be obtained.

[0071] Then, in step S3, output the longitude and latitude coordinates of the faulty photovoltaic module, and at the same time, use the longitude and latitude correspondence relationship to mark the defect position in the panoramic view of the photovoltaic field.

[0072] In a specific application scenario, in step S1, the drone 101 is equipped with a wide-angle camera with a high-definition 4056*3050 image quality and 30 frames of imaging. The camera field of view angle ∠FOV is 82.9° (degrees). During the aerial photography process, the gimbal of the drone 101 is stabilized to ensure that the camera takes pictures of the photovoltaic field at a vertically downward angle. Moreover, the flight altitude of the drone 101 is continuously increased until the picture taken by the drone 101 can cover the entire photovoltaic field. Then, a panoramic picture is taken to obtain a panoramic picture of the photovoltaic field. Suppose the flight altitude h of the drone 101 when taking the panoramic picture at this time is 212.373 m (meters), and the latitude and longitude of the drone 101 when taking pictures are approximately (31.430358°N, 121.170034°E). According to the drone flight parameters, the clockwise azimuth angle θ between the drone flight direction and the north at this time can be obtained as 138.1°. The corresponding latitude and longitude range of the photovoltaic field can be obtained according to equations (1), (2), (3), and (4) for fault marking.

[0073] Then, in step S2, an infrared camera with a resolution of 688*556 and a camera field of view angle of 40.6° is carried by the drone. In the actual operation and maintenance scenario, the route information is loaded according to the requirements and the flight mission is carried out according to the corresponding data. This route planning can ensure that all photovoltaic components (photovoltaic panels) in the photovoltaic field can be completely photographed in one aerial photography. Preferably, the inspection is carried out under the conditions of relatively clear weather, no wind or gentle breeze, and good light. The flight speed should be less than 1 m / s (meters per second) to prevent blurring in the photographed images. The vertical distance between the drone and the photovoltaic components is about 5 - 20 meters. After zooming, the area of the photovoltaic panel in the picture should be as large as possible, but at least one complete photovoltaic panel should be able to be photographed. During the shooting process, the flight parameters (for example, the latitude and longitude information and flight altitude information of the drone) are added to each frame of the video as watermark information in real time, and the watermark information is added to the upper left corner of the image. The video data obtained by shooting is saved inside the camera.

[0074] The video data stored in the drone's camera is transmitted to the ground workstation. Based on the ground workstation, the video frames of the drone are extracted in real time and saved. In some optional implementation manners, the frame extraction interval is one frame per second. In other optional implementation manners, the frame extraction interval is reasonably set according to needs, and multiple frame extraction intervals can be set. The high-definition images obtained after frame extraction are read as component images. An artificial intelligence algorithm model is used to identify from these component images, extract the images containing faulty photovoltaic components, and identify the position of the fault defect in the image. Then, the faulty image and its next image (that is, the next frame image adjacent to the faulty image in the video data) are extracted for the next step of processing.

[0075] Generate a model for character recognition of the flight parameter watermark in the upper left corner of the image (OCR character recognition model) based on the YOLO V5 algorithm. The model parameters are as follows: The total number of character detection types in the model is 15, including the ten Arabic numerals from 0 to 9 and the characters ":", "-", "°E", "°N", "m" that appear regularly in the watermark information. The size of the input image is 608*608. 90% of the training dataset is used for training, and 10% is used for testing. The initial weight is 0, and the initial anchor boxes are set to [10,13,16,30,33,23], [30,61,62,45,59,119], [116,90,156,198,373,326], which can well adapt to the detection of foreign objects (i.e., faults) of various different sizes. The backbone network uses the Cross Stage Partial Network (CSPNet), which accelerates the algorithm running speed and reduces memory overhead. The batch size is 8, and the number of epochs is 100. Finally, when IoU = 0.5 is set, the accuracy of OCR recognition can reach over 90%. In a specific implementation scenario, the longitude and latitude (31.429928°N, 121.170097°E) at the time of the drone's shooting and the height of 49.406m at the time of shooting are recognized.

[0076] According to the change in longitude and latitude between the fault image and its next image, the clockwise azimuth angle between the drone and the north during the inspection at this time can be obtained as 47.6°. Assuming the installation height of the rooftop photovoltaic module is 30m, the distance between the drone and the photovoltaic module can be obtained as 19.406m. At this time, according to equations (1), (2), (3), and (4), the longitude and latitude coordinates of the faulty photovoltaic module can be obtained.

[0077] Then, based on the longitude and latitude range of the photovoltaic field and the longitude and latitude of the faulty module, mark the faulty module in the panoramic image. The inspection personnel can view the fault detection results in real time on the panoramic image. When any faulty module is found, the fault image can be traced back for review, or they can navigate to the location of the faulty module according to the longitude and latitude and perform corresponding processing on the corresponding faulty module.

[0078] The method for locating a faulty photovoltaic module in a photovoltaic power plant according to this embodiment can enhance the automation level of inspection by using the automatic cruising function of an unmanned aerial vehicle (UAV) to inspect the photovoltaic power plant. By first determining the longitude and latitude range of the photovoltaic power plant and then determining its position within the photovoltaic power plant based on the longitude and latitude of the fault, the workload of fault location is reduced. By first recording a high-definition inspection video and then exporting the saved video for processing, the demand for real-time data transmission is reduced. By using the method of character recognition and image watermarking to obtain the parameter data during flight, the dependence on flight data is reduced. When locating the faulty module, the height difference between the photovoltaic module and the ground is considered, resulting in higher accuracy. Moreover, the faulty module is marked in the panoramic view, making the result more intuitive.

[0079] The method for locating a faulty photovoltaic module in a photovoltaic power plant according to this embodiment greatly reduces the workload during the operation and maintenance process by automating the UAV automatic inspection and fault location, realizes regular inspection, and can achieve predictive maintenance of faults. It simplifies the longitude and latitude calculation work during the preliminary deployment of the inspection system and the construction of the network transmission system during the construction, and realizes the low-cost and rapid deployment of the UAV inspection system. It has a low dependence on network transmission and flight data, and the fault location has stronger robustness. By adding the method of calculating the height difference between the photovoltaic module and the ground, this location method can be applied to both centralized photovoltaic power plants and distributed photovoltaic power plants that are often installed on rooftops with a relatively high height from the ground. By directly marking the fault in the panoramic view, the operation and maintenance personnel can more intuitively determine the location of the faulty module.

[0080] Although the specific embodiments of the present invention have been described above, those skilled in the art should understand that this is only an example, and the protection scope of the present invention is defined by the appended claims. Without departing from the principles and essence of the present invention, those skilled in the art can make various changes or modifications to these embodiments, but these changes and modifications all fall within the protection scope of the present invention.

Claims

1. A method for locating a faulty photovoltaic module in a photovoltaic field, characterized in that, Including the following steps: Obtain a panoramic view of the photovoltaic field, and obtain the corresponding geographical coordinate range of the photovoltaic field based on the first position information of the first acquisition position, the first camera field of view angle of the unmanned aerial vehicle, and the panoramic view. The first acquisition position is the position where the first acquisition device is located when it acquires the panoramic view. Among them, the first acquisition device includes the unmanned aerial vehicle; the first camera field of view angle is the camera field of view angle of the unmanned aerial vehicle when acquiring the panoramic view of the photovoltaic field; Obtain a fault image, where the fault image is an image including at least the faulty photovoltaic module. Obtain the position coordinates of the faulty photovoltaic module based on the second position information of the second acquisition position, the second camera field of view angle of the unmanned aerial vehicle, and the fault image. The second acquisition position is the position where the second acquisition device is located when it acquires the fault image. Among them, the second position information includes the longitude, latitude, and altitude of the second acquisition position; the second acquisition device includes the unmanned aerial vehicle; the second camera field of view angle is the camera field of view angle of the unmanned aerial vehicle when acquiring the fault image; the unmanned aerial vehicle is used to acquire video data of the photovoltaic field and annotate watermark information in each frame image of the video data; the watermark information includes the second position information; The obtaining the position coordinates of the faulty photovoltaic module based on the second position information of the second acquisition position and the fault image includes: Extract the first longitude and latitude information from the watermark information of the fault image, and extract the second longitude and latitude information from the watermark information of the auxiliary image. The auxiliary image is the next frame image adjacent to the fault image in the video data; Obtain a first azimuth angle based on the first longitude and latitude information and the second longitude and latitude information. The first azimuth angle is the included angle between the flight direction of the unmanned aerial vehicle at the second acquisition position and the first preset reference direction; Wherein, the first azimuth angle The first longitude and latitude information is (X2, Y2), and the second longitude and latitude information is (X3, Y3); Extract the first altitude information from the watermark information of the fault image; Obtain the position coordinates of the faulty photovoltaic module based on the fault image, the first longitude and latitude information, the first azimuth angle, the first altitude information, the installation height of the photovoltaic panel of the photovoltaic module, and the second camera field of view angle; Calibrate the faulty photovoltaic module in the panoramic view based on the position coordinates and the geographical coordinate range; 2. The positioning method of a faulty photovoltaic module in a photovoltaic field according to claim 1, characterized in that, The obtaining the panoramic view of the photovoltaic field includes: obtaining the panoramic view based on the unmanned aerial vehicle; The obtaining the corresponding geographical coordinate range of the photovoltaic field based on the first position information of the first acquisition position, the first camera field of view angle of the unmanned aerial vehicle, and the panoramic view includes: Obtain the coordinate range based on the size of the panoramic view, the first camera field of view angle of the unmanned aerial vehicle, and the first position information. The first position information includes the longitude, latitude, flight altitude of the unmanned aerial vehicle when shooting, and the clockwise azimuth angle between the flight direction of the unmanned aerial vehicle and the preset reference direction.

3. The method for locating a faulty photovoltaic module in a photovoltaic field according to claim 2, wherein, The obtaining the coordinate range based on the size of the panoramic view, the first camera field of view angle of the unmanned aerial vehicle, and the first position information includes: Construct a first Cartesian coordinate system, and obtain the corresponding first coordinate range of the panoramic image in the first Cartesian coordinate system according to the size of the panoramic image, the first camera field of view angle, and the flight altitude; Obtain the corresponding geographical coordinate range of the photovoltaic field according to the first coordinate range, the longitude and latitude when the UAV takes pictures, and the clockwise azimuth angle.

4. The method for positioning a faulty photovoltaic module in a photovoltaic field according to claim 3, characterized in that, The first Cartesian coordinate system takes the projection point of the drone in the panoramic view as the origin, takes the direction from left to right horizontally along the panoramic view as the positive x-axis direction, and takes the direction from bottom to top vertically along the panoramic view as the positive y-axis direction. Then the coordinates (X ′ , Y ′ ) corresponding to the boundary corner points of the panoramic view are represented as: Where, ∠FOV represents the first camera field of view angle, h represents the flight altitude, l represents the length of the panoramic image, and w represents the width of the panoramic image.

5. The positioning method of a faulty photovoltaic module in a photovoltaic field according to claim 4, characterized in that, The geographical coordinates (X1, Y1) of the boundary corner point are represented as: X1 = X′cos(θ) - Y′sin(θ) + X, Y1 = Y′cos(θ) + X′sin(θ) + Y, Where, θ represents the clockwise azimuth angle; (X, Y) represents the longitude and latitude when the UAV takes pictures.

6. The method for locating a faulty photovoltaic module in a photovoltaic field according to claim 1, characterized in that, The obtaining of the fault image includes: Obtain a component image, where the component image includes at least one photovoltaic component in the photovoltaic field; Identify the component image to determine whether the photovoltaic component in the component image is the faulty photovoltaic component. If so, determine the component image as the fault image.

7. The method for positioning a faulty photovoltaic module in a photovoltaic field according to claim 6, wherein The obtaining of the component image includes: Based on the UAV, conduct inspection and shooting of the photovoltaic field to obtain the component image, and mark watermark information in the component image.

8. The method for positioning a faulty photovoltaic module in a photovoltaic field according to claim 7, characterized in that, The obtaining of the component image includes: Based on the UAV, conduct inspection and shooting of the photovoltaic field to traverse the photovoltaic field, and obtain the video data of the photovoltaic field; Perform frame extraction on the video data to obtain the component image.

9. The method for locating a faulty photovoltaic module in a photovoltaic field according to claim 8, characterized in that, The performing of frame extraction on the video data to obtain the component image includes: Based on the ground workstation, receive the video data from the UAV, and perform frame extraction on the video data based on the ground workstation to obtain the component image.

10. The method for locating a faulty photovoltaic module in a photovoltaic field according to claim 1, characterized in that, According to the fault image, the first longitude and latitude information, the first azimuth angle, the first height information, the installation height of the photovoltaic panel of the photovoltaic component, and the second camera field of view angle, obtain the position coordinates of the faulty photovoltaic component, including: Construct a second Cartesian coordinate system, and obtain the corresponding fault coordinates of the faulty photovoltaic component in the second Cartesian coordinate system according to the size of the fault image, the second camera field of view angle of the UAV, the first height information, and the installation height of the photovoltaic panel; Obtain the position coordinates of the faulty photovoltaic component according to the fault coordinates, the first longitude and latitude information, and the first azimuth angle.

11. The method for locating a faulty photovoltaic module in a photovoltaic field according to claim 10, characterized in that, The positioning method further includes: Output the position coordinates of the faulty photovoltaic component.

12. The method for locating a faulty photovoltaic module in a photovoltaic field according to claim 6, characterized in that, The identifying of the component image to determine whether the photovoltaic component in the component image is the faulty photovoltaic component includes: Based on the OCR character recognition model, identify the component image to determine whether the photovoltaic component in the component image is the faulty photovoltaic component.

13. The method for positioning a faulty photovoltaic module in a photovoltaic field according to claim 12, wherein, The OCR character recognition model is generated based on the YOLO V5 algorithm.

Images