Photovoltaic panel installation statistical method and system based on unmanned aerial vehicle and YOLO model

Abstract

The invention discloses a photovoltaic panel installation statistical method and system based on an unmanned aerial vehicle and a YOLO model, and belongs to the technical field of photovoltaic panel installation progress management. The method comprises the steps of shooting a photovoltaic panel installation site image in real time by using an unmanned aerial vehicle; inputting a photovoltaic panel installation site image into the trained YOLO model, outputting the position and number labels of the photovoltaic panels, and obtaining the installation progress of the photovoltaic panels in real time; wherein the YOLO model training comprises the steps of constructing an image data set of a photovoltaic panel installation site, selecting a YOLO model architecture, and inputting data to carry out model training; updating the weight matrix of the YOLO model by adopting a gradient descent method; environmental information is collected by integrating a meteorological API and a built-in sensor of the unmanned aerial vehicle, features are extracted based on the environmental information, parameters of the YOLO model are adjusted, and the trained YOLO model is obtained. The installation of the photovoltaic panel is detected through the unmanned aerial vehicle image and the YOLO model, the problem that the traditional ground coverage range is limited is solved, and the efficiency of photovoltaic panel installation progress management is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of photovoltaic panel installation progress management, and in particular relates to a photovoltaic panel installation statistics method and system based on an unmanned aerial vehicle and a YOLO model. Background Art

[0002] Photovoltaic power stations usually cover a large area and have a large number of components. Traditional inspection methods mainly rely on manual inspection. However, this method is not only inefficient and costly, but also for some difficult-to-reach locations, such as photovoltaic panels near mountains and waters, traditional inspection methods can hardly achieve full coverage. In addition, with the continuous expansion of the scale of photovoltaic power stations, higher requirements are placed on the timeliness and accuracy of inspection work. Although fixed cameras can make up for the shortcomings of manual inspection to a certain extent, due to their fixed installation position and limited viewing angle, they are difficult to adapt to the complex and changing construction site environment. Especially during the construction of large-scale photovoltaic power stations, fixed monitoring systems are difficult to provide sufficient flexibility to meet the needs of dynamic changes. In photovoltaic panel installation projects, real-time and accurate statistics of the number of photovoltaic panels installed are crucial for project progress management and quality control. Traditional manual counting methods are not only time-consuming and labor-intensive, but also prone to errors. As an emerging technical means, drones have the advantages of strong mobility, high cost-effectiveness, wide coverage, high data accuracy and good security, and can play a role in multiple application fields. However, there is currently no technical means for drones to obtain images in the field of photovoltaic power station inspection.

[0003] Therefore, the present invention proposes a photovoltaic panel installation statistics method and system based on drones and the YOLO model, which realizes the automatic detection and statistics of the photovoltaic panel installation progress through high-resolution images taken by drones and deep learning technology. Summary of the invention

[0004] The purpose of the present invention is to address the problems existing in the prior art and provide a photovoltaic panel installation statistics method and system based on drones and YOLO models. The drones and YOLO models are used to detect and identify the installation progress of photovoltaic panels in real time, which not only solves the problems of low efficiency and limited coverage of traditional ground inspections, but also improves the efficiency and accuracy of photovoltaic panel installation progress management.

[0005] According to one aspect of the present specification, a photovoltaic panel installation statistics method based on a drone and a YOLO model is provided, comprising:

[0006] Use drones to capture real-time images of photovoltaic panel installation sites;

[0007] The photovoltaic panel installation site image is input into the trained YOLO model, the location and quantity labels of the photovoltaic panels are output and displayed in real time on the image, and the photovoltaic panel installation progress is obtained; wherein the YOLO model training includes:

[0008] Construct an image dataset of photovoltaic panel installation sites, perform preprocessing, and manually annotate the preprocessed image data;

[0009] Select the YOLO model architecture, initialize the model parameters, input the preprocessed image data and annotation data into the YOLO model architecture, and perform model training;

[0010] Use the gradient descent method to update the weight matrix of the YOLO model;

[0011] Based on the environmental information of the photovoltaic panel installation site image, the YOLO model parameters are adjusted to obtain a trained YOLO model.

[0012] Furthermore, data preprocessing is performed on the photovoltaic panel installation site images, including:

[0013] The wavelet transform algorithm is used to denoise the images of photovoltaic panel installation sites;

[0014] Based on the denoised image, adaptive histogram equalization is used for image enhancement;

[0015] Based on the enhanced image, the size is cropped to obtain the preprocessed photovoltaic panel installation site image.

[0016] Furthermore, manual annotation is performed, including:

[0017] Annotate the photovoltaic panel area in the image data to obtain annotation information of the photovoltaic panel area;

[0018] Based on the annotation information of the photovoltaic panel area, a position label and a quantity label are added, where the position label includes the center coordinates and bounding box of the photovoltaic panel, and the quantity label includes the number of photovoltaic panels.

[0019] Furthermore, the weight matrix of the YOLO model is updated, including:

[0020] Set the initial weight matrix using a predefined or random initialization method;

[0021] Calculate the gradient of the initial weight matrix based on the loss function;

[0022] Based on the gradient of the initial weight matrix and combined with the learning rate, the gradient descent method is used to adjust the initial weight matrix to reduce the loss function, and then the optimization is continuously iterated to obtain the updated weight matrix.

[0023] Furthermore, the method for collecting environmental information of photovoltaic panel installation site images includes:

[0024] Environmental information is collected by integrating meteorological application programming interfaces and drone built-in sensors to image the photovoltaic panel installation site.

[0025] Furthermore, the YOLO model parameters are adjusted, including:

[0026] Extract key features based on the environmental information of the photovoltaic panel installation site image, and determine whether the environment has changed based on the key features;

[0027] When the environment changes, the online learning process is triggered, and the YOLO model parameters are automatically adjusted using rule-driven, machine learning or reinforcement learning algorithms;

[0028] Use Bayesian optimization algorithm to explore the model parameter space and find the parameters of the global optimal solution;

[0029] Based on the parameters of the global optimal solution, the final YOLO model parameters are obtained through further optimization through iteration.

[0030] According to one aspect of the present specification, a photovoltaic panel installation statistics system based on a drone and a YOLO model is provided, comprising:

[0031] Acquire data module and use drones to take real-time images of photovoltaic panel installation sites;

[0032] The photovoltaic panel recognition module inputs the photovoltaic panel installation site image into the trained YOLO model, outputs the location and quantity labels of the photovoltaic panels and displays them on the image in real time to obtain the photovoltaic panel installation progress; wherein the YOLO model training includes:

[0033] Construct an image dataset of photovoltaic panel installation sites, perform preprocessing, and manually annotate the preprocessed image data;

[0034] Select the YOLO model architecture, initialize the model parameters, input the preprocessed image data and annotation data into the YOLO model architecture, and perform model training;

[0035] Use the gradient descent method to update the weight matrix of the YOLO model;

[0036] Based on the environmental information of the photovoltaic panel installation site image, the YOLO model parameters are adjusted to obtain a trained YOLO model.

[0037] According to one aspect of the present specification, there is provided an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of the photovoltaic panel installation statistical method based on a drone and a YOLO model when executing the computer program.

[0038] According to one aspect of the present specification, there is provided a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of the photovoltaic panel installation statistical method based on drones and YOLO models are implemented.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] 1. The present invention proposes a photovoltaic panel installation statistics method and system based on drones and YOLO models. The photovoltaic panel installation progress is detected by using images taken by drones and YOLO models, which solves the problems of low efficiency and limited coverage of traditional ground inspections, and improves the efficiency and accuracy of photovoltaic panel installation progress management.

[0041] 2. The present invention proposes a photovoltaic panel installation statistics method and system based on drones and YOLO models. Through drones equipped with high-definition cameras and other sensors, it can not only adapt to the complex and changeable construction site environment and provide large-scale, high-precision photovoltaic panel location and quantity information, but also can quickly obtain image data of large-area photovoltaic panel arrays at high altitudes, avoiding the problem of difficulty in reaching certain areas during manual inspections.

[0042] 3. The present invention proposes a photovoltaic panel installation statistical method and system based on drones and YOLO models. Through regular drone patrols, problems in the installation process can be discovered in a timely manner, thereby improving the construction quality and progress control level. Compared with fixed monitoring systems, drones are not restricted by geographical locations and are more flexible, making them suitable for photovoltaic power station construction of various sizes and types. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 It is a flow chart of a photovoltaic panel installation statistical method based on a drone and a YOLO model according to an embodiment of the present invention;

[0045] Figure 2A photo of photovoltaic panels installed at the installation site taken by a drone in an embodiment of the present invention;

[0046] Figure 3 This is a photovoltaic panel inspection image taken by a drone in an embodiment of the present invention at the installation site; DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] The embodiment of the present invention provides a photovoltaic panel installation statistics method based on drones and YOLO models, such as Figure 1 As shown, it includes: using a drone to take real-time images of the photovoltaic panel installation site and perform data preprocessing. The preprocessed photovoltaic panel installation site images are input into the trained YOLO model, the location and quantity labels of the photovoltaic panels are output and displayed in real time on the image, and the photovoltaic panel installation progress is obtained. Among them, the YOLO model training includes: constructing an image data set of the photovoltaic panel installation site, preprocessing, and manually annotating the preprocessed image data; selecting a YOLO model architecture, initializing model parameters, inputting the preprocessed image data and annotated data into the YOLO model architecture, and training the model; using the gradient descent method to update the weight matrix of the YOLO model; collecting environmental information by integrating the meteorological application programming interface (Application Programming Interface, API) and the built-in sensors of the drone, extracting features based on the environmental information of the photovoltaic panel installation site images, adjusting the YOLO model parameters, and obtaining a trained YOLO model.

[0049] Specifically, at the photovoltaic panel installation site, drones equipped with high-precision positioning systems are used to capture high-resolution images (4K and above), such as Figure 2As shown in the figure, an image dataset containing photovoltaic panels is constructed to ensure that the image quality is sufficient to support subsequent accurate identification and counting operations. UAVs can flexibly capture images at different heights and angles to ensure the diversity and comprehensiveness of image data. UAVs have the following advantages: strong mobility, the ability to quickly respond to mission requirements in a short period of time, not restricted by geographical conditions, and the ability to easily cross various terrain obstacles such as hills and lakes, ensuring that all photovoltaic panels can be effectively monitored. High cost-effectiveness, which can significantly reduce labor costs, and no additional infrastructure is required, reducing overall costs. Wide coverage, photovoltaic panel images can be captured over a large range at high altitudes, and inspections of large areas can be completed at one time, greatly improving work efficiency. High data accuracy, with the help of advanced equipment such as high-definition cameras, clearer and more accurate data can be obtained than ground inspections, which helps to discover potential problems and take timely measures. Good safety, reducing the risk of workers being directly exposed to dangerous environments, especially when handling high-voltage electrical equipment.

[0050] Specifically, the collected image data is preprocessed to improve the image quality and model training effect, especially optimizing the lighting changes and angle differences that are unique to drone aerial photography. The following steps are included:

[0051] 1. Use the wavelet transform algorithm to denoise the photovoltaic panel images taken by the drone to highlight the important features. Perform wavelet transform on the collected images and decompose them into wavelet coefficients of different scales. Perform threshold processing on the wavelet coefficients to remove high-frequency noise; reconstruct the image through inverse wavelet transform to obtain the denoised image.

[0052] 2. Use adaptive histogram equalization to enhance the contrast and brightness of the image, highlight the contrast between photovoltaic panels, and enhance the visibility of fine structures. Divide the image into several small blocks; perform histogram equalization on each small block; use bilinear interpolation to stitch the processed small blocks into a complete image.

[0053] 3. Crop each collected photovoltaic panel image taken by the drone into multiple 640×640 images for easy labeling and training.

[0054] Specifically, the preprocessed image data is manually annotated to mark the location, outline, and number of photovoltaic panels. Considering the changes in the perspective of drone photography, a multi-perspective annotation strategy is adopted to enhance the robustness of the model. The specific steps include:

[0055] 1. Use the annotation assistant software to annotate the photovoltaic panel area. By "drawing a frame", the boundaries of the photovoltaic panel are delineated on the image to form a rectangular frame that accurately covers the photovoltaic panel area. Generate accurate photovoltaic panel area annotation information through manual operation.

[0056] 2. Label the photovoltaic panel area with location labels and quantity labels. The location label includes the center coordinates and bounding box of the photovoltaic panel, and the quantity label includes the number of photovoltaic panels. These labels will be used for model training to ensure that the model can accurately identify and count photovoltaic panels.

[0057] Specifically, the YOLO series network architecture is selected to initialize the model parameters. The selected YOLOv8x has high detection speed and accuracy, and is suitable for real-time processing of large amounts of image data. The cross-validation method is used to adjust the model parameters, optimize the model performance, and improve accuracy and real-time performance. First, the preprocessed image data and annotated data are input into the YOLOv8x model for model training. During the training process, the model will learn how to identify and count photovoltaic panels from images. Several annotated photovoltaic panel images are selected, of which 70% are used as training sets, 20% as validation sets, and 10% as test sets. The training set data is input into the YOLOv8x model for training; during the training process, not only the validation set is used to adjust the parameters of the pre-trained model, but also the weight matrix is ​​finely adjusted to ensure that the model can better adapt to the specific application scenario of photovoltaic panel installation progress statistics provided by drones, thereby optimizing the model performance and ensuring high accuracy and real-time performance.

[0058] Specifically, the weight matrix of the YOLO series model is updated and the model is optimized, and the initial weight matrix is ​​set by a predefined or random initialization method. Using the input samples in the training data set, the weight matrices of the query, key, and value are calculated, and the predicted output is generated. The loss function, such as the mean square error (MSE) or other appropriate loss function, is calculated based on the difference between the predicted output and the true label. Based on the result of the loss function, the gradient of the loss to the query, key, and value matrices is calculated, and then the gradient of the loss to the weight matrix is ​​calculated. Using the gradient descent method or other optimization algorithms (such as Adam), the weight matrix is ​​adjusted according to the learning rate to minimize the loss function. Repeat the above steps until the convergence condition is met or the predetermined maximum number of iterations is reached, and finally a YOLO model that has been fully trained and optimized for the photovoltaic panel recognition task is obtained.

[0059] Specifically, environmental information is collected by integrating the meteorological API and the built-in sensors of the drone, including but not limited to time, weather conditions, light intensity, etc. Key features are extracted based on the environmental information to describe the characteristics of the current scene. For example, light intensity, humidity level, etc., these features will serve as an important basis for subsequent adjustment of model parameters. The reason is that light intensity will directly affect the shooting quality of drone images. When the light is good or poor, the corresponding images taken are different. Humidity level will affect the formation of clouds and fog, and will also affect the quality of drone images. In addition, the characteristics of light, humidity, etc. when shooting training images are obtained through the meteorological API, and they are used as a variable during training. In actual application, the light and humidity during actual application are also obtained through the meteorological API and input into the trained model. Then the model recognizes photovoltaic panels based on the three variables of picture, light, and humidity. This is equivalent to adding light and humidity as new variables to the training of the model, which will make the model more adaptable to different meteorological environments, such as sunny or rainy days, and can complete image recognition well.

[0060] Specifically, the online learning process is triggered according to environmental changes, and the YOLO model parameters are automatically adjusted using rule-driven, machine learning or reinforcement learning algorithms. The Bayesian optimization algorithm is used to explore the parameter space. First, preliminary parameter performance data is collected through the initial sampling points, and then a Gaussian process is established as a proxy model to approximate the objective function. The maximum expected improvement acquisition function is then used to select the next most promising parameter combination for evaluation. By continuously iterating and updating the proxy model, the parameter space is efficiently explored, and the global optimal parameters are finally found. Through experimental verification and iterative improvement, the algorithm logic is continuously optimized to ensure that the ideal adaptive effect can be achieved in different environments. Specifically, the experimental verification was carried out in several representative photovoltaic power stations as pilots, recording the system response and performance indicator changes in different environments, and then making targeted adjustments to the model.

[0061] Specifically, the images taken in real time at the photovoltaic panel installation site are input into the trained YOLO series model. The trained YOLO series model outputs the location and quantity information of the photovoltaic panels, including the center coordinates, bounding boxes, and quantity labels of the photovoltaic panels, such as Figure 3 As shown. The recognition results are displayed in real time on the image, and project managers can make judgments and processes based on the annotation information. Specifically, OpenCV is used to load the photovoltaic panel detection video taken by the drone and obtain the video stream frame; the trained YOLOv8x model is called to detect the video frame, thereby outputting the location and number of photovoltaic panels, and the recognition results are displayed in real time on the image, providing intuitive progress feedback for project managers.

[0062] Specifically, export the trained YOLO series model to a format suitable for the terminal device, such as ONNX, TFLite, or TensorRT. Transfer the exported model file to the terminal device. Install the necessary operating environment and dependent libraries on the terminal device. Load the model file on the terminal device and configure the model parameters. Run the model on the terminal device, process the photovoltaic panel images taken by the drone in real time, and output the recognition results. Project managers can view the installation progress of photovoltaic panels in real time through the terminal device to improve management efficiency.

[0063] The implementation basis of each embodiment of the present invention is realized by programmed processing through a device with a processor function. Therefore, in engineering practice, the technical solutions and functions of each embodiment of the present invention are encapsulated into various modules. Based on this reality, on the basis of the above embodiments, an embodiment of the present invention provides an intelligent statistical system for photovoltaic panel installation progress based on drones and YOLO models, which is used to execute a photovoltaic panel installation statistical method based on drones and YOLO models in the above method embodiments.

[0064] The system includes: a data acquisition module, which uses a drone to take real-time images of a photovoltaic panel installation site and performs data preprocessing; a photovoltaic panel recognition module, which inputs the preprocessed photovoltaic panel installation site images into a trained YOLO model, outputs the location and quantity labels of the photovoltaic panels and displays them on the images in real time, and obtains the photovoltaic panel installation progress; wherein the YOLO model training includes: constructing an image data set of the photovoltaic panel installation site, performing preprocessing, and manually annotating the preprocessed image data; selecting a YOLO model architecture, initializing model parameters, inputting the preprocessed image data and annotated data into the YOLO model architecture, and performing model training; using a gradient descent method to update the weight matrix of the YOLO model; collecting environmental information by integrating a meteorological API and built-in sensors of the drone, extracting features based on the environmental information of the photovoltaic panel installation site images, and adjusting the YOLO model parameters to obtain a trained YOLO model.

[0065] The photovoltaic panel installation progress intelligent statistical system based on drones and YOLO models provided in the embodiments of the present invention aims at the situation that traditional manual statistical methods are time-consuming and labor-intensive and prone to errors. By adopting several modules in the system and using drones equipped with high-definition cameras and other sensors, the problems of low efficiency and limited coverage of traditional ground inspections are solved, thereby improving the efficiency and accuracy of photovoltaic panel installation progress management.

[0066] Based on the same inventive concept as the aforementioned embodiment, an embodiment of the present invention further provides an electronic device, including a memory and a processor, the memory being used to store computer-executable instructions, and the processor being used to execute computer-executable instructions, so as to implement an intelligent statistical method for photovoltaic panel installation progress based on a drone and a YOLO model as proposed in the aforementioned embodiment.

[0067] The embodiment of the present invention also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by the processor, it is used to detect the progress of photovoltaic panel installation through images taken by drones and YOLO models, improve the efficiency and accuracy of photovoltaic panel installation progress management, and improve the construction quality and progress control level. The storage medium can be any non-volatile storage device such as a hard disk, a solid-state hard disk, a flash drive, an optical disk, etc., for storing computer program codes and necessary data files, and the stored computer program includes: a data acquisition module and a photovoltaic panel identification module.

[0068] Finally, it should be pointed out that the above specific embodiments are only representative examples of the present invention. Obviously, the present invention is not limited to the above specific embodiments, and there are many variations. Any simple modification, equivalent changes and modifications made to the above specific embodiments based on the technical essence of the present invention should be considered to belong to the protection scope of the present invention.

Claims

1. A photovoltaic panel installation statistical method based on drones and YOLO model, characterized in that: include: Use drones to capture real-time images of photovoltaic panel installation sites; Input the photovoltaic panel installation site image into the trained YOLO model, output the location and quantity labels of the photovoltaic panels, and obtain the photovoltaic panel installation progress; wherein the YOLO model training includes: Construct an image dataset of photovoltaic panel installation sites, perform preprocessing, and manually annotate the preprocessed image data; Select the YOLO model architecture, initialize the model parameters, input the preprocessed image data and annotation data into the YOLO model architecture, and perform model training; Use the gradient descent method to update the weight matrix of the YOLO model; Based on the environmental information of the photovoltaic panel installation site image, the YOLO model parameters are adjusted to obtain a trained YOLO model.

2. According to claim 1, a photovoltaic panel installation statistical method based on drones and YOLO model is characterized in that: Data preprocessing is performed on the photovoltaic panel installation site images, including: The wavelet transform algorithm is used to denoise the images of photovoltaic panel installation sites; Based on the denoised image, adaptive histogram equalization is used for image enhancement; Based on the enhanced image, the size is cropped to obtain the preprocessed photovoltaic panel installation site image.

3. The photovoltaic panel installation statistical method based on drone and YOLO model according to claim 1 is characterized in that: Perform manual annotation, including: Annotate the photovoltaic panel area in the image data to obtain annotation information of the photovoltaic panel area; Based on the annotation information of the photovoltaic panel area, a position label and a quantity label are added, where the position label includes the center coordinates and bounding box of the photovoltaic panel, and the quantity label includes the number of photovoltaic panels.

4. The photovoltaic panel installation statistical method based on drone and YOLO model according to claim 1 is characterized in that: Update the weight matrix of the YOLO model, including: Set the initial weight matrix using a predefined or random initialization method; Calculate the gradient of the initial weight matrix based on the loss function; Based on the gradient of the initial weight matrix and combined with the learning rate, the gradient descent method is used to adjust the initial weight matrix to reduce the loss function, and then the optimization is continuously iterated to obtain the updated weight matrix.

5. The photovoltaic panel installation statistical method based on drone and YOLO model according to claim 1, characterized in that: The method for collecting environmental information of photovoltaic panel installation site images comprises: Environmental information is collected by integrating meteorological application programming interfaces and drone built-in sensors to image the photovoltaic panel installation site.

6. The photovoltaic panel installation statistical method based on drone and YOLO model according to claim 1, characterized in that: Adjust the YOLO model parameters, including: Extract key features based on the environmental information of the photovoltaic panel installation site image, and determine whether the environment has changed based on the key features; When the environment changes, the online learning process is triggered, and the YOLO model parameters are automatically adjusted using rule-driven, machine learning or reinforcement learning algorithms; Use Bayesian optimization algorithm to explore the model parameter space and find the parameters of the global optimal solution; Based on the parameters of the global optimal solution, the final YOLO model parameters are obtained through further optimization through iteration.

7. A photovoltaic panel installation statistics system based on drones and YOLO model, characterized in that: include: Acquire data module and use drones to take real-time images of photovoltaic panel installation sites; The photovoltaic panel recognition module inputs the photovoltaic panel installation site image into the trained YOLO model, outputs the location and quantity labels of the photovoltaic panels and displays them on the image in real time to obtain the photovoltaic panel installation progress; wherein the YOLO model training includes: Construct an image dataset of photovoltaic panel installation sites, perform preprocessing, and manually annotate the preprocessed image data; Select the YOLO model architecture, initialize the model parameters, input the preprocessed image data and annotation data into the YOLO model architecture, and perform model training; Use the gradient descent method to update the weight matrix of the YOLO model; Based on the environmental information of the photovoltaic panel installation site image, the YOLO model parameters are adjusted to obtain a trained YOLO model.

8. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the photovoltaic panel installation statistical method based on the drone and the YOLO model are implemented as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the photovoltaic panel installation statistical method based on drone and YOLO model described in any one of claims 1 to 6 are implemented.

Images