CN116809578A - UAV-based photovoltaic panel cleaning method, device and UAV - Google Patents

Abstract

The disclosure relates to a photovoltaic panel cleaning method and device based on an unmanned aerial vehicle. Wherein the method comprises the following steps: determining the installation position information of the first photovoltaic panel in response to the cleaning operation of the unmanned aerial vehicle entering the first photovoltaic panel, and determining the position information of a central point of the unmanned aerial vehicle to be hovered according to the installation position information; according to the position information of the central point, controlling the unmanned aerial vehicle to stop at the central point to be hovered; controlling a camera on the unmanned aerial vehicle to acquire an image of the first photovoltaic panel so as to acquire the image of the first photovoltaic panel; determining the position and the type of the pollutant on the first photovoltaic panel according to the image; according to the position and the kind of pollutant, combine multiple clean structure to carry out cleaning treatment to first photovoltaic board, wherein, multiple clean structure is the different clean structure of setting on unmanned aerial vehicle. The cleaning efficiency can be improved, the cost is saved, the safety problem of manual cleaning is avoided, and the targeted cleaning of pollutants can be realized.

Description

UAV-based photovoltaic panel cleaning method, device and UAV

Technical field

The present disclosure relates to the field of solar photovoltaic technology, and in particular to a drone-based photovoltaic panel cleaning method, device, drone and storage medium.

Background technique

As the proportion of society's demand for clean energy continues to increase, photovoltaic panels account for an increasing proportion of solar power generation, which is of great significance to human development. As we all know, photovoltaic panels are generally placed in remote areas or on the roofs of buildings, where they accumulate dust, leaves, bird droppings, etc. over a long period of time, seriously affecting power generation efficiency. However, manual cleaning of photovoltaic panels is inefficient and poses safety risks.

Contents of the invention

The present disclosure provides a drone-based photovoltaic panel cleaning method, device, drone and storage medium.

According to a first aspect of an embodiment of the present disclosure, a drone-based photovoltaic panel cleaning method is provided, including:

In response to the drone entering the cleaning operation of the first photovoltaic panel, determining the installation location information of the first photovoltaic panel, and determining the location information of the center point where the drone is to hover based on the installation location information;

According to the center point position information, control the drone to dock at the center point to be hovered;

Control the camera on the drone to collect images of the first photovoltaic panel to obtain an image of the first photovoltaic panel;

Determine the location and type of contaminants on the first photovoltaic panel based on the image;

According to the location and type of the contaminants, the first photovoltaic panel is cleaned using a combination of multiple cleaning structures, where the multiple cleaning structures are different cleaning structures provided on the drone.

According to a second aspect of the embodiment of the present disclosure, a drone-based photovoltaic panel cleaning device is provided, including:

The first determination module is used to determine the installation position information of the first photovoltaic panel when the drone enters the cleaning operation of the first photovoltaic panel, and determines that the drone is to hover based on the installation position information. center point location information;

A first control module, configured to control the drone to dock at the center point to be hovered based on the center point position information;

The second control module is used to control the camera on the drone to collect images of the first photovoltaic panel to obtain images of the first photovoltaic panel;

a second determination module, configured to determine the location and type of contaminants on the first photovoltaic panel based on the image;

A cleaning processing module, used to clean the first photovoltaic panel in combination with a variety of cleaning structures according to the location and type of the contaminants, wherein the multiple cleaning structures are different cleaning structures provided on the drone. Clean structure.

According to a third aspect of the embodiments of the present disclosure, a drone is provided, including:

UAV body;

A camera, which is arranged at the front end of the drone body and is used for image collection;

A variety of different cleaning structures, the multiple cleaning structures are arranged at different positions of the drone body;

At least one processor connected to the camera and the plurality of cleaning structures respectively;

A memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to cause the at least one process The device is capable of executing the method described in the first aspect.

According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided for storing instructions. When the instructions are executed by a processor, the method described in the first aspect is implemented.

The technical solution provided by the embodiments of the present disclosure can include the following beneficial effects: using drones to clean photovoltaic panels can improve cleaning efficiency and save costs; for photovoltaic panels installed in remote areas, using drones to clean photovoltaic panels can reduce safety hazards. This avoids the safety issues associated with manual cleaning. In addition, this disclosure uses the camera on the drone to accurately detect the type and location of pollutants on the photovoltaic panel, achieve targeted cleaning of pollutants, and adopt targeted solutions for different pollutants to further improve efficiency and reduce costs. In addition, by controlling the drone through programs, more precise control can be achieved, which is stable, safe and reliable.

It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only, and do not limit the present disclosure.

Description of the drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description serve to explain the principles of the invention.

Figure 1 is a schematic diagram of the application environment of a drone-based photovoltaic panel cleaning method proposed in this disclosure.

Figure 2 is a flow chart of a drone-based photovoltaic panel cleaning method provided by an embodiment of the present disclosure.

Figure 3 is a flow chart of another drone-based photovoltaic panel cleaning method provided by an embodiment of the present disclosure.

Figure 4a is a flow chart of yet another drone-based photovoltaic panel cleaning method provided by an embodiment of the present disclosure.

Figure 4b is an example diagram of the mapping relationship between the contaminants on the photovoltaic panel in the image and the contaminants on the actual photovoltaic panel provided by the embodiment of the present disclosure.

Figure 5 is a block diagram of a drone-based photovoltaic panel cleaning device provided by an embodiment of the present disclosure.

Figure 6 is a block diagram of another drone-based photovoltaic panel cleaning device provided by an embodiment of the present disclosure.

Figure 7 is a schematic diagram of a drone provided by an embodiment of the present disclosure.

Detailed ways

Exemplary embodiments will be described in detail herein, examples of which are illustrated in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings refer to the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the invention. Rather, they are merely examples of apparatus and methods consistent with aspects of the invention as detailed in the appended claims.

As the proportion of society's demand for clean energy continues to increase, photovoltaic panels account for an increasing proportion of solar power generation, which is of great significance to human development. As we all know, photovoltaic panels are generally placed in remote areas or on the roofs of buildings, where they accumulate dust, leaves, bird droppings, etc. over a long period of time, seriously affecting power generation efficiency. However, manual cleaning of photovoltaic panels is inefficient and poses safety risks.

To this end, the present disclosure provides a UAV-based photovoltaic panel cleaning method. In this method, when the UAV enters the cleaning operation of a certain photovoltaic panel along the cleaning planning path, it can be installed according to the installation of the photovoltaic panel. The position information controls the drone to dock at the center point to be hovered, and determine the location and type of contaminants on the photovoltaic panel based on the image of the photovoltaic panel, and based on the location and type of contaminants, combine a variety of cleaning structures to clean the Photovoltaic panels are cleaned. Therefore, using drones to clean photovoltaic panels can improve cleaning efficiency and save costs. For photovoltaic panels installed in remote areas, using drones to clean can reduce safety hazards and avoid the safety problems of manual cleaning. In addition, the present disclosure uses a deep learning model to accurately detect the types and locations of pollutants on photovoltaic panels, achieve targeted cleaning of pollutants, and adopt targeted solutions for different pollutants to improve efficiency and reduce costs. In addition, the drone can be controlled through programs to achieve more precise control, which is stable, safe and reliable. In order to facilitate a detailed description of the disclosed solution, the application environment in the embodiments of the present disclosure will be introduced below with reference to the accompanying drawings.

Please refer to Figure 1 , which is a schematic diagram of the application environment of a drone-based photovoltaic panel cleaning method proposed in this disclosure. As shown in Figure 1, the drone-based photovoltaic panel cleaning method provided by the present disclosure can be applied to the drone 100, so that the drone 100 can clean along the cleaning planning path of the photovoltaic array 200. When the cleaning planning path enters the cleaning operation of a certain photovoltaic panel, the photovoltaic panel can be independently cleaned according to the location and type of the detected pollutants on the photovoltaic panel, combined with a variety of cleaning structures on the drone 100; the photovoltaic panel The panel cleaning method can also be applied to the server. The server controls the drone to dock at the center point to be hovered, and determines the location and type of contaminants on the photovoltaic panel based on the image of the photovoltaic panel. and types, remotely control a variety of cleaning structures on the drone to clean the photovoltaic panels.

It should be noted that the drone 100 in the embodiment of the present disclosure is used to clean the photovoltaic panels in the photovoltaic array. In some embodiments, the drone 100 in the embodiment of the present disclosure may include a drone body, a camera, and a variety of different cleaning structures. The UAV body may include but is not limited to a UAV casing, a flight control system, a positioning system and a communication module. The positioning system may include but is not limited to a GPS (Global Positioning System) positioning system. The communication module Modules include but are not limited to Bluetooth modules, etc.

In some embodiments, the plurality of different cleaning structures may include, but are not limited to, a water-washing cleaning structure. The water-washing cleaning structure may include, but is not limited to, a water tank, a pump body, and a water gun. The water tank may be installed on the drone through an arm. At the bottom of the body, the pump body is set on the drone body. The water tank is connected to the water inlet of the pump body. The water outlet of the pump body is connected to the water gun through the water supply pipe. The pump body draws water from the water tank to the water gun to complete the flushing operation. , which allows the drone to use the water washing cleaning structure to perform water washing operations on photovoltaic panels. As an example, the pump body can be a brushless diaphragm pump, which has the characteristics of high pressure, large flow, long life, and low noise, making it easier to accurately control the flow of water.

In some embodiments, the plurality of different cleaning structures may also include, but are not limited to, cleaning robot suction cups. The suction cup of the cleaning robot is arranged at the bottom of the drone body. The suction cup of the cleaning robot can include but is not limited to a motor. The suction cup of the cleaning robot can use the high-speed rotation of the motor to form a vacuum area to absorb dust, which makes it easier for the drone to use the suction cup of the cleaning robot to carry out operations. Adsorption of accumulated dust. The suction cup of the cleaning robot in the embodiment of the present disclosure adopts the principle of a sweeping robot and is made of rubber, thereby reducing damage to photovoltaic panels.

In some embodiments, the plurality of different cleaning structures may also include, but are not limited to, particulate matter cleaning structures. The particulate matter cleaning structure may include, but is not limited to, a brush, a driving motor and a rotating shaft. The driving motor is arranged on the drone body. The driving motor and the rotating shaft drive the brush to achieve cleaning operations, which allows the drone to use the particulate matter cleaning structure. Clean sand and other particles from photovoltaic panels.

It is worth noting that in some embodiments, the drone 100 in the embodiment of the present disclosure may also include but is not limited to an ultrasonic ranging module and a lidar (such as a mini lidar, also called a small lidar). Among them, the ultrasonic ranging module is set at the front end of the drone body, and the ultrasonic ranging module is used to measure the distance to the ground; the lidar is set in the same vertical direction as the center of the drone body, and the lidar is used to scan the photovoltaic plate. In one implementation, the ultrasonic ranging module can be used to measure the ground distance and determine the hovering height based on the return time of the sound wave and the actual temperature. The drone can obtain the geodetic coordinate system coordinates of the drone's center point based on the positioning system, use lidar to scan the photovoltaic panels, process the point cloud data, cluster and segment the shape of the photovoltaic panels, and select the four largest extreme point coordinates. Then the coordinate points are converted to the geodetic coordinate system, so that the drone and the photovoltaic panel are in the same coordinate system. The drone uses the four extreme point coordinates to determine the location information of the center point to be hovered.

It should also be noted that the present disclosure can also provide a background visual management platform, which can realize remote management and control of the UAV 100, including but not limited to the setting of the cleaning planning path of the photovoltaic array, etc. Here, This disclosure does not limit this.

The drone-based photovoltaic panel cleaning method, device, drone and storage medium provided by the present disclosure will be introduced in detail below with reference to the accompanying drawings.

Please refer to Figure 2. Figure 2 is a flow chart of a drone-based photovoltaic panel cleaning method provided by an embodiment of the present disclosure. As shown in Figure 2, this method can be applied to drones, which means that this method can The method can be executed by a drone, or the method can also be executed by a server, that is, the server controls the drone to clean the photovoltaic panels. The server can be independent of the drone. It should be understood that in some embodiments, the server The server can be installed on the drone, the server can be a drone processor, or the server can be a system server in a predetermined area, which is not limited in this disclosure. The method may include, but is not limited to, the following steps.

In step 201, in response to the drone entering the cleaning operation of the first photovoltaic panel, the installation position information of the first photovoltaic panel is determined, and based on the installation position information, the position information of the center point where the drone is to be hovered is determined.

In this embodiment of the present disclosure, the first photovoltaic panel can be understood as the first photovoltaic panel to be cleaned in the photovoltaic array, or it can also be a non-first photovoltaic panel to be cleaned in the photovoltaic array. In one implementation, the drone can clean the photovoltaic panels in the photovoltaic array along a preset cleaning planning path. For example, when the drone obtains the start instruction, the drone determines the first photovoltaic panel to be cleaned based on the cleaning planning path. At this time, the first photovoltaic panel to be cleaned can be called the first photovoltaic panel, and the control unit The human-machine enters the cleaning operation of the first photovoltaic panel to be cleaned; after the drone completes the cleaning operation of the first photovoltaic panel to be cleaned, the drone determines the next photovoltaic panel to be cleaned based on the cleaning planning path , at this time, the next photovoltaic panel to be cleaned can be called the first photovoltaic panel, and the drone is controlled to enter the cleaning operation of the photovoltaic panel.

In the embodiment of the present disclosure, when the drone enters the cleaning operation of the first photovoltaic panel, the drone can determine the first hovering height by measuring the ground distance, and determine when the drone is at the first hovering height. The geodetic coordinate system coordinates of the UAV are combined with the lidar to determine the installation position information of the first photovoltaic panel, so that the UAV can determine the center point position of the UAV to hover based on the installation position information. information.

In one implementation, the drone uses an ultrasonic ranging module to measure the distance on the ground based on the return time of the sound wave to determine the first hovering height. Ultrasonic ranging formula: L = C × T, where C is the propagation speed of sound waves in the air, which is 340m/s, and T is half of the propagation time. In order to improve the measurement accuracy, an ultrasonic ranging module with temperature compensation can be used to measure the distance on the ground. For example, the drone uses the ultrasonic ranging module to measure the distance on the ground based on the return time of the sound wave and the actual temperature to determine the first hovering height. .

In embodiments of the present disclosure, after determining the first hovering height of the drone, the installation position of the first photovoltaic panel may be determined first, so that the position of the drone can be calibrated based on the installation position of the first photovoltaic panel to determine The center point where the drone hovers. In one implementation, the coordinates of the geodetic coordinate system when the UAV is at the first hovering height can be determined, and the lidar on the UAV is controlled to scan the first photovoltaic panel to obtain lidar point cloud data; according to the lidar point Cloud data determines the coordinates of the first photovoltaic panel in the lidar coordinate system, and converts the coordinates of the first photovoltaic panel in the lidar coordinate system to the geodetic coordinate system to obtain the installation location information of the first photovoltaic panel.

As an example, after the drone hovers at the first hovering height, the geodetic coordinate system coordinates of the center point of the drone can be obtained according to the positioning system on the drone (such as the GPS system), and the laser on the drone can be used to The radar scans the first photovoltaic panel to obtain lidar point cloud data, and processes the lidar point cloud data to segment the shape of the first photovoltaic panel through clustering, selects the largest four extreme point coordinates, and then divides the The coordinates of the four extreme points are converted to the geodetic coordinate system, and the coordinates of the four extreme points converted to the geodetic coordinate system are determined as the installation position information of the first photovoltaic panel, so that the UAV and the first photovoltaic panel are in the same location. Coordinate System.

In an embodiment of the present disclosure, after the installation position information of the first photovoltaic panel is determined, the drone position calibration may be performed according to the installation position information of the first photovoltaic panel to determine the center point of the drone hovering. In one implementation, the coordinates of the center point to be corrected can be determined based on the installation position information of the first photovoltaic panel; the calibration values of each coordinate component of the drone in the geodetic coordinate system can be determined; and the calibration values of each coordinate component of the UAV can be determined. The corrected center point coordinates are calibrated, and the coordinates obtained after calibration are determined as the center point position information of the drone to be hovered.

As an example, the center point coordinates of the first photovoltaic panel are determined according to the installation position information of the first photovoltaic panel, and the center point coordinates of the first photovoltaic panel are determined as the center point coordinates to be corrected. Determine the calibration value of each coordinate component of the UAV in the geodetic coordinate system, calibrate the coordinates of the center point to be corrected based on the calibration value of each coordinate component, and determine the coordinates obtained after calibration as the position of the center point to be hovered by the UAV. information. For example, the coordinates of the four extreme points converted to the geodetic coordinate system are (x ₁ ,y ₁ ,z ₁ ), (x ₂ ,y ₂ ,z ₂ ), (x ₃ ,y ₃ ,z ₃ ), (x ₄ , y ₄ , z ₄ ), take the arithmetic mean of the four coordinate points as the center point of the first photovoltaic panel The center point/> It can be used as the coordinates of the center point to be corrected. The calibration value of the X-axis coordinate component of the UAV in the geodetic coordinate system is -3, the calibration value of the Y-axis coordinate component in the geodetic coordinate system is -3, and the calibration value of the Z-axis coordinate component in the geodetic coordinate system is 3. , the unit is meters (m), then sum the coordinates of the center point to be corrected and the calibration values of the corresponding coordinate components to obtain the coordinates after calibration/> As the center point location information for the drone to hover.

In step 202, the drone is controlled to dock at the center point to be hovered based on the center point location information.

In one implementation, based on the position information of the center point to be hovered by the UAV, the flight control system on the UAV body is used to control the UAV to dock at the center point to be hovered.

In step 203, the camera on the drone is controlled to collect images of the first photovoltaic panel to obtain an image of the first photovoltaic panel.

In one implementation, after the drone docks at the center point to be hovered, the camera on the drone can be controlled to collect images. For example, the camera can rotate 180 degrees, take pictures of the first photovoltaic panel every 30 degrees for three rounds, and obtain image data of the first photovoltaic panel according to the preset scanning range. The image data can be stored to facilitate detection using the image data. Whether there are contaminants on the first photovoltaic panel.

In step 204, the location and type of contaminants on the first photovoltaic panel are determined based on the image.

In embodiments of the present disclosure, a trained deep learning model can be used to predict the image to detect whether there are contaminants on the first photovoltaic panel. If it is detected that there are no contaminants on the first photovoltaic panel, then The next photovoltaic panel to be cleaned can be determined according to the cleaning planning path, and the drone can be controlled to enter the cleaning operation of the next photovoltaic panel. If a contaminant is detected on the first photovoltaic panel, and the deep learning model outputs the location and type of the contaminant. The position of the pollutant output by the deep learning model can be understood as the position of the pollutant on the first photovoltaic panel in the image. The types of pollutants may include at least one of the following: accumulated dust, particulate matter (sand, leaves, etc.), and lumps (such as bird droppings, etc.).

In step 205, the first photovoltaic panel is cleaned according to the location and type of the contaminants by combining multiple cleaning structures, where the multiple cleaning structures are different cleaning structures provided on the drone.

In one implementation, a corresponding cleaning plan can be determined according to the type of pollutants; a cleaning structure corresponding to the cleaning plan is selected from a variety of cleaning structures; according to the location of the pollutants, the selected cleaning structure is used to clean the first photovoltaic The board is cleaned.

It is worth noting that since the positions of different cleaning structures on the drone will be different, when cleaning the first photovoltaic panel, it is necessary to determine the cleaning structure of the drone based on the location of the contaminants and the selected cleaning structure. When the pollutant is detected, the drone is controlled to dock at the operating docking point, and the selected cleaning structure is controlled to clean the first photovoltaic panel. As an example, the drone hovers at the center point to be hovered, and after determining the location and type of the pollutant, the corresponding cleaning plan can be determined based on the type of pollutant, and the cleaning structure can be selected from a variety of cleaning structures. Clean structure corresponding to the scheme. According to the location of the pollutant and the selected cleaning structure, the position of the drone is fine-tuned to determine the operating stop point when the drone cleans the pollutant, so that the drone can be located at the operating stop point to clean the first photovoltaic The panels are cleaned so that they can be cleaned to the greatest extent without damaging the photovoltaic panels.

By implementing the embodiments of the present disclosure, the use of drones to clean photovoltaic panels can improve cleaning efficiency and save costs; for photovoltaic panels installed in remote areas, using drones to clean photovoltaic panels can reduce safety hazards and avoid the safety problems of manual cleaning. . In addition, this disclosure uses the camera on the drone to accurately detect the type and location of pollutants on the photovoltaic panel, achieve targeted cleaning of pollutants, and adopt targeted solutions for different pollutants to further improve efficiency and reduce costs. In addition, by controlling the drone through programs, more precise control can be achieved, which is stable, safe and reliable.

Please refer to Figure 3. Figure 3 is a flow chart of another drone-based photovoltaic panel cleaning method provided by an embodiment of the present disclosure. As shown in Figure 3, the method may include but is not limited to the following steps.

In step 301, in response to the drone entering the cleaning operation of the first photovoltaic panel, the installation position information of the first photovoltaic panel is determined, and based on the installation position information, the position information of the center point where the drone is to be hovered is determined.

In the embodiments of the present disclosure, step 301 can be implemented in any manner in the embodiments of the present disclosure. The embodiments of the present disclosure do not limit this and will not be described again.

In step 302, the drone is controlled to dock at the center point to be hovered based on the center point location information.

In the embodiments of the present disclosure, step 302 can be implemented in any manner in the embodiments of the present disclosure. The embodiments of the present disclosure do not limit this and will not be described again.

In step 303, the camera on the drone is controlled to collect images of the first photovoltaic panel to obtain an image of the first photovoltaic panel.

In the embodiments of the present disclosure, step 303 can be implemented in any manner in the embodiments of the present disclosure. The embodiments of the present disclosure do not limit this and will not be described again.

In step 304, the image is input to a preset contaminant detection model, and the type of contaminant on the first photovoltaic panel and the position of the contaminant on the first photovoltaic panel in the image are obtained.

Among them, in the embodiment of the present disclosure, the pollutant detection model has learned the ability to predict the probability and location of various types of pollutants on the photovoltaic panel.

It should be noted that the pollutant detection model can be trained using a deep learning model. In one implementation, training data can be obtained, and the training data can include image samples of photovoltaic panels and labels of the image samples. The labels can include whether there are contaminants, the location of the contaminants, the types of contaminants, etc. The image samples may include images of photovoltaic panels with contaminants present and images of photovoltaic panels without contaminants present. Input the training data into the deep learning model. The deep learning model extracts features from the image samples and makes predictions based on the extracted features to output prediction results. A loss value is generated based on the prediction results and labels, and the deep learning model is trained based on the loss value. In this way, the deep learning model is continuously trained until the end condition is met, and the model at this time is determined as a pollutant detection model, so that the pollutant detection model has learned to predict the probability and location of various types of pollutants on the photovoltaic panel. Ability.

For example, take photos of photovoltaic panel pollutants in different environments, and perform data enhancement through cropping, stretching, rotation, translation, etc. Then the data set is produced and labels are produced for classifying pollutants. The data set format adopts the coco data set format to facilitate YOLO (target detection model) training and learning. Optionally, the ratio of data set training and testing can be 9:1. Build a neural network based on the collected data, conduct deep learning pollutant feature extraction, learn the characteristics of pollutants, conduct model training, and obtain the optimal model to improve the accuracy of pollutant detection.

For example, the model can use the one-stage improved version of YOLOv5, which can achieve almost real-time levels and maintain accuracy comparable to two-stage, which makes it easier for drones to clean photovoltaic panels and detect contaminants on photovoltaic panels. Sensitive scenes are at a premium. The input to the model can be a data set of pollutant images after data augmentation using optimal data augmentation with an automatic learning data strategy. The backbone network of this model uses slicing operations to change image features, and extracts features after a convolution, normalization operation and activation function. The CSP architecture module is used to learn the residual features, and then the maximum pooling layer operation is performed to improve the receptive field. Then the FPN (Feature Pyramid Network) + PAN (Pixel Aggregation Network) module is used to extract network-level features and multi-scale positioning capabilities respectively. Finally, the feature map is output. The model loss function uses the loss function of YOLOv5. The number of training rounds is set to 300 rounds, the batch-size (batch processing size) is set to 16, and the number-works (number of central processing unit CPU threads) is eight-thread training. Finally, the training model is obtained and the accuracy of the test model reaches more than 90% as the final result of the training. The final trained model is determined as a pollutant detection model and can be applied to UAVs. It is worth noting that the CSP structure of YOLOv5s is to divide the original input into two branches, perform convolution operations respectively to reduce the number of channels by half, then perform Bottleneck*N operation on one branch, and then concat (connect) the two branches, making BottlenneckCSP The input and output are the same size so that the model can learn more features. It should be noted that the pollutant detection model can be pre-trained on electronic equipment (such as a server), so that the trained model can be deployed on the drone.

In step 305, according to the position of the contaminant on the first photovoltaic panel in the image, a coordinate system transformation is performed according to the proportion of the contaminant in the image and the first photovoltaic panel to determine the position of the contaminant on the first photovoltaic panel.

In one implementation, the actual position of the pollutant in the geodetic coordinate system can be converted according to the position of the pollutant on the first photovoltaic panel in the image and the proportion of the image.

In step 306, the first photovoltaic panel is cleaned according to the location and type of the contaminants by combining multiple cleaning structures, where the multiple cleaning structures are different cleaning structures provided on the drone.

In the embodiments of the present disclosure, step 306 can be implemented in any manner in the embodiments of the present disclosure. The embodiments of the present disclosure do not limit this and will not be described again.

In some embodiments, after all the contaminants on the first photovoltaic panel are cleaned, or it is determined based on the image that there are no contaminants on the first photovoltaic panel, the cleaning planning path of the photovoltaic array where the first photovoltaic panel is located is determined; according to The cleaning planning path determines the next photovoltaic panel to be cleaned; the drone is controlled to enter the cleaning operation of the next photovoltaic panel to be cleaned, and the next photovoltaic panel to be cleaned is regarded as the new first photovoltaic panel, and returns to step 301, and so on. , until the cleaning operation of all photovoltaic panels in the photovoltaic array is completed, the drone can return to the preset position. For example, the preset position can be the initial position when the drone receives startup, or it can also be the cleaning planned path The initial position is not limited in this disclosure.

In order to facilitate those skilled in the art to understand the present disclosure more clearly, a detailed introduction will be provided below.

As shown in Figure 4a, the UAV obtains instructions (such as cleaning instructions) (step 401) and starts. The starting position can be directly in front of the photovoltaic array. The ultrasonic ranging module is used to measure the ground based on the return time of the sound waves and the actual temperature. distance to determine the hover height. Calibrate the position of the drone based on the installation position of the photovoltaic panels to determine the center point of the drone's hovering. For example, the drone obtains the geodetic coordinate system coordinates of the drone's center point based on GPS positioning, uses lidar to scan the photovoltaic panels, processes point cloud data, clusters and segments the shape of the photovoltaic panels, and selects the four largest extreme points. coordinates, and then the coordinate points are converted to the geodetic coordinate system so that the drone and the photovoltaic panel are in the same coordinate system. The arithmetic average of the four extreme point coordinates is used as the center point of the photovoltaic panel. The center point coordinates of the photovoltaic panel are calibrated according to the calibration value of each coordinate component. The calibrated coordinates are used as the center stop point of the UAV ( Step 402). Afterwards, the camera rotates 180 degrees and takes pictures of photovoltaic panels every 30 degrees for three rounds. The photovoltaic panel number image data is obtained according to the preset scanning range (step 403). After the data is stored, the trained deep learning model is used to detect pollutants. (step 404).

When it is detected that there is no contaminant on the photovoltaic panel, the next photovoltaic panel is detected, that is, the cleaning operation of the next photovoltaic panel to be cleaned is entered. For example, the drone selects the docking center point for the next photovoltaic panel to be cleaned. When contaminants are detected on the photovoltaic panel, the position of the contaminant on the photovoltaic panel in the image is converted to the actual location of the contaminant on the photovoltaic panel according to the image proportion. For example, in an image taken at 30 degrees, the pollutants are proportional to the photovoltaic panel, and the photovoltaic panel is also proportional to the size of the image. The position is determined based on the proportional relationship mapped to the actual photovoltaic panel. The pollutants in the data are converted to the point cloud coordinate system according to the proportion of the image, and then transferred to the geodetic coordinate system. The center point of the pollutant target is input to the UAV flight control system. The UAV moves according to the point position, and then starts Clean. As shown in Figure 4b, the relative position of the actual pollutant on the photovoltaic panel is determined through the proportion and relative position of the image pollutant and the photovoltaic panel. The vertices of the pollutant rectangular frame are passed to the drone, and the drone goes to this position to start. Operation. Among them, the coordinate system is converted. The lidar point cloud data is relative to the position of the IMU (Inertial Measurement Unit) sensor, and then the point cloud coordinates are converted to the lidar coordinate system. The lidar coordinate system is converted to the UAV coordinate system, and then converted to the geodetic coordinate system, and then the UAV coordinate center point is calibrated with the converted coordinate data of the point cloud.

After obtaining the type and location of the contaminants on the photovoltaic panel, the drone can perform the cleaning process (step 405). Optionally, the drone can use different cleaning solutions based on the types of contaminants on the photovoltaic panels. For example, based on the location of pollutants, a cleaning robot suction cup is used to clean accumulated dust, and the high-speed rotation of the motor is used to form a vacuum zone to adsorb accumulated dust. Use a high-pressure water gun to clean lumps such as bird droppings, use a brushless diaphragm pump to extract water from the water tank, and use a water gun to rinse. There are also sand and other materials that are handled with a brush and then rinsed with a water tank. After the cleaning process is completed, the drone returns to the original center point and scans 180 degrees again to check whether the pollutants have been cleaned. After cleaning is completed, proceed to the next photovoltaic panel. After all operations are completed, the drone returns autonomously. Achieve unmanned autonomous, efficient and accurate cleaning of photovoltaic panels.

In the embodiments of the present disclosure, the deep learning model can be used to accurately detect the types and locations of pollutants, achieve targeted cleaning of pollutants, and adopt targeted solutions for different pollutants to improve efficiency and reduce costs. In addition, the use of drones combined with the working principle of cleaning robot suction cups is more efficient and has stronger cleaning capabilities. This disclosure uses drones to clean photovoltaic panels, which can improve cleaning efficiency and save costs. It can reduce safety hazards when installed in remote areas and avoid safety problems existing in manual cleaning. In addition, the present disclosure achieves more precise control by controlling the drone through programs, which is stable, safe and reliable.

The present disclosure also provides a drone-based photovoltaic panel cleaning device. Figure 5 is a block diagram of a drone-based photovoltaic panel cleaning device provided by an embodiment of the present disclosure. As shown in Figure 5, the drone-based photovoltaic panel cleaning device may include a first determination module 501, a first control module 502, the second control module 503, the second determination module 504 and the cleaning processing module 505.

Among them, the first determination module 501 is used to determine the installation position information of the first photovoltaic panel when the drone enters the cleaning operation of the first photovoltaic panel, and determine the center point position of the drone to hover based on the installation position information. information.

In one implementation, the implementation process of the first determination module 501 determining the installation position information of the first photovoltaic panel may be as follows: determining the coordinates of the geodetic coordinate system when the drone is at the first hovering height; controlling the laser on the drone The radar scans the first photovoltaic panel to obtain lidar point cloud data; determines the coordinates of the first photovoltaic panel in the lidar coordinate system based on the lidar point cloud data; converts the coordinates of the first photovoltaic panel in the lidar coordinate system to Under the geodetic coordinate system, the installation position information of the first photovoltaic panel is obtained.

In one implementation, the first determination module 501 determines the location information of the center point to be hovered by the drone based on the installation location information. The implementation process can be as follows: determine the coordinates of the center point to be corrected based on the installation location information; determine whether The calibration value of each coordinate component of the human-machine in the geodetic coordinate system; the coordinates of the center point to be corrected are calibrated based on the calibration values of each coordinate component, and the coordinates obtained after calibration are determined as the position information of the center point to be hovered by the UAV.

The first control module 502 is used to control the drone to dock at the center point to be hovered based on the center point location information.

The second control module 503 is used to control the camera on the drone to collect images of the first photovoltaic panel to obtain images of the first photovoltaic panel.

The second determination module 504 is used to determine the location and type of contaminants on the first photovoltaic panel according to the image. In one implementation, the second determination module 504 is specifically configured to: input the image into a preset contaminant detection model, obtain the type of contaminants on the first photovoltaic panel and the location of the contaminants on the first photovoltaic panel in the image. position; among them, the pollutant detection model has learned the ability to predict the probability and position of various types of pollutants on the photovoltaic panel; according to the position of the pollutant on the first photovoltaic panel in the image, according to the pollutant in the image and the first photovoltaic panel The proportion of the panel is transformed into a coordinate system to determine the location of the contaminants on the first photovoltaic panel.

The cleaning module 505 is used to clean the first photovoltaic panel in combination with multiple cleaning structures based on the location and type of contaminants, where the multiple cleaning structures are different cleaning structures provided on the drone.

In one implementation, the cleaning processing module 505 is specifically used to: determine a corresponding cleaning plan according to the type of pollutants; select a cleaning structure corresponding to the cleaning plan from a variety of cleaning structures; use the selected cleaning structure according to the location of the pollutants. The cleaning structure performs cleaning processing on the first photovoltaic panel.

In a possible implementation, the cleaning processing module 505 is specifically used to: determine the operating stopping point when the drone cleans the pollutants based on the location of the contaminants and the selected cleaning structure; control the drone to dock at the operating stopping point , and control the selected cleaning structure to clean the first photovoltaic panel.

Optionally, in some embodiments, as shown in FIG. 6 , the drone-based photovoltaic panel cleaning device may also include a third determination module 606 , a fourth determination module 607 and a third control module 608 . Among them, the third determination module 606 is used to determine the cleaning plan of the photovoltaic array where the first photovoltaic panel is located after all the contaminants on the first photovoltaic panel are cleaned, or if it is determined based on the image that there are no contaminants on the first photovoltaic panel. path; the fourth determination module 607 is used to determine the next photovoltaic panel to be cleaned according to the cleaning planning path; the third control module 608 is used to control the drone to enter the cleaning operation of the next photovoltaic panel to be cleaned. Among them, 601-605 in Figure 6 and 501-505 in Figure 5 have the same functions and structures.

Regarding the devices in the above embodiments, the specific manner in which each module performs operations has been described in detail in the embodiments related to the method, and will not be described in detail here.

In the embodiments of the present disclosure, the deep learning model can be used to accurately detect the types and locations of pollutants, achieve targeted cleaning of pollutants, and adopt targeted solutions for different pollutants to improve efficiency and reduce costs. In addition, the use of drones combined with the working principle of cleaning robot suction cups is more efficient and has stronger cleaning capabilities. This disclosure uses drones to clean photovoltaic panels, which can improve cleaning efficiency and save costs. It can reduce safety hazards when installed in remote areas and avoid safety problems existing in manual cleaning. In addition, the present disclosure achieves more precise control by controlling the drone through programs, which is stable, safe and reliable.

The present disclosure also provides a drone and a computer-readable storage medium. Figure 7 is a schematic diagram of a drone provided by an embodiment of the present disclosure. As shown in Figure 7, the drone may include: a drone body 710, a camera 720, and a variety of different cleaning structures (not shown in Figure 7). The UAV body may include but is not limited to a UAV casing, a flight control system, a positioning system and a communication module. The positioning system may include but is not limited to a GPS (Global Positioning System, Global Positioning System) positioning system. The communication module Including but not limited to Bluetooth modules, etc.

In some embodiments, the plurality of different cleaning structures may include, but are not limited to, water-wash cleaning structures. Among them, as shown in Figure 7, the water washing cleaning structure may include but is not limited to a water tank 731, a pump body (not shown in Figure 7) and a water gun 732, wherein the water tank 731 is installed on the machine arm (not shown in Figure 7). At the bottom of the drone body 710, the pump body (not shown in Figure 7) is arranged on the drone body 710. The water tank 731 is connected to the water inlet of the pump body (not shown in Figure 7). The pump body (not shown in Figure 7 The water outlet (not shown in FIG. 7 ) is connected to the water gun 732 through a water supply pipe (not shown in FIG. 7 ). Among them, the pump body draws water from the water tank to the water gun to complete the flushing operation, which allows the drone to use the water cleaning structure to flush the photovoltaic panels. As an example, the pump body can be a brushless diaphragm pump, which has the characteristics of high pressure, large flow, long life, and low noise, making it easier to accurately control the flow of water.

In some embodiments, the plurality of different cleaning structures may also include, but are not limited to, cleaning robot suction cups. As shown in FIG. 7 , the cleaning robot suction cup 740 is provided at the bottom of the drone body 710 . Among them, the suction cup of the cleaning robot can include but is not limited to a motor. The suction cup of the cleaning robot can use the high-speed rotation of the motor to form a vacuum area to absorb dust, which allows the drone to use the suction cup of the cleaning robot to adsorb accumulated dust. The suction cup of the cleaning robot in the embodiment of the present disclosure adopts the principle of a sweeping robot and is made of rubber, thereby reducing damage to photovoltaic panels.

In some embodiments, the plurality of different cleaning structures may also include, but are not limited to, particulate matter cleaning structures. As shown in Figure 7, the particle cleaning structure may include but is not limited to a brush 751, a driving motor (not shown in Figure 7) and a rotating shaft 752. The driving motor (not shown in Figure 7) is provided on the drone body 710. On the above, the driving motor (not shown in Figure 7) and the rotating shaft 752 drive the brush 751 to implement the cleaning operation, which allows the drone to use the particle cleaning structure to clean sand and other particles on the photovoltaic panel.

It is worth noting that in some embodiments, the drone in the embodiment of the present disclosure may also include but is not limited to an ultrasonic ranging module 760 and a lidar (such as a mini lidar, also called a small lidar) 770. Among them, the ultrasonic ranging module 760 is set at the front end of the drone body, and the ultrasonic ranging module is used to measure the distance from the ground; the lidar 770 is set in the same vertical direction as the center of the drone body, and the lidar 770 is For scanning photovoltaic panels. In one implementation, the ultrasonic ranging module 760 can be used to measure the ground distance and determine the hovering height based on the return time of the sound wave and the actual temperature. The UAV can obtain the geodetic coordinate system coordinates of the UAV center point based on the positioning system, use the LiDAR 770 to scan the photovoltaic panels, process the point cloud data, cluster and segment the shape of the photovoltaic panels, and select the coordinates of the four largest extreme points. , and then the coordinate points are converted to the geodetic coordinate system, so that the drone and the photovoltaic panel are in the same coordinate system. The drone uses the four extreme point coordinates to determine the position information of the center point to be hovered.

The drone may also include one or more of the following components: a processing component, a memory, a power component, a multimedia component, an input/output (I/O) interface, a sensor component, and a communications component.

The processing component typically controls the overall operation of the drone, such as operations associated with data communications, flight control-related operations, and cleaning-related operations. The processing component may include one or more processors to execute instructions to complete all or part of the steps of the above method. Additionally, a processing component may include one or more modules that facilitate interaction between the processing component and other components. For example, a processing component may include a multimedia module to facilitate interaction between the multimedia component and the processing component.

The memory is configured to store various types of data to support the operation of the drone. Examples of this data include instructions, pictures, videos, etc. for any application or method used to operate on the drone. Memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable memory Read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk.

The electrical components provide power to various components of the drone. The power component may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the drone.

In some embodiments, the multimedia component includes a front-facing camera and/or a rear-facing camera. When the drone is in an operating mode, such as shooting mode or video mode, the front camera and/or rear camera can receive external multimedia data. Each front-facing camera and rear-facing camera can be a fixed optical lens system or have a focal length and optical zoom capabilities.

The I/O interface provides an interface between the processing component and the peripheral interface module. The above-mentioned peripheral interface module can be a keyboard, a click wheel, a button, etc. These buttons may include, but are not limited to: Home button, Volume buttons, Start button, and Lock button.

The sensor component includes one or more sensors that provide various aspects of status assessment for the drone. For example, the sensor component can detect the open/closed status of the drone, the relative positioning of the components, the sensor component can also detect the position change of the drone or a component of the drone, the orientation of the drone or the acceleration/deceleration and the drone. machine temperature changes. The sensor assembly may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

The communication component is configured to facilitate wired or wireless communication between the drone and other devices. Drones can access wireless networks based on communication standards such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component further includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

In an exemplary embodiment, the drone may be powered by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable A programmable gate array (FPGA), controller, microcontroller, microprocessor or other electronic component implementation is used to perform the above method.

In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions, such as a memory including instructions, is also provided, and the instructions can be executed by a processor of the drone to complete the above method. For example, the non-transitory computer-readable storage medium may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.

In an exemplary embodiment, a computer program product is also provided, including a computer program that implements the above method when executed by a processor of a drone.

Other embodiments of the invention will be readily apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary technical means in the technical field that are not disclosed in the present disclosure. . It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the invention being indicated by the following claims.

It is to be understood that the present invention is not limited to the precise construction described above and illustrated in the accompanying drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the invention is limited only by the appended claims.

Claims (12)

1. A drone-based photovoltaic panel cleaning method, characterized by including: In response to the drone entering the cleaning operation of the first photovoltaic panel, determining the installation location information of the first photovoltaic panel, and determining the location information of the center point where the drone is to hover based on the installation location information; According to the center point position information, control the drone to dock at the center point to be hovered; Control the camera on the drone to collect images of the first photovoltaic panel to obtain an image of the first photovoltaic panel; Determine the location and type of contaminants on the first photovoltaic panel based on the image; According to the location and type of the contaminants, the first photovoltaic panel is cleaned using a combination of multiple cleaning structures, where the multiple cleaning structures are different cleaning structures provided on the drone. 2. The method of claim 1, wherein determining the installation location information of the first photovoltaic panel includes: Determine the coordinates of the geodetic coordinate system when the UAV is at the first hovering height; Control the lidar on the drone to scan the first photovoltaic panel to obtain lidar point cloud data; Determine the coordinates of the first photovoltaic panel in the lidar coordinate system according to the lidar point cloud data; The coordinates of the first photovoltaic panel in the lidar coordinate system are converted into the geodetic coordinate system to obtain the installation position information of the first photovoltaic panel. 3. The method of claim 1 or 2, wherein determining the location information of the center point where the drone is to be hovered based on the installation location information includes: According to the installation position information, determine the coordinates of the center point to be corrected; Determine the calibration value of each coordinate component of the UAV in the geodetic coordinate system; The coordinates of the center point to be corrected are calibrated according to the calibration values of each coordinate component, and the coordinates obtained after calibration are determined as the position information of the center point to be hovered by the UAV. 4. The method of claim 1 or 2, wherein determining the location and type of contaminants on the first photovoltaic panel based on the image includes: Input the image into a preset contaminant detection model to obtain the type of contaminants on the first photovoltaic panel and the location of the contaminants on the first photovoltaic panel in the image; wherein, the contaminants The detection model has learned the ability to predict the probability and location of various types of pollutants on photovoltaic panels; According to the position of the contaminant on the first photovoltaic panel in the image, coordinate system transformation is performed according to the proportion of the contaminant and the first photovoltaic panel in the image to determine the concentration of the contaminant on the first photovoltaic panel. Location. 5. The method of claim 1 or 2, wherein cleaning the first photovoltaic panel in combination with the plurality of cleaning structures according to the location and type of the contaminants includes: Determine the corresponding cleaning plan according to the type of pollutants; Select a cleaning structure corresponding to the cleaning scheme from the plurality of cleaning structures; The first photovoltaic panel is cleaned using the selected cleaning structure according to the location of the contaminants. 6. The method of claim 5, wherein using the selected cleaning structure to clean the first photovoltaic panel according to the location of the contaminants includes: According to the location of the contaminant and the selected cleaning structure, determine the operating stop point when the drone cleans the contaminant; The drone is controlled to dock to the operation docking point, and the selected cleaning structure is controlled to clean the first photovoltaic panel. 7. The method of claim 1 or 2, characterized in that the method further comprises: After all the contaminants on the first photovoltaic panel are cleaned, or it is determined based on the image that there are no contaminants on the first photovoltaic panel, the cleaning planning path of the photovoltaic array where the first photovoltaic panel is located is determined. ; Determine the next photovoltaic panel to be cleaned according to the cleaning planning path; Control the drone to enter the cleaning operation of the next photovoltaic panel to be cleaned. 8. A drone-based photovoltaic panel cleaning device, characterized by including: The first determination module is used to determine the installation position information of the first photovoltaic panel when the drone enters the cleaning operation of the first photovoltaic panel, and determines that the drone is to hover based on the installation position information. center point location information; A first control module, configured to control the drone to dock at the center point to be hovered based on the center point position information; The second control module is used to control the camera on the drone to collect images of the first photovoltaic panel to obtain images of the first photovoltaic panel; a second determination module, configured to determine the location and type of contaminants on the first photovoltaic panel based on the image; A cleaning processing module, used to clean the first photovoltaic panel in combination with a variety of cleaning structures according to the location and type of the contaminants, wherein the multiple cleaning structures are different cleaning structures provided on the drone. Clean structure. 9. A drone, characterized in that it includes: UAV body; A camera, which is arranged at the front end of the drone body and is used for image collection; A variety of different cleaning structures, the multiple cleaning structures are arranged at different positions of the drone body; At least one processor connected to the camera and the plurality of cleaning structures respectively; A memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to cause the at least one process The device is capable of performing the method as claimed in any one of claims 1 to 7. 10. The UAV according to claim 9, further comprising: Ultrasonic ranging module, the ultrasonic ranging module is arranged at the front end of the drone body, and the ultrasonic ranging module is used to measure the distance to the ground; LiDAR, the LiDAR is arranged in the same vertical direction as the center position of the UAV body, and the LiDAR is used to scan photovoltaic panels. 11. The drone of claim 9, wherein the plurality of cleaning structures include: The water washing cleaning structure includes a water tank, a pump body and a water gun, wherein the water tank is installed at the bottom of the drone body through an arm, the pump body is arranged on the drone body, and the water tank is connected to the drone body. The water inlet of the pump body is connected, and the water outlet of the pump body is connected to the water gun through a water supply pipe, wherein the pump body draws water in the water tank to the water gun to implement flushing operations; A cleaning robot suction cup, the cleaning robot suction cup is arranged at the bottom of the drone body, wherein the cleaning robot suction cup includes a motor, and the cleaning robot suction cup uses the high-speed rotation of the motor to form a vacuum zone to absorb dust; The particulate matter cleaning structure includes a brush, a driving motor and a rotating shaft. The driving motor is arranged on the drone body, wherein the driving motor and the rotating shaft drive the brush to achieve cleaning operations. 12. A computer-readable storage medium for storing instructions, characterized in that when the instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.