Method and system for aligning power receiving racks of multiple types of UAVs based on image processing
Through image processing technology, the rotation image of the drone is obtained, the charging contact characteristics are segmented, and the three-dimensional reconstruction is carried out, which solves the problems of inaccurate alignment of the power frame of the traditional drone and large manpower investment, and realizes the adaptability of multiple models and Efficient automatic charging system.
Patent Information
- Application Number
- CN202411002503.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-25
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-07-25
AI Technical Summary
The existing drone power receiving rack is difficult to adapt to various types of drones. Traditional charging systems rely on physical displacement sliding tables and communication identification, resulting in inaccurate alignment, large vibration, low hangar utilization, and a large amount of manpower investment. .
Using a multi-model drone power frame alignment method based on image processing, the charging contact feature image is segmented, point cloud data is obtained, and the charging contacts are calculated in the world coordinates are calculated. and aligned charging is achieved using mechanical devices.
It has achieved adaptability to various types of drones, reduced physical pressure and vibration, improved alignment accuracy and hangar usage, reduced manpower investment, and improved drone operation and maintenance efficiency.
Smart Images

Figure CN118521646B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of image processing, and in particular to a method and system for aligning power receiving racks of multiple types of unmanned aerial vehicles based on image processing. Background Art
[0002] With the widespread application of drones in the fields of automatic inspection, monitoring, data collection, etc., the downtime maintenance and charging of drones have become an important guarantee for their automated and efficient operation. Especially for aircraft that need to perform long-term, high-frequency flight missions, frequent charging and maintenance are required, and traditional operation and maintenance methods require huge manpower investment. Therefore, automatic charging technology based on drone hangars or aprons is currently used. That is, after the drone lands, the positioning mechanism is used to make the drone reach the designated position, and then the charging connector in the hangar will actively dock with the charging interface on the drone, thereby realizing automatic charging.
[0003] However, this method also has many technical problems: First, due to the wide variety of drones and their different structures, the designed power receiving mechanism is difficult to meet the needs of various types of drones; for example, patent 202021982347.8 proposes a bayonet-type power receiving structure that can be fixed to the drone tripod. Patent 202321424539.0 also proposes a charging beam fixed on the crossbar of the drone bracket; these two structures are only applicable to drones with bottom brackets; secondly, the charging contacts of the power receiving mechanism are usually designed for specific battery voltages and only include positive and negative pole wiring; for example, patent 202210313580.4 designs a charging contact with a clamping mechanism with positive and negative pole contacts; currently, most charging contacts cannot achieve balanced charging control of battery series connection; thirdly, the charging position alignment system of the drone is used for contact alignment of the power receiving contacts; the system usually adopts a rigid displacement slide, such as patent 202021905492.6 proposes a position adjustment system based on a 2D positioning platform; during the rigid alignment process, the drone tripod or power receiving frame will be deformed, and additional vibration will cause components to fall off; finally, the automatic apron and hangar's recognition of drones depends on communication with the drone. The current solution usually requires the apron to be equipped with a single type of drone; this leads to a low utilization rate of the hangar.
[0004] To this end, the present invention proposes a method and system for aligning power receiving racks of multiple types of UAVs based on image processing. Summary of the invention
[0005] In view of the technical problems existing in the prior art, the present invention provides a method and system for aligning power receiving racks of multiple types of unmanned aerial vehicles based on image processing.
[0006] The technical solution of the present invention to solve the above technical problems is as follows: a method and system for aligning power receiving racks of multiple types of UAVs based on image processing;
[0007] The method for aligning power receiving racks of multiple types of UAVs based on image processing includes the following steps:
[0008] S1: After the UAV is parked on the power rack, a rotating image of the UAV is obtained;
[0009] S2: Segment the characteristic image of the charging contact according to the rotation image of the drone and obtain the point cloud data of the charging contact;
[0010] S3: Process the point cloud data and perform 3D reconstruction;
[0011] S4: Calculate the coordinate point of the charging contact in the world coordinate using the relationship between the image coordinate, the camera coordinate and the preset world coordinate;
[0012] S5: The charging head is connected to the charging contact point according to the coordinate point and completes alignment charging.
[0013] Furthermore, the method for aligning power receiving racks of multiple types of UAVs based on image processing comprises in step S1:
[0014] The power receiving frame is provided with a rotating platform and a camera. After the UAV is parked on the rotating platform, the rotating platform is started, and the camera takes a rotating image of the UAV rotating 360 degrees.
[0015] Furthermore, the method for aligning multiple types of UAV power receiving racks based on image processing includes extracting features of the charging contacts according to a feature extraction algorithm in step S2:
[0016] Perform scale space transformation on the image of suspected charging contacts on the drone to capture local extreme points in the image;
[0017] The position of the key point is determined by fitting a two-dimensional Gaussian function around the local extreme point, and its scale and direction are calculated;
[0018] Calculate the gradient direction histogram of the image area around the key point and select the dominant gradient direction as the direction of the key point;
[0019] Establish a coordinate system in the area around the key point, and describe the pixels in the area through feature vectors according to the gradient direction;
[0020] The feature vectors of the charging contacts in the preset library are matched, and when the similarity is greater than a preset value, the suspected charging contact image is the charging contact image, otherwise the image is discarded.
[0021] Furthermore, the method for aligning multiple types of UAV power racks based on image processing includes point cloud registration in step S3:
[0022] Set target point cloud And the point cloud set to be registered , represents three-dimensional space, and Respectively represent the coordinate information of each point cloud. The two point cloud sets are n*3 and m*3 matrices. By calculating the rotation matrix R and displacement matrix T between the two matrices:
[0023] .
[0024] Furthermore, the multi-model UAV power rack alignment method based on image processing, point cloud registration also requires coordinate transformation of the point cloud data under each viewing angle into a unified coordinate system by using the mapping matrix H:
[0025] ;
[0026] S is the scaling factor of the deformation. Since the point cloud data only has rotation and translation, S=1 here. The rotation matrix and translation matrix are:
[0027] ;
[0028] ;
[0029] in, , and They are the angles of rotation of the point cloud on the x, y and z axes, , and are the displacements of the point cloud on the x, y, and z axes respectively.
[0030] Furthermore, the method for aligning power receiving racks of multiple types of UAVs based on image processing includes in step S4:
[0031] Transformation from image coordinate system to camera coordinate system. Let the pixel at any point in the image be (u, v), then:
[0032] ;
[0033] in is the camera's internal parameter, , and are the camera coordinates.
[0034] Furthermore, in the method for aligning the power receiving racks of multiple types of UAVs based on image processing, the transformation from the camera coordinate system to the world coordinate system includes:
[0035] ;
[0036] in, , and is the world coordinate, is the external parameter of the camera.
[0037] Furthermore, in the method for aligning the power receiving racks of multiple drones based on image processing, after obtaining the position coordinates of the charging contacts, the rotation matrix can reflect the rotation of the charging head and the charging contacts. express:
[0038] ;
[0039] Since the charging head and the charging contacts are parallel on the z-axis, the rotation matrix of the charging head around the x-axis and y-axis can be expressed as a matrix product, that is, rotating around the x-axis first and then around the y-axis or vice versa:
[0040] ;
[0041] Here and are the rotation angles around the x-axis and y-axis respectively;
[0042] .
[0043] The multi-model UAV power rack alignment system based on image processing is applied to realize the multi-model UAV power rack alignment method based on image processing, and the alignment system includes:
[0044] Camera unit: After the UAV is parked on the power receiving rack, the rotating image of the UAV is obtained;
[0045] Acquisition unit: segment the characteristic image of the charging contact according to the rotating image of the drone and obtain the point cloud data of the charging contact;
[0046] Processing unit: Process the point cloud data and then perform 3D reconstruction;
[0047] Calculation unit: calculates the coordinate point of the charging contact in the world coordinate using the relationship between the image coordinate, the camera coordinate and the preset world coordinate;
[0048] Alignment unit: The charging head is connected with the charging contacts according to the coordinate points and completes alignment charging.
[0049] The beneficial effects of the present invention are:
[0050] Multi-model adaptability: It can adapt to various types of drones without bottom brackets or with differentiated designs, and is not limited to specific structures or models, thus solving the problem of dependence on specific models of drones in existing technologies;
[0051] Flexible alignment mechanism: Through image acquisition and processing, the charging port on the drone can be accurately located without relying on a physical displacement slide; compared with a rigid alignment system, it reduces the physical pressure and possible additional vibration on the drone and the charging rack;
[0052] No need for communication and identification: Since this method does not rely on communication with the drone for identification and alignment, it can reduce dependence on a single type of drone and improve hangar utilization and flexibility. The automated alignment process reduces manpower input and reduces operating costs in the operation and maintenance of drones that perform long-term, high-frequency flight missions. This has positive significance for accelerating the application and promotion of drones in the fields of automatic inspection, monitoring and data collection. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] Figure 1 It is a schematic diagram of the flow of the method for aligning power receiving racks of multiple types of UAVs based on image processing according to the present invention;
[0054] Figure 2 The present invention is a flowchart of feature extraction in the method for aligning power receiving racks of multiple types of UAVs based on image processing. DETAILED DESCRIPTION
[0055] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.
[0056] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, "plurality" means two or more, unless otherwise clearly and specifically defined.
[0057] In the description of the present application, the term "for example" is used to mean "used as an example, illustration or description". Any embodiment described as "for example" in the present application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid unnecessary details to obscure the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application. Example 1
[0058] The method for aligning power receiving racks of multiple types of UAVs based on image processing includes the following steps:
[0059] S1: After the UAV is parked on the power rack, a rotating image of the UAV is obtained;
[0060] S2: Segment the characteristic image of the charging contact according to the rotation image of the drone and obtain the point cloud data of the charging contact;
[0061] S3: Process the point cloud data and perform 3D reconstruction;
[0062] S4: Calculate the coordinate point of the charging contact in the world coordinate using the relationship between the image coordinate, the camera coordinate and the preset world coordinate;
[0063] S5: The charging head is connected to the charging contact point according to the coordinate point and completes alignment charging.
[0064] Step S1 includes:
[0065] The power receiving frame is provided with a rotating platform and a camera. After the UAV is parked on the rotating platform, the rotating platform is started, and the camera takes a rotating image of the UAV rotating 360 degrees.
[0066] In this embodiment, before starting to align the drone for charging, it is first necessary to obtain an image of the drone on the power receiving rack. After the drone docks on the power receiving rack equipped with a rotating platform, the platform will rotate the drone to a certain angle (usually a 360-degree panoramic rotation), and the camera mounted on the power receiving rack will capture images of the drone at different angles. Through these images, the charging contacts (charging interface) and other related features on the drone can be fully observed. After collecting all-round image information, the next step is to use image processing technology to separate the characteristic images of the charging contacts from the image, and further obtain the point cloud data of these features; point cloud data is a data set composed of a series of points, and the distribution of these points in space defines the shape boundary of the object; after having the point cloud data of the charging contacts, it can be processed and the point cloud information can be used for three-dimensional reconstruction; this process includes point cloud cleaning (removing noise points), registration (alignment and merging of multi-angle point clouds), and The model is constructed, so that the accurate three-dimensional model of the charging contact of the drone can be obtained. The acquired image coordinates (i.e., the pixel position of each point in the image) and camera coordinates (the position of the object in the camera coordinate system when the camera is shooting) are used, and combined with the preset world coordinate system (usually the coordinate system of the power receiving rack or charging station), the exact coordinate position of the charging contact in the world coordinate system can be calculated. Finally, after obtaining the specific position of the charging contact of the drone in the world coordinate system, a mechanical device (such as a charging head carried by a mechanical arm) can be used to move according to this position, so as to align and connect with the charging contact on the drone to complete the charging alignment. After the charging alignment is successful, the drone battery can be charged; according to the above method, it can be applied to the charging of various types of drones (such as whether it has a tripod, different models, etc.). At the same time, the parking of the drone does not require communication connection, only the position of the power receiving rack needs to be set in the drone. When the drone is low on power, it can automatically fly to the power receiving rack and accept it to charge the drone. In addition, the parking angle of the drone does not need to be adjusted, which reduces the drone's stay time in the air and increases the utilization rate of the power receiving rack.
[0067] In one embodiment, step S2 includes extracting features of the charging contacts according to a feature extraction algorithm:
[0068] S21: Perform scale space transformation on the image of suspected charging contacts on the drone to capture local extreme points in the image;
[0069] S22: Determine the position of the key point by fitting a two-dimensional Gaussian function around the local extreme point, and calculate its scale and direction;
[0070] S23: Calculate the gradient direction histogram for the image area around the key point, and select the dominant gradient direction as the direction of the key point;
[0071] S24: Establishing a coordinate system in the surrounding area of the key point, and describing the pixels in the area by a feature vector according to the gradient direction;
[0072] S25: matching the feature vectors of the charging contacts in the preset library, and when the similarity is greater than a preset value, the suspected charging contact image is the charging contact image, otherwise the image is discarded.
[0073] In this embodiment, feature extraction is performed using the SIFT algorithm. First, the SIFT algorithm performs a Gaussian blur operation on the input image at different scales, which is achieved by constructing a Gaussian pyramid. By constructing a Gaussian difference pyramid (DoG) at different scales, local extreme points with significant gradient changes in the image can be captured. After the local extreme points are detected, the SIFT algorithm uses a method of fitting a two-dimensional Gaussian function to determine the exact position of the key points, and calculates the scale and dominant direction of the key points. In this step, the position and scale of the key points are determined, while ensuring invariance to scale and rotation. By calculating the gradient direction histogram in the image area around the key points, the SIFT algorithm selects the dominant gradient. The direction of the degree is taken as the direction of the key point; this direction makes the subsequent feature descriptor construction invariant to the rotation of the image; after determining the position, scale and direction of the key point, the SIFT algorithm constructs a gradient histogram containing multiple directions in the image area around the key point; these histograms are finally combined into a high-dimensional feature vector to describe the local texture information in the area, and have a certain robustness; the extracted feature vector of the charging contact is matched with the feature vector in the preset library, usually using some similarity measurement methods (such as Euclidean distance, Hamming distance, etc.), when the matching similarity is greater than the preset threshold, it can be confirmed that the suspected charging contact image is the charging contact image, otherwise the image is discarded.
[0074] Image processing and computer vision libraries such as OpenCV can be used in the specific implementation steps; the following is a simplified example to illustrate the process of implementing the SIFT algorithm in Python using the OpenCV library:
[0075] # Read the image
[0076] image = cv2.imread('drone_image.jpg')
[0077] gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
[0078] # Initialize SIFT detector
[0079] sift = cv2.SIFT_create()
[0080] # Detect key points and calculate descriptors
[0081] keypoints, descriptors = sift.detectAndCompute(gray, None)
[0082] # Draw key points
[0083] output_img = cv2.drawKeypoints(image, keypoints, None)
[0084] # Display results
[0085] cv2.imshow('SIFT Keypoints', output_img)
[0086] cv2.waitKey(0)
[0087] cv2.destroyAllWindows()
[0088] In the above example, the image representing the drone image is first read through the OpenCV library. Then, the key points of the image are detected using the SIFT algorithm and the descriptors of these key points are calculated. Finally, the positions of the key points are plotted on the image and displayed.
[0089] In one embodiment, step S3 includes point cloud registration:
[0090] Set target point cloud And the point cloud set to be registered , represents three-dimensional space, and Respectively represent the coordinate information of each point cloud. The two point cloud sets are n*3 and m*3 matrices. By calculating the rotation matrix R and displacement matrix T between the two matrices:
[0091] .
[0092] Furthermore, point cloud registration also requires the coordinate transformation of the point cloud data at each perspective into a unified coordinate system by using the mapping matrix H:
[0093] ;
[0094] S is the scaling factor of the deformation. Since the point cloud data only has rotation and translation, S=1 here. The rotation matrix and translation matrix are:
[0095] ;
[0096] ;
[0097] in, , and They are the angles of rotation of the point cloud on the x, y and z axes, , and are the displacements of the point cloud on the x, y, and z axes respectively.
[0098] By calculating the rotation matrix and the displacement matrix, the point cloud data obtained from different perspectives can be unified into the same three-dimensional space coordinate system; this provides the necessary accuracy and reliability for the precise positioning of the drone components, especially the charging contacts, and the subsequent charging docking operations; this also makes the automated charging process possible, because the charging head needs to be accurately aligned with the drone's charging contacts for smooth charging operations.
[0099] In one embodiment, step S4 includes:
[0100] Transformation from image coordinate system to camera coordinate system. Let the pixel at any point in the image be (u, v), then:
[0101] ;
[0102] in is the camera's internal parameter, , and are the camera coordinates.
[0103] Furthermore, the transformation from the camera coordinate system to the world coordinate system includes:
[0104] ;
[0105] in, , and is the world coordinate, is the external parameter of the camera.
[0106] In one embodiment, after obtaining the position coordinates of the charging contacts, the rotation matrix can reflect the rotation of the charging head and the charging contacts. express:
[0107] ;
[0108] A rotation matrix is a mathematical representation used to describe the rotation of an object around an arbitrary axis in three-dimensional space. In this embodiment, the rotation matrix is used to represent the relative rotation between the charging head and the charging contacts; because real systems often require that the charging head and the charging contacts need to maintain a certain posture when docking, for example, the two may need to maintain a certain direction when connected; for example, if the charging head and the charging contacts are designed to be in the same plane and have a specific direction, then the charging head may need to rotate around the x-axis and y-axis to achieve the correct direction.
[0109] Typically, a rotation matrix looks like this:
[0110] ;
[0111] This matrix describes the transformation of rotating clockwise around the origin by an angle of θ in two-dimensional space; for three-dimensional space, there are three such matrices representing rotations around the x-axis, y-axis, and z-axis respectively;
[0112] ;
[0113] ;
[0114] ;
[0115] Since the charging head and the charging contacts are parallel on the z-axis, the rotation matrix of the charging head around the x-axis and y-axis can be expressed as a matrix product, that is, rotating around the x-axis first and then around the y-axis or vice versa:
[0116] ;
[0117] Here and are the rotation angles around the x-axis and y-axis respectively;
[0118] .
[0119] This rotation matrix can be used to determine whether the charging head needs to be rotated in the x-axis and y-axis directions to align with the direction of the charging contacts; the calculated angles can be directly used to control the stepper motor or servo motor of the charging head to achieve precise physical alignment and ensure that the physical connection during charging is correct. Example 2
[0120] The multi-model UAV power rack alignment system based on image processing is applied to realize the multi-model UAV power rack alignment method based on image processing, and the alignment system includes:
[0121] Camera unit: After the UAV is parked on the power receiving rack, the rotating image of the UAV is obtained;
[0122] Acquisition unit: segment the characteristic image of the charging contact according to the rotating image of the drone and obtain the point cloud data of the charging contact;
[0123] Processing unit: Process the point cloud data and then perform 3D reconstruction;
[0124] Calculation unit: calculates the coordinate point of the charging contact in the world coordinate using the relationship between the image coordinate, the camera coordinate and the preset world coordinate;
[0125] Alignment unit: The charging head is connected with the charging contacts according to the coordinate points and completes alignment charging.
[0126] Camera unit: This part includes one or more cameras to capture images of the drone while it is parked on the power rack. The camera unit may include lights or other lighting facilities to ensure clear pictures, especially in low light conditions.
[0127] Acquisition unit: Uses image processing technology to identify and extract the features of the charging contacts when the drone rotates; is responsible for segmenting the relevant areas in the image and converting them into point cloud data in preparation for further 3D reconstruction.
[0128] Processing unit: The function of this unit is to perform necessary preprocessing on the point cloud data provided by the acquisition unit, such as filtering and noise removal, and perform three-dimensional reconstruction; this may also include point cloud registration, that is, fusing point cloud data from multiple perspectives together.
[0129] Computing unit: The main computing facility, responsible for executing the transformation algorithm, converting the image coordinates obtained by the camera unit into the camera coordinate system, and then into the world coordinate system; the computing unit needs to know the camera's intrinsic and extrinsic parameter data in order to correctly perform the coordinate transformation.
[0130] Alignment unit: includes a robotic arm and charging head that can be moved and precisely positioned.
[0131] Receive the output coordinate data of the computing unit and use this data to control the charging head to align with the charging contacts of the drone; after the alignment is completed, ensure that the charging connection is properly contacted to start the charging process.
[0132] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.
[0133] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.
Claims
1. A method for aligning power receiving racks of multiple types of UAVs based on image processing, characterized in that: The following steps are involved: S1: After the UAV is parked on the power rack, a rotating image of the UAV is obtained; S2: Segment the characteristic image of the charging contact according to the rotation image of the drone and obtain the point cloud data of the charging contact; S2 includes extracting features of the charging contacts according to a feature extraction algorithm: Perform scale space transformation on the image of suspected charging contacts on the drone to capture local extreme points in the image; The position of the key point is determined by fitting a two-dimensional Gaussian function around the local extreme point, and its scale and direction are calculated; Calculate the gradient direction histogram of the image area around the key point and select the dominant gradient direction as the direction of the key point; Establish a coordinate system in the area around the key point, and describe the pixels in the area through feature vectors according to the gradient direction; Match the feature vectors of the charging contacts in the preset library, and when the similarity is greater than a preset value, the suspected charging contact image is the charging contact image, otherwise the image is discarded; S3: Process the point cloud data and perform 3D reconstruction; S4: Calculate the coordinate point of the charging contact in the world coordinate using the relationship between the image coordinate, the camera coordinate and the preset world coordinate; S5: The charging head is connected to the charging contact point according to the coordinate point and completes alignment charging; After obtaining the position coordinates of the charging contacts, the rotation matrix can reflect the rotation of the charging head and the charging contacts. express: ; Since the charging head and the charging contacts are parallel on the z-axis, the rotation matrix of the charging head around the x-axis and y-axis can be expressed as a matrix product, that is, rotating around the x-axis first and then around the y-axis or vice versa: ; Here and are the rotation angles around the x-axis and y-axis respectively; 。 2. The method for aligning power receiving racks of multiple types of unmanned aerial vehicles based on image processing according to claim 1 is characterized in that: Step S1 includes: The power receiving frame is provided with a rotating platform and a camera. After the UAV is parked on the rotating platform, the rotating platform is started, and the camera takes a rotating image of the UAV rotating 360 degrees.
3. The method for aligning power receiving racks of multiple types of unmanned aerial vehicles based on image processing according to claim 1, characterized in that: In step S3, point cloud registration is included: Set target point cloud And the point cloud set to be registered , represents three-dimensional space, and Respectively represent the coordinate information of each point cloud. The two point cloud sets are n*3 and m*3 matrices. By calculating the rotation matrix R and displacement matrix T between the two matrices: .
4. The method for aligning power receiving racks of multiple types of unmanned aerial vehicles based on image processing according to claim 3 is characterized in that: Point cloud registration also requires the coordinate transformation of the point cloud data at each perspective into a unified coordinate system by using the mapping matrix H: ; S is the scaling factor of the deformation. Since the point cloud data only has rotation and translation, S=1 here. The rotation matrix and translation matrix are: ; ; in, , and They are the angles of rotation of the point cloud on the x, y and z axes, , and are the displacements of the point cloud on the x, y, and z axes respectively.
5. The method for aligning power receiving racks of multiple types of unmanned aerial vehicles based on image processing according to claim 1, characterized in that: Step S4 includes: Transformation from image coordinate system to camera coordinate system. Let the pixel at any point in the image be (u, v), then: ; in is the camera's internal parameter, , and are the camera coordinates.
6. The method for aligning power receiving racks of multiple types of unmanned aerial vehicles based on image processing according to claim 5 is characterized in that: The transformation from the camera coordinate system to the world coordinate system includes: ; in, , and is the world coordinate, is the external parameter of the camera.
7. The multi-model UAV power rack alignment system based on image processing is characterized by: Applied to implement the multi-type UAV power receiving rack alignment method based on image processing as described in any one of claims 1 to 6, the alignment system comprises: Camera unit: After the UAV is parked on the power receiving rack, the rotating image of the UAV is obtained; Acquisition unit: segment the characteristic image of the charging contact according to the rotating image of the drone and obtain the point cloud data of the charging contact; Processing unit: Process the point cloud data and then perform 3D reconstruction; Calculation unit: calculates the coordinate point of the charging contact in the world coordinate using the relationship between the image coordinate, the camera coordinate and the preset world coordinate; Alignment unit: The charging head is connected with the charging contacts according to the coordinate points and completes alignment charging.
