Automatic Plugging and Unplugging Method of Charging Gun for Mobile Charging Robot Based on Active Vision Positioning Technology

Abstract

The present invention relates to the field of visual positioning technology, and specifically, to a method for automatically plugging and unplugging a charging gun of a mobile charging robot based on active vision positioning technology. The method includes: capturing an image of a vehicle and inputting it into a cloud large model to automatically identify the vehicle model; using a multi-modal SLAM pose estimation algorithm to locate the vehicle pose, generating a three-dimensional model of the vehicle pose and the position coordinates of the charging gun; capturing an image of the charging port and using an advanced active vision positioning algorithm to real-time locate the position coordinates of the charging port; introducing the influence of the robotic arm vibration, using the RRT algorithm to calculate the movement path of the robotic arm, and completing the plugging and unplugging operation through real-time feedback control of the robotic arm. The method for automatically plugging and unplugging a charging gun of a mobile charging robot based on active vision positioning technology solves the problem of precise positioning of the vehicle pose and the charging gun position through a multi-modal SLAM pose estimation algorithm, and real-time corrects the position of the charging port through an advanced active vision positioning algorithm, calculates the influence of the robotic arm vibration, and realizes precise automatic plugging and unplugging of the charging gun.

Description

Technical Field

[0001] The present invention relates to the technical field of visual positioning, and more specifically, to a method for automatically plugging and unplugging a charging gun of a mobile charging robot based on active visual positioning technology. Background Art

[0002] The method for automatically plugging and unplugging a charging gun of a mobile charging robot based on active visual positioning technology aims to improve the accuracy and stability of the charging process. By using a multi-modal SLAM pose estimation algorithm and an advanced active visual positioning algorithm, the movement path of the robotic arm is controlled to achieve high-precision real-time positioning of the charging port and automatic plugging and unplugging operations of the charging gun.

[0003] Existing methods for automatically plugging and unplugging a charging gun usually have difficulty in accurately identifying the charging port in a complex environment. Moreover, since the vibration of the robotic arm of the robot will affect the movement path of the robot, it will lead to inaccurate plugging and unplugging and charging failure problems. In summary, a method for automatically plugging and unplugging a charging gun of a mobile charging robot based on active visual positioning technology is provided. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for automatically plugging and unplugging a charging gun of a mobile charging robot based on active visual positioning technology, so as to solve the problems of inaccurate plugging and unplugging and charging failure caused by the vibration of the robotic arm of the robot affecting the movement path of the robot as mentioned in the above background art.

[0005] To achieve the above purpose, the present invention provides a method for automatically plugging and unplugging a charging gun of a mobile charging robot based on active visual positioning technology, including the following steps:

[0006] S1. Take a vehicle image and input it into the cloud large model to automatically identify the vehicle model;

[0007] S2. Use a multi-modal SLAM pose estimation algorithm to locate the vehicle pose, and determine the charging gun position information according to the vehicle model to generate a three-dimensional model of the vehicle pose and the charging gun position coordinates;

[0008] S3. Take an image of the charging port and use an advanced active visual positioning algorithm to real-time locate the charging port position coordinates;

[0009] S4. Based on the charging gun position information and the charging port position information, introduce the influence of the robotic arm vibration, use the RRT algorithm to calculate the movement path of the robotic arm, and control the robotic arm to complete the plugging and unplugging operation through the real-time feedback of the active vision system.

[0010] In the above S1, taking a vehicle image and inputting it into the cloud large model to automatically identify the vehicle model, the specific method is as follows:

[0011] S1.1. Use a camera to capture multi-angle vehicle images of the license plate and vehicle body identification;

[0012] S1.2. Denoise the vehicle image and use the Canny edge detection algorithm to extract the edge information in the image;

[0013] S1.3. Compress the preprocessed vehicle image and upload it to the cloud large model to automatically identify the vehicle model.

[0014] As a further improvement of this technical solution, in S2, the multi-modal SLAM pose estimation algorithm is based on the ORB-SLAM algorithm and the multi-modal data fusion algorithm. By fusing the vehicle image data and the point cloud data, a three-dimensional vehicle pose model and the charging gun position coordinates are generated. The specific method is as follows:

[0015] S2.1. Obtain the vehicle image data I(x, y) from S1 and use the lidar to obtain the three-dimensional point cloud data P of the vehicle cloud ;

[0016] S2.2. Use the ORB-SLAM algorithm to process the vehicle image data and the three-dimensional point cloud data to generate a preliminary vehicle pose estimation and an environmental map;

[0017] S2.3. Combine the preliminary vehicle pose estimation and the environmental map, use the HRNet deep learning model to process the vehicle image data, and use the PointNet++ algorithm to process the three-dimensional point cloud data to generate a three-dimensional vehicle model;

[0018] S2.4. Use the Kalman filter and the BA optimization algorithm to generate the accurate pose of the vehicle

[0019] [R opt , t opt and the charging gun position coordinates

[0020] where, R opt is the rotation matrix of the vehicle in three-dimensional space, which is a 3x3 orthogonal matrix; t opt is the position vector of the vehicle in three-dimensional space, which is a 3x1 vector; is the position coordinate of the charging gun.

[0021] As a further improvement of this technical solution, in S3, the advanced active vision positioning algorithm is based on the active vision system and uses the image processing method and the deep learning model to real-time locate the charging port position information. The specific method is as follows:

[0022] S3.1. Use the camera to capture the real-time image of the charging port and perform denoising and edge enhancement processing on the image;

[0023] S3.2. Determine the internal parameter matrix K and external parameter matrix [R|t] of the camera, and convert the three-dimensional world coordinates into two-dimensional image coordinates;

[0024]

[0025]

[0026] Among them, K is the camera internal parameter matrix; f x is the focal length of the camera in the x direction; f y is the focal length of the camera in the y direction; c x is the principal point offset of the camera in the x direction; c y is the principal point offset of the camera in the y direction; R is the rotation matrix; t is the translation vector; is the homogeneous coordinate in the two-dimensional image coordinate; is the homogeneous coordinate in the three-dimensional world coordinate;

[0027] S3.3. Apply the distortion correction formula to minimize the total reprojection error E(R, t) to obtain the optimized R and t;

[0028] S3.4. Apply the inverse perspective transformation and obtain the charging port position coordinates through the transformation matrix T wc to obtain the charging port position coordinates

[0029] As a further improvement of this technical solution, in the S3.3, when applying the distortion correction formula to minimize the total reprojection error, the specific method is as follows:

[0030] (u d , v d ) = distort(u, v; k1, k2, p1, p2);

[0031] Among them, (u d , v d ) are the image coordinates after distortion correction; (u, v) are the original image coordinates without distortion correction; k1 is the quadratic term radial distortion coefficient, k2 is the quartic term radial distortion coefficient; p1, p2 are the tangential distortion coefficients; E(R, t) represents the total reprojection error under the given rotation matrix R and translation vector t; q i is the observed coordinate of the i-th feature point in the image coordinate system; Q i is the three-dimensional coordinate of the i-th feature point in the world coordinate system; N is the total number of feature points;

[0032] As a further improvement of this technical solution, in the S3.4, apply the inverse perspective transformation and obtain the charging port position coordinates through the transformation matrix T wc to obtain the charging port position coordinates The specific method is as follows:

[0033]

[0034] Among them, is the position coordinate of the charging port; T wc is the transformation matrix; [R|t] -1 is the inverse matrix of the extrinsic parameter matrix [R|t]; K -1 is the inverse matrix of the intrinsic parameter matrix K; is the image coordinate after distortion correction.

[0035] As a further improvement of this technical solution, in S4, based on the charging gun position information and the charging port position information, calculate the movement path of the robotic arm. The specific method is as follows:

[0036] S4.1.1. Define the position coordinate of the charging gun The position coordinate of the charging port

[0037] The initial pose of the robotic arm The target pose of the robotic arm

[0038] Among them, R0 is the initial rotation matrix; t0 is the initial translation vector; R d is the target rotation matrix; t d is the target translation vector;

[0039] S4.1.2. Use the RRT algorithm to calculate the movement path of the robotic arm from the initial pose T0 to the target pose T d of the robotic arm {T0, T1,..., T n}.

[0040] As a further improvement of this technical solution, in S4, through the real-time feedback of the active vision system and calculating the influence of the vibration of the robotic arm, the robotic arm automatically completes the plugging and unplugging operation. The specific method is as follows:

[0041] S4.2.1. Calculate the position P of the end effector of the current robotic arm current and the error e between the target insertion position P c :

[0042] e = P c - P current

[0043] Among them, P c is the target insertion position, that is, the position coordinate of the charging port; P current is the position of the end effector of the current robotic arm; e is the position P of the end effector of the current robotic arm current and the target insertion position P cError;

[0044] S4.2.2. Measure the vibration and attitude changes of the robotic arm, and collect the vibration data of the robotic arm in real time, including the robotic arm acceleration a(t) and the robotic arm angular velocity ω(t):

[0045]

[0046] where t is time; a x (t) is the acceleration of the robotic arm in the x-axis direction; a y (t) is the acceleration of the robotic arm in the y-axis direction; a z (t) is the acceleration of the robotic arm in the z-axis direction; ω x (t) is the angular velocity of the robotic arm around the x-axis; ω y (t) is the angular velocity of the robotic arm around the y-axis; ω z (t) is the angular velocity of the robotic arm around

[0047] the z-axis;

[0048] S4.2.3. Establish the vibration model V(t) of the robotic arm;

[0049] S4.2.4. According to the current vibration model V(t) and the position error e, adjust the robotic arm movement path {T0, T1,..., T n} in real time.

[0050] As a further improvement of this technical solution, in S4.2.3, the specific method for establishing the vibration model V(t) of the robotic arm is as follows:

[0051]

[0052] where t is time; Δx(t) is the displacement change of the robotic arm in the x-axis direction; Δy(t) is the displacement change of the robotic arm in the y-axis direction; Δz(t) is the displacement change of the robotic arm in the z-axis direction; Δθ x (t) is the angular change of the robotic arm around the x-axis; Δθ y (t) is the angular change of the robotic arm around the y-axis; Δθ z (t) is the angular change of the robotic arm around the z-axis.

[0053] As a further improvement of this technical solution, in S4.2.4, according to the current vibration model V(t) and the position error e, adjust the robotic arm movement path T0, T1,..., T n}, the specific method is as follows:

[0054] P adjusted (t) = P current (t) - V(t) + e;

[0055] Among them, P adjusted (t) adjusts the position of the robotic arm at time t in real time.

[0056] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0057] 1. In the method for automatically plugging and unplugging the charging gun of the mobile charging robot based on the active vision positioning technology, the accurate positioning problems of the vehicle attitude and the position of the charging gun are solved through the multi-modal SLAM attitude estimation algorithm, realizing efficient and reliable automatic plugging and unplugging of the charging gun operation.

[0058] 2. In the method for automatically plugging and unplugging the charging gun of the mobile charging robot based on the active vision positioning technology, the position of the charging port is corrected in real time through the advanced active vision positioning algorithm, and the influence brought by the vibration of the calculation robotic arm is introduced to realize the precise docking and plugging and unplugging of the charging gun. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 is the overall method flow chart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0060] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0061] Embodiment:

[0062] Please refer to Figure 1 As shown, this embodiment provides a method for automatically plugging and unplugging the charging gun of a mobile charging robot based on the active vision positioning technology, including the following steps:

[0063] S1. Take a vehicle image and input it into the cloud large model to automatically identify the vehicle model;

[0064] Taking a vehicle image and inputting it into the cloud large model to automatically identify the vehicle model, the specific method is as follows:

[0065] S1.1. Use a camera to capture multi-angle vehicle images of the license plate and vehicle body markings;

[0066] S1.2. Denoise the vehicle image and use the Canny edge detection algorithm to extract the edge information in the image;

[0067] S1.3. Compress the preprocessed vehicle image and upload it to the cloud large model to automatically identify the vehicle model.

[0068] S2. Use a multi-modal SLAM pose estimation algorithm to locate the vehicle's pose, determine the position information of the charging gun according to the vehicle model, and generate a three-dimensional model of the vehicle's pose and the position coordinates of the charging gun;

[0069] The multi-modal SLAM pose estimation algorithm is based on the ORB-SLAM algorithm and the multi-modal data fusion algorithm. By fusing the vehicle image data and the point cloud data, a three-dimensional model of the vehicle's pose and the position coordinates of the charging gun are generated. The specific method is as follows:

[0070] S2.1. Obtain the vehicle image data I(x, y) from S1, and use a lidar to obtain the three-dimensional point cloud data P of the vehicle cloud ;

[0071] S2.2. Use the ORB-SLAM algorithm to process the vehicle image data and the three-dimensional point cloud data to generate a preliminary vehicle pose estimation and an environmental map;

[0072] S2.3. Combine the preliminary vehicle pose estimation and the environmental map, use the HRNet deep learning model to process the vehicle image data, and use the PointNet++ algorithm to process the three-dimensional point cloud data to generate a three-dimensional model of the vehicle;

[0073] S2.4. Use the Kalman filter and the BA optimization algorithm to generate the accurate pose [R opt , t opt of the vehicle and the position coordinates of the charging gun

[0074] where I(x, y) is the vehicle image data; [R opt , t opt is the accurate pose of the vehicle; R opt is the rotation matrix of the vehicle in three-dimensional space, which is a 3x3 orthogonal matrix; t opt is the position vector of the vehicle in three-dimensional space, which is a 3x1 vector;

[0075] are the position coordinates of the charging gun;

[0076] where the ORB-SLAM algorithm is a feature-based SLAM algorithm for real-time construction of an environmental map and positioning; the multi-modal data fusion algorithm is used to improve the accuracy of positioning and mapping by fusing data from different sensors;

[0077] Among them, the HRNet deep learning model is a deep learning model for image processing tasks, especially suitable for tasks such as pose estimation, object detection, and image segmentation; the PointNet++ algorithm is a deep learning algorithm for processing three-dimensional point cloud data, which is an extended version of PointNet. The PointNet++ algorithm captures local structural information of point cloud data better through hierarchical feature extraction and aggregation;

[0078] Among them, the Kalman filter is a recursive filter for linear dynamic systems, which can estimate the state of the system in noisy measurement data; the BA optimization algorithm is a method for simultaneously optimizing the poses of multiple cameras and the positions of three-dimensional points, and is commonly used in multi-view geometry problems. The goal of BA is to optimize camera parameters and the positions of three-dimensional points by minimizing the total reprojection error;

[0079] S3. Take an image of the charging port and use an advanced active vision positioning algorithm to real-time locate the position coordinates of the charging port;

[0080] The advanced active vision positioning algorithm is based on an active vision system, uses image processing methods and deep learning models to real-time locate the position information of the charging port. The specific method is as follows:

[0081] S3.1. Use a camera to capture a real-time image of the charging port and perform denoising and edge enhancement processing on the image;

[0082] S3.2. Determine the internal parameter matrix K and external parameter matrix [R|t] of the camera, and convert the three-dimensional world coordinates into two-dimensional image coordinates;

[0083]

[0084] [R|t]

[0085]

[0086] Among them, K is the camera internal parameter matrix; f x is the focal length of the camera in the x direction; f y is the focal length of the camera in the y direction; c x is the offset of the principal point of the camera in the x direction; c y is the offset of the principal point of the camera in the y direction;

[0087] R is the rotation matrix; t is the translation vector; is the homogeneous coordinate in the two-dimensional image coordinate; is the homogeneous coordinate in the three-dimensional world coordinate;

[0088] S3.3. Apply the distortion correction formula to minimize the total reprojection error E(R, t) to obtain the optimized R and t;

[0089] S3.4. Apply the inverse perspective transformation and obtain the position coordinates of the charging port through the transformation matrix T wc Obtain the position coordinates of the charging port

[0090] Apply the distortion correction formula to minimize the total reprojection error. The specific method is as follows:

[0091] (u d , v d ) = distort(u, v; k1, k2, p1, p2)

[0092]

[0093] where (u d , v d ) are the image coordinates after distortion correction; (u, v) are the original image coordinates without distortion correction; k1 is the quadratic term radial distortion coefficient, k2 is the quartic term radial distortion coefficient; p1, p2 are the tangential distortion coefficients; E(R, t) represents the total reprojection error under the given rotation matrix R and translation vector t;

[0094] q i is the observed coordinate of the i-th feature point in the image coordinate system; Q i is the three-dimensional coordinate of the i-th feature point in the world coordinate system; N is the total number of feature points;

[0095] where the distortion coefficients k1, k2, p1, p2 are obtained through the camera calibration process, and these coefficients are used to correct the distortion in the image. The steps for the distortion coefficients k1, k2, p1, p2 include:

[0096] 1. Take calibration board images: Use the camera to take multiple images of the calibration board.

[0097] 2. Extract feature points: Identify feature points such as corner points or dot points on the calibration board.

[0098] 3. Optimization: Use a non-linear optimization method to minimize the difference between the image coordinates and the true coordinates of the calibration board, and calculate the distortion coefficients.

[0099] Apply the inverse perspective transformation and obtain the position coordinates of the charging port through the transformation matrix T wc Obtain the position coordinates of the charging port

[0100] The specific method is as follows:

[0101]

[0102] where are the position coordinates of the charging port; T wcis the transformation matrix; [R|t] -1 is the inverse matrix of the external parameter matrix [R|t}; K -1 is the inverse matrix of the internal parameter matrix K; represents the image coordinates after distortion correction.

[0103] S4. Based on the charging gun position information and the charging port position information, introduce the influence of the robotic arm vibration, use the RRT algorithm to calculate the robotic arm movement path, and control the robotic arm to complete the plugging and unplugging operation through the real-time feedback of the active vision system;

[0104] Calculate the robotic arm movement path based on the charging gun position information and the charging port position information. The specific method is as follows:

[0105] S4.1.1. Define the charging gun position coordinates Charging port position coordinates

[0106] The initial pose of the robotic arm The target pose of the robotic arm

[0107] Among them, R0 is the initial rotation matrix; t0 is the initial translation vector; R d is the target rotation matrix; t d is the target translation vector;

[0108] S4.1.2. Use the RRT algorithm to calculate the robotic arm movement path from the initial pose T0 to the target pose T d {T0, T1,..., T n};

[0109] Among them, the RRT algorithm is an algorithm for path planning, used for fast path search in high-dimensional space. The RRT algorithm constructs a random tree, gradually expands the search space, and finds a feasible path from the starting point to the target point.

[0110] Through the real-time feedback of the active vision system and calculate the influence of the robotic arm vibration, realize the robotic arm to automatically complete the plugging and unplugging operation. The specific method is as follows:

[0111] S4.2.1. Calculate the current position P of the end effector of the robotic arm current and the error e between the target insertion position P c :

[0112] e = P c - P current

[0113] Among them, P c is the target insertion position, that is, the charging port position coordinates; P currentis the position of the end effector of the current robotic arm; e is the position P of the end effector of the current robotic arm current and the target insertion position P c error;

[0114] S4.2.2. Measure the vibration and attitude changes of the robotic arm, and collect the vibration data of the robotic arm in real time, including the robotic arm acceleration a(t) and the robotic arm angular velocity ω(t):

[0115]

[0116] where t is time; a x (t) is the acceleration of the robotic arm in the x-axis direction; a y (t) is the acceleration of the robotic arm in the y-axis direction; a z (t) is the acceleration of the robotic arm in the z-axis direction; ω x (t) is the angular velocity of the robotic arm around the x-axis; ω y (t) is the angular velocity of the robotic arm around the y-axis; ω z (t) is the angular velocity of the robotic arm around the z-axis;

[0117] S4.2.3. Establish the vibration model V(t) of the robotic arm;

[0118] S4.2.4. According to the current vibration model V(t) and the position error e, adjust the robotic arm movement path {T0, T1,..., T n} in real time;

[0119] The specific method for establishing the vibration model V(t) of the robotic arm is as follows:

[0120]

[0121] where t is time; Δx(t) is the displacement change of the robotic arm in the x-axis direction; Δy(t) is the displacement change of the robotic arm in the y-axis direction; Δz(t) is the displacement change of the robotic arm in the z-axis direction; Δθ x (t) is the angular change of the robotic arm around the x-axis; Δθ y (t) is the angular change of the robotic arm around the y-axis; Δθ z (t) is the angular change of the robotic arm around the z-axis.

[0122] According to the current vibration model V(t) and the position error e, adjust the robotic arm movement path {T0, T1,..., T n} in real time, and the specific method is as follows:

[0123] P adjusted (t) = P current (t) - V(t) + e;

[0124] Among them, P adjusted (t) adjusts the position of the robotic arm at time t in real time.

[0125] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed.

Claims

1. An automatic plugging and unplugging method of a charging gun for a mobile charging robot based on active vision positioning technology, characterized in that, It includes the following steps: S1. Take a vehicle image and input it into the cloud large model to automatically identify the vehicle model; S2. Use the multi-modal SLAM pose estimation algorithm to locate the vehicle pose, determine the charging gun position information according to the vehicle model, and generate a 3D model of the vehicle pose and the charging gun position coordinates; S3. Take an image of the charging port and use the advanced active vision positioning algorithm to real-time locate the charging port position coordinates; S4. Based on the charging gun position information and the charging port position information, introduce the influence of the manipulator vibration, use the RRT algorithm to calculate the manipulator movement path, and control the manipulator to complete the plugging and unplugging operation through the real-time feedback of the active vision system; In S3, the advanced active vision positioning algorithm is based on the active vision system, uses image processing methods and deep learning models to real-time locate the charging port position information. The specific method is as follows: S3.

1. Use a camera to capture the real-time image of the charging port, and perform denoising and edge enhancement processing on the image; S3.

2. Determine the intrinsic matrix and the extrinsic matrix of the camera, and convert the three-dimensional world coordinates into two-dimensional image coordinates; ； ； wherein is the focal length of the camera in the direction; is the focal length of the camera in the direction; is the principal point offset of the camera in the direction; is the principal point offset of the camera in the direction; is the rotation matrix; is the translation vector; is the homogeneous coordinate in the two-dimensional image coordinate; is the homogeneous coordinate in the three-dimensional world coordinate; S3.

3. Apply the distortion correction formula to minimize the total reprojection error , to obtain the optimized and ; S3.

4. Apply inverse perspective transformation and obtain the position coordinates of the charging port through the transformation matrix ;​ In S3.3, apply the distortion correction formula to minimize the reprojection error. The specific method is as follows: ； ； Among them, is the image coordinate after distortion correction; is the original image coordinate without distortion correction; is the quadratic term radial distortion coefficient, is the quartic term radial distortion coefficient; is the tangential distortion coefficient; represents the total reprojection error under the given rotation matrix and translation vector ; is the -th feature point's coordinate in the two-dimensional image coordinate; is the -th feature point's coordinate in the three-dimensional world coordinate; is the total number of feature points; In the step S3.4, an inverse perspective transformation is applied, and the position coordinates of the charging port are obtained through the transformation matrix as follows: Specifically, the method is as follows: ； Among them, is the inverse matrix of the external parameter matrix ; is the inverse matrix of the internal parameter matrix ; The specific method in S4 is as follows: S4.2.

1. Calculate the position of the end effector of the current robotic arm and the target insertion position to obtain the position error : ； S4.2.

2. Measure the vibration and attitude changes of the robotic arm, and collect the vibration data of the robotic arm in real time, including the acceleration of the robotic arm and the angular velocity of the robotic arm : ； ； Among them, is time; is the acceleration of the robotic arm in the axis direction; is the acceleration of the robotic arm in the axis direction; is the acceleration of the robotic arm in the axis direction; is the angular velocity of the robotic arm around the axis; is the angular velocity of the robotic arm around the axis; is the angular velocity of the robotic arm around the axis; S4.2.

3. Establish the vibration model of the robotic arm ; S4.2.

4. Adjust the moving path of the robotic arm in real time according to the current vibration model and the position error . .

2. The method for automatically plugging and unplugging a charging gun of a mobile charging robot based on the active vision positioning technology according to claim 1, wherein: In S1, take a vehicle image and input it into the cloud large model to automatically identify the vehicle model. The specific method is as follows: S1.

1. Use a camera to capture multi-angle vehicle images of the license plate and vehicle body identification; S1.

2. Perform denoising processing on the vehicle image, and use the Canny edge detection algorithm to extract the edge information in the image; S1.

3. Compress the preprocessed vehicle image and upload it to the cloud large model to automatically identify the vehicle model.

3. The method for automatically plugging and unplugging a charging gun of a mobile charging robot based on an active vision positioning technology according to claim 2, wherein: In S2, the multi-modal SLAM pose estimation algorithm is based on the ORB-SLAM algorithm and the multi-modal data fusion algorithm. By fusing the vehicle image data and the point cloud data, generate a 3D model of the vehicle pose and the charging gun position coordinates. The specific method is as follows: S2.

1. Obtain vehicle image data from S1 , and obtain the three-dimensional point cloud data of the vehicle using lidar ; S2.

2. Use the ORB-SLAM algorithm to process the vehicle image data and the 3D point cloud data to generate a preliminary vehicle pose estimation and an environmental map; S2.

3. Combine the preliminary vehicle pose estimation and the environmental map, use the HRNet deep learning model to process the vehicle image data, and use the PointNet++ algorithm to process the 3D point cloud data to generate a 3D vehicle model; S2.

4. Generate the accurate attitude of the vehicle and the position coordinates of the charging gun using the Kalman filter and the BA optimization algorithm and the position coordinates of the charging gun ; Among them, is the rotation matrix of the vehicle in three-dimensional space, which is a 3x3 orthogonal matrix; is the position vector of the vehicle in three-dimensional space, which is a 3x1 vector; is the position coordinate of the charging gun.

4. The method for automatically plugging and unplugging a charging gun of a mobile charging robot based on an active vision positioning technology according to claim 3, characterized in that: In S4, calculate the manipulator movement path based on the charging gun position information and the charging port position information. The specific method is as follows: S4.1.

1. Define the position coordinates of the charging gun , the position coordinates of the charging port , the initial pose of the robotic arm , the target pose of the robotic arm ; Among them, is the initial rotation matrix; is the initial translation vector; is the target rotation matrix; is the target translation vector; S4.1.

2. Calculate the manipulator movement path from the initial pose to the target pose using the RRT algorithm. .

5. The method for automatically plugging and unplugging a charging gun of a mobile charging robot based on an active vision positioning technology according to claim 4, wherein: In the above S4.2.3, establish the vibration model of the robotic arm The specific method is as follows: ； Among them, is the time; is the displacement change of the robotic arm in the axis direction; is the displacement change of the robotic arm in the axis direction; is the displacement change of the robotic arm in the axis direction; is the angular change of the robotic arm around the axis; is the angular change of the robotic arm around the axis; is the angular change of the robotic arm around the axis.

6. The method for automatically plugging and unplugging a charging gun of a mobile charging robot based on an active vision positioning technology according to claim 5, wherein: In the above S4.2.4, according to the current vibration model and the position error , the moving path of the robotic arm is adjusted in real time , and the specific method is as follows: ； Among them, real-time adjustment of the position of the robotic arm at moment.

Images