CN108453738A - A kind of quadrotor based on Opencv image procossings independently captures the control method of operation in the air - Google Patents

Abstract

The invention proposes a control method for autonomous aerial grasping operation of a four-rotor aircraft based on Opencv image processing. The quadrotor aircraft operating system is composed of the quadrotor aircraft and the operating device, wherein the operating device is composed of a mechanical arm and a center-of-gravity balance mechanism installed symmetrically under the quadrotor aircraft. The present invention utilizes the camera on the quadrotor aircraft to obtain images of captured targets and artificial signs, uses Opencv to process the images and solves the camera pose information, thereby obtaining the pose information of the quadrotor aircraft, and classifies according to the pose information of the quadrotor aircraft Control the quadrotor aircraft, robotic arm and center of gravity balance mechanism. The invention broadens the application field of the four-rotor aircraft, and can carry out aerial work both indoors and outdoors with dense surrounding buildings, thereby improving the precision of aerial work.

Description

A self-grasping operation of quadrotor aircraft in the air based on Opencv image processing control method

technical field

The invention relates to the technical field of aerial robot control, in particular to a control method for an autonomous aerial grasping operation of a quadrotor aircraft based on Opencv image processing.

Background technique

Quadrotor aircraft has always been a research hotspot in the military and commercial fields due to its maneuverability and portability. It is widely used in aerial surveys, pesticide spraying and plant protection, and target tracking and positioning. Compared with the mobile operation of ground mobile robots, unmanned aerial vehicles with aerial operation capabilities expand the operating space to three dimensions, which broadens the application field of unmanned aerial vehicles, handling dangerous objects in complex environments, placing and retrieving survey equipment, It has a broad application prospect in the capture of air targets and so on.

In the research of quadrotor aircraft, the pose estimation problem, that is, the localization problem, has always been the core problem of the research. The accurate estimation of the quadrotor's own position and orientation is an important basis and premise for research directions such as quadrotor control, obstacle avoidance, path planning, temporary habitat, and grasping. At present, there are mainly two types of methods for the positioning of quadrotor aircraft: one is to obtain position information based on the vision of the quadrotor aircraft itself. The other is to provide location information based on external devices, such as the global satellite positioning system (GPS), motion capture system. However, GPS is easily affected by the surrounding environment and climate. In the case of indoor or dense surrounding buildings, it is difficult for drones to rely on GPS to obtain accurate location information. The motion capture system can only be used in specific research places such as laboratories, which limits the application range of quadrotor aircraft.

Contents of the invention

In order to solve the above-mentioned existing problems, the invention provides a kind of Opencv image processing based on Opencv image processing in view of the characteristics of high precision, high stability, simple operation and flexible operation when the quadrotor aircraft uses its own mechanical arm to grab objects. The control method of the quadrotor aircraft's aerial autonomous grasping operation has broadened the application field of the quadrotor aircraft, and can perform air operations indoors and outdoors in dense surrounding buildings, improving the accuracy of aerial operations. In order to achieve this purpose, The present invention provides a kind of control method based on Opencv image processing four-rotor aircraft air autonomous grasping operation, concrete steps are as follows:

Step 1: Build a system implementation platform, which is composed of a quadrotor aircraft and an aerial operation device, wherein the aerial operation device includes a three-degree-of-freedom mechanical arm and a single-degree-of-freedom center-of-gravity balance mechanism; Derivation of kinematics and dynamics equations for operating devices and quadrotors;

Step 2: After calibrating the camera, select any sign in the ArUco library, use the recognition algorithm based on OpenCV_ArUco to identify the positioning sign, and convert the position and attitude information of the sign in the camera coordinate system to the pose information of the camera in the world coordinate system;

Step 3: The entire control system is divided into three levels, which are aircraft control, mechanical arm control, and center of gravity balance mechanism control; set an air operation range value S, and judge the quadrotor aircraft based on the position and orientation information of the quadrotor aircraft obtained by image processing If the quadrotor aircraft does not enter the air operation range, control the quadrotor aircraft to approach or stay away from the artificial mark until it enters the air operation range; otherwise, the quadrotor aircraft hovers, plans the angles of the joints of the manipulator and calculates the center of gravity balance mechanism joints The movement angle can reduce the disturbance of the quadrotor aircraft caused by the movement of the robotic arm, and control the robotic arm to grab the target autonomously.

As a further improvement of the present invention, in step one, the steps of deriving kinematics and dynamic equations for the operating device and the quadrotor aircraft are as follows:

Step 1: Assuming that the quadrotor aircraft, the mechanical arm and the center of gravity balance mechanism are all rigid bodies, the relevant coordinate system is defined as follows: the body coordinate system {O _b }-x _b y _b z _b takes the centroid of the quadrotor aircraft as the origin , the coordinate axis x _b is consistent with the forward direction of the quadrotor aircraft, the coordinate axis z _b is perpendicular to the propeller rotation plane and along the ascent direction of the quadrotor aircraft, {O _c }-x _c y _c z _c and {O _i }-x _i y _i z _i represent the coordinate system of the center of gravity balance mechanism and the coordinate system of the manipulator, respectively. The camera is fixed on the quadrotor aircraft, above the manipulator with the lens facing down; the captured target is a regular-shaped lightweight cuboid or Cube or cylinder, the size of which does not exceed the size range that the manipulator can grab and is placed horizontally on the ground;

Step 2: Carry out kinematic modeling of the mechanical arm and the center of gravity balance mechanism by the D-H method, and the modeling parameters are as follows:

Among them, θ _i , d _i , α _i-1 , a _i-1 are the modeling parameters of the DH method, i=1, 2, 3 are the joint numbers of the mechanical arm, c is the joint number of the center of gravity balance mechanism, 0, 0' is the installation reference frame number, H=[θ ₁ ,θ ₂ ,θ ₃ ] ^T is the joint angle of the manipulator, and θ _c is the joint angle of the center of gravity balance mechanism. It can be seen that the homogeneous transformation matrix from the joint coordinate system 3 of the manipulator to the installation reference coordinate system 0, that is, the positive kinematic equation is:

⁰ T ₃ = A ₁ A ₂ A ₃ (1);

Among them, the expression of A _i is as follows:

After solving the inverse kinematics of the forward motion equation, the value of each joint can be determined, so that the mechanical arm can reach the desired pose;

Step 3: Model the kinematics of the quadrotor:

The relationship between the body angular velocity and the Euler angle change rate between the world coordinate system and the body coordinate system is as follows:

Among them, P=[x,y,z] ^T represents the origin of the coordinate system {O _b } in the coordinate system {O _w }, Φ=[φ,θ,ψ] ^T represents the coordinate system {O _b } in The roll angle, pitch angle and yaw angle measured in the coordinate system {O _w }, V=[u,v,w] ^T represents the linear velocity of the coordinate system {O _b } relative to the coordinate system {O _w }, Ω=[p,q,r] ^T represents the angular velocity of the quadrotor aircraft, R is the rotation matrix rotating in ZYX order, and Q is the conversion matrix between the angular velocity of the body and the rate of change of the Euler angle;

Step 4: According to the Newton-Euler method, the dynamic modeling of the composite system of the quadrotor aircraft with the aerial work device can be carried out:

For a composite system consisting of a quadrotor aircraft, a robotic arm, and a center-of-gravity balance mechanism, it is assumed that the rotation of each joint of the robot arm and the center-of-gravity balance mechanism is smooth and slow, and this slow change only affects the center of gravity and moment of inertia of the system. The dynamic equations of the composite system at the center of gravity in the carrier coordinate system can be derived through the Newton-Euler dynamic equations, and the dynamic equations can be obtained:

Among them, m represents the mass of the compound system, G=[0,0,-mg] ^T represents the gravity vector of the compound system, I _m represents the inertia tensor at the center of gravity of the compound system, and the center of gravity offset vector r _G =[x _G ,y _G , z _G ] ^T represents the offset from the center of gravity of the composite system to the body coordinate system {O _b }. F _p ＝[0,0,f _p ] ^T represents the propulsion vector of the blade acting on the compound system, τ _p ＝[τ _x ,τ _y ,τ _z ]T is the torque generated by the propulsive force acting on the compound system vector, is the torque generated by the interaction force between the system and the environment. Considering the relatively slow motion of the manipulator and the state of the manipulator before or after grasping, the interaction force F _e of the composite system and the environment is 0.

Further improvement of the present invention, the detection of artificial mark in step 2 and the conversion process steps between coordinate systems are as follows:

The artificial logo selects the corresponding logo in the ArUco library, and the camera imaging uses the small hole imaging model:

P=K[R|T]Q(5);

Among them, Q is the coordinates of the logo in the three-dimensional space, [R|T] is the camera external parameter matrix, which is used to transform the logo in the world coordinates into the camera coordinates, K is the camera internal parameter, which is used to convert a point in the camera coordinate projected onto the image plane. P is the pixel coordinates after the sign projection. Use the recognition algorithm based on OpenCV_ArUco to identify and locate artificial signs. The detection of artificial signs and the conversion between coordinate systems have the following steps:

Step 1: Use the chessboard to calibrate the camera, and determine the internal parameter matrix K of the camera;

Step 2: Detect signs, including:

Step 2.1: Search for all candidate landmarks in the image, use adaptive thresholding to segment the landmarks, and then extract contours from the thresholded image, and discard those that are not convex polygons, and those that are not square, and use some additional filtering;

Step 2.2: Analyze the internal codes of candidate signs to determine if they are indeed signs, including:

Step 2.2.1: Perform perspective transformation on the image to obtain its normal form;

Step 2.2.2: Thresholding the normalized image with Ossu to separate white and black bits, extracting the flag bits of each flag and analyzing the flag number;

Step 3: After the sub-pixel refinement at the corner, according to the pinhole imaging model, using the pixel coordinates of the four vertices marked by ArUco, the rotation matrix R and translation matrix T of the camera can be obtained by the solvePnP method;

Step 4: Obtain the position and attitude information of the camera according to the Euclidean transformation relationship between the camera coordinate system and the world coordinate system. The solution formula is as follows:

P _c ＝-R ^-1 T

yaw(ψ)=atan(r ₂₁ /r ₁₁ )(6);

pitch(θ)=asin(-r ₃₁ )

roll(φ)=atan(r ₃₂ /r ₃₃ )

Among them, P _c represents the coordinates of the camera in the world coordinate system.

Further improvement of the present invention, in step 3, the hierarchical control quadrotor aircraft and operating device steps are as follows:

Assuming that the operating range of the quadrotor aircraft to grab the target with known coordinates is S, the pose information of the quadrotor aircraft is obtained by calculating the camera pose information after image processing shown in Figure 4. If the quadrotor aircraft does not enter the operating range S, the quadrotor aircraft The rotorcraft is close to or far away from the artificial marker, and the target trajectory generator outputs the target position P ^* = [x ^* , y ^* , z ^* ] ^T and the yaw angle ψ _d to the position controller and the attitude stability controller respectively, and the position control The main body of the controller is the PD controller, and its output is calculated by the system dynamics equation to obtain the attitude angle θ _d , φ _d required for attitude control. The attitude stability controller is designed by the inversion method, and its output is decoupled from the motor The rotational speed ω _i (i=1,2,3,4) is output to the motor until the quadrotor aircraft enters the operating range S;

If the quadrotor aircraft enters the operating range S, the quadrotor aircraft hovers, and the pose information of the quadrotor aircraft is output to the trajectory generator of the manipulator. The trajectory generator of the manipulator first converts the pose information of the quadrotor aircraft into the target pose information and then plans Each joint movement angle θ _i (i=1,2,3) is output to the center of gravity balance controller, and the center of gravity balance controller calculates the joint movement angle θ _c of the center of gravity balance mechanism according to the formula of the center of gravity balance equation, and finally sends all joint angle signals wirelessly When the steering gear controller controls the operating device to complete the grabbing, the center of gravity balance equation formula is:

∑ m _i x _i + m _c x _c + m _t x _t = 0(7);

Among them, m _i , m _c , m _t are the mass of each part of the manipulator arm, the center of gravity balance mechanism and the grasped object, and _xi , x _c , x _t are the mass of each part of the manipulator arm, the center of gravity balance mechanism and the grasped object. The position of the center of gravity in the machine system {O _b }.

The present invention provides a kind of control method based on Opencv image processing four-rotor aircraft aerial autonomous grasping operation, beneficial effect is as follows:

1. Overcome the application limitations of GPS and motion capture systems;

2. Improve the speed and grasping accuracy of aiming at the target, reduce battery consumption and prolong battery life;

3. The hierarchical control of the quadrotor aircraft and the operating device not only ensures the flight stability of the quadrotor aircraft system, but also ensures the stability of the grasping operation process.

Description of drawings

Fig. 1 is a schematic diagram of the system structure of the present invention;

Fig. 2 is a schematic diagram of a quadrotor aircraft coordinate system;

Fig. 3 is a schematic diagram of the world coordinate system of the captured target and artificial markers;

Fig. 4 is the flow chart of Opencv image processing;

Fig. 5 is a structural diagram of the control system;

Figure 6 is a flow chart of the host computer control.

Detailed ways

Below in conjunction with accompanying drawing and specific embodiment the present invention is described in further detail:

The present invention provides a control method based on Opencv image processing for four-rotor aircraft autonomous grabbing operations in the air, which broadens the application field of four-rotor aircraft, and can perform air operations indoors and outdoors in dense surrounding buildings, improving Aerial work accuracy.

As shown in Figure 1, the system implementation platform of the present invention mainly consists of a quadrotor aircraft 1, a flight controller 2, a wireless communication module, a steering gear controller 3, a camera 4, and a three-degree-of-freedom mechanical arm including a manipulator 5, a joint-6, Joint three 7, joint two 8, center of gravity balance joint 9, battery 10, PC 11, captured target 12 and artificial sign 13 form. The working device includes a three-degree-of-freedom robot arm and a single-degree-of-freedom center-of-gravity balance mechanism, which are symmetrically installed under the quadrotor aircraft; in the system design, the influence of the movement of the robot arm on the quadrotor aircraft is regarded as a disturbance to the flight control , using the center of gravity balance mechanism to reduce the disturbance effect of the mechanical arm movement on the quadrotor aircraft.

Specific steps are as follows:

Step 1. Deriving the kinematics and dynamics equations for the working device and the quadrotor aircraft respectively

Step 1: The following definitions are made in this patent:

Assuming that the quadrotor aircraft, the mechanical arm, and the center-of-gravity balance mechanism are all rigid bodies, the relevant coordinate systems are defined as follows: As shown in Figure 2, the body coordinate system {O _b }-x _b y _b z _b takes the shape The center is the origin, the coordinate axis x _b is consistent with the forward direction of the quadrotor aircraft, the coordinate axis z _b is perpendicular to the propeller rotation plane and along the rising direction of the quadrotor aircraft, {O _c }-x _c y _c z _c and {O _i } -x _i y _i z _i represent the coordinate system of the center of gravity balance mechanism and the coordinate system of the manipulator, respectively. The camera is fixed on the quadrotor aircraft, above the manipulator with the lens facing down; A quality cuboid or cube or cylinder whose size does not exceed the size range that the manipulator can grab and is placed horizontally on the ground. This patent takes a lightweight cuboid as an example, as shown in Figure 3, the world coordinate system {O _w }-x _w y _w z _w takes the center of the artificial mark 13 on the captured target 12 as the origin, and z _w is opposite to the direction of gravity.

Step 2: Carry out kinematic modeling of the mechanical arm and the center of gravity balance mechanism by the D-H method, and the modeling parameters are as follows:

Table 1 Modeling parameters of manipulator and center of gravity balance mechanism D-H method;

Among them, θ _i , d _i , α _i-1 , a _i-1 are the modeling parameters of the DH method, i=1, 2, 3 are the joint numbers of the mechanical arm, c is the joint number of the center of gravity balance mechanism, 0, 0' is the installation reference system number, H=[θ ₁ ,θ ₂ ,θ ₃ ]T is the joint angle of the manipulator, and θ _c is the joint angle of the center of gravity balance mechanism. It can be seen that the homogeneous transformation matrix from the joint coordinate system 3 of the manipulator to the installation reference coordinate system 0, that is, the positive kinematic equation is:

⁰ T ₃ = A ₁ A ₂ A ₃ (1);

Among them, the expression of A _i is as follows:

After solving the inverse kinematics of the forward motion equation, the value of each joint can be determined, so that the robot arm can reach the desired pose.

Step 3: Model the kinematics of the quadrotor:

The relationship between the body angular velocity and the Euler angle change rate between the world coordinate system and the body coordinate system is as follows:

Among them, P=[x,y,z] ^T represents the origin of the coordinate system {O _b } in the coordinate system {O _w }, Φ=[φ,θ,ψ] ^T represents the coordinate system {O _b } in The roll angle, pitch angle and yaw angle measured in the coordinate system {O _w }, V=[u,v,w] ^T represents the linear velocity of the coordinate system {O _b } relative to the coordinate system {O _w }, Ω=[p,q,r] ^T represents the angular velocity of the quadrotor aircraft, R is the rotation matrix rotating in the order of ZYX, and Q is the conversion matrix between the angular velocity of the body and the rate of change of Euler angle.

Step 4: According to the Newton-Euler method, the dynamic modeling of the composite system of the quadrotor aircraft with the aerial work device can be carried out:

For a composite system consisting of a quadrotor aircraft, a robotic arm, and a center-of-gravity balance mechanism, it is assumed that the rotation of each joint of the robot arm and the center-of-gravity balance mechanism is smooth and slow, and this slow change only affects the center of gravity and moment of inertia of the system. The dynamic equations of the composite system at the center of gravity in the carrier coordinate system can be derived through the Newton-Euler dynamic equations, and the dynamic equations can be obtained:

Among them, m represents the mass of the compound system, G=[0,0,-mg] ^T represents the gravity vector of the compound system, I _m represents the inertia tensor at the center of gravity of the compound system, and the center of gravity offset vector r _G =[x _G ,y _G , z _G ] ^T represents the offset from the center of gravity of the composite system to the body coordinate system {O _b }. F _p ＝[0,0,f _p ] ^T represents the propulsion vector of the blade acting on the compound system, τ _p ＝[τ _x ,τ _y ,τ _z ] ^T is the torque generated by the propulsive force acting on the compound system vector, is the torque generated by the interaction force between the system and the environment. Considering the relatively slow motion of the manipulator and the state of the manipulator before or after grasping, the interaction force F _e of the composite system and the environment is 0.

Step 2, the detection of artificial marks and the conversion process between coordinate systems;

The artificial sign is selected from the ArUco library as shown in Figure 3, the size is 5cm*5cm, and the number is 0. Camera imaging uses a pinhole imaging model:

P=K[R|T]Q(5);

where Q is the coordinates of the marker in 3D space (relative to the world coordinate system). [R|T] is the camera extrinsic parameter matrix, which is used to transform the logo in the world coordinates to the camera coordinates. K is the internal reference of the camera, which is used to project a point in the camera coordinates onto the image plane. P is the pixel coordinate after logo projection. As shown in Figure 4, using the recognition algorithm based on OpenCV_ArUco to identify and locate artificial signs, the detection of artificial signs and the conversion between coordinate systems have the following steps:

Step 1: Use the chessboard to calibrate the camera, and determine the internal parameter matrix K of the camera;

Step 2: Detect signs, including:

Step 2.1: Search for all candidate landmarks in the image, use adaptive thresholding to segment the landmarks, and then extract contours from the thresholded image, and discard those that are not convex polygons, and those that are not square, and use some additional Filtering (to remove those contours that are too small or too large, too similar convex polygons, etc.);

Step 2.2: Analyze the internal codes of candidate signs to determine if they are indeed signs, including:

Step 2.2.1: Perform perspective transformation on the image to obtain its normal form (front view);

Step 2.2.2: Thresholding the normalized image with Ossu to separate white and black bits, extracting the flag bits of each flag and analyzing the flag number;

Step 3: After the sub-pixel refinement at the corner, according to the pinhole imaging model, using the pixel coordinates of the four vertices marked by ArUco, the rotation matrix R and translation matrix T of the camera can be obtained by the solvePnP method;

Step 4: Obtain the position and attitude information of the camera according to the Euclidean transformation relationship between the camera coordinate system and the world coordinate system. The solution formula is as follows:

P _c ＝-R ^-1 T

yaw(ψ)=atan(r ₂₁ /r ₁₁ )(6);

pitch(θ)=asin(-r ₃₁ )

roll(φ)=atan(r ₃₂ /r ₃₃ )

Among them, Pc represents the coordinates of the camera in the world coordinate system.

Step 3, according to Fig. 5, the control method of Fig. 6, hierarchically control quadrotor aircraft and operating device

Assuming that the operating range of the quadrotor aircraft to grab the target with known coordinates is S, the pose information of the quadrotor aircraft is obtained by calculating the camera pose information after image processing shown in Figure 4. If the quadrotor aircraft does not enter the operating range S, the quadrotor aircraft The rotorcraft is close to or far away from the artificial marker, and the target trajectory generator outputs the target position P ^* = [x ^* , y ^* , z ^* ] ^T and the yaw angle ψ _d to the position controller and the attitude stability controller respectively, and the position control The main body of the controller is the PD controller, and its output is calculated by the system dynamics equation to obtain the attitude angle θ _d , φ _d required for attitude control. The attitude stability controller is designed by the inversion method, and its output is decoupled from the motor The rotational speed ω _i (i=1,2,3,4) is output to the motor until the quadrotor aircraft enters the operating range S;

If the quadrotor aircraft enters the operating range S, the quadrotor aircraft hovers, and the pose information of the quadrotor aircraft is output to the trajectory generator of the manipulator. The trajectory generator of the manipulator first converts the pose information of the quadrotor aircraft into the target pose information and then plans Each joint movement angle θ _i (i=1,2,3) is output to the center of gravity balance controller, and the center of gravity balance controller calculates the joint movement angle θ _c of the center of gravity balance mechanism according to the formula of the center of gravity balance equation, and finally sends all joint angle signals wirelessly Go to the steering gear controller to control the operating device to complete the grabbing. Center of gravity balance equation formula:

∑ m _i x _i + m _c x _c + m _t x _t = 0(7);

Among them, m _i , m _c , m _t are the mass of each part of the manipulator arm, the center of gravity balance mechanism and the grasped object, and _xi , x _c , x _t are the mass of each part of the manipulator arm, the center of gravity balance mechanism and the grasped object. The position of the center of gravity in the machine system {O _b }.

The above is only a preferred embodiment of the present invention, and is not intended to limit the present invention in any other form, and any modification or equivalent change made according to the technical essence of the present invention still belongs to the scope of protection claimed by the present invention .

Claims (4)

1. a kind of control method based on Opencv image processing four-rotor aircraft autonomous grab operation in the air, concrete steps are as follows, it is characterized in that: Step 1: Build a system implementation platform, which is composed of a quadrotor aircraft and an aerial operation device, wherein the aerial operation device includes a three-degree-of-freedom mechanical arm and a single-degree-of-freedom center-of-gravity balance mechanism; Derivation of kinematics and dynamics equations for operating devices and quadrotors; Step 2: After calibrating the camera, select any sign in the ArUco library, use the recognition algorithm based on OpenCV_ArUco to identify the positioning sign, and convert the position and attitude information of the sign in the camera coordinate system to the pose information of the camera in the world coordinate system; Step 3: The entire control system is divided into three levels, which are aircraft control, mechanical arm control, and center of gravity balance mechanism control; set an air operation range value S, and judge the quadrotor aircraft based on the position and orientation information of the quadrotor aircraft obtained by image processing If the quadrotor aircraft does not enter the air operation range, control the quadrotor aircraft to approach or stay away from the artificial mark until it enters the air operation range; otherwise, the quadrotor aircraft hovers, plans the angles of the joints of the manipulator and calculates the center of gravity balance mechanism joints The movement angle can reduce the disturbance of the quadrotor aircraft caused by the movement of the robotic arm, and control the robotic arm to grab the target autonomously. 2. according to claim 1, a kind of control method based on Opencv image processing quadrotor aircraft air autonomous grabbing operation is characterized in that: in the step 1, deriving kinematics and dynamic equation step to operating device and quadrotor aircraft as follows: Step 1: Assuming that the quadrotor aircraft, the mechanical arm and the center of gravity balance mechanism are all rigid bodies, the relevant coordinate system is defined as follows: the body coordinate system {O _b }-x _b y _b z _b takes the centroid of the quadrotor aircraft as the origin , the coordinate axis x _b is consistent with the forward direction of the quadrotor aircraft, the coordinate axis z _b is perpendicular to the propeller rotation plane and along the ascent direction of the quadrotor aircraft, {O _c }-x _c y _c z _c and {O _i }-x _i y _i z _i represent the coordinate system of the center of gravity balance mechanism and the coordinate system of the manipulator, respectively. The camera is fixed on the quadrotor aircraft, above the manipulator with the lens facing down; the captured target is a regular-shaped lightweight cuboid or Cube or cylinder, the size of which does not exceed the size range that the manipulator can grab and is placed horizontally on the ground; Step 2: Carry out kinematic modeling of the mechanical arm and the center of gravity balance mechanism by the D-H method, and the modeling parameters are as follows: Among them, θ _i , d _i , α _i-1 , a _i-1 are the modeling parameters of the DH method, i=1, 2, 3 are the joint numbers of the mechanical arm, c is the joint number of the center of gravity balance mechanism, 0, 0' is the installation reference frame number, H=[θ ₁ ,θ ₂ ,θ ₃ ] ^T is the joint angle of the manipulator, and θ _c is the joint angle of the center of gravity balance mechanism. It can be seen that the homogeneous transformation matrix from the joint coordinate system 3 of the manipulator to the installation reference coordinate system 0, that is, the positive kinematic equation is: ⁰ T ₃ = A ₁ A ₂ A ₃ (1); Among them, the expression of A _i is as follows: After solving the inverse kinematics of the forward motion equation, the value of each joint can be determined, so that the mechanical arm can reach the desired pose; Step 3: Model the kinematics of the quadrotor: The relationship between the body angular velocity and the Euler angle change rate between the world coordinate system and the body coordinate system is as follows: Among them, P=[x,y,z] ^T represents the origin of the coordinate system {O _b } in the coordinate system {O _w }, Φ=[φ,θ,ψ] ^T represents the coordinate system {O _b } in The roll angle, pitch angle and yaw angle measured in the coordinate system {O _w }, V=[u,v,w] ^T represents the linear velocity of the coordinate system {O _b } relative to the coordinate system {O _w }, Ω=[p,q,r] ^T represents the angular velocity of the quadrotor aircraft, R is the rotation matrix rotating in ZYX order, and Q is the conversion matrix between the angular velocity of the body and the rate of change of the Euler angle; Step 4: According to the Newton-Euler method, the dynamic modeling of the composite system of the quadrotor aircraft with the aerial work device can be carried out: For a composite system consisting of a quadrotor aircraft, a robotic arm, and a center-of-gravity balance mechanism, it is assumed that the rotation of each joint of the robot arm and the center-of-gravity balance mechanism is smooth and slow, and this slow change only affects the center of gravity and moment of inertia of the system. The dynamic equations of the composite system at the center of gravity in the carrier coordinate system can be derived through the Newton-Euler dynamic equations, and the dynamic equations can be obtained: Among them, m represents the mass of the compound system, G=[0,0,-mg]T represents the gravity vector of the compound system, I _m represents the inertia tensor at the center of gravity of the compound system, and the center of gravity offset vector r _G =[x _G ,y _G , z _G ] ^T represents the offset from the center of gravity of the composite system to the body coordinate system {O _b }. F _p ＝[0,0,f _p ] ^T represents the propulsion vector of the blade acting on the compound system, τ _p ＝[τ _x ,τ _y ,τ _z ] ^T is the torque generated by the propulsive force acting on the compound system vector, is the torque generated by the interaction force between the system and the environment. Considering the relatively slow motion of the manipulator and the state of the manipulator before or after grasping, the interaction force F _e of the composite system and the environment is 0. 3. a kind of control method based on the four-rotor aircraft air autonomous grasping operation of Opencv image processing according to claim 1, it is characterized in that: the conversion process step between the detection of artificial sign and the coordinate system in step 2 is as follows: The artificial logo selects the corresponding logo in the ArUco library, and the camera imaging uses the small hole imaging model: P=K[R|T]Q(5); Among them, Q is the coordinates of the logo in the three-dimensional space, [R|T] is the camera external parameter matrix, which is used to transform the logo in the world coordinates into the camera coordinates, K is the camera internal parameter, which is used to convert a point in the camera coordinate projected onto the image plane. P is the pixel coordinates after the sign projection. Use the recognition algorithm based on OpenCV_ArUco to identify and locate artificial signs. The detection of artificial signs and the conversion between coordinate systems have the following steps: Step 1: Use the chessboard to calibrate the camera, and determine the internal parameter matrix K of the camera; Step 2: Detect signs, including: Step 2.1: Search for all candidate landmarks in the image, use adaptive thresholding to segment the landmarks, and then extract contours from the thresholded image, and discard those that are not convex polygons, and those that are not square, and use some additional filtering; Step 2.2: Analyze the internal codes of candidate signs to determine if they are indeed signs, including: Step 2.2.1: Perform perspective transformation on the image to obtain its normal form; Step 2.2.2: Thresholding the normalized image with Ossu to separate white and black bits, extracting the flag bits of each flag and analyzing the flag number; Step 3: After the sub-pixel refinement at the corner, according to the pinhole imaging model, using the pixel coordinates of the four vertices marked by ArUco, the rotation matrix R and translation matrix T of the camera can be obtained by the solvePnP method; Step 4: Obtain the position and attitude information of the camera according to the Euclidean transformation relationship between the camera coordinate system and the world coordinate system. The solution formula is as follows: Among them, P _c represents the coordinates of the camera in the world coordinate system. 4. a kind of control method based on the four-rotor aircraft air autonomous grasping operation of Opencv image processing according to claim 1, it is characterized in that: in the step 3, hierarchical control four-rotor aircraft and operating device step are as follows: Assuming that the operating range of the quadrotor aircraft to grab the target with known coordinates is S, the pose information of the quadrotor aircraft is obtained by calculating the camera pose information after image processing shown in Figure 4. If the quadrotor aircraft does not enter the operating range S, the quadrotor aircraft The rotorcraft is close to or far away from the artificial marker, and the target trajectory generator outputs the target position P ^* = [x ^* , y ^* , z ^* ] ^T and the yaw angle ψ _d to the position controller and the attitude stability controller respectively, and the position control The main body of the controller is the PD controller, and its output is calculated by the system dynamics equation to obtain the attitude angle θ _d , φ _d required for attitude control. The attitude stability controller is designed by the inversion method, and its output is decoupled from the motor The rotational speed ω _i (i=1,2,3,4) is output to the motor until the quadrotor aircraft enters the operating range S; If the quadrotor aircraft enters the operating range S, the quadrotor aircraft hovers, and the pose information of the quadrotor aircraft is output to the trajectory generator of the manipulator. The trajectory generator of the manipulator first converts the pose information of the quadrotor aircraft into the target pose information and then plans Each joint movement angle θ _i (i=1,2,3) is output to the center of gravity balance controller, and the center of gravity balance controller calculates the joint movement angle θ _c of the center of gravity balance mechanism according to the formula of the center of gravity balance equation, and finally sends all joint angle signals wirelessly When the steering gear controller controls the operating device to complete the grabbing, the center of gravity balance equation formula is: ∑m _i x _i + m _c x _c + m _t x _t = 0 (7); Among them, m _i , m _c , m _t are the mass of each part of the manipulator arm, the center of gravity balance mechanism and the grasped object, and _xi , x _c , x _t are the mass of each part of the manipulator arm, the center of gravity balance mechanism and the grasped object. The position of the center of gravity in the machine system {O _b }.