Double-roller parallel safflower picking robot and control method thereof

Abstract

The invention discloses a double-roller parallel safflower picking robot and a control method thereof.The double-roller parallel safflower picking robot comprises a rack, linear sliding table modules arranged in parallel are installed at the top of the rack, a motion platform is slidably connected between the linear sliding table modules on the two sides, a collecting box is fixed to the top of the motion platform, and a parallel mechanism is installed at the bottom of the motion platform; the parallel mechanism comprises a plurality of mechanical arms, the bottoms of the mechanical arms are jointly hinged to a collecting box, the collecting box is communicated with the collecting box, and rollers rotating in the opposite directions are installed in the collecting box. An actual operation area is divided through the remote terminal, global path planning is automatically generated, data fusion is carried out by adopting GPS and inertial navigation, and extremely accurate navigation can be achieved; the motion platform is automatically leveled through front and rear gyroscopes, so that the motion platform is parallel to the ground, and the difficulty of visual positioning is greatly reduced; based on an improved YOLOv8 target detection model, rapid deployment is realized, and the picking efficiency is improved.

Description

Technical Field

[0001] The invention relates to a safflower picking robot and a control method thereof, and in particular to a double-roller parallel safflower picking robot and a control method thereof. Background Art

[0002] Safflower is an important traditional Chinese medicine and dye plant with extensive medicinal and economic value. The petals of safflower are rich in pigments and active ingredients and are widely used in medicine, food, cosmetics and other fields. However, safflower petals fall off easily and the picking period is short. With the development of satellite navigation counting, robotics and artificial intelligence technology, harvesting robots based on visual positioning and automatic control have gradually become a research hotspot.

[0003] Chinese patent document CN112640664A discloses a safflower pulling and harvesting test bench based on visual control. The picking part of the test bench is composed of a pulling cylinder and a connecting rod. When the camera recognizes safflower, the shield of the picking part moves to the vicinity of the filaments. The filaments enter the area to be picked for the filaments under the action of negative pressure, and the pulling cylinder works. The clamping plate clamps the filaments, and the pulling cylinder moves upward to complete the harvesting of the filaments. Under the action of negative pressure air suction, the filaments are sucked along the hose into the flower collection box. Although the above-mentioned safflower harvesting device can realize automatic identification, picking and collection of safflower. However, due to the large difference in the size of the safflower balls, the use of a fixed-size flower collection port cannot well limit whether the ball can enter, resulting in poor filament picking effect.

[0004] The safflower global multi-module intelligent harvesting robot and its control method disclosed in Chinese patent document CN113796204A adopts multiple screw linear guide modules, obtains the planned path during planting through the satellite positioning system, uses a monocular camera to identify the planting row spacing in real time, adjusts the spacing of the harvesting modules, and realizes efficient multi-row picking. However, since safflower does not grow straight, there will be many directions at different angles, and it is difficult for the linear harvesting module to achieve accurate picking. The picking module needs to be further optimized to meet the needs of picking safflowers at different angles. Summary of the invention

[0005] Purpose of the invention: The purpose of the invention is to propose a double-roller parallel safflower picking robot and a control method thereof to improve the efficiency of safflower picking.

[0006] Technical solution: The present invention includes a frame, on the top of which are installed parallel linear slide modules, a moving platform is slidably connected between the linear slide modules on both sides, a collection box is fixed on the top of the moving platform, a parallel mechanism is installed on the bottom, the parallel mechanism includes multiple mechanical arms, a collection box is hinged at the bottom of the multiple mechanical arms, the collection box is connected to the collection box, and sticks rotating in opposite directions are installed in the collection box.

[0007] A depth camera is installed at the front of the motion platform to identify the safflower and determine the position of the safflower in three-dimensional space.

[0008] The mechanical arm comprises a large arm and a small arm, one end of the large arm is connected to a stepper motor, the stepper motor is fixed at the bottom of the motion platform, the other end of the large arm is hinged to the small arm, and the other end of the small arm is hinged to the collection box.

[0009] The collection box is connected to the collection box through a filament conveying hose, and a negative pressure fan is installed in the collection box. After the filaments are harvested, they enter the collection box along the filament conveying hose under the action of negative pressure, completing the filament harvesting.

[0010] One end of the stick is connected to the reduction motor, and the other end is supported by a bearing.

[0011] A slider is slidably connected inside the linear slide module, a motion platform is connected between the linear slide modules on both sides, and the linear slide modules on both sides are respectively connected to the motor drive.

[0012] The linear slide module is provided with a front limiter at the front and a rear limiter at the rear, which are used to limit the displacement of the motion platform.

[0013] A front gyroscope is installed at the front of the frame, and a rear gyroscope is installed at the rear. A GPS module is fixed at the front of the frame for receiving transmission information from a satellite and determining real-time longitude and latitude information.

[0014] An electric push rod is installed at the bottom of the frame, and a wheel hub is installed at the bottom of each electric push rod to ensure the levelness of the motion platform, improve the positioning accuracy of the red flower and adapt to the terrain of various environments.

[0015] A control method for a roller-type parallel safflower picking robot comprises the following steps:

[0016] S1. Initialize the parallel roller safflower picking robot, perform global path planning, generate the optimal operation path, and drive to the operation starting point;

[0017] S2. The parallel roller safflower picking robot travels along the globally planned path. The main controller guides the vehicle sound control system to move forward and adjust the posture according to the real-time longitude and latitude analyzed by GNSS combined with the inertial module, and the depth camera identifies and locates the spatial position of the safflower. When the safflower is identified, the walking device stops moving, and the leveling controller adjusts the electric push rod in combination with the inclination values ​​of the front and rear gyroscopes to keep the motion platform parallel to the ground. The recognition algorithm uses the improved YOLOv8 algorithm for target detection, and the depth camera captures the RGB image and the depth map at the same time. The main controller sends the coordinates required for the actual picking mechanism to the harvesting device controller, solves the number of pulses required for the actual motion platform and each stepper motor of the parallel mechanism, and decomposes its motion into the first motion instruction and the second motion instruction for execution.

[0018] S3, the first motion instruction is given to the motion platform. After the motion platform moves to the instruction position, the parallel mechanism, under the control of the second motion instruction, the collection box reaches the safflower to-be-picked area; the roller rotates to grind the safflower into the collection box, and under the action of the negative pressure fan, the safflower enters the collection box along the safflower conveying hose; in the case of multiple targets in the camera, the task scheduling is performed through the x, y coordinates in the actual space of the multiple targets to achieve the optimal picking strategy until the safflower in the camera image is picked; after the image camera is picked, the motion platform moves forward, and when the safflower is recognized again, the above actions are repeated until the motion platform touches the induction switch at the front end of the linear slide module, and the safflower picking in the area is completed;

[0019] S4. After completing the picking of an area of ​​safflower, the motion platform returns to the rear part of the frame. There are induction switches at the front and rear of the linear slide rail module. When the motion platform moves to the sensing point, it stops moving, and the walking device continues to move in the direction of travel. The main controller combines the inertial navigation module with the GPS data to compare the current position with the planned path, calculates the deviation, and controls the differential speed of the two wheels of the hub electronics through the PID control algorithm. Repeat the above steps until all picking is completed.

[0020] Beneficial effects: The present invention has the following advantages:

[0021] (1) The actual operation area is divided by the remote terminal, and the global path planning is automatically generated. The GPS and inertial navigation are used for data fusion to achieve extremely accurate navigation. The motion platform is automatically leveled by the front and rear gyroscopes to ensure that the motion platform is parallel to the ground, which greatly reduces the difficulty of visual positioning. Based on the improved YOLOv8 target detection model, it can be deployed quickly without high requirements on equipment. The overall control system can be promoted and applied in a wide range of safflower automated picking scenarios, reducing labor costs and improving picking efficiency.

[0022] (2) It can adapt to flower ball types of different sizes, reducing the requirements for visual positioning. The parallel design has more degrees of freedom and can reach more directions in space, which can meet the actual growth and picking conditions of safflower. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a schematic diagram of the overall structure of the present invention;

[0024] Figure 2 for Figure 1 The main view of

[0025] Figure 3 for Figure 1 A top view of

[0026] Figure 4 It is a schematic diagram of the structure of the picking part of the present invention;

[0027] Figure 5 It is a schematic diagram of the roller mechanism of the present invention;

[0028] Figure 6 is an overall block diagram of the control system of the present invention;

[0029] Figure 7 It is an operation flow chart of the present invention. DETAILED DESCRIPTION

[0030] The present invention will be further described below in conjunction with the accompanying drawings.

[0031] like Figure 1 to Figure 5 As shown, the roller-type parallel safflower picking robot of the present invention comprises a frame 1, a linear slide module 5 arranged in parallel is installed on the top of the frame 1, a slider 25 is slidably connected in the linear slide module 5, the linear slide modules 5 on both sides are respectively connected to the motor 24 for driving, a motion platform 7 is connected between the linear slide modules 5 on both sides, the bottom of the motion platform 7 is respectively fixed to the slider 25 on the corresponding side, and the slider 25 slides to drive the motion platform 7 to move along the linear slide module 5. A collection box 6 is fixed on the top of the motion platform 7, and a parallel mechanism is installed on the bottom to drive the collection box 11 to the accurate target position.

[0032] like Figure 4 As shown, the parallel mechanism includes a plurality of stepper motors 14 arrayed at the bottom of the motion platform 7. In this embodiment, three stepper motors 14 are installed. The three stepper motors 14 are respectively hinged to the collection box 11 through a mechanical arm. The mechanical arm includes a large arm and a small arm 10. One end of the large arm is connected to the stepper motor 14, and the other end is hinged to the small arm 10 through a hinge 15. The other end of the small arm 10 is hinged to the connectors 18 on both sides of the collection box 11. The collection box 11 is connected to the collection box 6 through a filament delivery hose 12. A negative pressure fan 13 is installed in the collection box 6. Figure 2 shown.

[0033] like Figure 5 As shown, a pair of rollers 19 rotating in opposite directions are installed in the collection box 11, an arc surface is formed between the rollers 19, and the surface of the rollers 19 has stripes to increase friction. One end of the roller 19 is connected to the reduction motor 17, and the other end is supported by the deep groove ball bearing 16. When the safflower enters the collection box 11, it will be rolled into the collection box 11 by the roller 19, and the filaments and the flower balls will be quickly separated. After the harvesting is completed, the filaments will enter the collection box 6 along the filament conveying hose 12 under the action of negative pressure, and the filaments are harvested.

[0034] like Figure 1 As shown, four electric push rods 2 are installed at the bottom of the frame 1 to ensure the level of the motion platform 7, improve the positioning accuracy of the red flower and adapt to the terrain of various environments. A wheel hub is installed at the bottom of each electric push rod 2. The wheel hub located in the front of the motion direction is the driving wheel, which is equipped with a wheel hub motor 3 with an encoder, and the wheel hub located in the rear of the motion direction adopts a universal wheel 4.

[0035] like Figure 3 As shown, a depth camera 22 is installed at the front of the motion platform 7 for identifying safflowers and determining the position of safflowers in three-dimensional space. A front limit 21 is provided at the front of the linear slide module 5, and a rear limit 20 is provided at the rear to limit the displacement of the motion platform 7. A front gyroscope 23 is installed at the front of the frame 1, and a rear gyroscope 9 is installed at the rear to determine the posture information and motion information of the vehicle body. A GPS module 8 is also fixed at the front of the frame 1 to receive transmission information from the satellite and determine real-time longitude and latitude information.

[0036] like Figure 6 and Figure 7 As shown, the control method of the roller-type parallel safflower picking robot of the present invention comprises the following steps:

[0037] S1. Initialize the parallel roller safflower picking robot, perform global path planning, generate the optimal operation path, and drive to the operation starting point.

[0038] The control system of the roller-type parallel safflower picking robot is powered on and initialized. It receives the differential signal solved by the GPS and the ground base station to obtain real-time longitude and latitude information, and combines the inertial sensor on the body to obtain real-time posture information. After the initialization is completed, the main controller wirelessly receives the operation area and operation tasks from the terminal scheduling system, performs global path planning, generates the optimal operation path, and the walking equipment drives to the operation starting point according to the task requirements.

[0039] S2, the parallel roller safflower picking robot travels along the globally planned path. The main controller guides the vehicle sound control system to move forward and adjust the posture according to the real-time longitude and latitude analyzed by GNSS combined with the inertial module. During the movement, the PID control algorithm is used to ensure that the vehicle body moves along the preset path. The depth camera identifies and locates the spatial position of safflower; when the safflower is identified, the walking device stops moving, and the leveling controller adjusts the electric push rod in combination with the inclination value of the front and rear gyroscopes to keep the motion platform parallel to the ground, reducing the error of subsequent safflower positioning.

[0040] The recognition algorithm uses the improved YOLOv8 algorithm for target detection. The depth camera captures RGB images and depth maps at the same time. The RGB images are processed using YOLOv8 to obtain the 2D bounding box of the target object. The bounding box provides the x, y coordinates, width and height of the object in the image. The depth map provides the distance information to the camera for each pixel. By calibrating and using the camera's internal parameters, the 2D pixel coordinates and depth information are converted into coordinates in the actual camera coordinate system. Since the positions of the depth camera and the parallel mechanism moving platform are relatively fixed, the 3D coordinates in the camera coordinate system can be converted into the coordinate system of the parallel mechanism. The main controller sends the coordinates required for the actual picking mechanism to move to the harvesting device controller. After receiving the instructions, the harvesting controller performs the kinematic inverse solution for the coordinate points that need to be moved to, solves the number of pulses required for the actual moving platform and each stepper motor of the parallel mechanism, and decomposes its movement into the first motion instruction and the second motion instruction for execution.

[0041] S3. The first motion instruction is given to the motion platform. After the motion platform moves to the instruction position, the parallel mechanism, under the control of the second motion instruction, the collection box reaches the safflower picking area. The DC reduction motor in the collection box drives the roller to rotate, and the safflower is crushed into the collection box. Under the action of the negative pressure fan, it enters the collection box along the safflower conveying hose. In the case of multiple targets in the camera, the task scheduling is carried out through the x and y coordinates in the actual space of the multiple targets to achieve the optimal picking strategy until the safflower picking in the camera image is completed. After the image camera is picked, the motion platform moves forward. When the safflower is recognized again, the above action is repeated until the motion platform touches the sensor switch at the front end of the linear slide module to complete the safflower picking in this area.

[0042] S4. After completing the picking of an area of ​​safflower, the motion platform returns to the rear part of the frame. There are induction switches at the front and rear of the linear slide rail module. When the motion platform moves to the induction point, stopping the movement is equivalent to returning to the initial position. The walking device continues to move in the direction of travel. The main controller combines the inertial navigation module with the GPS data to compare the current position with the planned path, calculate the deviation, and control the differential speed of the two wheels of the hub electronics through the PID control algorithm to ensure that the safflower picking robot travels on the predetermined path. Repeat steps two and three to achieve identification, leveling, positioning, picking, and collection until all work areas are completed.

[0043] The control system main controller communicates with the navigation controller, harvesting controller, and vehicle body controller via the RS485 bus using the standard MODBUS protocol to ensure the reliability and stability of communication transmission.

Claims

1. A roller-type parallel safflower picking robot, characterized in that: It includes a frame, on the top of which are installed linear slide modules arranged in parallel, a moving platform is slidably connected between the linear slide modules on both sides, a collection box is fixed on the top of the moving platform, a parallel mechanism is installed on the bottom, the parallel mechanism includes multiple mechanical arms, a collection box is hinged at the bottom of the multiple mechanical arms, the collection box is connected to the collection box, and sticks rotating in opposite directions are installed in the collection box.

2. The roller-type parallel safflower picking robot according to claim 1, characterized in that: A depth camera is installed at the front of the motion platform.

3. The roller-type parallel safflower picking robot according to claim 1, characterized in that: The mechanical arm comprises a large arm and a small arm, one end of the large arm is connected to a stepper motor, the stepper motor is fixed at the bottom of the motion platform, the other end of the large arm is hinged to the small arm, and the other end of the small arm is hinged to the collection box.

4. The roller-type parallel safflower picking robot according to claim 3, characterized in that: The collection box is connected to the collection box through a filament conveying hose, and a negative pressure fan is installed in the collection box.

5. The roller-type parallel safflower picking robot according to claim 1, characterized in that: One end of the stick is connected to the reduction motor, and the other end is supported by a bearing.

6. The roller-type parallel safflower picking robot according to claim 1, characterized in that: A slider is slidably connected inside the linear slide module, a motion platform is connected between the linear slide modules on both sides, and the linear slide modules on both sides are respectively connected to the motor drive.

7. The roller-type parallel safflower picking robot according to claim 6, characterized in that: The linear slide module is provided with a front limiter at the front and a rear limiter at the rear.

8. The roller-type parallel safflower picking robot according to claim 1, characterized in that: A front gyroscope is installed at the front of the frame, a rear gyroscope is installed at the rear, and a GPS module is fixed at the front of the frame.

9. The roller-type parallel safflower picking robot according to claim 8, characterized in that: An electric push rod is installed at the bottom of the frame, and a wheel hub is installed at the bottom of each electric push rod.

10. A control method for the roller-type parallel safflower picking robot according to any one of claims 1 to 9, characterized in that: The following steps are involved: S1. Initialize the parallel roller safflower picking robot, perform global path planning, generate the optimal operation path, and drive to the operation starting point; S2, the parallel roller safflower picking robot travels along the globally planned path. The main controller guides the vehicle sound control system to move forward and adjust the posture according to the real-time longitude and latitude analyzed by GNSS combined with the inertial module, and the depth camera identifies and locates the spatial position of the safflower; when the safflower is identified, the walking device stops moving, and the leveling controller adjusts the electric push rod according to the inclination value of the front and rear gyroscopes to keep the motion platform parallel to the ground; The recognition algorithm uses an improved YOLOv8 algorithm for target detection. The depth camera captures RGB images and depth maps at the same time. The main controller sends the coordinates of the actual picking mechanism to the harvesting device controller, solves the number of pulses required by the actual motion platform and each stepper motor of the parallel mechanism, and decomposes its motion into the first motion instruction and the second motion instruction for execution. S3, the first motion instruction is given to the motion platform. After the motion platform moves to the instruction position, the parallel mechanism, under the control of the second motion instruction, the collection box reaches the safflower to-be-picked area; the roller rotates to grind the safflower into the collection box, and under the action of the negative pressure fan, the safflower enters the collection box along the safflower conveying hose; in the case of multiple targets in the camera, the task scheduling is performed through the x, y coordinates in the actual space of the multiple targets to achieve the optimal picking strategy until the safflower in the camera image is picked; after the image camera is picked, the motion platform moves forward, and when the safflower is recognized again, the above actions are repeated until the motion platform touches the induction switch at the front end of the linear slide module, and the safflower picking in the area is completed; S4. After completing the picking of an area of ​​safflower, the motion platform returns to the rear part of the frame. There are induction switches at the front and rear of the linear slide rail module. When the motion platform moves to the sensing point, it stops moving, and the walking device continues to move in the direction of travel. The main controller combines the inertial navigation module with the GPS data to compare the current position with the planned path, calculates the deviation, and controls the differential speed of the two wheels of the hub electronics through the PID control algorithm. Repeat the above steps until all picking is completed.

Images