Large-scale apple picking robots and picking methods

Abstract

This invention discloses a long-stroke apple-harvesting robot, comprising a self-balancing tracked chassis and a working platform. The working platform includes an extension platform, a control box, a detection device, a sorting device, and a visual navigation system, as well as several collection boxes. The detection device, sorting device, and collection boxes are arranged sequentially according to the apple detection, sorting, and collection process. A harvesting robotic arm unit is mounted on the extension platform, with a harvesting gripper unit and a fruit conveying tube unit connected to the harvesting robotic arm unit. The control box includes a first controller and a second controller. The extension platform, harvesting robotic arm unit, and harvesting gripper unit are all connected to the first controller; the detection device and sorting device are both connected to the second controller; and the self-balancing tracked chassis is connected to the visual navigation system. This robot can effectively reduce apple damage during harvesting and grade the harvested apples according to quality. A long-stroke apple-harvesting method is also disclosed.

Description

Technical Field

[0001] This invention belongs to the field of apple picking robot technology, specifically relating to a long-stroke apple picking robot and a long-stroke apple picking method. Background Technology

[0002] With the shortage of labor resources and the rapid development of mechanical automation technology, robotic harvesting has become an inevitable trend to replace manual harvesting, and apple harvesting robots have become a research hotspot both domestically and internationally. However, apple harvesting robots still face many problems: a. Limited by the arm span, the working space of the harvesting robot cannot cover all the apples in the current field of vision; b. The robotic arm of the apple harvesting robot has few degrees of freedom and poor adaptability. Faced with complex environments, the robotic arm with few degrees of freedom cannot effectively avoid obstacles such as tree trunks, branches, and leaves, which leads to the end effector being unable to harvest in a suitable posture; c. The opening and closing angle and pressure adaptability of the end effector of the apple harvesting robot are poor, which can easily damage the apples during harvesting; d. Each harvesting operation of the apple harvesting robot involves multiple adjustments, which can easily result in a slow harvesting speed; e. The transportation and sorting of harvested apples still require a large amount of manpower.

[0003] Furthermore, existing related technologies still have shortcomings.

[0004] As described in patent number 202120258834.8, a high-degree-of-freedom apple-picking robot can adjust the height of its gripper according to the height of the apple tree to pick apples at different locations, but it cannot flexibly adjust the posture of its end effector, making it difficult to pick apples that are hidden behind branches.

[0005] The lightweight dual-arm apple picking robot described in patent number 202220407779.9 simplifies the structure of the robotic arm and expands the working space of the apple picking robot by using dual arms, which can achieve efficient single-sided picking operations.

[0006] The automatic apple picking robot described in patent number 201821166649.0 uses a hemispherical blade picker to cut the fruit stem to achieve picking. Although it will not damage the apple due to excessive pressure, it may cut the apple due to the positioning error of the blade.

[0007] The apple picking robot described in patent number 201720703300.5 has the function of picking and recycling at the same time. It can pick and collect apples on a branch at one time, but manual sorting is still required.

[0008] The negative pressure apple picking machine described in patent number 202220752780.5 is equipped with a screening barrel to directly screen the picked apples, but it only screens by the size of the apple diameter and cannot grade the apples. Summary of the Invention

[0009] The first objective of this invention is to provide a long-stroke apple-picking robot that expands the range of motion of the picking robotic arm, improves the flexibility and obstacle avoidance capabilities of the picking robotic arm, effectively reduces apple damage during picking, and can also grade the picked apples according to their quality.

[0010] The second objective of this invention is to provide a method for harvesting long-distance apples.

[0011] The first technical solution adopted in this invention is a long-stroke apple picking robot, which includes a self-balancing tracked chassis and a working platform on the self-balancing tracked chassis. The working platform is a stepped platform, with an extension platform, control box, detection device, sorting device and visual navigation system on the upper platform, and several collection boxes on the lower platform. The detection device, sorting device and several collection boxes are arranged in sequence according to the apple detection, sorting and collection process.

[0012] The platform is equipped with a harvesting robotic arm unit, which is equipped with a harvesting gripper unit and a fruit conveying tube unit, which is connected to the harvesting robotic arm unit.

[0013] The control box includes controller No. 1 and controller No. 2;

[0014] The expansion platform, the picking robotic arm unit, and the picking gripper unit are all connected to controller No. 1;

[0015] Both the detection device and the sorting device are connected to controller No. 2;

[0016] The self-balancing tracked chassis is connected to the visual navigation system.

[0017] The invention is further characterized in that:

[0018] The self-balancing tracked chassis includes an electric tracked chassis and a frame platform. The electric tracked chassis and the frame platform are connected by six hydraulic outriggers, which are arranged in two rows of three evenly distributed on the bottom of the frame platform. The bottom of the frame platform is also equipped with a hydraulic pump station, control valve group and hydraulic controller. Tilt sensors are arranged on the side of the frame platform.

[0019] The control valve group includes interconnected electro-hydraulic directional valves and multi-way directional valves; the hydraulic pump station's oil circuit, electro-hydraulic directional valves, and multi-way directional valves are sequentially connected to form a closed loop;

[0020] Each hydraulic outrigger includes a hydraulic lifting cylinder, and a pressure sensor is installed at the bottom of the hydraulic lifting cylinder. The oil inlet of the hydraulic lifting cylinder is connected to the oil outlet of the multi-way directional valve through an oil pipeline, and the oil outlet of the hydraulic lifting cylinder is connected to the oil inlet of the multi-way directional valve through an oil pipeline. A one-way hydraulic lock and a two-position electro-hydraulic valve are installed on the oil pipeline between the oil inlet of the hydraulic lifting cylinder and the oil outlet of the multi-way directional valve.

[0021] Each two-position electro-hydraulic valve and pressure sensor is connected to the hydraulic controller.

[0022] The electro-hydraulic directional valve, multi-way directional valve, and tilt sensor are all connected to the hydraulic controller.

[0023] The expansion platform includes a Z-axis expansion platform, coupling I, and an X-axis expansion platform;

[0024] The Z-axis extension platform includes a Z-axis linear module 1, a Z-axis linear module 2, and a Z-axis extension platform servo motor. The Z-axis linear modules 1 and 2 are vertically mounted on the upper platform of the work platform. The synchronous pulley of the Z-axis linear module 1 is connected to the synchronous pulley of the Z-axis linear module 2 through coupling I. The Z-axis extension platform servo motor is connected to the synchronous pulley of the Z-axis linear module 2. The Z-axis extension platform servo motor drives the synchronous pulley of the Z-axis linear module 2 and drives the synchronous pulley of the Z-axis linear module 1 to rotate through coupling I.

[0025] The X-axis extension platform includes an X-axis linear module and an X-axis extension platform servo motor; the sliders of Z-axis linear module 1 and Z-axis linear module 2 are respectively connected to the profile of the X-axis linear module by bolts, the X-axis extension platform servo motor is connected to the synchronous wheel of the X-axis linear module, and the X-axis linear module is driven by the X-axis extension platform servo motor.

[0026] The Z-axis extension platform servo motor and the X-axis extension platform servo motor are connected to controller No. 1.

[0027] The harvesting robotic arm unit is connected to the slider of the X-axis linear module.

[0028] The testing device includes a testing conveyor belt installed on a high platform of the work platform. A testing conveyor belt motor is installed at one end of the testing conveyor belt. A U-shaped testing chamber shell is fitted on the testing conveyor belt. Black curtains are installed on the two opposite side walls of the U-shaped testing chamber shell. A partition with apple passage holes is installed inside the U-shaped testing chamber shell. The partition is fitted above the testing conveyor belt. Several appearance testing modules and several internal quality testing modules are evenly distributed on both sides of the testing conveyor belt. Several appearance testing modules are located in the testing chamber shell on one side of the partition, and several internal quality testing modules are located in the testing chamber shell on the other side of the partition.

[0029] Shadowless lamps are installed on two opposite side walls of the testing chamber housing where several appearance inspection modules are located;

[0030] Halogen lamps are installed on the two opposite side walls of the testing chamber shell where several internal quality testing modules are located;

[0031] Z-axis linear module 1 and Z-axis linear module 2 are located on both sides of the entrance end of the detection conveyor belt, respectively.

[0032] Several appearance inspection modules are industrial cameras, and several appearance inspection modules are connected to controller No. 2;

[0033] Several internal quality inspection modules are infrared spectrometers; several internal quality inspection modules are connected to controller No. 2.

[0034] The visual navigation system includes a binocular camera base and a navigation controller. The binocular camera base is fixedly installed on a high platform of the working platform. The binocular camera is mounted on the binocular camera base, and the lens of the binocular camera is facing the direction of travel of the electric tracked chassis. Both the binocular camera and the electric tracked chassis are connected to the navigation controller.

[0035] The sorting device includes several sorting conveyor belts and collection boxes. The sorting conveyor belts are evenly divided into two groups and arranged on both sides of the inspection conveyor belt. Each sorting conveyor belt has a collection box at its end. Each sorting conveyor belt has a sorting conveyor belt servo motor installed on one side. All sorting conveyor belt servo motors are connected to controller No. 2.

[0036] Four vertical supports are evenly distributed in two groups on the outside of several sorting conveyor belts;

[0037] The Y-axis module of sorting robot arm No. 1 is installed on two vertical supports on one side of the inspection conveyor belt, and the Y-axis module of sorting robot arm No. 2 is installed on two vertical supports on the other side of the inspection conveyor belt. Both the Y-axis modules of sorting robot arm No. 1 and No. 2 are synchronous belt linear modules. The synchronous pulleys of the Y-axis modules of sorting robot arm No. 1 and No. 2 are connected by coupling II. The Y-axis module also includes a sorting robot arm Y-axis servo motor. The sorting robot arm Y-axis servo motor is connected to the synchronous pulley of the Y-axis module of sorting robot arm No. 1. The sorting robot arm Y-axis servo motor drives the synchronous pulley of the Y-axis module of sorting robot arm No. 1 and drives the synchronous pulley of the Y-axis module of sorting robot arm No. 2 to rotate synchronously through coupling II.

[0038] It also includes a sorting robot arm X-axis module, a sorting robot arm X-axis servo motor, and a sorting robot arm Z-axis module. The sorting robot arm X-axis module and the sorting robot arm Z-axis module are both synchronous belt linear modules. The sliders of sorting robot arm Y-axis module 1 and sorting robot arm Y-axis module 2 are respectively connected to the sorting robot arm X-axis module by bolts. The sorting robot arm X-axis servo motor is connected to the synchronous pulley of the sorting robot arm X-axis module. The sorting robot arm Z-axis module is mounted on the slider of the sorting robot arm X-axis module by bolts. The sorting robot arm Z-axis servo motor is installed at the synchronous pulley of the sorting robot arm Z-axis module. The sorting robot arm Y-axis servo motor, sorting robot arm X-axis servo motor, and sorting robot arm Z-axis servo motor are all connected to controller 2.

[0039] The sorting robot arm's Z-axis module is equipped with a flexible sorting gripper at its end. The flexible sorting gripper has the same structure as the flexible picking gripper. An industrial camera is installed on the side of the end of the sorting robot arm's Z-axis module and connected to controller No. 2. The industrial camera takes pictures of the apples after secondary grading on the inspection conveyor belt and sends the pictures to controller No. 2. The controller No. 2 uses a predetermined algorithm to identify and locate the apples after secondary grading. Controller No. 2 drives the sorting robot arm to stack the graded apples onto the corresponding sorting conveyor belt.

[0040] The harvesting robotic arm unit includes a redundant robotic arm base and a redundant robotic arm connected in sequence. The redundant robotic arm base is connected to the slider of the X-axis linear module. A depth camera base is installed on the arm body at the end of the redundant robotic arm away from the redundant robotic arm base, and a depth camera is installed on the depth camera base. The redundant robotic arm is connected to the harvesting gripper unit.

[0041] The depth camera and redundant robotic arm are both connected to controller number 1.

[0042] The picking gripper unit includes a flexible picking gripper, and a thin-film pressure sensor is provided on the inner side of the flexible picking gripper. The flexible picking gripper is connected to the end of the redundant robotic arm away from the base of the redundant robotic arm through connector I. Connector I includes a flange, and two connecting protrusions are provided on the edge of the flange. The flange is sleeved on the redundant robotic arm, and the two connecting protrusions are connected to the end of the flexible picking gripper.

[0043] The flexible picking gripper and the membrane pressure sensor are both connected to controller No. 1.

[0044] The fruit conveying tube unit includes a funnel-shaped inlet, a telescopic hose, and a connecting pipe connected in sequence. The upper end of the connecting pipe is connected to the telescopic hose, and the lower end of the connecting pipe is fixed to the end cap of the detection conveyor belt. A pad is provided at the outlet of the connecting pipe. It also includes a connector II, which includes a connecting rod with an open retaining ring and a connecting rod with a through hole that are connected to each other. The connecting rod with the through hole is sleeved with a redundant robotic arm, and the connecting rod with the open retaining ring is sleeved with the funnel-shaped inlet.

[0045] The funnel-shaped inlet end is equipped with a one-way valve.

[0046] The second technical solution adopted in this invention is a long-stroke apple picking method using the aforementioned robot. Specifically, a navigation controller controls an electric tracked chassis to enable the long-stroke picking robot to move within the orchard. Once the robot reaches a designated location near the target tree, a hydraulic controller adjusts the hydraulic outriggers to keep the work platform level. A depth camera captures images of the fruit trees, and controller 1 processes these images to plan the picking sequence and position of the target apples. Controller 1 drives the picking robotic arm and flexible picking gripper to pick the apples in sequence, and performs preliminary coding and grading based on data from the depth camera and thin-film pressure sensor. After picking, the apples are transported to a detection device via a fruit conveying tube unit. The apples are then further coded and graded based on data from various appearance and internal quality detection modules within the detection device. Finally, they are transported to collection boxes of the corresponding grade via a sorting device.

[0047] Specifically, the following steps are included:

[0048] Step S1: The binocular camera takes photos of the orchard roads and sends them to the navigation controller. The navigation controller uses machine vision and deep learning autonomous navigation algorithms to send control commands to the electric tracked chassis, thereby enabling the long-range picking robot to move in the orchard.

[0049] Step S2: The long-stroke picking robot moves to a designated position near the target fruit tree. The hydraulic controller receives data from the tilt sensor and various pressure sensors. The hydraulic controller sends control commands to the hydraulic pump station, control valve group and various two-position electro-hydraulic valves, thereby realizing the adjustment of each hydraulic outrigger and the relative horizontal control of the chassis platform and the working platform.

[0050] Step S3: Controller 1 controls the movement of the extension platform and the picking robotic arm unit, uses a depth camera to acquire images of the target fruit tree, and sends the images to Controller 1. YOLOv5 is used to identify all target apples, and then 3D point cloud processing is used to realize the picking sequence planning and pose positioning of the target apples.

[0051] Step S4: Controller 1 drives the extension platform 3 and the redundant robotic arm to the picking position according to the target apple picking order, and drives the flexible picking gripper to wrap around the target apple; Controller 1 feeds back pressure information through the thin film pressure sensor, controls the motor in real time, and then adjusts the opening and closing degree of the bionic gripper.

[0052] Step S5: Controller 1 uses YOLOv5 to initially acquire information on the target apple's diameter, shape, color, surface defects, and hardness based on data from the depth camera and thin-film pressure sensor. Controller 1 then performs initial grading and coding of the target apple based on this information, and controls the redundant robotic arm and flexible picking gripper to pick the target apple.

[0053] Step S6: After being picked, the apples are transported through the fruit conveyor unit to the inspection conveyor belt in the inspection device;

[0054] Step S7: Controller No. 2 uses predetermined algorithms and detection models to perform secondary detection, grading and coding of apples based on data from each appearance detection module and each internal quality detection module in the detection device, according to fruit diameter, fruit shape, color, fruit surface defects, soluble solids and internal lesion indicators; the coded apples are then transported to the sorting device via the detection conveyor belt.

[0055] Step S8: Controller 2 uses the YOLOv5 algorithm to identify the coded apples based on the photos taken by the camera and calculates their positions; Controller 2 drives the sorting robotic arm and flexible sorting gripper to transport the secondary graded apples to the corresponding graded collection boxes;

[0056] Step S9: After the target apples have been sorted, repeat steps S4-S7;

[0057] Step S10: After all the target apples on the fruit tree have been harvested, repeat steps S2-S9.

[0058] In step S5, the method by which controller 1 encodes the target apple is as follows: The encoding order of the apple is xxxx-dxx-sxx-rxx-fxx-hxx, where xxxx represents the picking order of the apple, numbered 0001-9999; dxx represents the diameter of the picked apple, which is the maximum of the apple's transverse and longitudinal diameters, and xx is the apple's diameter size in mm; sx.xx represents the shape of the picked apple, with the shape index being the ratio of the fruit's longitudinal to transverse diameters, and x.xx being the apple's shape index; rxx represents the color of the picked apple, and xx is the apple's coloring ratio; fxx represents the surface defects of the picked apple, and xx represents the number of surface defects; hxx represents the firmness of the picked apple, and xx is the apple's firmness in mm. ;

[0059] In step S7, the method by which controller 2 encodes the target apple is as follows: The secondary encoding sequence of the apple is xxxx-dxx-sxx-rxx-fxx-hxx-bxx-mx-lx, where xxxx represents the picking order of the apple, numbered 0001-9999; dxx represents the diameter of the picked apple, which is the maximum of the apple's transverse and longitudinal diameters, and xx is the apple's diameter size in mm; sx.xx represents the shape of the picked apple, with the shape index being the ratio of the fruit's longitudinal to transverse diameters, and x.xx being the apple's shape index; rxx represents the color of the picked apple, and xx is the apple's coloring ratio; fxx represents the surface defects of the picked apple, and xx represents the number of apple surface damages; hxx represents the firmness of the picked apple, and xx is the apple's firmness in mm. bxx represents the soluble solids content of the harvested apples, and xx is the soluble solids content of the apples; mx represents the internal lesions of the harvested apples, x=0 represents no internal lesions, x=1 represents internal lesions; lx represents the grade of the harvested apples, x=1 represents superior grade, x=2 represents first grade, x=3 represents second grade, and x=4 represents inferior grade.

[0060] The beneficial effects of this invention are:

[0061] The self-balancing tracked chassis of this invention not only enables the apple-picking robot to move within the orchard via its tracks, but also allows the working platform to be kept relatively level through a hydraulic self-balancing system. An extension platform is positioned above the working platform to increase the working range of the picking robotic arm unit. The picking robotic arm unit is mounted on the extension platform, with a depth camera and a picking gripper unit at its end, allowing the flexible picking gripper to reach the target apple in any posture. A thin-film pressure sensor is mounted on the flexible picking gripper, which can control the posture and opening / closing angle of the flexible picking gripper based on real-time feedback signals from visual and tactile sensors. The robot's control system enables flexible grading and coding of apples during harvesting. It adaptively picks apples with different diameters, shapes, colors, and surface defects, effectively reducing damage and allowing for grading based on quality. After harvesting, the apples are transported by a conveyor pipe, shortening the return time of traditional robotic arms and improving harvesting efficiency, enabling continuous apple harvesting. The apple harvesting robot is also equipped with a secondary inspection and grading device to perform secondary inspections on the diameter, shape, color, surface defects, and internal quality of the coded apples. This, combined with a sorting device, sorts the graded apples into the appropriate collection boxes. Attached Figure Description

[0062] Figure 1 This is a two-dimensional structural schematic diagram of the harvesting robot of the present invention;

[0063] Figure 2 This is a three-dimensional structural diagram of the harvesting robot body of the present invention;

[0064] Figure 3 This is a three-dimensional structural diagram of the self-balancing tracked chassis of the harvesting robot of this invention;

[0065] Figure 4 This is the hydraulic circuit diagram of the self-balancing tracked chassis of the harvesting robot of this invention;

[0066] Figure 5 This is a top view of the components of the harvesting robot's working platform according to the present invention;

[0067] Figure 6 This is a three-dimensional structural diagram of the harvesting robot connector I of the present invention;

[0068] Figure 7 This is a three-dimensional structural diagram of the harvesting robot connector II of the present invention;

[0069] Figure 8 This is a schematic diagram showing the connection between the harvesting robotic arm unit and the harvesting gripper unit of the harvesting robot of the present invention.

[0070] In the diagram, 1. Self-balancing tracked chassis, 1-1. Electric tracked chassis, 1-2. Hydraulic pump station, 1-3. Control valve group, 1-3-1. Electro-hydraulic directional valve, 1-3-2. Multi-way directional valve;

[0071] 1-4. No. 1 hydraulic outrigger, 1-4-1.1 hydraulic lifting cylinder, 1-4-2. No. 1 two-position electro-hydraulic valve, 1-4-3.1 pressure sensor, 1-4-4.1 one-way hydraulic lock;

[0072] Hydraulic outrigger 1-5.2, hydraulic lifting cylinder 1-5-1.2, two-position electro-hydraulic valve 1-5-2.2, pressure sensor 1-5-3.2, one-way hydraulic lock 1-5-4.2;

[0073] 1-6. No. 3 hydraulic outrigger, 1-6-1.3 hydraulic lifting cylinder, 1-6-2.3 two-position electro-hydraulic valve, 1-6-3.3 pressure sensor, 1-6-4.3 one-way hydraulic lock;

[0074] 1-7. No. 4 hydraulic outrigger, 1-7-1.4 hydraulic lifting cylinder, 1-7-2.4 two-position electro-hydraulic valve, 1-7-3.4 pressure sensor, 1-7-4.4 one-way hydraulic lock;

[0075] Hydraulic outriggers 1-8.5, hydraulic lifting cylinder 1-8-1.5, two-position electro-hydraulic valve 1-8-2.5, pressure sensor 1-8-3.3, one-way hydraulic locks 1-8-4 and 1-5;

[0076] Hydraulic outrigger 1-9.6, hydraulic lifting cylinder 1-9-1.6, two-position electro-hydraulic valve 1-9-2.6, pressure sensor 1-9-3.6, one-way hydraulic lock 1-9-4.6;

[0077] 1-10. Chassis platform; 1-11. Tilt sensor; 1-12. Hydraulic controller;

[0078] 2. Work platform;

[0079] 3. Extension platform, 3-1. Z-axis extension platform, 3-1-1.1 Z-axis linear module, 3-1-2.2 Z-axis linear module, 3-1-3. Z-axis extension platform servo motor, 3-2. Coupling I, 3-3. X-axis extension platform, 3-3-1. X-axis linear module, 3-3-2. X-axis extension platform servo motor;

[0080] 4. Harvesting robotic arm unit; 4-1. Depth camera base; 4-2. Depth camera; 4-3. Redundant robotic arm; 4-4. Redundant robotic arm base;

[0081] 5. Harvesting gripper unit; 5-1. Connector I; 5-2. Flexible harvesting gripper; 5-3. Thin-film pressure sensor; 5-1-1. Flange; 5-1-2. Connecting protrusion;

[0082] 6. Control box, 6-1. Controller No. 1, 6-2. Controller No. 2;

[0083] 7. Fruit delivery tube unit, 7-1. Connector II, 7-2. Funnel-shaped inlet, 7-3. Telescopic hose, 7-4. Connecting tube, 7-5. One-way valve, 7-6. Pad; 7-1-1. Connecting rod with open retaining ring, 7-1-2. Connecting rod with through hole;

[0084] 8. Testing equipment; 8-1. Black curtain; 8-2.1 Appearance inspection module; 8-3.2 Appearance inspection module; 8-4.3 Appearance inspection module; 8-5.4 Appearance inspection module; 8-6.5 Appearance inspection module; 8-7.6 Appearance inspection module; 8-8.1 Internal quality inspection module; 8-9.2 Internal quality inspection module; 8-10.3 Internal quality inspection module; 8-11.4 Internal quality inspection module; 8-12-1. Inspection conveyor belt; 8-12-2. Inspection conveyor belt motor; 8-13. Partition; 8-14. Shadowless lamp; 8-15. Halogen lamp; 8-16. Testing chamber shell;

[0085] 9. Sorting device, sorting conveyor belt 9-1-1.1, sorting conveyor belt 9-2-1.2, sorting conveyor belt 9-3-1.3, sorting conveyor belt 9-4-1.4, servo motor for sorting conveyor belt 9-1-2.1, servo motor for sorting conveyor belt 9-2-2.2, servo motor for sorting conveyor belt 9-3-2.3, servo motor for sorting conveyor belt 9-4-2.4, sorting mechanism 9-5-1.1 Y-axis module of sorting robot arm, 9-5-2.2 Y-axis module of sorting robot arm, 9-5-3 Y-axis servo motor of sorting robot arm, 9-6-1 X-axis module of sorting robot arm, 9-6-2 X-axis servo motor of sorting robot arm, 9-7-1 Z-axis module of sorting robot arm, 9-7-2 Z-axis servo motor of sorting robot arm, 9-8 Flexible sorting gripper, 9-9 Industrial camera, 9-10 Bracket, 9-11 Coupling II;

[0086] 10. Collection box;

[0087] 11. Visual navigation system, 11-1. Binocular camera base, 11-2. Binocular camera, 11-3. Navigation controller. Detailed Implementation

[0088] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0089] This invention provides a long-range apple-picking robot, such as... Figure 1-8 As shown, it includes a self-balancing tracked chassis 1, on which a working platform 2 is mounted. The working platform 2 is a stepped platform, divided into a high platform and a low platform. The high platform is equipped with an extension platform 3, a control box 6, a detection device 8, a sorting device 9, and a vision navigation system 11. The low platform is equipped with several collection boxes 10. The detection device 8, the sorting device 9, and the several collection boxes 10 on the working platform 2 are arranged in sequence according to the apple detection, sorting, and collection process.

[0090] The expansion platform 3 is equipped with a picking robotic arm unit 4, a picking gripper unit 5, and a fruit conveying tube unit 7, which is connected to the picking robotic arm unit 4.

[0091] Control box 6 includes controller 6-1 (No. 1) and controller 6-2 (No. 2);

[0092] The expansion platform 3, the picking robotic arm unit 4, and the picking gripper unit 5 are all connected to controller 6-1.

[0093] Both the detection device 8 and the sorting device 9 are connected to controller 6-2 (number 2).

[0094] The self-balancing tracked chassis 1 is connected to the visual navigation system 11.

[0095] The self-balancing tracked chassis 1 includes an electric tracked chassis 1-1 and a chassis platform 1-10. The electric tracked chassis 1-1 and the chassis platform 1-10 are connected by six hydraulic outriggers. The six hydraulic outriggers are divided into two rows, and three outriggers are evenly arranged at the bottom of the chassis platform 1-10 in each row.

[0096] The six hydraulic outriggers are hydraulic outrigger 1-4, hydraulic outrigger 1-5, hydraulic outrigger 3-6, hydraulic outrigger 4-7, hydraulic outrigger 5-8, and hydraulic outrigger 6-9.

[0097] Hydraulic outrigger 1-4 includes hydraulic lifting cylinder 1-4-1, two-position electro-hydraulic valve 1-4-2, pressure sensor 1-4-3, and one-way hydraulic lock 1-4-4.

[0098] Hydraulic outrigger 1-5 includes hydraulic lifting cylinder 2-5-1, two-position electro-hydraulic valve 2-5-2, pressure sensor 2-5-3, and one-way hydraulic lock 2-5-4.

[0099] Hydraulic outrigger 1-6 includes hydraulic lifting cylinder 1-6-1, two-position electro-hydraulic valve 1-6-2, pressure sensor 1-6-3, and one-way hydraulic lock 1-6-4.

[0100] Hydraulic outrigger 4 (1-7) includes hydraulic lifting cylinder 4 (1-7-1), two-position electro-hydraulic valve 4 (1-7-2), pressure sensor 4 (1-7-3), and one-way hydraulic lock 4 (1-7-4).

[0101] The No. 5 hydraulic outrigger 1-8 includes the No. 5 hydraulic lifting cylinder 1-8-1, the No. 5 two-position electro-hydraulic valve 1-8-2, the No. 5 pressure sensor 1-8-3, and the No. 5 one-way hydraulic lock 1-8-4.

[0102] Hydraulic outrigger 1-9 includes hydraulic lifting cylinder 1-9-1, two-position electro-hydraulic valve 1-9-2, pressure sensor 1-9-3, and one-way hydraulic lock 1-9-4.

[0103] The bottom of the chassis platform 1-10 is also equipped with a hydraulic pump station 1-2, a control valve group 1-3 and a hydraulic controller 1-12; tilt sensors 1-11 are installed on the side of the chassis platform 1-10.

[0104] The control valve group 1-3 includes an electro-hydraulic directional valve 1-3-1 and a multi-way directional valve 1-3-2 that are interconnected; the oil circuit of the hydraulic pump station 1-2, the electro-hydraulic directional valve 1-3-1 and the multi-way directional valve 1-3-2 are connected in sequence to form a closed circuit;

[0105] Each hydraulic outrigger includes a hydraulic lifting cylinder, and a pressure sensor is installed at the bottom of the hydraulic lifting cylinder. The oil inlet of the hydraulic lifting cylinder is connected to the oil outlet of the multi-way directional valve 1-3-2 through an oil pipeline, and the oil outlet of the hydraulic lifting cylinder is connected to the oil inlet of the multi-way directional valve 1-3-2 through an oil pipeline. A one-way hydraulic lock and a two-position electro-hydraulic valve are installed on the oil pipeline between the oil inlet of the hydraulic lifting cylinder and the oil outlet of the multi-way directional valve 1-3-2.

[0106] Each two-position electro-hydraulic valve and pressure sensor is connected to hydraulic controller 1-12;

[0107] The electro-hydraulic directional valve 1-3-1, the multi-way directional valve 1-3-2, and the tilt sensor 1-11 are all connected to the hydraulic controller 1-12.

[0108] The expansion platform 3 is an I-shaped expansion platform, which includes the Z-direction expansion platform 3-1, coupling I 3-2 and the X-direction expansion platform 3-3; where Z is the vertical direction and X is the forward direction of the picking robot.

[0109] Z-axis extension platform 3-1 includes Z-axis linear module 1 3-1-1, Z-axis linear module 2 3-1-2, and Z-axis extension platform servo motor 3-1-3. Z-axis linear module 1 3-1-1 and Z-axis linear module 2 3-1-2 are vertically mounted on the upper platform of work platform 2. The synchronous pulley of Z-axis linear module 1 3-1-1 is connected to the synchronous pulley of Z-axis linear module 2 3-1-2 through coupling I 3-2. Z-axis extension platform servo motor 3-1-3 is connected to the synchronous pulley of Z-axis linear module 2 3-1-2. Z-axis extension platform servo motor 3-1-3 drives the synchronous pulley of Z-axis linear module 2 3-1-2 and drives the synchronous pulley of Z-axis linear module 1 3-1-1 to rotate through coupling I 3-2.

[0110] The Z-axis extension platform can expand the working range of the picking robotic arm unit in the Z-axis. The effective length of the No. 1 Z-axis linear module and the No. 2 Z-axis linear module is 2m, so that the working range of the picking robotic arm unit can cover the entire fruit tree.

[0111] The X-axis extension platform 3-3 includes an X-axis linear module 3-3-1 and an X-axis extension platform servo motor 3-3-2; the sliders of Z-axis linear module 1 3-1-1 and Z-axis linear module 2 3-1-2 are respectively connected to the profile of X-axis linear module 3-3-1 by bolts; the X-axis extension platform servo motor 3-3-2 is connected to the synchronous pulley of X-axis linear module 3-3-1; X-axis linear module 3-3-1 is driven by the X-axis extension platform servo motor.

[0112] The X-axis extension platform can expand the working range of the picking robotic arm unit in the X direction. The effective length of the X-axis linear module is 1.5m, which allows the working range of the picking robotic arm unit to cover multiple fruit trees.

[0113] Z-axis extension platform servo motor 3-1-3 and X-axis extension platform servo motor 3-3-2 are connected to controller 6-1. Controller 6-1 can control Z-axis extension platform servo motor 3-1-3 and X-axis extension platform servo motor 3-3-2, thereby realizing motion control of Z-axis extension platform 3-1 and X-axis extension platform 3-3.

[0114] The picking robotic arm unit 4 is connected to the slider of the X-direction linear module 3-3-1.

[0115] The testing device 8 includes a testing conveyor belt 8-12-1 installed on a high platform of the working platform 2. A testing conveyor belt motor 8-12-2 is installed at one end of the testing conveyor belt 8-12-1. A U-shaped testing chamber shell 8-16 is fitted on the testing conveyor belt 8-12-1. Black curtains 8-1 are installed on the two opposite side walls of the U-shaped testing chamber shell 8-16. A partition 8-13 with apple passage holes is installed inside the U-shaped testing chamber shell 8-16. The partition 8-13 is fitted above the testing conveyor belt 8-12-1. Six appearance testing modules and four internal quality testing modules are evenly distributed on both sides of the testing conveyor belt 8-12-1. The six appearance testing modules are located inside the testing chamber shell 8-16 on one side of the partition 8-13, and the four internal quality testing modules are located inside the testing chamber shell 8-16 on the other side of the partition 8-13.

[0116] The six appearance inspection modules are: Appearance Inspection Module 1 (8-2), Appearance Inspection Module 2 (8-3), Appearance Inspection Module 3 (8-4), Appearance Inspection Module 4 (8-5), Appearance Inspection Module 5 (8-6), and Appearance Inspection Module 6 (8-7).

[0117] The four internal quality inspection modules are: Internal Quality Inspection Module 1 (8-8), Internal Quality Inspection Module 2 (8-9), Internal Quality Inspection Module 3 (8-10), and Internal Quality Inspection Module 4 (8-11).

[0118] Shadowless lamps 8-14 are installed on two opposite side walls of the inspection chamber shell 8-16 where several appearance inspection modules are located;

[0119] Halogen lamps 8-15 are installed on the two opposite side walls of the testing chamber shell 8-16 where several internal quality testing modules are located;

[0120] Z-axis linear module 3-1-1 and Z-axis linear module 3-1-2 are located on both sides of the entrance end of the detection conveyor belt 8-12-1, respectively.

[0121] Apples are conveyed into the inspection chamber housing 8-16 by inspection conveyor belt 8-12-1, which is driven by inspection conveyor belt motor 8-12-2. The front and rear ends of the inspection chamber housing 8-16 are covered with black curtains 8-1 to create a dark environment. Six appearance inspection modules and four internal quality inspection modules are arranged on both sides of the inspection conveyor belt 8-12-1. The internal space of the inspection chamber housing 8-16 is divided into two parts: the appearance inspection module area and the internal quality inspection module area, separated by a partition 8-13. Shadowless lamps 8-14 are installed on both sides of the inner side walls of the appearance inspection module area, and their brightness can be freely adjusted. Halogen lamps 8-15 are installed on both sides of the inner side walls of the internal quality inspection module area.

[0122] All six appearance inspection modules are industrial cameras, and all six are connected to controller 2 (6-2) to acquire images in real time and send them to controller 2 (6-2). Among them, appearance inspection module 1 (8-2) and appearance inspection module 2 (8-3) are used to detect the diameter and shape of apples, appearance inspection module 3 (8-4) and appearance inspection module 4 (8-5) are used to detect the color of apples, and appearance inspection module 5 (8-6) and appearance inspection module 6 (8-7) are used to detect defects on the surface of apples.

[0123] All four internal quality inspection modules are infrared spectrometers, which can collect apple spectral data in real time. All four internal quality inspection modules are connected to controller 6-2.

[0124] The spectral data is then sent to controller 6-2 (number 2). Controller 6-2 inputs the feedback spectral data into the apple soluble solids analysis model, which outputs the soluble solids results, thus obtaining the apple's soluble solids information. Controller 6-2 also inputs the feedback spectral data into the apple internal lesion analysis model, which outputs the internal lesion results, thus obtaining the apple's internal lesion information.

[0125] The detection device 8 performs secondary testing and grading of apples according to indicators such as fruit diameter, shape, color, surface defects, soluble solids, and internal lesions, classifying them into four grades: superior, first-class, second-class, and inferior. Each apple's grade is the lowest among all indicators. Controller 6-2 performs secondary coding on the tested apples.

[0126] The visual navigation system 11 includes a binocular camera base 11-1 and a navigation controller 11-3. The binocular camera base 11-1 is fixedly installed on the upper platform of the working platform 2. A binocular camera 11-2 is installed on the binocular camera base 11-1, and the lens of the binocular camera 11-2 faces the direction of movement of the electric tracked chassis 1-1. Both the binocular camera 11-2 and the electric tracked chassis 1-1 are connected to the navigation controller 11-3.

[0127] The sorting device 9 includes four sorting conveyor belts and collection boxes. The four sorting conveyor belts are evenly divided into two groups and arranged on both sides of the detection conveyor belt 8-12-1. Each sorting conveyor belt has a collection box 10 at its end. Each sorting conveyor belt has a sorting conveyor belt servo motor installed on one side. All sorting conveyor belt servo motors are connected to controller 6-2.

[0128] The four sorting conveyor belts are: sorting conveyor belt 1 (9-1-1), sorting conveyor belt 2 (9-2-1), sorting conveyor belt 3 (9-3-1), and sorting conveyor belt 4 (9-4-1); each sorting conveyor belt has a collection box 10 at its end, and there are four collection boxes 10 in total.

[0129] Each sorting conveyor belt is equipped with a sorting conveyor belt servo motor on one side, specifically including: sorting conveyor belt motor No. 1 9-1-2, sorting conveyor belt motor No. 2 9-2-2, sorting conveyor belt motor No. 3 9-3-2 and sorting conveyor belt motor No. 4.

[0130] Four vertical supports, numbered 9-10, are evenly distributed in two groups on the outside of several sorting conveyor belts.

[0131] The Y-axis module 9-5-1 of sorting robot arm 1 is installed on two vertical supports 9-10 located on one side of the inspection conveyor belt 8-12-1. The Y-axis module 9-5-2 of sorting robot arm 2 is installed on two vertical supports 9-10 located on the other side of the inspection conveyor belt 8-12-1. Both the Y-axis modules 9-5-1 and 9-5-2 of sorting robot arm 1 are synchronous belt linear modules. The synchronous pulley of Y-axis module 9-5-2 is connected through coupling II 9-11. It also includes a sorting robot arm Y-axis servo motor 9-5-3. The sorting robot arm Y-axis servo motor 9-5-3 is connected to the synchronous pulley of sorting robot arm Y-axis module 9-5-1. The sorting robot arm Y-axis servo motor 9-5-3 drives the synchronous pulley of sorting robot arm Y-axis module 9-5-1 and drives the synchronous pulley of sorting robot arm Y-axis module 9-5-2 to rotate synchronously through coupling II 9-11.

[0132] It also includes a sorting robot arm X-axis module 9-6-1, a sorting robot arm X-axis servo motor 9-6-2, and a sorting robot arm Z-axis module 9-7-1. Both the X-axis module 9-6-1 and the Z-axis module 9-7-1 are synchronous belt linear modules. The sliders of the sorting robot arm Y-axis modules 9-5-1 (module 1) and 9-5-2 (module 2) are connected to the sorting robot arm X-axis module 9-6-1 via bolts. The sorting robot arm X-axis servo motor 9-6-... 2. The sorting robot arm X-axis module 9-6-1 is connected to the synchronous pulley. The sorting robot arm Z-axis module 9-7-1 is bolted to the slider of the sorting robot arm X-axis module 9-6-1. The sorting robot arm Z-axis servo motor 9-7-2 is installed at the synchronous pulley of the sorting robot arm Z-axis module 9-7-1. The sorting robot arm Y-axis servo motor 9-5-3, sorting robot arm X-axis servo motor 9-6-2 and sorting robot arm Z-axis servo motor 9-7-2 are all connected to controller 6-2.

[0133] The end of the sorting robot arm's Z-axis module 9-7-1 is equipped with a flexible sorting gripper 9-8. The flexible sorting gripper has the same structure as the flexible picking gripper. An industrial camera 9-9 is installed on the side of the end of the sorting robot arm's Z-axis module 9-7-1. The industrial camera 9-9 is connected to controller 6-2. The camera takes pictures of the apples after secondary grading on the inspection conveyor belt (1-12-1) and sends the pictures to controller 2. The secondary grading apples are identified and located by a predetermined algorithm. Controller 2 drives the sorting robot arm to stack the graded apples onto the corresponding sorting conveyor belt.

[0134] The harvesting robotic arm unit 4 includes a redundant robotic arm base 4-4 and a redundant robotic arm 4-3 connected in sequence. The redundant robotic arm base 4-4 is connected to the slider of the X-direction linear module 3-3-1. A depth camera base 4-1 is installed on the arm body of the redundant robotic arm 4-3 away from the redundant robotic arm base 4-4. A depth camera 4-2 is installed on the depth camera base 4-1. The redundant robotic arm 4-3 is connected to the harvesting gripper unit 5.

[0135] The depth camera 4-2 and the redundant robotic arm 4-3 are both connected to controller 6-1.

[0136] The picking gripper unit 5 includes a flexible picking gripper 5-2. A thin-film pressure sensor 5-3 is provided on the inner side of the flexible picking gripper 5-2. The flexible picking gripper 5-2 is connected to the end of the redundant robotic arm 4-3 away from the base 4-4 of the redundant robotic arm through a connector I 5-1. The connector I 5-1 includes a flange 5-1-1. Two connecting protrusions 5-1-2 are provided on the edge of the flange 5-1-1. The flange 5-1-1 is sleeved on the redundant robotic arm 4-3. The two connecting protrusions 5-1-2 are connected to the end of the flexible picking gripper 5-2.

[0137] The flexible picking gripper 5-2 and the thin-film pressure sensor 5-3 are both connected to controller 6-1.

[0138] The fruit conveying tube unit 7 includes a funnel-shaped inlet 7-2, a telescopic hose 7-3, and a connecting pipe 7-4 connected in sequence. The upper end of the connecting pipe 7-4 is connected to the telescopic hose 7-3, and the lower end of the connecting pipe 7-4 is fixed to the end cap of the detection conveyor belt 8-12. A support pad 7-6 is provided at the outlet of the connecting pipe 7-4. It also includes a connector II 7-1, which includes a connecting rod 7-1-1 with an open retaining ring and a connecting rod 7-1-2 with a through hole that are connected to each other. The connecting rod 7-1-2 with the through hole is sleeved with the redundant robotic arm 4-3, and the connecting rod 7-1-1 with the open retaining ring is sleeved with the funnel-shaped inlet 7-2.

[0139] The funnel-shaped inlet 7-2 is fitted with a one-way valve 7-5 at its end;

[0140] in,

[0141] The electric tracked chassis in the self-balancing tracked chassis 1 is an existing structure. The hydraulic pump station 1-2 and the control valve group 1-3 are fixedly connected to the bottom of the chassis platform 1-10. An angle sensor A1-11 is installed on the side of the chassis platform 1-10, and a hydraulic controller 1-12 is installed at the bottom of the chassis platform 1-10.

[0142] The tops of hydraulic lifting cylinders 1-4-1, 1-5-1, 1-6-1, 1-7-1, 1-8-1, and 1-9-1 are fixedly connected to the chassis platform 1-10; the bottoms of hydraulic lifting cylinders 1-4-1, 1-5-1, 1-6-1, 1-7-1, 1-8-1, and 1-9-1 are fixedly connected to the chassis of the electric tracked chassis 1-1.

[0143] Pressure sensor 1-4-3 is installed at the bottom of hydraulic lifting cylinder 1-4-1; pressure sensor 1-5-3 is installed at the bottom of hydraulic lifting cylinder 2-5-1; pressure sensor 1-6-3 is installed at the bottom of hydraulic lifting cylinder 3-6-1; pressure sensor 1-7-3 is installed at the bottom of hydraulic outrigger column 4-7-2; pressure sensor 1-8-3 is installed at the bottom of hydraulic lifting cylinder 5-8-1; and pressure sensor 1-9-3 is installed at the bottom of hydraulic lifting cylinder 6-9-1.

[0144] The hydraulic circuit of hydraulic lifting cylinder 1-4-1 is equipped with hydraulic lock 1-4-4 and electro-hydraulic valve 1-4-2. The hydraulic circuit of hydraulic lifting cylinder 2-5-1 is equipped with hydraulic lock 1-5-4 and electro-hydraulic valve 1-5-2. The hydraulic circuit of hydraulic lifting cylinder 3-6-1 is equipped with hydraulic lock 1-6-4 and electro-hydraulic valve 1-6-2. The hydraulic circuit of hydraulic lifting cylinder 4-7-1 is equipped with hydraulic lock 1-7-4 and electro-hydraulic valve 1-7-2. The hydraulic circuit of hydraulic lifting cylinder 5-8-1 is equipped with hydraulic lock 1-8-4 and electro-hydraulic valve 1-8-2. The hydraulic circuit of hydraulic lifting cylinder 6-9-1 is equipped with hydraulic lock 1-9-4 and electro-hydraulic valve 1-9-2.

[0145] The oil circuit of hydraulic pump station 1-2 is connected in series with electro-hydraulic directional valve 1-3-1 and multi-way directional valve 1-3-2. The oil circuits of hydraulic outriggers 1-4, 2-5, 3-6, 4-7, 5-8 and 6-9 are connected in parallel with multi-way directional valve 1-3-2.

[0146] The hydraulic controller 1-12 is connected to the electro-hydraulic directional valve 1-3-1, the multi-way directional valve 1-3-2, the No. 1 two-position electro-hydraulic valve 1-4-2, the No. 2 two-position electro-hydraulic valve 1-5-2, the No. 3 two-position electro-hydraulic valve 1-6-2, the No. 4 two-position electro-hydraulic valve 1-7-2, the No. 5 two-position electro-hydraulic valve 1-8-2, the No. 6 two-position electro-hydraulic valve 1-9-2, and the tilt sensor A1-11.

[0147] Hydraulic pump station 1-2 serves as the hydraulic source, transmitting hydraulic oil to electro-hydraulic directional valve 1-3-1. Electro-hydraulic directional valve 1-3-1 reverses the hydraulic flow to facilitate the extension and retraction of each hydraulic outrigger. Multi-way directional valve 1-3-2 divides the oil circuit into six routes, with the oil circuits of the six hydraulic outriggers connected in parallel via multi-way directional valve 1-3-2. Each hydraulic lifting cylinder's oil circuit is equipped with a two-position electro-hydraulic valve, acting as a switch to allow each hydraulic lifting cylinder to operate independently for easy leveling. Each hydraulic lifting cylinder's oil circuit is also equipped with a one-way hydraulic lock to prevent the vertical hydraulic cylinder from naturally settling and causing a "soft leg" phenomenon due to circuit leakage.

[0148] The tilt sensor A1-11 measures the tilt angle of the chassis platform relative to the horizontal plane and converts the angle signal into a digital signal, which is then transmitted to the hydraulic controller 1-12 via a wireless network. Pressure sensors 1-4-3, 1-5-3, 1-6-3, 1-7-3, 1-8-3, and 1-9-3 can respectively detect the axial pressure of hydraulic lifting cylinders 1-4-1, 1-5-1, 1-6-1, 1-7-1, 1-8-1, and 1-9-1, and send feedback signals to the hydraulic controller 1-12.

[0149] Based on the feedback signals from tilt sensor A1-11 and various pressure sensors, hydraulic controller 1-12 sends commands to hydraulic pump station 1-2, electro-hydraulic directional valve 1-3-1, multi-way directional valve 1-3-2, No. 1 two-position electro-hydraulic valve 1-4-5, No. 2 two-position electro-hydraulic valve 1-5-5, No. 3 two-position electro-hydraulic valve 1-6-5, No. 4 two-position electro-hydraulic valve 1-7-5, No. 5 two-position electro-hydraulic valve 1-8-5, and No. 6 two-position electro-hydraulic valve 1-9-5 according to the leveling algorithm. This adjusts the lifting height of hydraulic lifting cylinders 1-4-1, 1-5-1, 1-6-1, 1-7-1, 1-8-1, and 1-9-1 to ensure relative pressure balance at the bottom of each hydraulic outrigger and achieve relative horizontal control of the frame platform 1-10.

[0150] The bottom of the work platform 2 is fixedly connected to the chassis platform 1-10. The work platform 2 is a stepped platform, divided into two levels, high and low. The upper level is equipped with the extension platform 3, control box 6, detection device 8, sorting device 9, and visual navigation system 11. The lower level is equipped with collection boxes 10.

[0151] An extension platform 3 is installed on the side of the working platform 2. The extension platform 3 has an I-shaped structure and is divided into a Z-axis extension platform 3-1 and an X-axis extension platform 3-3, where Z is the vertical direction and X is the forward direction of the picking robot. The Z-axis extension platform 3-1 is divided into a Z-axis linear module 3-1-1 and a Z-axis extension module 3-1-2. Both Z-axis linear modules 3-1-1 and 3-1-2 are synchronous belt linear modules, which is the existing structure. The lower ends of Z-axis linear modules 3-1-1 and 3-1-2 are fixed to the working platform 2. A Z-axis extension platform servo motor 3-1-3 is installed at the synchronous pulley at the upper end of Z-axis linear module 3-1-1. The other side of the synchronous pulley of Z-axis linear module 3-1-1 is connected to the upper synchronous pulley of Z-axis linear module 3-1-2 via coupling I 3-2. The Z-axis extension platform 3-1 is driven by the Z-axis extension platform servo motor 3-1-3. The X-axis extension platform 3-3 is bolted onto the sliders of Z-axis linear module 3-1-1 (module 1) and Z-axis linear module 3-1-2 (module 2). X-axis linear module 3-3-1 is a synchronous belt linear module. The X-axis extension platform servo motor 3-3-2 is mounted on the synchronous pulley of X-axis linear module 3-3-1. The X-axis extension platform 3-1 is driven by the X-axis extension platform servo motor 3-3-2.

[0152] The expansion platform 3 can expand the working range of the harvesting robotic arm unit 4. The Z-axis expansion platform 3-1 can expand the working range of the harvesting robotic arm unit 4 in the Z-axis. The effective length of the Z-axis linear module 3-1-1 and the Z-axis linear module 3-1-2 is 2m, so that the working range of the harvesting robotic arm unit 4 can cover the entire fruit tree. The X-axis expansion platform 3-3 can expand the working range of the harvesting robotic arm unit 4 in the X-axis. The effective length of the X-axis linear module 3-3-1 is 1.5m, so that the working range of the harvesting robotic arm unit 4 can cover multiple fruit trees. The Z-axis expansion platform servo motor 3-1-3 and the X-axis expansion platform servo motor 3-3-2 are connected to the controller 6-1. The controller 6-1 can control the Z-axis expansion platform servo motor 3-1-3 and the X-axis expansion platform servo motor 3-3-2, thereby realizing motion control of the Z-axis expansion platform 3-1 and the X-axis expansion platform 3-3.

[0153] The redundant robotic arm base 4-4 is bolted onto the slider of the X-direction linear module 3-3-1.

[0154] A redundant robotic arm 4-3 is bolted onto the redundant robotic arm base 4-4. Taking a serial seven-DOF redundant robotic arm as an example, the redundant robotic arm 4-3 has more degrees of freedom than the three-dimensional spatial degrees of freedom (e.g., a serial seven-DOF redundant robotic arm). Utilizing redundant joints, obstacle avoidance within the workspace can be achieved. Furthermore, while ensuring the position and posture of the picking gripper unit 5 remain unchanged, various postures can be controlled by adjusting the redundant joints, further enhancing the flexibility of the picking robot. It also improves the obstacle avoidance capability of the picking gripper unit 5; the redundant robotic arm 4-3 can freely bypass obstacles such as tree trunks, branches, and leaves, and can pick apples obscured by obstacles.

[0155] The redundant robotic arm 4-3 is connected to controller 6-1, which can realize motion planning and control of the redundant robotic arm 4-3.

[0156] The depth camera base 4-1 is fixed near the end of the redundant robotic arm 4-3. A depth camera 4-2 is mounted on the depth camera base 4-1 and connected to controller 6-1. The depth camera 4-2 collects information on the fruit tree and the position, posture, shape, and color of each apple on the tree.

[0157] The depth camera 4-2 sends the captured images to controller 6-1. Controller 6-1 uses the YOLO algorithm to identify the target apple and then uses the RGB-ICP algorithm to locate the apple's pose. Based on the target apple's pose, controller 6-1 performs motion planning and control on the extension platform 3 and the picking robotic arm unit 4, transporting the picking gripper unit 5 to the target apple picking position.

[0158] A flexible picking gripper unit 5 is installed at the end of the redundant robotic arm 4-3, and the flexible picking gripper 5-2 and the redundant robotic arm 4-3 are connected by bolts through connector I 5-1.

[0159] The flexible picking gripper 5-2 is connected to the I / O module of controller 6-1. The thin-film pressure sensor 5-3 can collect the surface pressure of the fingers of the flexible picking gripper 5-2 in real time and convert it into a digital signal and send it to controller 6-1. Controller 6-1 controls the opening and closing degree of the flexible picking gripper 5-2 according to the feedback signal of the thin-film pressure sensor 5-3. When a certain pressure is reached, the bionic gripper 5-2 will no longer close, ensuring that the apple will not fall or be damaged by squeezing during the picking process, thus realizing flexible apple picking.

[0160] The thin-film pressure sensor 5-3 collects information on the fruit's diameter and firmness, which is combined with the depth camera 4-2 to form a visual-tactile coupling system. The visual-tactile coupling system sends the collected information to controller 1 1-6, which then uses the RGB-YOLOv5 algorithm to encode and initially grade the apples according to the apple quality grading requirements specified in GB / T10651-2008, enabling online real-time grading during the harvesting process for subsequent grading and sorting by the sorting device 9.

[0161] To address the problems of limited working range, poor obstacle avoidance, high apple damage, slow harvesting speed, and difficulty in apple grading in current apple harvesting robots, this invention provides a long-stroke apple harvesting robot equipped with an extension platform to expand the working space of the harvesting robotic arm. The robot features a redundant degree of freedom, enhancing its flexibility and obstacle avoidance capabilities. A flexible harvesting gripper at the end of the robotic arm effectively reduces apple damage. After harvesting, apples can be transported via a fruit conveyor, shortening the return time of traditional robotic arms and improving harvesting efficiency, enabling continuous apple harvesting. Furthermore, the harvested apples can be graded according to quality.

[0162] The following method is used to code the harvested apples.

[0163] The coding sequence for apples is xxxx-dxx-sxx-rxx-fxx-hxx, where xxxx represents the picking order of the apples, numbered 0001-9999; dxx represents the diameter of the apple at picking, which is the maximum of the apple's width and length, with xx being the apple's diameter size in mm; sx.xx represents the shape of the apple at picking, with the shape index being the ratio of the fruit's width to its length, and x.xx being the apple's shape index; rxx represents the color of the apple at picking, with xx being the apple's coloring ratio; fxx represents the surface defects of the apple at picking, with xx representing the number of surface defects; and hxx represents the firmness of the apple at picking, with xx being the apple's firmness in mm. .

[0164] Controller 1 (6-1) controls the expansion platform 3, redundant robotic arm 4, and flexible picking gripper 5; Controller 2 (6-2) controls the detection device 8 and sorting device 9.

[0165] A connector II 7-1 is installed at the end of the redundant robotic arm 4. The redundant robotic arm 4 is bolted to the funnel-shaped inlet 7-2 of the fruit conveying tube unit 7 via connector II 7-1. The upper end of the fruit conveying tube unit 7 is the funnel-shaped inlet 7-2, and a one-way valve is installed at the end of the funnel-shaped inlet 7-2. The upper half of the fruit conveying tube unit 7 is a flexible telescopic hose 7-3, made of corrugated pipe, which is fixedly connected at both ends to the funnel-shaped inlet 7-2 and the connecting pipe 7-4, respectively. A one-way valve 7-5 is installed at the connection between the flexible telescopic hose 7-3 and the connecting pipe 7-4. The lower half is the connecting pipe 7-4, made of aluminum alloy. The lower end of the connecting pipe 7-4 is fixedly connected to the working platform 2, and a support pad 7-6 is installed at the connection point.

[0166] The funnel-shaped inlet 7-2 is made of lightweight and soft silicone, reducing the load on the redundant robotic arm 4. To prevent damage to falling apples, the funnel-shaped inlet 7-2 has a slight curvature and is lined with a layer of sponge material. The one-way valve 7-5 at the end of the funnel-shaped inlet 7-2 reduces the impact of falling apples. The inside of the fruit conveying tube unit 7 is lined with wool material to reduce damage from rolling apples and to wipe away dust from the apple surface. A structure using a telescopic hose 7-3 and a connecting tube 7-4 prevents the fruit conveying tube unit 7 from sagging due to gravity, which could cause apples to get stuck inside. The pad 7-6 at the connection point between the connecting tube 7-4 and the detection conveyor belt 8-12-1 reduces the impact of rolling apples.

[0167] Inspection devices 8 are installed on the work platform 2. Apples are transported to the interior of the thin-walled outer shell 8-15 by inspection conveyor belt 8-12-1, which is driven by inspection conveyor belt motor 8-12-2. The front and rear ends of the thin-walled outer shell 8-16 are covered with black curtains 8-1 to create a dark environment. On one side of the inspection conveyor belt 8-12-1, there are No. 1 appearance inspection module 8-2, No. 3 appearance inspection module 8-4, No. 5 appearance inspection module 8-6, No. 1 internal quality inspection module 8-8, and No. 3 internal quality inspection module 8-10; on the other side, there are No. 2 appearance inspection module 8-3, No. 4 appearance inspection module 8-5, No. 6 appearance inspection module 8-7, No. 2 internal quality inspection module 8-9, and No. 4 internal quality inspection module 8-11. The thin-walled outer shell 8-16 has an internal space divided into two parts: an appearance inspection module area and an internal quality inspection module area, separated by a partition 8-13. Shadowless lamps 8-14 are installed on both sides of the internal sidewalls of the appearance inspection module area, and their brightness can be freely adjusted. Halogen lamps 8-15 are installed on both sides of the sidewalls of the internal quality inspection module area.

[0168] Appearance inspection modules 1 (8-2), 3 (8-4), 5 (8-6), 2 (8-3), 4 (8-5), and 6 (8-7) are all industrial cameras connected to controller 2 (6-2) for real-time image acquisition and transmission. Modules 1 (8-2) and 2 (8-3) are used to inspect the apple's diameter and shape; modules 3 (8-4) and 4 (8-5) are used to inspect the apple's color; and modules 5 (8-6) and 6 (8-7) are used to inspect the apple's surface defects.

[0169] Appearance inspection modules 1 (8-2) and 2 (8-3) send the captured images to controller 2 (6-2). Controller 2 uses OpenCV to perform image segmentation, contour extraction, and size and shape calculations, thereby obtaining and recording the apple's diameter and shape information. Appearance inspection modules 3 (8-4) and 4 (8-5) send the captured images to controller 2, calculating the ratio of the total area of ​​red pixels to the total area of ​​pixels on the apple's surface, thus obtaining the apple's color information. Appearance inspection modules 5 (8-6) and 6 (8-7) send the captured images to controller 2. Controller 2 uses YOLOv5 software to train a surface defect detection model, and then uses this model to detect surface defects, thereby obtaining the apple's surface defect information.

[0170] Internal quality inspection modules 1 (8-8), 3 (8-10), 2 (8-9), and 4 (8-11) are all infrared spectrometers, capable of acquiring apple spectral data in real time and sending it to controller 2 (6-2). Controller 2 inputs the feedback spectral data into the apple soluble solids analysis model, which then outputs the soluble solids results, thus obtaining the apple's soluble solids information. Controller 2 also inputs the feedback spectral data into the apple internal lesion analysis model, which outputs the internal lesion results, thus obtaining the apple's internal lesion information.

[0171] The testing device 8 performs secondary testing and grading on apples according to indicators such as fruit diameter, fruit shape, color, fruit surface defects, soluble solids, and internal lesions. The apples are divided into four grades: superior, first-class, second-class, and inferior. Each apple's grade is the lowest grade for each indicator.

[0172] According to GB / T10651-2008, apples can be classified into four grades based on fruit diameter: large apples with a diameter ≥70mm and medium-sized apples with a diameter ≥60mm are classified as superior and first grade; large apples with a diameter ≥65mm and medium-sized apples with a diameter ≥55mm are classified as second grade; the remaining apples are classified as substandard. Based on fruit shape, apples can be classified into four grades: those possessing the characteristics of the variety are superior; those with minor shape defects are first grade; those with shape defects but still maintaining the basic characteristics of the variety are second grade; the remaining apples are classified as inferior. Based on apple color, apples can be classified into four grades; red varieties are graded according to the proportion of skin color, while other varieties should possess the color they should have when ripe. Based on the degree of surface defects, apples can be classified into four grades: superior and first grade apples have no surface defects; second grade apples have no more than four minor skin damages that do not significantly harm the flesh; and apples with obvious punctures, cracks, insect damage, or spots are classified as substandard.

[0173] According to GB / T10651-2008, the physicochemical index of soluble solids in apples is referenced. Apples with values ​​below the reference value are considered inferior fruits, as are apples with internal lesions.

[0174] Controller 2 (6-2) performs secondary encoding on the detected apples, following the method described below.

[0175] The secondary encoding sequence of the apple is xxxx-dxx-sxx-rxx-fxx-hxx-bxx-mx-lx.

[0176] Where xxxx represents the picking order of the apples, numbered 0001-9999; dxx represents the diameter of the picked apples, which is the maximum of the apple's transverse and longitudinal diameters, with xx representing the apple's diameter size in mm; sx.xx represents the shape of the picked apples, with the shape index being the ratio of the fruit's longitudinal to transverse diameters, and x.xx representing the apple's shape index; rxx represents the color of the picked apples, with xx representing the apple's coloring ratio; fxx represents the surface defects of the picked apples, with xx representing the number of surface defects; and hxx represents the firmness of the picked apples, with xx representing the apple's firmness in mm. bxx represents the soluble solids content of the harvested apples, and xx is the soluble solids content of the apples; mx represents the internal lesions of the harvested apples, x=0 represents no internal lesions, x=1 represents internal lesions; lx represents the grade of the harvested apples, x=1 represents superior grade, x=2 represents first grade, x=3 represents second grade, and x=4 represents inferior grade.

[0177] A sorting device 9 is installed on the work platform 2. The sorting device includes a three-degree-of-freedom sorting robot arm, four sorting conveyor belts, and a support frame for the sorting robot arm. The support frame 9-10 consists of four aluminum alloy profiles and is fixedly connected to the work platform 2. The three-degree-of-freedom sorting robot arm is a three-degree-of-freedom Cartesian coordinate type robot arm, consisting of sorting robot arm 1 (Y-axis 9-5-1), sorting robot arm 2 (Y-axis 9-5-2), sorting robot arm X-axis 9-6-1, and sorting robot arm Z-axis 9-7-1. Each sorting robot arm uses a synchronous belt linear module and is driven by a servo motor.

[0178] Sorting conveyors 1 (9-1-1), 2 (9-2-1), 3 (9-3-1), and 4 (9-4-1) are existing equipment. These conveyors are made of rubber to reduce damage during transport and have side skirts to prevent apples from falling sideways. Their bottoms are fixedly connected to the work platform 2. Sorting conveyors 1 (9-1-1), 2 (9-2-1), 3 (9-3-1), and 4 (9-4-1) are driven by servo motors 9-1-2 (9-1-2), 9-2-2 (9-2-2), 9-3-2 (9-3-2), and 9-4-2 (9-4-2), respectively.

[0179] The Y-axis of sorting robot arm 1 (9-5-1) and the Y-axis of sorting robot arm 2 (9-5-2) are fixedly connected to the bracket (9-10). A Y-axis servo motor (9-5-3) is installed at the synchronous pulley of Y-axis 1. The other side of the synchronous pulley of Y-axis 1 is connected to the synchronous pulley of Y-axis 2 of sorting robot arm 9-5-2 via coupling II (9-11). Both Y-axis of sorting robot arm 1 (9-5-1) and Y-axis of sorting robot arm 2 (9-5-2) are driven by the Y-axis servo motor (9-5-3). The sorting robot arm's X-axis 9-6-1 is bolted to the sliders of sorting robot arm Y-axis 9-5-1 (position 1) and sorting robot arm Y-axis 9-5-2 (position 2). A servo motor for sorting robot arm X-axis 9-6-2 is mounted on one side of the synchronous pulley of sorting robot arm X-axis 9-6-1, and sorting robot arm X-axis 9-6-1 is driven by servo motor X-axis 9-6-2. The sorting robot arm's Z-axis 9-7-1 is bolted to the slider of sorting robot arm X-axis 9-6-1. A servo motor for sorting robot arm Z-axis 9-7-1 is mounted on the synchronous pulley of sorting robot arm Z-axis 9-7-1, and sorting robot arm Z-axis 9-7-1 is driven by servo motor Z-axis 9-7-2. Servo motors 9-5-3 (position 2), 9-6-2 (position 3), and 9-7-2 (position 4) are connected in series with controller 6-2 (position 2) via TCP / IP. Controller 2 (serial number 6-2) controls the movement of the sorting robot arm by controlling the Y-axis servo motor 9-5-3, the X-axis servo motor 9-6-2, and the Z-axis servo motor 9-7-2. The flexible sorting gripper 9-8 is installed at the end of the Z-axis of the sorting robot arm (9-7-1), and its structure is the same as the flexible picking gripper. Camera 9-9 is installed on the side of the end of the Z-axis of the sorting robot arm (9-7-1).

[0180] Camera 9-9 is connected to controller 6-2. Camera 9-9 takes pictures of the apples after secondary grading on the detection conveyor belt 8-12 and sends the pictures to controller 6-2. The secondary grading apples are identified and located by a predetermined algorithm. Controller 6-2 drives the sorting robot arm to stack the graded apples onto the corresponding sorting conveyor belt.

[0181] Four collection boxes 10 are set at the end of the work platform 2. The four collection boxes 10 are connected to the No. 1 sorting conveyor belt 9-1-1, the No. 2 sorting conveyor belt 9-2-1, the No. 3 sorting conveyor belt 9-3-1, and the No. 4 sorting conveyor belt 9-4-1, respectively.

[0182] After sorting, the apples are transported to the corresponding collection boxes 10 via a sorting conveyor belt. The interior walls of each collection box 10 are lined with foam material to reduce apple damage. The external structure is fixed in place by grooves, allowing for free installation and removal to prevent the collection box 10 from tipping over due to vibration during robot movement. The apple grades in the four collection boxes 10 are: superior, first-class, second-class, and inferior.

[0183] A visual navigation system is deployed on the working platform 2. A binocular camera base 11-1 is fixedly connected to the working platform 2. A binocular camera 11-2 is mounted on top of the binocular camera base 11-1, and a navigation controller 11-2 is mounted on the side of the binocular camera base 11-1. The binocular camera 11-2 is connected to the navigation controller 11-2. The binocular camera 11-2 sends the captured road images to the navigation controller 11-2. Based on machine vision and deep learning-based autonomous navigation algorithms, the navigation controller 11-2 sends control commands to the electric tracked chassis 1-1, enabling the apple-picking robot to navigate autonomously in the orchard.

[0184] The models of each device in the apple-picking robot of this invention are as follows:

[0185] The navigation controller 11-3 is an STM32F103 microcontroller, the controller 1 is a Zhichang general-purpose controller ROB200, and the controller 2 is a Zhichang controller ROB100N; all of the above controllers are connected to an external computer.

[0186] The electric tracked chassis 1-1 is model KRT-2000; the tilt sensor 1-11 is model RS-DIP-N01-1H; the Z-axis linear module 1 (3-1-1) is model RXP50; the Z-axis linear module 2 (3-1-2) is model RXP50; the Z-axis extended platform servo motor 3-1-3 is model SDGA-08C12PD; the X-axis linear module 3-3-1 is model RXP50; the X-axis extended platform servo motor 3-3-2 is model SDGA-08C11PD; and the depth camera 4-2 is model The Intel®. The following components are used: RealSense™ D415, redundant robotic arm 4-3 (model TC7-R3), appearance inspection module (model MV-CE013-50GM), internal quality inspection module (model SOC710-VP), sorting robotic arm Y-axis module 9-5-1 (model RXP30), sorting robotic arm Y-axis module 9-5-2 (model RXP30), sorting robotic arm Y-axis servo motor 9-5-3 (model SDGA-04C11PD), sorting robotic arm X-axis module 9-6-1 (model RXP30), sorting robotic arm X-axis servo motor 9-6-2 (model SDGA-04C11PD), sorting robotic arm Z-axis module 9-7-1 (model RXP30), sorting robotic arm Z-axis servo motor 9-7-2 (model SDGA-04C12PD), and industrial camera 9-9 (model HBVCAM-W202011HD). The V33 and binocular camera 11-2 are model Basler Stereo; the flexible picking gripper 5-2 and flexible sorting gripper 9-8 both use Wheeltec's flexible bionic mechanical grippers; the hydraulic controller 1-12 is Jingfeng Pneumatic's hydraulic intelligent programmable controller.

[0187] This invention also provides a long-stroke apple picking method using the aforementioned robot. Specifically, a navigation controller controls an electric tracked chassis to enable the long-stroke picking robot to move within the orchard. Once the robot reaches a designated location near the target tree, a hydraulic controller adjusts the hydraulic outriggers to keep the work platform level. A depth camera captures images of the fruit trees, and controller 1 processes these images to plan the picking sequence and position of the target apples. Controller 1 drives the picking robotic arm and flexible picking gripper to pick the apples in sequence, and performs preliminary coding and grading based on data from the depth camera and thin-film pressure sensor. After picking, the apples are transported to a detection device via a fruit conveying tube unit. The apples are then further coded and graded based on data from various appearance and internal quality detection modules within the detection device. Finally, they are transported to collection boxes of the appropriate grade via a sorting device.

[0188] The large-scale apple picking method is implemented according to the following steps: Step S1: The binocular camera 11-2 takes photos of the orchard road and sends them to the navigation controller 11-3. Using machine vision and deep learning autonomous navigation algorithms, the controller sends control commands to the electric tracked chassis 1-1, thereby enabling the picking robot to move in the orchard.

[0189] Step S2: The picking robot moves to a designated position near the target fruit tree. The hydraulic controller 1-12 receives data from the tilt sensor A1-11 and various pressure sensors. The hydraulic controller 1-12 sends control commands to the hydraulic pump station 1-2, the control valve group 1-3, and various two-position electro-hydraulic valves, thereby realizing the adjustment of each hydraulic outrigger and the relative horizontal control of the chassis platform 1-10 and the working platform 2.

[0190] Step S3: Controller 1 6-1 controls the movement of the extension platform 3 and the picking robotic arm unit 4, uses the depth camera 4-2 to acquire images of the target fruit tree, and sends the images to Controller 1 6-1. Using the YOLOv5 algorithm, all target apples are identified, and then the picking sequence and pose positioning of the target apples are planned through 3D point cloud processing.

[0191] Step S4: Controller 1 6-1 drives the extension platform 3 and redundant robotic arm 4-3 to the picking position according to the target apple picking order, and drives the flexible picking gripper 5-2 to wrap around the target apple. Controller 1 6-1 uses the thin-film pressure sensor 5-4-3 to feed back pressure information and control the motor 5-2-1 in real time, thereby adjusting the opening and closing degree of the bionic gripper 5-4.

[0192] Step S5: Controller 1 6-1, based on data from depth camera 4-2 and thin-film pressure sensor, uses the YOLOv5 algorithm to initially acquire information on the target apple's diameter, shape, color, surface defects, and firmness. Controller 1 6-1 then performs initial grading and encoding of the target apple based on this information, controlling the redundant robotic arm 4 and the flexible picking gripper 5 to pick the target apple.

[0193] Step S6: After harvesting, the apples are transported through the fruit conveying tube unit 7 to the detection device 8 and the detection conveyor belt 8-12-1.

[0194] Step S7: Controller 6-2, based on data from the appearance inspection modules and internal quality inspection modules in the inspection device 8, uses a predetermined algorithm and inspection model to perform secondary inspection, grading, and coding of the apples according to indicators such as fruit diameter, shape, color, surface defects, soluble solids, and internal lesions. The coded apples are then transported to the sorting device 9 via the inspection conveyor belt 8-12-1.

[0195] Step S8: Controller 2 (6-2) takes photos using camera 9-9, uses the YOLOv5 algorithm to identify the coded apples, and calculates their positions. Controller 2 (6-2) then drives the sorting robotic arm and flexible sorting gripper to transport the secondary-graded apples to the corresponding graded collection boxes.

[0196] Step S9: After the target apples have been sorted, repeat steps S4-S7;

[0197] Step S10: After all the target apples on the fruit tree have been harvested, repeat steps S2-S9.

[0198] Example 1

[0199] Large-scale apple picking robots, such as Figure 1-7 As shown, it includes a self-balancing tracked chassis 1, on which a working platform 2 is mounted. The working platform 2 is a stepped platform, divided into a high platform and a low platform. The high platform is equipped with an extension platform 3, a control box 6, a detection device 8, a sorting device 9, and a vision navigation system 11. The low platform is equipped with several collection boxes 10. The detection device 8, the sorting device 9, and the several collection boxes 10 on the working platform 2 are arranged in sequence according to the apple detection, sorting, and collection process.

[0200] The expansion platform 3 is equipped with a picking robotic arm unit 4, a picking gripper unit 5, and a fruit conveying tube unit 7, which is connected to the picking robotic arm unit 4.

[0201] Control box 6 includes controller 6-1 (No. 1) and controller 6-2 (No. 2);

[0202] The expansion platform 3, the picking robotic arm unit 4, and the picking gripper unit 5 are all connected to controller 6-1.

[0203] Both the detection device 8 and the sorting device 9 are connected to controller 6-2 (number 2).

[0204] The self-balancing tracked chassis 1 is connected to the visual navigation system 11.

[0205] Example 2

[0206] Based on Embodiment 1, the self-balancing tracked chassis 1 includes an electric tracked chassis 1-1 and a chassis platform 1-10. The electric tracked chassis 1-1 and the chassis platform 1-10 are connected by six hydraulic outriggers. The six hydraulic outriggers are divided into two rows, and three outriggers are evenly arranged at the bottom of the chassis platform 1-10 in each row.

[0207] The six hydraulic outriggers are hydraulic outrigger 1-4, hydraulic outrigger 1-5, hydraulic outrigger 3-6, hydraulic outrigger 4-7, hydraulic outrigger 5-8, and hydraulic outrigger 6-9.

[0208] Hydraulic outrigger 1-4 includes hydraulic lifting cylinder 1-4-1, two-position electro-hydraulic valve 1-4-2, pressure sensor 1-4-3, and one-way hydraulic lock 1-4-4.

[0209] Example 3

[0210] Based on Example 2, hydraulic outrigger 1-5 includes hydraulic lifting cylinder 2-5-1, two-position electro-hydraulic valve 2-5-2, pressure sensor 2-5-3, and one-way hydraulic lock 2-5-4.

[0211] Example 4

[0212] Based on Example 3, the No. 3 hydraulic outrigger 1-6 includes the No. 3 hydraulic lifting cylinder 1-6-1, the No. 3 two-position electro-hydraulic valve 1-6-2, the No. 3 pressure sensor 1-6-3, and the No. 3 one-way hydraulic lock 1-6-4.

[0213] Example 5

[0214] Based on Example 4, the No. 4 hydraulic outrigger 1-7 includes the No. 4 hydraulic lifting cylinder 1-7-1, the No. 4 two-position electro-hydraulic valve 1-7-2, the No. 4 pressure sensor 1-7-3, and the No. 4 one-way hydraulic lock 1-7-4.

[0215] Example 6

[0216] Based on Example 5, the No. 5 hydraulic outrigger 1-8 includes the No. 5 hydraulic lifting cylinder 1-8-1, the No. 5 two-position electro-hydraulic valve 1-8-2, the No. 5 pressure sensor 1-8-3, and the No. 5 one-way hydraulic lock 1-8-4.

Claims

1. A long-range apple picking robot, characterized in that: It includes a self-balancing tracked chassis (1), on which a working platform (2) is set; the working platform (2) is a stepped platform, with an extension platform (3), a control box (6), a detection device (8), a sorting device (9) and a visual navigation system (11) on the upper platform, and several collection boxes (10) on the lower platform; the detection device (8), the sorting device (9) and several collection boxes (10) are arranged in sequence according to the apple detection, sorting and collection process; The expansion platform (3) is equipped with a picking robotic arm unit (4), a picking gripper unit (5) is installed on the picking robotic arm unit (4), and also includes a fruit conveying tube unit (7), which is connected to the picking robotic arm unit (4); The control box (6) includes controller No. 1 (6-1) and controller No. 2 (6-2); The expansion platform (3), the picking robotic arm unit (4), and the picking gripper unit (5) are all connected to controller No. 1 (6-1); Both the detection device (8) and the sorting device (9) are connected to controller No. 2 (6-2); The self-balancing tracked chassis (1) is connected to the visual navigation system (11); The detection device (8) includes a detection conveyor belt (8-12-1) installed on a high-level platform of the work platform (2). Several appearance detection modules and several internal quality detection modules are evenly distributed on both sides of the detection conveyor belt (8-12-1). The picking robotic arm unit (4) includes a redundant robotic arm base (4-4) and a redundant robotic arm (4-3) connected in sequence. A depth camera base (4-1) is installed on the arm body of the redundant robotic arm (4-3) away from the redundant robotic arm base (4-4). A depth camera (4-2) is installed on the depth camera base (4-1). The picking gripper unit (5) includes a flexible picking gripper (5-2). A thin film pressure sensor (5-3) is provided on the inner side of the flexible picking gripper (5-2). Controller 1 (6-1) uses YOLOv5 to initially acquire information on the target apple's diameter, shape, color, surface defects, and hardness based on data from the depth camera (4-2) and the thin-film pressure sensor (5-3). Controller 1 (6-1) then performs initial grading and coding of the target apple based on this information, and controls the redundant robotic arm (4-3) and the flexible picking gripper (5-2) to pick the target apple. Controller 2 (6-2) performs secondary grading and coding of the apple based on data from the appearance detection module and internal quality detection module in the detection device (8), according to the indicators of fruit diameter, shape, color, surface defects, soluble solids, and internal lesions.

2. The long-stroke apple-picking robot according to claim 1, characterized in that, The self-balancing tracked chassis (1) includes an electric tracked chassis (1-1) and a frame platform (1-10). The electric tracked chassis (1-1) and the frame platform (1-10) are connected by six hydraulic outriggers. The six hydraulic outriggers are arranged in two rows, with three outriggers evenly distributed at the bottom of the frame platform (1-10). A hydraulic pump station (1-2), a control valve group (1-3), and a hydraulic controller (1-12) are also provided at the bottom of the frame platform (1-10). Tilt sensors (1-11) are arranged on the side of the frame platform (1-10). The control valve group (1-3) includes an electro-hydraulic directional valve (1-3-1) and a multi-way directional valve (1-3-2) that are connected to each other; the oil circuit of the hydraulic pump station (1-2), the electro-hydraulic directional valve (1-3-1) and the multi-way directional valve (1-3-2) are connected in sequence to form a closed circuit; Each hydraulic outrigger includes a hydraulic lifting cylinder, and a pressure sensor is installed at the bottom of the hydraulic lifting cylinder. The oil inlet of the hydraulic lifting cylinder is connected to the oil outlet of the multi-way directional valve (1-3-2) through an oil pipeline, and the oil outlet of the hydraulic lifting cylinder is connected to the oil inlet of the multi-way directional valve (1-3-2) through an oil pipeline. A one-way hydraulic lock and a two-position electro-hydraulic valve are installed on the oil pipeline between the oil inlet of the hydraulic lifting cylinder and the oil outlet of the multi-way directional valve (1-3-2). Each two-position electro-hydraulic valve and pressure sensor is connected to the hydraulic controller (1-12); The electro-hydraulic directional valve (1-3-1), the multi-way directional valve (1-3-2), and the tilt sensor (1-11) are all connected to the hydraulic controller (1-12).

3. The long-stroke apple picking robot according to claim 2, characterized in that, The expansion platform (3) includes a Z-axis expansion platform (3-1), a coupling I (3-2), and an X-axis expansion platform (3-3). The Z-axis extension platform (3-1) includes a Z-axis linear module 1 (3-1-1), a Z-axis linear module 2 (3-1-2), and a Z-axis extension platform servo motor (3-1-3). The Z-axis linear module 1 (3-1-1) and the Z-axis linear module 2 (3-1-2) are vertically installed on the upper platform of the working platform (2). The synchronous pulley of the Z-axis linear module 1 (3-1-1) is connected to the synchronous pulley of the Z-axis linear module 2 (3-1-2) through coupling I (3-2). The Z-axis extension platform servo motor (3-1-3) is connected to the synchronous pulley of the Z-axis linear module 2 (3-1-2). The Z-axis extension platform servo motor (3-1-3) drives the synchronous pulley of the Z-axis linear module 2 (3-1-2) and drives the synchronous pulley of the Z-axis linear module 1 (3-1-1) to rotate through coupling I (3-2). The X-axis extension platform (3-3) includes an X-axis linear module (3-3-1) and an X-axis extension platform servo motor (3-3-2); the sliders of Z-axis linear module 1 (3-1-1) and Z-axis linear module 2 (3-1-2) are respectively connected to the profile of X-axis linear module (3-3-1) by bolts; the X-axis extension platform servo motor (3-3-2) is connected to the synchronous pulley of X-axis linear module (3-3-1); and the X-axis linear module (3-3-1) is driven by the X-axis extension platform servo motor. The Z-axis extension platform servo motor (3-1-3) and the X-axis extension platform servo motor (3-3-2) are connected to controller No. 1 (6-1); The picking robotic arm unit (4) is connected to the slider of the X-direction linear module (3-3-1).

4. The long-stroke apple picking robot according to claim 2, characterized in that, A conveyor motor (8-12-2) is installed at one end of the inspection conveyor belt (8-12-1). A U-shaped inspection chamber housing (8-16) is fitted on the inspection conveyor belt (8-12-1). Black curtains (8-1) are installed on the two opposite side walls of the U-shaped inspection chamber housing (8-16). A partition (8-13) with apple passage holes is installed inside the U-shaped inspection chamber housing (8-16). The partition (8-13) is fitted above the inspection conveyor belt (8-12-1). Several appearance inspection modules are set in the inspection chamber housing (8-16) on one side of the partition (8-13), and several internal quality inspection modules are set in the inspection chamber housing (8-16) on the other side of the partition (8-13). Shadowless lamps (8-14) are installed on two opposite side walls of the inspection chamber shell (8-16) where several appearance inspection modules are located. Halogen lamps (8-15) are installed on two opposite side walls of the testing chamber shell (8-16) where several internal quality testing modules are located. Z-axis linear module 1 (3-1-1) and Z-axis linear module 2 (3-1-2) are located on both sides of the entrance end of the detection conveyor belt (8-12-1); Several appearance inspection modules are industrial cameras, and several appearance inspection modules are connected to controller No. 2 (6-2); Several internal quality inspection modules are infrared spectrometers; several internal quality inspection modules are connected to controller 2 (6-2); The visual navigation system (11) includes a binocular camera base (11-1) and a navigation controller (11-3). The binocular camera base (11-1) is fixedly installed on the upper platform of the working platform (2). A binocular camera (11-2) is installed on the binocular camera base (11-1). The lens of the binocular camera (11-2) faces the direction of the electric tracked chassis (1-1) moving. Both the binocular camera (11-2) and the electric tracked chassis (1-1) are connected to the navigation controller (11-3).

5. The long-stroke apple picking robot according to claim 4, characterized in that, The sorting device (9) includes several sorting conveyor belts and collection boxes. The sorting conveyor belts are evenly divided into two groups and arranged on both sides of the detection conveyor belt (8-12-1). Each sorting conveyor belt has a collection box (10) at its end. Each sorting conveyor belt has a sorting conveyor belt servo motor installed on one side. All sorting conveyor belt servo motors are connected to controller No. 2 (6-2). Four vertical supports (9-10) are evenly distributed in two groups on the outside of several sorting conveyor belts; The Y-axis module (9-5-1) of sorting robot arm 1 is installed on two vertical supports (9-10) on one side of the inspection conveyor belt (8-12-1), and the Y-axis module (9-5-2) of sorting robot arm 2 is installed on two vertical supports (9-10) on the other side of the inspection conveyor belt (8-12-1). Both the Y-axis module (9-5-1) and the Y-axis module (9-5-2) of sorting robot arm 1 are synchronous belt linear modules. The synchronous pulley of the Y-axis module (9-5-2) is connected through coupling II (9-11). It also includes a sorting robot arm Y-axis servo motor (9-5-3). The sorting robot arm Y-axis servo motor (9-5-3) is connected to the synchronous pulley of the No. 1 sorting robot arm Y-axis module (9-5-1). The sorting robot arm Y-axis servo motor (9-5-3) drives the synchronous pulley of the No. 1 sorting robot arm Y-axis module (9-5-1) and drives the synchronous pulley of the No. 2 sorting robot arm Y-axis module (9-5-2) to rotate synchronously through coupling II (9-11). It also includes a sorting robot arm X-axis module (9-6-1), a sorting robot arm X-axis servo motor (9-6-2), and a sorting robot arm Z-axis module (9-7-1). Both the X-axis module (9-6-1) and the Z-axis module (9-7-1) are synchronous belt linear modules. The sliders of sorting robot arm Y-axis modules 1 (9-5-1) and 2 (9-5-2) are connected to the sorting robot arm X-axis module (9-6-1) via bolts. The sorting robot arm X-axis servo motor (9-6-2)... The sorting robot arm's X-axis module (9-6-1) is connected to the synchronous wheel of the sorting robot arm. The sorting robot arm's Z-axis module (9-7-1) is bolted to the slider of the sorting robot arm's X-axis module (9-6-1). The sorting robot arm's Z-axis servo motor (9-7-2) is installed at the synchronous wheel of the sorting robot arm's Z-axis module (9-7-1). The sorting robot arm's Y-axis servo motor (9-5-3), X-axis servo motor (9-6-2), and Z-axis servo motor (9-7-2) are all connected to controller 2 (6-2). The sorting robot arm's Z-axis module (9-7-1) is equipped with a flexible sorting gripper (9-8) at its end. The flexible sorting gripper has the same structure as the flexible picking gripper. An industrial camera (9-9) is installed on the side of the end of the sorting robot arm's Z-axis module (9-7-1). The industrial camera (9-9) is connected to controller 2 (6-2). The camera takes pictures of the apples after secondary grading on the inspection conveyor belt and sends the pictures to controller 2. The secondary grading apples are identified and located by a predetermined algorithm. Controller 2 drives the sorting robot arm to stack the graded apples onto the corresponding sorting conveyor belt.

6. The long-stroke apple-picking robot according to claim 5, characterized in that, The redundant robotic arm base (4-4) is connected to the slider of the X-direction linear module (3-3-1); the redundant robotic arm (4-3) is connected to the picking gripper unit (5); The depth camera (4-2) and the redundant robotic arm (4-3) are both connected to controller 1 (6-1).

7. The long-stroke apple-picking robot according to claim 6, characterized in that, The flexible picking gripper (5-2) is connected to the end of the redundant robotic arm (4-3) away from the base (4-4) of the redundant robotic arm via connector I (5-1); connector I (5-1) includes a flange (5-1-1), and two connecting protrusions (5-1-2) are provided on the edge of the flange (5-1-1). The flange (5-1-1) is sleeved on the redundant robotic arm (4-3), and the two connecting protrusions (5-1-2) are connected to the end of the flexible picking gripper (5-2); The flexible picking gripper (5-2) and the membrane pressure sensor (5-3) are both connected to controller 1 (6-1).

8. The long-stroke apple-picking robot according to claim 7, characterized in that, The fruit conveying tube unit (7) includes a funnel-shaped inlet (7-2), a telescopic hose (7-3), and a connecting pipe (7-4) connected in sequence. The upper end of the connecting pipe (7-4) is connected to the telescopic hose (7-3), and the lower end of the connecting pipe (7-4) is fixedly connected to the end cap of the detection conveyor belt (8-12-1). A pad (7-6) is provided at the outlet of the connecting pipe (7-4). It also includes a connector II (7-1). The connector II (7-1) includes a connecting rod (7-1-1) with an open retaining ring and a connecting rod (7-1-2) with a through hole connected to each other. The connecting rod (7-1-2) with the through hole is sleeved with the redundant robotic arm (4-3), and the connecting rod (7-1-1) with the open retaining ring is sleeved with the funnel-shaped inlet (7-2). The funnel-shaped inlet (7-2) is fitted with a one-way valve (7-5) at its end.

9. A long-distance apple harvesting method, characterized in that, The robot described in claim 8 is specifically used as follows: the navigation controller (11-3) controls the electric tracked chassis (1-1) to enable the long-stroke picking robot to move in the orchard. After the long-stroke picking robot moves to a designated position near the target fruit tree, the hydraulic controller (1-12) adjusts the hydraulic outriggers to keep the working platform (2) horizontal. The depth camera (4-2) collects images of the fruit tree, and the No. 1 controller (6-1) processes the images to realize the picking sequence planning and pose positioning of the target apples. The No. 1 controller (6-1) drives the picking robotic arm unit (4) and the flexible picking gripper (5-2) to pick in sequence, and performs preliminary coding and grading of the apples according to the data of the depth camera (4-2) and the thin film pressure sensor (5-3). The picked apples are transported to the detection device (8) through the fruit conveying tube unit (7). The apples are then coded and graded a second time according to the data of each appearance detection module and each internal quality detection module in the detection device (8). Finally, the apples are transported to the corresponding grade collection box (10) through the sorting device (9).

10. The long-distance apple harvesting method according to claim 9, characterized in that, Specifically, the following steps are included: Step S1: The binocular camera (11-2) takes photos of the orchard road and sends them to the navigation controller (11-3). The navigation controller (11-3) uses machine vision and deep learning autonomous navigation algorithms to send control commands to the electric tracked chassis (1-1), thereby enabling the picking robot to move in the orchard. Step S2: The long-stroke picking robot moves to a designated position near the target fruit tree. The hydraulic controller (1-12) receives data from the tilt sensor (1-11) and each pressure sensor. The hydraulic controller (1-12) sends control commands to the hydraulic pump station (1-2), the control valve group (1-3), and each two-position electro-hydraulic valve, thereby realizing the adjustment of each hydraulic outrigger and the relative horizontal control of the chassis platform (1-10) and the working platform (2). Step S3: Controller 1 (6-1) controls the movement of the extension platform (3) and the picking robotic arm unit (4), uses a depth camera (4-2) to collect images of the target fruit tree, and sends the images to Controller 1 (6-1). Using YOLOv5, all target apples are identified, and then three-dimensional point cloud processing is used to realize the picking sequence planning and pose positioning of the target apples. Step S4: Controller 1 (6-1) drives the extension platform (3) and redundant robotic arm (4-3) to the picking position according to the target apple picking order, and drives the flexible picking gripper (5-2) to wrap the target apple; Controller 1 (6-1) feeds back pressure information through the membrane pressure sensor (5-3) and adjusts the opening and closing degree of the flexible picking gripper (5-2) in real time. Step S5: Controller 1 (6-1) uses YOLOv5 to initially acquire information on the target apple's diameter, shape, color, surface defects, and hardness based on data from the depth camera (4-2) and the thin-film pressure sensor (5-3). Controller 1 (6-1) then performs initial grading and coding of the target apple based on this information, and controls the redundant robotic arm (4-3) and the flexible picking gripper (5-2) to pick the target apple. Step S6: After picking, the apples are transported through the fruit conveying tube unit (7) to the detection device (8) and the detection conveyor belt (8-12-1). Step S7: Controller No. 2 (6-2) performs secondary testing, grading and coding of apples according to the data of each appearance detection module and each internal quality detection module in the detection device (8), based on the indicators of fruit diameter, fruit shape, color, fruit surface defects, soluble solids and internal lesions; the coded apples are transported to the sorting device (9) through the detection conveyor belt (8-12-1). Step S8: Controller 2 (6-2) takes a picture based on the industrial camera (9-9), uses the YOLOv5 algorithm to identify the coded apple, and calculates its position; The No. 2 controller (6-2) drives the sorting robotic arm and the flexible sorting gripper (9-8) to transport the apples after secondary grading to the corresponding graded collection box (10); Step S9: After the target apples have been sorted, repeat steps S4-S7; Step S10: After all the target apples on the fruit tree have been harvested, repeat steps S2-S9. In step S5, the method by which controller 1 (6-1) encodes the target apple is as follows: The encoding order of the apple is xxxx-dxx-sxx-rxx-fxx-hxx, where xxxx represents the picking order of the apple, numbered 0001-9999; dxx represents the diameter of the picked apple, which is the maximum value of the apple's transverse and longitudinal diameters, and xx is the apple's diameter size in mm; sx.xx represents the shape of the picked apple, with the shape index being the ratio of the fruit's longitudinal to transverse diameters, and x.xx being the apple's shape index; rxx represents the color of the picked apple, and xx is the apple's coloring ratio; fxx represents the surface defects of the picked apple, and xx represents the number of surface damages; hxx represents the firmness of the picked apple, and xx is the apple's firmness in mm. ; In step S7, the method by which controller 2 (6-2) encodes the target apple is as follows: The secondary encoding sequence of the apple is xxxx-dxx-sxx-rxx-fxx-hxx-bxx-mx-lx, where xxxx represents the picking order of the apple, numbered 0001-9999; dxx represents the diameter of the picked apple, which is the maximum value of the apple's transverse and longitudinal diameters, and xx is the apple's diameter size in mm; sx.xx represents the shape of the picked apple, with the shape index being the ratio of the fruit's longitudinal to transverse diameters, and x.xx being the apple's shape index; rxx represents the color of the picked apple, and xx is the apple's coloring ratio; fxx represents the surface defects of the picked apple, and xx represents the number of surface damages; hxx represents the firmness of the picked apple, and xx is the apple's firmness in mm. bxx represents the soluble solids content of the harvested apples, and xx is the soluble solids content of the apples; mx represents the internal lesions of the harvested apples, x=0 represents no internal lesions, x=1 represents internal lesions; lx represents the grade of the harvested apples, x=1 represents superior grade, x=2 represents first grade, x=3 represents second grade, and x=4 represents inferior grade.

Images