A mobile picking platform based on visual recognition

Abstract

The utility model discloses a kind of mobile picking platform based on visual identification, it is related to automatic picking field, including shear type hydraulic elevator, the top of shear type hydraulic elevator is provided with protective platform, guide rail is provided on the protective platform along its length direction distribution, trolley is provided on the protective platform, the trolley is adapted with guide rail and can move along guide rail, the lower end middle part of the trolley is provided with several picking assemblies, visual acquisition module is provided on the outer side of picking assembly on the trolley, the visual acquisition module includes the identification component for identifying and marking picking object position, the picking assembly includes mechanical arm and picking unit arranged on the mechanical arm dynamic platform;The utility model can adapt to different height truss mushroom bed by the height adjusting function of shear type hydraulic elevator, solve the problem that only single-layer mushroom bed can be adapted by existing equipment, significantly expand the scope of application, improve the versatility of equipment.

Description

Technical Field

[0001] This utility model belongs to the field of automatic harvesting, specifically a mobile harvesting platform based on visual recognition. Background Technology

[0002] In the field of automated harvesting in modern agriculture, mobile platforms are the core equipment for achieving efficient crop harvesting, especially suitable for large-scale fruit and vegetable harvesting scenarios. Currently, most automated mobile harvesting platforms use a fixed-height frame structure, which can only be adapted to single-layer mushroom beds or planting areas of specific heights. Harvesting operations are completed through the cooperation of robotic arms and vision modules, which improves harvesting efficiency to some extent.

[0003] With the development of cultivation technology, truss-style multi-layer mushroom beds have been widely used due to their high space utilization, especially for the cultivation of mushrooms such as button mushrooms. However, the fixed-height harvesting platform cannot be adjusted according to the height of different layers of the mushroom bed, and can only harvest crops from a single layer. This cannot meet the harvesting needs of multi-layer truss-style mushroom beds, resulting in poor adaptability. To harvest other layers, the equipment must be moved manually or the platform must be replaced, which is not only cumbersome but also significantly reduces harvesting efficiency. This limitation in adapting to single-layer mushroom beds makes it difficult to meet the harvesting needs of multi-layer, large-scale cultivation, thus restricting the application scope of automated harvesting equipment. Utility Model Content

[0004] The purpose of this invention is to provide a mobile harvesting platform based on visual recognition to solve the problem that the harvesting platform in the prior art has a fixed height, cannot be adapted to multi-layer mushroom beds, resulting in low harvesting efficiency and difficulty in meeting the needs of large-scale planting.

[0005] To achieve the above objectives, this utility model provides the following technical solution: a mobile harvesting platform based on visual recognition, including a scissor-type hydraulic lift, wherein a protective platform is provided on the top of the scissor-type hydraulic lift, and guide rails distributed along its length are provided on the protective platform.

[0006] The protective platform is equipped with a trolley, which is adapted to the guide rail and can move along the guide rail;

[0007] The lower middle part of the trolley is provided with several picking components, and the trolley is provided with a vision acquisition module, which is located on the outside of the picking components.

[0008] The visual acquisition module includes an identification component for identifying and marking the location of the picking object, and the picking component includes a robotic arm and a picking unit set on the robotic arm's moving platform.

[0009] Preferably, the protective platform includes a tray and a fence, the fence being disposed on the edge of the tray, and the guide rails being disposed on the tray and located on three sides not adjacent to the mushroom bed.

[0010] Preferably, the trolley includes a frame and a drive wheel, the drive wheel being located at the lower end of the frame and adapted to the guide rail.

[0011] Preferably, the vision acquisition module further includes a lead screw, a movable seat, and a guide rod. The lead screw is arranged along the width direction of the vehicle frame, the movable seat is connected to the lead screw via a transmission, and the recognition component is a camera that is hinged to the movable seat.

[0012] Preferably, the guide rod passes through the movable seat and is arranged parallel to the lead screw, and a drive motor is provided on the frame, the drive motor being connected to the lead screw via a transmission.

[0013] Preferably, the harvesting unit includes a three-claw mechanical gripper, each of the three gripper segments of which is provided with a flexible clamping plate. A rotary drive is provided on the frame, and the robotic arm is mounted on the rotary drive. The robotic arm is a delta robotic arm.

[0014] Preferably, a feeding conveyor belt is provided at the end of the trolley away from the vision acquisition module, and baffles are provided on both sides of the feeding conveyor belt. Several notches and slots are provided on the baffles on the side closer to the picking component.

[0015] Preferably, the movable base is provided with an angle sensor, which is connected to the hinge axis of the camera, and an LED ring light source is provided on the outside of the camera.

[0016] Compared with the prior art, the present invention has the following beneficial effects:

[0017] 1. This utility model, through the height adjustment function of the scissor-type hydraulic lift, can be adapted to truss mushroom beds of different heights, solving the problem that existing equipment can only be adapted to single-layer mushroom beds, significantly expanding the scope of application and improving the versatility of the equipment.

[0018] 2. This utility model uses the lead screw transmission structure of the vision acquisition module to drive the camera to move along the width of the frame. Combined with the movement of the trolley along the guide rail, it can achieve full coverage shooting of the mushroom bed, reduce edge missed detection, and provide a reliable positioning basis for accurate harvesting.

[0019] 3. This utility model adopts a picking component that combines a delta robotic arm with a three-jaw mechanical gripper. The delta structure has a fast response speed and high positioning accuracy. Combined with the flexible gripper, it can ensure picking efficiency, avoid damage to the picking object, and improve the picking quality. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this utility model. For those skilled in the art, other drawings can be obtained based on these drawings.

[0021] Figure 1 A schematic diagram of the overall structure provided for an embodiment of this utility model;

[0022] Figure 2 A three-dimensional structural diagram of the protective platform, trolley, and their associated components provided in an embodiment of this utility model;

[0023] Figure 3 A three-dimensional structural diagram of the trolley and its associated components provided for an embodiment of this utility model;

[0024] Figure 4 A three-dimensional structural diagram of the harvesting component provided in an embodiment of this utility model.

[0025] In the picture:

[0026] 1. Scissor lift; 2. Protective platform; 201. Pallet; 202. Fence; 203. Guide rail; 3. Trolley; 301. Frame; 302. Drive wheel; 4. Vision acquisition module; 401. Lead screw; 402. Moving seat; 403. Camera; 5. Harvesting component; 501. Robotic arm; 502. Three-jaw mechanical gripper; 503. Flexible clamp; 6. Feeding conveyor belt; 7. Drive motor; 8. Guide rod; 9. Rotary drive component. Detailed Implementation

[0027] To enable those skilled in the art to better understand the technical solution of this utility model, the present utility model will be further described in detail below with reference to the accompanying drawings.

[0028] As attached Figure 1 To be continued Figure 4 As shown:

[0029] This utility model provides a mobile harvesting platform based on visual recognition, including a scissor lift 1. A protective platform 2 is provided on the top of the scissor lift 1, and guide rails 203 distributed along its length are provided on the protective platform 2. A trolley 3 is provided on the protective platform 2, and the trolley 3 is adapted to the guide rails 203 and can move along the guide rails 203. Several harvesting components 5 are provided in the middle of the lower end of the trolley 3. A visual acquisition module 4 is provided on the trolley 3, and the visual acquisition module 4 is located outside the harvesting components 5. The visual acquisition module 4 includes an identification component for identifying and marking the position of the harvested object. The harvesting components 5 include a robotic arm 501 and harvesting units provided on the moving platform of the robotic arm 501.

[0030] The scissor lift 1 has a rated load of 500kg and a lifting height of 0.3-5m. It is equipped with casters with braking devices at the bottom. The support plate 201 of the protective platform 2 is made of 8mm thick Q235 steel plate (2m×1.5m), with a 30cm high fence 202 (30mm×30mm square steel pipe, reinforced with fillet welds) welded to the edge. Two I-shaped guide rails 203 (1m in length, parallelism error ≤0.5mm / m) are installed along the length of the support plate 201, which are fixed every 50cm by M10 expansion bolts. The inner side of the guide rails is adapted to the drive wheel. The frame 301 of the small car 3 is constructed of 2024 aluminum alloy profiles (40mm×40mm profiles, corner bracket connection, weight ≤50kg). Four polyurethane drive wheels 302 (diameter 100mm, width 50mm, inner groove depth 15mm, fit clearance 0.5-1mm) are mounted on both sides of the lower end via bearing seats. The drive wheels are connected to servo motors (positioning accuracy ±1mm). The lead screw 401 of the vision acquisition module 4 is set along the width direction of the frame (fixed at both ends by bearing seats), with one end connected to the servo motor via a coupling. The moving seat 402 is driven by the lead screw via a threaded sleeve, and the guide rod 8 passes through the guide hole of the moving seat (parallel to the lead screw). A Pixy camera 403 (with a damped rotating shaft, adjustable angle) is hinged to the lower end of the moving seat, and a white LED ring light source (illuminance uniformity ≥85%) is installed on the outside. The robotic arm 501 of the harvesting component 5 is a delta parallel robotic arm (working space diameter 0.5m, positioning accuracy ±0.5mm), with a moving platform flange connected to a three-jaw mechanical gripper 502 (driven by a gripper servo motor); a 2mm thick flexible clamping plate 503 of polyurethane-based material (with anti-slip texture, Shore hardness 60A) is bonded to the inner side of the gripper, and a pressure sensor (range 0-10N, accuracy ±0.1N) is installed on the inner side. A food-grade feeding conveyor belt 6 (with flexible raised strips) is installed at the other end of the trolley 3, with notches and slots on both sides near the harvesting component (adapted to the gripper trajectory). A lithium battery pack (range ≥8 hours) is installed on the frame 301 to power all components.

[0031] During operation, the worker first pushes the platform to the side of the mushroom bed and locks the casters. The protective platform 2 is adjusted to the target height using a scissor-type hydraulic lift 1, and the trolley 3 moves along guide rail 203 to the work area. The drive motor 7 of the vision acquisition module 4 drives the lead screw 401 to rotate, the moving seat 402 moves along guide rod 8, and the camera 403 captures images of the mushroom bed and identifies the location of the mushroom to be harvested. The robotic arm 501 drives the three-jaw mechanical gripper 502 to the target location. A pressure sensor provides feedback on the force, and the rotary drive component 9 drives the robotic arm 501 to rotate and pull out the mushroom. The robotic arm then transfers the harvested mushroom to the unloading conveyor belt 6, which transports it to the storage device. Through visual recognition positioning, multi-dimensional mobile coverage, and flexible gripping, efficient and precise harvesting is achieved, adaptable to multi-layer mushroom beds.

[0032] As attached Figure 1 As shown: In one embodiment of the present invention, the protective platform 2 includes a tray 201 and a fence 202. The fence 202 is disposed on the edge of the tray 201, and the guide rail 203 is disposed on the tray 201 and located on three sides that are not adjacent to the mushroom bed.

[0033] The edge of the tray 201 is fully welded to the fence 202 (weld leg height 5mm), and the top of the fence 202 is rounded (radius 5mm) to prevent bumps. The guide rail 203 forms a closed track on the three non-adjacent edges of the tray and the mushroom bed, and the guide rail interface is smoothly transitioned (height difference ≤0.2mm).

[0034] During operation, the trolley 3 moves along the closed guide rail to cover the entire area of ​​the pallet, eliminating blind spots; the guardrail 202 prevents platform components from falling, protecting operational safety. The three-sided guide rail design expands the operating range, while the guardrail enhances equipment safety.

[0035] As attached Figure 1 To be continued Figure 3 As shown: In one embodiment of the present invention, the trolley 3 includes a frame 301 and a drive wheel 302. The drive wheel 302 is disposed at the lower end of the frame 301 and is adapted to the guide rail 203.

[0036] The bolt preload torque at the connection of the 301 frame profile is 25-30 N·m to ensure structural rigidity. The rubber rim of the 302 drive wheel has a hardness of 65 Shore A and a wear life of ≥1000 km. The drive shaft is connected to the servo motor via a reducer (reduction ratio 10:1).

[0037] During operation, the servo motor first drives the drive wheel 302 to move along the guide rail 203, and the frame 301 stably supports all components; the groove of the drive wheel fits tightly with the guide rail, without lateral swaying. Through precise driving and wheel-rail adaptation, the trolley achieves high-precision movement and stable load-bearing.

[0038] As attached Figure 2 To be continued Figure 3 As shown: In one embodiment of this utility model, the visual acquisition module 4 further includes a lead screw 401, a movable seat 402, and a guide rod 8. The lead screw 401 is arranged along the width direction of the frame 301, and the movable seat 402 is throttle-connected to the lead screw 401. The recognition component is a camera 403 and is hinged to the movable seat 402. The guide rod 8 passes through the movable seat 402 and is arranged parallel to the lead screw 401. A drive motor 7 is provided on the frame 301, and the drive motor 7 is throttle-connected to the lead screw 401.

[0039] The lead screw 401 is a high-precision ball screw (5mm lead, positioning accuracy ±0.01mm), and the clearance between the threaded sleeve of the moving seat 402 and the lead screw is ≤0.02mm. The guide rod 8 is made of 45# steel with chrome plating (12mm diameter, straightness error ≤0.05mm / m), and the clearance between it and the guide hole of the moving seat is 0.05-0.1mm. The drive motor 7 is equipped with an encoder (1000 lines resolution) to achieve closed-loop position control.

[0040] During operation, the drive motor 7 first rotates the lead screw 401, and the moving seat 402 moves smoothly along the guide rod 8. The camera 403 follows the moving seat to complete image acquisition in the width direction of the mushroom bed. The hinged structure allows for adjustment of the camera angle, ensuring full coverage of the shooting range. Through lead screw transmission and guide design, visual acquisition without blind spots and accurate positioning are achieved.

[0041] As attached Figure 2 To be continued Figure 4 As shown: In one embodiment of this utility model, the harvesting unit includes a three-jaw mechanical gripper 502, each of the three gripper segments of which is provided with a flexible clamping plate 503. A rotary drive component 9 is provided on the frame 301, and a robotic arm 501 is mounted on the rotary drive component 9. The robotic arm 501 is a delta robotic arm. A pressure sensor is provided on the inner side of the gripper of the three-jaw mechanical gripper 502, and the flexible clamping plate 503 is made of polyurethane-based material with anti-slip texture on its surface.

[0042] The three connecting rods of the Delta robotic arm are made of carbon fiber (10mm diameter, 200mm length, weight ≤500g), and the dynamic platform response speed is ≥1m / s. The three-jaw mechanical gripper 502 has an opening angle of 0-90° and a closing time ≤0.5s; the flexible clamping plate 503 is bonded to the gripper with a polyurethane-based adhesive (peel strength ≥5N / cm). The pressure sensor is a thin-film pressure sensor (response time ≤10ms), and the signal is transmitted to the main control board via an amplification circuit (sampling frequency 100Hz). The flexible clamping plate 503 has a 0.5mm deep mesh anti-slip texture with a friction coefficient ≥0.8. The rotary drive component 9 can be a stepper motor or a rotary cylinder, and its output end is connected to the stationary platform of the Delta robotic arm via a connecting flange.

[0043] During operation, the delta robotic arm first rapidly moves to the visually marked position. The three-jaw mechanical gripper 502 opens and then closes, while the flexible clamping plate 503 flexibly holds the object to be picked. Once successfully gripped, the rotary drive component 9 controls the delta robotic arm to rotate horizontally, which in turn causes the three-jaw mechanical gripper 502 to twist horizontally. Simultaneously, the servo drive linkage on the delta robotic arm rotates upward, causing the three-jaw mechanical gripper 502 to lift vertically. A pressure sensor monitors the clamping force to prevent damage from excessive force. Through the cooperation of the high-speed robotic arm and the flexible gripper, efficient and damage-free picking is achieved.

[0044] As attached Figure 2 To be continued Figure 3 As shown: In one embodiment of this utility model, a feeding conveyor belt 6 is provided at the end of the trolley 3 away from the visual acquisition module 4. Baffles are provided on both sides of the feeding conveyor belt 6, and several notches are provided on the baffle on the side closer to the picking component 5.

[0045] The feeding conveyor belt 6 has a belt thickness of 2mm (food-grade PU material) and an adjustable conveying speed of 0.1-0.5m / s; the baffle height is 10cm (transparent acrylic material), and the notch size is adapted to the maximum opening state of the grippers (gap ≥5mm).

[0046] During operation, the picking component first transfers the picked object to the unloading conveyor belt 6, which then transports the object to the end. Baffles prevent the object from falling, and notched slots avoid the gripper's unloading action. Through directional conveying and structural avoidance, automatic transfer after picking is achieved.

[0047] In one embodiment of this utility model, an angle sensor is provided on the movable base 402, the angle sensor is connected to the hinge axis of the camera 403, and an LED ring light source is provided on the outside of the camera 403.

[0048] The angle sensor is an incremental encoder (1024 lines resolution, ±0.1° angle accuracy), providing real-time feedback on the rotation angle of the camera's hinge axis. The LED ring light source has a power of 10W, a color temperature of 5500K, and a beam angle of 120°, with brightness adjustable in 3 levels via a dimming module.

[0049] During operation, the angle sensor first monitors the shooting angle of camera 403 and feeds it back to the control system to ensure accurate image acquisition angle; the LED light source provides supplementary lighting in dim environments, improving image contrast. Through closed-loop angle control and precise supplementary lighting, the reliability of visual recognition is guaranteed.

[0050] Working Principle: During operation, the operator first adjusts the protective platform 2 to the same height as the target mushroom bed using the scissor-type hydraulic lift 1, and sets the harvesting parameters using the control system. The trolley 3 moves along the guide rail 203 to the initial harvesting area. The lead screw 401 of the vision acquisition module 4 rotates, driving the moving seat 402 to move the camera 403 along the width of the frame 301. Simultaneously, the camera 403 adjusts its shooting angle and, in conjunction with the LED light source, captures images of the mushroom bed in sections. The image data is transmitted to the control system, which uses the OpenCV algorithm to identify the location, maturity, and quality of the button mushrooms, generating harvesting coordinates. The robotic arm 501 drives the three-claw mechanical gripper 502 to move above the target button mushroom, adjusts its angle based on the positioning information, and then descends. The three-claw mechanical gripper 502 opens and closes, and the flexible clamping plate 503 flexibly clamps the harvested object. The pressure sensor provides real-time feedback on the force, and the clamping stops when the set value is reached. Once the mushroom is successfully grasped, the rotary drive 9 controls the delta robotic arm to rotate horizontally, which in turn drives the three-jaw mechanical gripper 502 to twist horizontally. Simultaneously, the servo drive linkage on the delta robotic arm rotates upward, which in turn drives the three-jaw mechanical gripper 502 to lift vertically. The robotic arm 501 is lifted and moved to the notch groove of the feeding conveyor belt 6. The gripper opens and places the mushroom on the conveyor belt. The conveyor belt transports the mushroom to the storage device, and the counting sensor completes the counting. The trolley 3 moves to the next picking area, and the above process is repeated until the current layer is picked. The scissor lift 1 adjusts its height to start the next layer picking operation.

[0051] The foregoing description only illustrates certain exemplary embodiments of the present invention. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A mobile harvesting platform based on visual recognition, characterized in that: The scissor lift (1) is provided with a protective platform (2) on its top, and the protective platform (2) is provided with guide rails (203) distributed along its length. The protective platform (2) is equipped with a trolley (3), which is adapted to the guide rail (203) and can move along the guide rail (203); The lower middle part of the trolley (3) is provided with several picking components (5), and the trolley (3) is provided with a visual acquisition module (4), and the visual acquisition module (4) is located on the outside of the picking components (5). The visual acquisition module (4) includes an identification component for identifying and marking the location of the picking object, and the picking component (5) includes a robotic arm (501) and a picking unit set on the moving platform of the robotic arm (501).

2. The mobile harvesting platform based on visual recognition according to claim 1, characterized in that: The protective platform (2) includes a tray (201) and a fence (202). The fence (202) is set on the edge of the tray (201), and the guide rail (203) is set on the tray (201) and located on three sides that are not adjacent to the mushroom bed.

3. The mobile harvesting platform based on visual recognition according to claim 1, characterized in that: The trolley (3) includes a frame (301) and a drive wheel (302), the drive wheel (302) being located at the lower end of the frame (301) and adapted to the guide rail (203).

4. The mobile harvesting platform based on visual recognition according to claim 1, characterized in that: The visual acquisition module (4) also includes a lead screw (401), a movable seat (402) and a guide rod (8). The lead screw (401) is arranged along the width direction of the frame (301). The movable seat (402) is connected to the lead screw (401) in a transmission manner. The recognition component is a camera (403) and is hinged on the movable seat (402).

5. A mobile harvesting platform based on visual recognition according to claim 4, characterized in that: The guide rod (8) passes through the movable seat (402) and is arranged parallel to the lead screw (401). The frame (301) is equipped with a drive motor (7), which is connected to the lead screw (401) in a transmission.

6. A mobile harvesting platform based on visual recognition according to claim 5, characterized in that: The picking unit includes a three-claw mechanical gripper (502), and each of the three gripper segments of the three-claw mechanical gripper (502) is provided with a flexible clamping plate (503). A rotary drive (9) is provided on the frame (301), and the robotic arm (501) is mounted on the rotary drive (9). The robotic arm (501) is a delta robotic arm.

7. A mobile harvesting platform based on visual recognition according to claim 4, characterized in that: The trolley (3) is equipped with a feeding conveyor belt (6) at the end away from the visual acquisition module (4). Both sides of the feeding conveyor belt (6) are equipped with baffles, and the baffle near the picking component (5) has several notches and slots.

8. A mobile harvesting platform based on visual recognition according to claim 5, characterized in that: An angle sensor is provided on the movable base (402), and the angle sensor is connected to the hinge axis of the camera (403). An LED ring light source is provided on the outside of the camera (403).

Images