A multi-vision perception right-angle hybrid multi-arm robot cotton harvesting device and method
By using a right-angle hybrid multi-arm robot device and an improved YOLOV8 model, the problem of low picking efficiency of existing six-degree-of-freedom robots has been solved, achieving efficient and precise cotton picking, reducing costs and improving picking efficiency and accuracy.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- XINJIANG UNIVERSITY
- Filing Date
- 2024-07-30
- Publication Date
- 2026-05-01
AI Technical Summary
Existing six-degree-of-freedom robotic cotton picking devices suffer from problems such as high cost, low efficiency, insufficient suction force of the end mechanism, and failure to consider the picking range and posture angle calculation of cotton plants, resulting in insufficient picking efficiency and accuracy.
A right-angle hybrid multi-arm robot device is adopted, which combines vision sensors and deep learning models. It achieves rapid motion control through a three-axis linear guide rail in a right-angle coordinate system. The harvesting range is designed in conjunction with agronomic planting methods. Bionic grippers and an improved YOLOV8 model are used for cotton recognition and posture control, thereby improving harvesting efficiency and accuracy.
It has improved cotton picking efficiency, shortened picking time to 3-4 seconds, increased picking rate to 60-80%, reduced picking costs, and can effectively identify and handle cotton in different postures, thus improving picking accuracy.
Smart Images

Figure CN118830404B_ABST
Abstract
Description
A multi-vision perception right-angle hybrid multi-arm robot cotton harvesting device and method Technical Field
[0001] This invention relates to the field of agricultural robot technology, specifically to a multi-vision perception right-angle hybrid multi-arm robot cotton harvesting device and method. Background Technology
[0002] Long-staple cotton is a premium cotton variety that requires manual harvesting, which is costly and labor-intensive. While ordinary cotton can be harvested using cotton harvesters, these machines mix a small amount of leaf stalks, cotton bolls, and cotton fibers together during harvesting. This mixture is then separated, but it cannot separate dry leaf dust particles and finely broken cotton bolls, resulting in higher impurity levels and lower post-harvest quality. This method is unsuitable for high-end and long-staple cotton. Furthermore, before using cotton harvesters, manual harvesting is necessary at field edges and corners. Additionally, manual harvesting is required in areas where a 15-18 meter wide path for the harvester is needed. Therefore, manual harvesting is also necessary in some areas where cotton harvesters operate. Cotton plants typically grow to a height of 0.8-1.3 meters. Wide and narrow row spacing is commonly used, with a wide spacing of approximately 100-120 cm and a row spacing of approximately 40-50 cm. Cotton fibers generally face the sun, with the majority (60-80%) facing upwards. Picking is easier when the fibers are pointed downwards. However, a small number of cotton fibers have a larger angle of inclination relative to their axis, requiring a flexible multi-degree-of-freedom robot for harvesting. A six-degree-of-freedom serial robot can achieve harvesting from cotton fibers at different angles. Since a single harvesting cycle for a six-degree-of-freedom serial robot takes at least 8 seconds even in fast mode, and typically requires visual recognition and obstacle avoidance control, the total time required is 12-18 seconds.
[0003] Currently, there is an existing invention patent for a cotton harvesting robot using a six-degree-of-freedom serial robotic arm, ZL202110576476.X, which is costly and inefficient. Several problems remain: 1. It does not consider the combination structure of the Cartesian coordinate robotic arm and the six-degree-of-freedom design, nor its control methods and devices; 2. The end effector uses fingertip suction cups to simulate manual pulling force, but the suction is insufficient due to the small size of the fingertips; 3. This patent does not provide effective calculations or formulas for the harvesting range of the cotton plant; 4. Because the posture angle of a small number of cotton fluffs is large and their axis is relatively large, and some flower axes have a certain angle, the six-degree-of-freedom robot needs to identify and locate the posture angle of the cotton fluffs and determine their center during harvesting. This patent does not consider the calculation of posture space parameters for cotton fluffs with different orientations, nor does it propose a method for calculating the center of the cotton fluffs and the radial axis, making it difficult to obtain precise posture control of its robotic arm.
[0004] For the reasons mentioned above, the low efficiency and difficulty in controlling the picking, positioning, and attitude of six-degree-of-freedom robots make it difficult to apply six-degree-of-freedom robots in current applications. Summary of the Invention
[0005] The purpose of this invention is to provide a cotton harvesting device and method using a right-angle hybrid multi-arm robot with multiple visual perceptions, in order to solve the problems mentioned in the background art.
[0006] To achieve the above objectives, the present invention provides the following technical solution: A track is included, a mobile chassis is disposed between two tracks, a vehicle body support is disposed on the top of the mobile chassis, multiple mounting brackets are distributed on the top of the vehicle body support, a robotic arm mounting seat for mounting a six-degree-of-freedom robotic arm is connected to the inner wall of the mounting bracket, a control box is disposed above the robotic arm mounting seat, and a connecting seat is disposed on the top of the mounting bracket; an X-axis slide rail is disposed on the top of the connecting seat, an X-axis motor is disposed at one end of the X-axis slide rail, two X-axis slide rails are connected to a right-angle robotic arm support at their opposite ends, an X-axis slide plate is movably disposed above the support, an X-axis connecting seat is connected to the top of the X-axis slide plate, a Y-axis slide rail is connected to the side wall of the X-axis connecting seat, a Y-axis motor is disposed at one end of the Y-axis slide rail, a Y-axis slide plate is movably disposed on the side wall of the Y-axis motor, and a sleeve is connected to the side wall of the Y-axis slide plate.
[0007] Preferably, the tail end of the right-angle and six-degree-of-freedom robotic arm end effector, the mobile chassis and the front end of the vehicle body support are all provided with visual communication components, the end of the X-axis connecting seat is provided with a D radar module, the end of the sleeve is provided with a connecting block, and the side wall of the connecting block is provided with a Beidou module.
[0008] Preferably, the other end of the sleeve is connected to a suction device, the bottom of the suction device is provided with a cylinder, the bottom of the cylinder is provided with a push rod, one end of the push rod is fixedly connected to a lower fixed plate, and one end of the cylinder is fixedly provided with an upper fixed plate; the outer ring of the upper fixed plate is connected to a connecting frame, the top of the connecting frame is provided with an adjusting motor, the output end of the adjusting motor is connected to a first support rod, the bottom of the first support rod is fixedly provided with a strut, one end of the strut is rotatably connected to a first U-shaped frame, the outer ring of the lower fixed plate is connected to a second support rod, the bottom of the second support rod is provided with a shaft, the outer ring of the shaft is rotatably connected to a second U-shaped frame, one end of the second U-shaped frame is rotatably connected to a pull rod, one end of the pull rod is rotatably connected to a third U-shaped frame, one end of the third U-shaped frame is connected to a telescopic rod, one end of the telescopic rod is provided with a gripper; a limit frame is fixedly provided above the vehicle body bracket, a receiving box is movably provided in the inner cavity of the limit frame, and a handle is fixedly connected to the side wall of the receiving box.
[0009] Preferably, the top two sides of the X-axis slide rail are provided with a slide frame, the outer ring of the slide frame is provided with a curtain, the inner cavity of the curtain is provided with a plurality of through holes, the inner cavity of the through holes is provided with a protective ring, one end of the curtain is provided with a connecting plate, and one end of the connecting plate is fixedly connected to the inner wall of the X-axis slide plate.
[0010] Compared with the prior art, the beneficial effects of the present invention are:
[0011] This multi-vision perception rectangular hybrid multi-arm robot cotton harvesting device and method uses a three-axis linear guide rail with a domestically produced integrated drive and control system to achieve the motion control of the robotic arm moving in rectangular coordinates. Because the three-degree-of-freedom rectangular coordinate robotic arm is fast and accurate, it can complete a cycle in 3-4 seconds, which is 2-3 times faster than the commonly used six-degree-of-freedom serial harvesting robot. It can quickly harvest 60-80% of the cotton fibers with the cotton fibrous axis facing upwards. The remaining small amount of cotton fibers with larger fibrous axis posture angles can be harvested by a flexible six-degree-of-freedom robot, improving efficiency and harvesting rate. At the same time, the combination structure of the rectangular coordinate robotic arm and the serial robotic arm can reduce costs and expand the harvesting operation space.
[0012] This multi-vision perception right-angle hybrid multi-arm robot cotton harvesting device and method, by combining agronomic planting methods to design and calculate the harvesting range, can derive a formula for calculating the harvesting area of the right-angle robotic arm, providing calculation parameters for control, robotic arm structure design, and visual detection range.
[0013] This multi-vision perception right-angle hybrid multi-arm robot cotton harvesting device and method integrates deep learning models, color models and image geometry calculations to locate the cotton fluff posture and its geometric center, and expands the field of view according to different spacing to improve recognition accuracy. Attached Figure Description
[0014] Figure 1 is a schematic diagram of the structure of the present invention;
[0015] Figure 2 is a top view of the present invention;
[0016] Figure 3 is a schematic diagram of the curtain structure of the present invention;
[0017] Figure 4 is a schematic diagram of the finger clamping structure of the present invention;
[0018] Figure 5 is a bottom view of the structure of Figure 4 of the present invention;
[0019] Figure 6 is a schematic diagram of the YOLOV8 model structure of the present invention;
[0020] Figure 7 is a schematic diagram of the principle of solving for the flower core and the calculation method of the present invention.
[0021] In the diagram: 1. Track; 2. Mobile chassis; 3. Vehicle body support; 4. Mounting bracket; 5. Robotic arm mounting base; 6. Control box; 7. Connecting seat; 8. X-axis slide rail; 9. X-axis motor; 10. Right-angle robotic arm support; 11. Carriage; 12. Curtain; 13. Through hole; 14. Protective ring; 15. Connecting plate; 16. X-axis connecting seat; 17. X-axis sliding plate; 18. Y-axis slide rail; 19. Y-axis motor; 20. Y-axis sliding plate; 21. Sleeve; 22. Connecting block; 23. Beidou module 24. Visual communication component; 25. 3D radar module; 27. Air suction device; 28. Cylinder; 29. Push rod; 30. Lower fixed plate; 31. Upper fixed plate; 32. Connecting frame; 33. Adjusting motor; 34. First support rod; 35. Support rod; 36. First U-shaped frame; 37. Second support rod; 38. Shaft rod; 39. Second U-shaped frame; 40. Pull rod; 41. Third U-shaped frame; 42. Telescopic rod; 43. Finger clamp; 44. Limiting frame; 45. Receiving box; 46. Handle. Detailed Implementation
[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0023] Example 1: Technical solution of a multi-vision perception right-angle hybrid multi-arm robot cotton harvesting device and method according to the present invention:
[0024] As shown in Figures 1, 2, 4, 5, 6, and 7, the system includes tracks 1, a mobile chassis 2 positioned between two tracks 1, a vehicle frame 3 on top of the mobile chassis 2, multiple mounting brackets 4 distributed on the top of the vehicle frame 3, and a robotic arm mounting base 5 for mounting a six-degree-of-freedom robotic arm connected to the inner wall of the mounting bracket 4. A control box 6 is positioned above the robotic arm mounting base 5, and a connecting seat 7 is positioned on the top of the mounting bracket 4. An X-axis slide rail 8 is positioned on the top of the connecting seat 7. An X-axis motor 9 is installed at one end of the slide rail 8. The two X-axis slide rails 8 and the opposite ends of the X-axis motor 9 are connected to a right-angle robotic arm bracket 10. An X-axis slide plate 17 is movably installed above it. An X-axis connecting seat 16 is connected to the top of the X-axis slide plate 17. A Y-axis slide rail 18 is connected to the side wall of the X-axis connecting seat 16. A Y-axis motor 19 is installed at one end of the Y-axis slide rail 18. A Y-axis slide plate 20 is movably installed on the side wall of the Y-axis motor 19. A sleeve 21 is connected to the side wall of the Y-axis slide plate 20.
[0025] The end effector of the right-angle and six-degree-of-freedom robotic arm is equipped with a visual communication component 24 at the tail, the mobile chassis 2 and the front of the vehicle body support 3. The end of the X-axis connecting seat 16 is equipped with a 3D radar module 25, the end of the sleeve 21 is equipped with a connecting block 22, and the side wall of the connecting block 22 is equipped with a Beidou module 23.
[0026] As shown in Figure 1, the three-axis rectangular coordinate robotic arm includes three linear guides (X, Y, and Z) and motors. There are two sets of rectangular coordinate robotic arms, enabling the device to simultaneously harvest cotton from both sides within a 120cm row spacing in the field, resulting in high efficiency. However, if too many cotton plants are covered (more than two plants), the rigidity of the X-axis slide rail 8, being a cantilever beam, is affected, causing vibration during harvesting. Referring to Figure 1, a parallel X-axis slide rail 8 (referred to as a double X-axis) is added, and the Y-axis slide rail 18 is installed in the middle of the parallel X-axis slide rail 8. This improves rigidity and stability. The Y-axis slide rail 18 can slide axially along the double X-axis slide rail 8. The end mechanism (shown in Figures 4 and 5) is installed on the Y-axis, allowing it to move along the Y-axis. The end mechanism shaft can move up and down within the sleeve 21, thereby adjusting the harvesting height.
[0027] The right-angle harvesting robotic arm of this structure has two decision-making and control methods:
[0028] The first scenario is that if the mobile chassis 2 moves while picking, the end mechanism can be fixed in a fixed position on the y-axis. The y-axis moves on the double X-axis, and the end mechanism can move up and down within the Z-axis bushing. In this case, since the mobile chassis 2 can be considered as one degree of freedom in the Y-direction, only two degrees of freedom of the rectangular coordinate arm participate in the movement.
[0029] The second method involves the mobile chassis 2 stopping, the right-angle robotic arm completing the picking of one cotton plant, and then the mobile chassis 2 moving forward a certain distance to pick the second cotton plant. In this case, the end axis is not fixed on the y-axis and can slide up and down on the y-axis to complete the picking of one cotton plant. After that, the robot moves forward (in the y-axis direction). The calculation method for the area covered by these movements is shown in the following formula:
[0030] Based on the planting height of Xinjiang long-staple cotton, the vertical movement range of the robotic arm is approximately 0.8 to 1.2 meters.
[0031] Maximum picking height: ZHmax = H1max - H0 = 1.2 - 0.3 = 0.9m
[0032] Minimum picking height: ZHmin = H1max - H0 = 0.8 - 0.3 = 0.5m
[0033] That is, the picking height range from ZHmin to ZHmax is 0.5-0.9 meters or 500-900 mm.
[0034] In the formula, H1max and H1min represent the highest and lowest plant heights of cotton, respectively, H0 is the height from the bottom of the cotton stalk to the lowest cotton boll (empirical data obtained from big data statistics), H0 = 1 / 3 H1 cotton plant height, and H1 is the cotton boll height.
[0035] Typically, two cotton plants are planted in a narrow row, with a width ranging from 230 to 260 mm.
[0036] The distance in the XY plane is related to the plant spacing and cotton growth. Assuming a wide-narrow row arrangement, with a maximum plant spacing of 260mm, and using this size as a square, it basically covers the growth range of each cotton plant.
[0037] In summary, the maximum harvesting area for XZ is M, and the maximum harvesting area for XY is N. Therefore:
[0038] Mmax = Hmax * 2 * Xmax = 900 * 520
[0039] Nmax = 2Xmax * Xmax = 520 * 260;
[0040] The dimensions above are the design parameters for the picking range of the robot and its end effector for each cotton plant;
[0041] The remaining small portion of cotton fibers that were not picked, with their axis pointing downwards or having a large angle with the Z-axis, were picked by a flexible six-degree-of-freedom serial robotic arm.
[0042] Cotton identification method: First, collect big data on cotton, including multi-target image samples of cotton fibers, cotton bolls, the intersection of cotton bolls and their center (which is also the center of the petals), axial diameter, etc. The dataset should contain no less than 5,000-10,000 images, each containing different cotton varieties.
[0043] When the YOLOv8 deep learning network model detects multiple targets in cotton, it preprocesses the image (such as resizing and cropping) to adapt to the model's input size requirements. Then, it extracts target information through the backbone network to adapt to the model's input size requirements. Next, it enhances the features through the feature fusion network (Neck). Finally, it detects and classifies multiple targets in the head part. This model can quickly identify cotton fibers, husks, and center points, but its detection accuracy is not high for small targets such as the relatively small husks and center points in the middle of the cotton fibers.
[0044] To improve the accuracy of small targets in multi-objective cotton analysis, this invention proposes an improved YOLOv8 network model (see Figure 6). The main improvements are as follows:
[0045] 1. Replace the original PAN structure of the YOLOv8 model with a pyramid structure. First, fuse cotton husk features to improve the model's accuracy in recognizing small targets. A progressive approach is adopted, fusing lower-level features first, then higher-level features. Therefore, the original PAN structure of the YOLOv8 model is replaced with a pyramid model (AFPN) that features progressive fusion, thus improving the detection accuracy of small targets.
[0046] 2. In the multi-layer feature fusion process of the pyramid AFPN, an adaptive spatial feature fusion structure (ASFF) is introduced to assign spatial weights to features of different layers. This enhances the feature representation of key layers and improves the feature fusion effect of the cotton lint center, the cotton boll and its center. By fusing feature information from three layers, the recognition accuracy of small targets is improved. The feature fusion calculation is shown in the following formula:
[0047]
[0048] In the formula This represents the result of the fused feature vector calculation. Let represent the feature vector at (x, y) from the nth to the 1st layer. , , , where represents the spatial weight of each level. The above formula contains: ,
[0049] Adding a CBAM (Concentration-Based Attention) module to the network model makes it pay more attention to key information such as cotton husks, confluence centers, and small diameter rods within the cotton fibers in the image input when extracting multiple features of cotton.
[0050] 3. Add a detection head to the YOLOv8 model. In the original YOLOv8 model, there are three detection heads to detect multi-target features at different scales during cotton harvesting. When there are small targets in the visually obtained image, defects such as missed detection often occur. Therefore, this invention adds a detection head to the original three detection heads, that is, adds a small detection head layer, so there are four detection heads. This increases the sensitivity and detection accuracy of the model for small targets.
[0051] As shown in Figure 7, the cotton fiber axis center parameter is a key geometric parameter for the cotton picking robot's picking posture and pose control. Traditional methods do not consider solving for the cotton fiber axis center point and its normal parameters. This invention proposes a combination of geometric and deep learning methods to identify the cotton fiber and use geometric calculation methods to solve for the flower stamen axis.
[0052] After classifying and recognizing multiple targets in cotton using an improved YOLOv8 model, segmentation methods were then used to accurately segment cotton fibers, hulls, and axial diameter. Following image segmentation, a circle was fitted to approximate a circle around the outer edge of the white cotton fibers using Hough transform. The center point of the fitted circle was then used to extract the largest connected region of the petal's main contour as the bud contour, and the center parameters of the cotton fibers were calculated.
[0053] A two-dimensional image of the bud region is used to establish a two-dimensional Cartesian coordinate system. After image processing, the cotton husk region inside the cotton fibers is extracted, resulting in cotton petal partitions Qx. Assuming the partition angle is... Then we have: = The centroid and radius of the upper part were calculated separately.
[0054] By selecting the upper centroid Rc1(xc1, yc1) and the lower centroid Rc2(xc2, yc2) and connecting Rc1 and Rc2, the problem of extracting the radial axis information of the lower part of the cotton is transformed into the problem of extracting the binary mask of the flower radial axis. The cotton radial axis is obtained. By taking the point of the radial axis close to the lower part of the flower and connecting Rc1 and Rc2, the angle, parameters and attitude of the axis can be obtained, which provides information for the attitude control of the robotic arm and its end effector.
[0055] A visual communication component 24, including a depth camera, is installed on the vehicle body support 3 to assist in the initial judgment of the road and cotton. If cotton is found, the picking control is activated. At the same time, it judges and identifies high-level obstacles, providing pre-decision for the movement and turning of the mobile chassis 2. A visual communication component 24 is installed at the front end of the mobile chassis 2 for accurate identification of the road ahead and obstacle identification, trajectory planning and decision-making.
[0056] Example 2
[0057] As shown in Figures 4 and 5, the other end of the sleeve 21 is connected to a suction device 27. A cylinder 28 is located at the bottom of the suction device 27, and a push rod 29 is located at the bottom of the cylinder 28. One end of the push rod 29 is fixedly connected to a lower fixed plate 30, and one end of the cylinder 28 is fixedly connected to an upper fixed plate 31. A connecting frame 32 is connected to the outer ring of the upper fixed plate 31. An adjusting motor 33 is located at the top of the connecting frame 32. The output end of the adjusting motor 33 is connected to a first support rod 34. A support rod 35 is fixedly located at the bottom of the first support rod 34, and one end of the support rod 35 is rotatably connected to a first U-shaped... The frame 36 has a second support rod 37 connected to the outer ring of the lower fixed plate 30. The bottom of the second support rod 37 is provided with a shaft 38. The outer ring of the shaft 38 is rotatably connected to a second U-shaped frame 39. One end of the second U-shaped frame 39 is rotatably connected to a pull rod 40. One end of the pull rod 40 is rotatably connected to a third U-shaped frame 41. One end of the third U-shaped frame 41 is connected to a telescopic rod 42. One end of the telescopic rod 42 is provided with a clamping finger 43. A limit frame 44 is fixedly provided above the vehicle body support 3. A receiving box 45 is movably provided in the inner cavity of the limit frame 44. A handle 46 is fixedly connected to the side wall of the receiving box 45.
[0058] In this embodiment, since the main cotton varieties have 4 petals, and some have 3 and 5, based on the largest petal, 5 bionic clamping fingers 43 are designed and extended and retracted by telescopic rod 42. Each bionic semi-flexible clamping finger 43 can deform according to the number of petals of the variety, retaining the same number of clamping fingers as the number of petals, and the average angle of each clamping finger = 360 degrees / clamping finger.
[0059] When the bionic finger has 5 fingers, the angle between each fingertip is 72 degrees, and the fingertips are evenly distributed. Similarly, when there are 4 fingers, the angle between each fingertip is 90 degrees. When one petal is missing, there are 4 fingers. One of the gripping fingers 43 is retracted via the telescopic rod 43, and the first support rod 34 is rotated by the adjusting motor 33. At this time, the first support rod 34 drives the first U-shaped frame 36 to rotate via the support rod 35. The first U-shaped frame 36 then drives the gripping finger 43 to make a circular motion around the axis 38, thereby deforming the remaining 4 fingers into an equiangular distribution. The same applies to 3 fingers. In use, the cylinder 28 drives... Push rod 29 moves upward, and push rod 29 drives the second support rod 37 to move upward through the lower fixed plate 30. At this time, the second U-shaped rod 37 drives the second U-shaped frame 39 to move upward through the shaft 38. The second U-shaped frame 39 then drives one end of the pull rod 40 to move upward. Subsequently, the other end of the pull rod 40 pulls one end of the third U-shaped frame 41 upward. At this time, the third U-shaped frame 41 makes a circular motion with the end of the first U-shaped frame 36 as the center point, so as to drive the clamping finger 43 to clamp the cotton. This method is simple, completed by control, does not require manual replacement, and is highly practical.
[0060] Example 3
[0061] As shown in Figures 1, 2 and 3, the top two sides of the X-axis slide rail 8 are provided with slide frames 11, the outer ring of the slide frame 11 is fitted with a curtain 12, the inner cavity of the curtain 12 is provided with several through holes 13, the inner cavity of the through holes 13 is fitted with a protective ring 14, one end of the curtain 12 is provided with a connecting plate 15, and one end of the connecting plate 15 is fixedly connected to the inner wall of the X-axis slide plate 17.
[0062] In this embodiment, since the cotton planting environment is dusty and cotton plant leaves and other materials are easily dropped when the device is moved, a curtain 12 is designed as a protective structure to prevent these foreign objects from falling onto the X-axis slide rail 8 and affecting the movement of the X-axis slide plate 17 on the X-axis slide rail 8. When the X-axis slide plate 17 moves, the connecting plate 15 connected to its side wall simultaneously drives the curtain 12 to unfold, covering the blank area on the X-axis slide rail 8. The slide frame 11 is used to limit the curtain 12. The curtain 12 slides on the surface of the slide frame 11 through the through hole 13. A metal protective ring 14 is provided in the through hole 13 to prevent the curtain 12 from being damaged during the pushing and pulling process of the X-axis slide plate 17.
[0063] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A multi-vision perception right-angle hybrid multi-arm robot cotton harvesting device, comprising tracks (1), characterized in that: A mobile chassis (2) is provided between the two tracks (1). A vehicle frame (3) is provided on the top of the mobile chassis (2). Multiple mounting brackets (4) are distributed on the top of the vehicle frame (3). A robotic arm mounting seat (5) for mounting a six-degree-of-freedom robotic arm is connected to the inner wall of the mounting bracket (4). A control box (6) is provided above the robotic arm mounting seat (5). A connecting seat (7) is provided on the top of the mounting bracket (4). An X-axis slide rail (8) is provided on the top of the connecting seat (7). An X-axis motor (9) is provided at one end of the X-axis slide rail (8). The two X-axis slide rails (8) are opposite to the X-axis motor (9). One end of the device is connected to a right-angle robotic arm bracket (10), and an X-axis slide plate (17) is movably mounted on top of it. The top of the X-axis slide plate (17) is connected to an X-axis connecting seat (16), and the side wall of the X-axis connecting seat (16) is connected to a Y-axis slide rail (18). One end of the Y-axis slide rail (18) is equipped with a Y-axis motor (19), and the side wall of the Y-axis motor (19) is movably mounted with a Y-axis slide plate (20). The side wall of the Y-axis slide plate (20) is connected to a sleeve (21). The other end of the sleeve (21) is connected to an air suction device (27), and the bottom of the air suction device (27) is equipped with a cylinder (28). A push rod (29) is provided at the bottom of the cylinder (28), and a lower fixed plate (30) is fixedly connected to one end of the push rod (29). An upper fixed plate (31) is fixedly provided at one end of the cylinder (28). A connecting frame (32) is connected to the outer ring of the upper fixed plate (31). An adjusting motor (33) is provided at the top of the connecting frame (32). A first support rod (34) is connected to the output end of the adjusting motor (33). A support rod (35) is fixedly provided at the bottom of the first support rod (34). A first U-shaped frame (36) is rotatably connected to one end of the support rod (35). A second support rod (37) is connected to the outer ring of the lower fixed plate (30). A shaft (38) is provided at the bottom of the support rod (37). The outer ring of the shaft (38) is rotatably connected to a second U-shaped frame (39). One end of the second U-shaped frame (39) is rotatably connected to a pull rod (40). One end of the pull rod (40) is rotatably connected to a third U-shaped frame (41). One end of the third U-shaped frame (41) is connected to a telescopic rod (42). One end of the telescopic rod (42) is provided with a finger clamp (43). A limit frame (44) is fixedly provided above the vehicle body support (3). A receiving box (45) is movably provided in the inner cavity of the limit frame (44). A handle (46) is fixedly connected to the side wall of the receiving box (45).
2. The multi-vision perception right-angle hybrid multi-arm robot cotton harvesting device according to claim 1, characterized in that: The tail end of the right-angle and six-degree-of-freedom robotic arm end mechanism, the front end of the mobile chassis (2) and the vehicle body support (3) are all provided with visual communication components (24), the end of the X-axis connecting seat (16) is provided with a 3D radar module (25), the end of the sleeve (21) is provided with a connecting block (22), and the side wall of the connecting block (22) is provided with a Beidou module (23).
3. The multi-vision perception right-angle hybrid multi-arm robot cotton harvesting device according to claim 1, characterized in that: The top two sides of the X-axis slide rail (8) are provided with slide frames (11), and the outer ring of the slide frame (11) is provided with a curtain (12). The inner cavity of the curtain (12) is provided with several through holes (13), and the inner cavity of the through holes (13) is provided with a protective ring (14). One end of the curtain (12) is provided with a connecting plate (15), and one end of the connecting plate (15) is fixedly connected to the inner wall of the X-axis slide plate (17).
Citation Information
Patent Citations
Self-adaptive cotton harvesting method based on binocular vision recognition and intelligent mechanical harvesting device
CN113330915A
Pepper picking robot and control method
CN116117807A
Slitting assembly and machining device
CN218285711U