CN113785694A

CN113785694A - Tea intelligent picking equipment based on deep learning and computer vision recognition

Info

Publication number: CN113785694A
Application number: CN202111236676.7A
Authority: CN
Inventors: 吴罗发; 叶春; 舒时富; 吴衍麒; 罗翔
Original assignee: Agricultural Engineering Research Institute Jiangxi Academy Of Agricultural Sciences
Current assignee: Agricultural Engineering Research Institute Jiangxi Academy Of Agricultural Sciences
Priority date: 2021-10-23
Filing date: 2021-10-23
Publication date: 2021-12-14
Anticipated expiration: 2041-10-23
Also published as: CN113785694B

Abstract

The invention belongs to the technical field of picking robots, in particular to intelligent tea leaf picking equipment based on depth learning and computer vision recognition, which comprises a main body and a picking mechanical arm, wherein a blade is arranged in the rotating shaft in a sleeved mode, the other end of the rotating shaft is meshed and connected with a transmission strip in the picking mechanical arm through a tooth space, the arrangement is used for picking tea leaves on tea trees cleanly and effectively improving picking effect and efficiency, one end of a sliding block is connected with a T-shaped rod through a square block through the arrangement of the sliding block, the other end of the sliding block is connected with a second gear groove through a second gear mechanism, the other end of the second gear mechanism is connected with a fixed rod through a rack mechanism, the arrangement drives the sliding block to slide through rotating the T-shaped rod, the sliding block drives the rack mechanism to rotate when sliding, so that the fixed rod is tightly connected with a square ring groove, and further fixes an interface of a storage device at the square ring groove, the tea leaves picked are effectively prevented from being lost, and the storage device and the picking mechanical arm are convenient to assemble and disassemble.

Description

Tea intelligent picking equipment based on deep learning and computer vision recognition

Technical Field

The invention belongs to the technical field of picking robots, and particularly relates to intelligent tea picking equipment based on deep learning and computer vision recognition.

Background

The fresh leaf picking is the harvesting process of tea tree cultivation, and is also the key link of adjusting tree vigor and processing. The quality of fresh leaves picked not only directly affects the yield, quality and economic benefit of tea, but also relates to the growth and development of tea trees and the economic life; picking fresh leaves is also the most time-consuming and most seasonal work in tea garden management, and the working hours of the common manual tea picking account for 50% -65% of the labor used for tea garden management all the year round; in recent years, with the development of social economy, a great amount of labor force in tea areas shifts to the second and third-generation industries in cities and towns, and the contradiction of shortage of tea picking workers is very prominent due to poor environment and high labor intensity of tea picking operation, so that the shortage of labor force and labor force shortage in tea areas occur, and the increasing labor force cost becomes a bottleneck restricting the development of tea industries. The tea leaf machine realizes the whole-process mechanized farming, fertilizing, trimming, picking and processing of tea leaves, and rapidly promotes the tea leaf industry to change people.

Original tealeaves intelligence based on degree of depth study and computer vision discernment is picked and is equipped there are some defects in the course of the work, there are some differences in the height of tea tree growth, and original equipment height is unified, effect in picking the in-process is not good, storage device sets up in mechanical arm department, original connection is complicated, be not convenient for load and unload storage device, efficiency is influenced, in picking the in-process, mechanical transmission is more in the equipment, produce vibrations, rock, precision to picking influences to some extent, in picking the in-process, original picking blade simple structure, tea tree on the tea that can not be fine enucleats totally.

Disclosure of Invention

In order to solve the problems in the prior art, the invention provides tea leaf intelligent picking equipment based on deep learning and computer vision recognition, and the equipment has the characteristics of simple structure, compact connection, better working effect, better effect and stronger applicability.

In order to achieve the purpose, the invention provides the following technical scheme: the intelligent tea leaf picking equipment based on deep learning and computer vision recognition comprises a main body and a picking mechanical arm, wherein the lower end of the main body is connected with a base through a telescopic rod, a walking crawler belt for moving the main body is arranged at the lower end of the base, the picking mechanical arm for picking is arranged at two ends of the main body, a vision sensor is arranged at the position of the picking mechanical arm, the main body is connected with the picking mechanical arm through an electrical property, a limiting sleeve for connecting the main body with a sliding function is arranged on the picking mechanical arm, a second spring and a first spring for buffering are arranged in the limiting sleeve, a square annular groove for positioning and mounting a storage device interface is arranged at the side end of the lower end of the picking mechanical arm, a sliding groove for sliding a sliding block is arranged at the position of the square annular groove, a second gear groove for transmission is arranged in the sliding groove, and a rectangular groove for limiting the tea tree branches is arranged at the lower end of the picking mechanical arm, and a blade for picking action is arranged in the rectangular groove of the picking mechanical arm.

As the preferable technical scheme of the intelligent tea picking equipment based on deep learning and computer vision recognition, a transmission strip for transmission is arranged in the picking mechanical arm, the picking mechanical arm is connected with a rotating shaft through the transmission strip, a blade for picking is sleeved at the rotating shaft, and the blade is arc-shaped and flaky.

As the preferable technical scheme of the intelligent tea picking equipment based on deep learning and computer vision recognition, a T-shaped rod used for transmission is arranged at the side end of a picking mechanical arm, the T-shaped rod is meshed and connected with a square tooth groove through the tooth groove, the other end of the square block is fixedly connected with a sliding block, a second gear mechanism used for being connected with a second gear groove is arranged at the sliding block, the other end of the second gear mechanism is meshed and connected with a rack mechanism, and the rack mechanism is meshed and connected with a fixed rod through the tooth groove.

As the preferable technical scheme of the intelligent tea leaf picking equipment based on the deep learning and the computer vision recognition, the base is internally connected with the first gear mechanism through the motor, the first gear mechanism is connected with the telescopic rod through the square groove, and the first gear groove for meshing transmission with the first gear mechanism is arranged in the telescopic rod.

According to the optimal technical scheme of the intelligent tea leaf picking equipment based on the deep learning and the computer vision recognition, one end of the picking mechanical arm is provided with a limiting sleeve, the other end of the picking mechanical arm is provided with a connecting rod, the connecting rod is inserted into the limiting sleeve through a limiting groove, and the outer end of the connecting rod is sleeved with a first spring for buffering.

As the optimal technical scheme of the intelligent tea picking equipment based on deep learning and computer vision recognition, a limiting groove for the sliding action of a connecting rod is arranged in the limiting sleeve, and a second spring for the buffering action is arranged in the limiting groove.

As the optimal technical scheme of the intelligent tea picking equipment based on deep learning and computer vision recognition, one end of the fixed rod is provided with a tooth socket for transmission, and the other end of the fixed rod is provided with a thread for connection with a sliding block.

As the optimal technical scheme of the intelligent tea picking equipment based on deep learning and computer vision recognition, a controller communicated with a vision sensor is loaded in the main body, and a neural network model is loaded in the controller and used for recognizing the tea tree height and the tea branches loaded in the image fed back by the vision sensor.

Compared with the prior art, the invention has the beneficial effects that: through the arrangement of the blades, the blades are sleeved in the rotating shafts, the other ends of the rotating shafts are meshed and connected with a transmission strip in a picking mechanical arm through tooth grooves, tea leaves on tea trees are picked cleanly through the arrangement, the picking effect and efficiency are effectively improved, through the arrangement of the sliding blocks, one ends of the sliding blocks are connected with the T-shaped rods through the square blocks, the other ends of the sliding blocks are connected with the second gear grooves through the second gear mechanism, the other ends of the second gear mechanism are connected with the fixed rods through the rack mechanism, the arrangement drives the sliding blocks to slide through rotating the T-shaped rods, the sliding blocks drive the rack mechanism to rotate when sliding, so that the fixed rods are tightly connected with the square ring grooves, the interfaces of the storage device are fixed at the square ring grooves, the picked tea leaves are effectively prevented from being lost, the storage device is convenient to be assembled and disassembled with the picking mechanical arm, through the arrangement of the telescopic rods, the base is connected with the first gear mechanism through the inner motor, a gear mechanism is connected with the telescopic link through a gear groove, should set up the meshing transmission through a gear mechanism and a gear groove, the tea tree height of accessible difference, the height of appropriate regulation main part, thereby improve the effect of picking, and the suitability of equipment, setting through the stop collar, the stop collar sets up in picking the arm, and the stop collar is connected with the connecting rod through No. two springs, this sets up the elastic action through a spring and No. two springs, play the effect of buffering and inhaling the shake to the vibrations that the internal plant work produced, the effectual work precision of picking the arm that improves.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention and not to limit the invention. In the drawings:

FIG. 1 is a schematic structural view of the present invention;

FIG. 2 is a schematic view of the picking arm according to the present invention;

FIG. 3 is a schematic structural view of the telescopic rod of the present invention;

FIG. 4 is a schematic structural diagram of a slider according to the present invention;

FIG. 5 is a schematic structural view of a position-limiting sleeve according to the present invention;

in the figure: 1. a main body; 2. picking mechanical arms; 3. a telescopic rod; 4. a square groove; 5. a slider; 6. a first spring; 7. a second spring; 8. a square ring groove; 9. a T-shaped rod; 10. a base; 11. a sliding groove; 12. a rotating shaft; 13. a blade; 14. a first gear groove; 15. a first gear mechanism; 16. a limiting groove; 17. fixing the rod; 18. a rack mechanism; 19. a second gear mechanism; 20. a second gear groove; 21. a square block; 22. a connecting rod; 23. a limiting sleeve.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Examples

Referring to fig. 1-5, the present invention provides the following technical solutions: the intelligent tea picking equipment based on deep learning and computer vision recognition comprises a main body 1 and a picking mechanical arm 2, wherein the lower end of the main body 1 is connected with a base 10 through a telescopic rod 3, the lower end of the base 10 is provided with a walking crawler for moving the main body 1, the two ends of the main body 1 are provided with the picking mechanical arm 2 for picking, a vision sensor is arranged at the picking mechanical arm 2, the main body 1 is connected with the picking mechanical arm 2 through an electrical property, the picking mechanical arm 2 is provided with a limiting sleeve 23 for sliding a connecting rod 22, a second spring 7 and a first spring 6 for buffering are arranged in the limiting sleeve 23, a square annular groove 8 for positioning and mounting a storage device interface is arranged at the lower end side end of the picking mechanical arm 2, a sliding groove 11 for sliding a sliding block 5 is arranged at the square annular groove 8, and a second gear groove 20 for transmission is arranged in the sliding groove 11, the lower end of the picking mechanical arm 2 is provided with a rectangular groove for limiting the tea tree branches, the rectangular groove of the picking mechanical arm 2 is internally provided with a blade 13 for picking, and a rubber pad for contacting with a storage device and having an anti-skid effect is arranged in the square ring groove 8 in the embodiment.

Specifically, be equipped with the transmission strip that is used for the transmission effect in the picking arm 2, picking arm 2 is connected with axis of rotation 12 through the transmission strip, and axis of rotation 12 department cover is equipped with the blade 13 that is used for picking the effect, and blade 13 is the arc form, and is the slice, and blade 13 plays and picks the tealeaves on the tea tree totally in this embodiment, and effectual improvement is picked effect and efficiency.

Specifically, a T-shaped rod 9 for transmission is arranged at the side end of the picking mechanical arm 2, the T-shaped rod 9 is meshed with a tooth groove of a square block 21 through the tooth groove, the other end of the square block 21 is fixedly connected with a sliding block 5, a second gear mechanism 19 for connection with a second gear groove 20 is arranged at the position of the sliding block 5, the other end of the second gear mechanism 19 is meshed with a rack mechanism 18, the rack mechanism 18 is meshed with a fixed rod 17 through the tooth groove, in the embodiment, the sliding block 5 is driven to slide by rotating the T-shaped rod 9, the rack mechanism 18 is driven to rotate when the sliding block 5 slides, so that the fixed rod 17 is tightly connected with the square ring groove 8, an interface of a storage device is fixed at the square ring groove 8, picked tea leaves are effectively prevented from being left, and the storage device and the picking mechanical arm 2 are convenient to assemble and disassemble.

Specifically, the base 10 is connected with a first gear mechanism 15 through a motor, the first gear mechanism 15 is connected with the telescopic rod 3 through a square groove 4, a first gear groove 14 for meshing transmission with the first gear mechanism 15 is arranged in the telescopic rod 3, and in the embodiment, the height of the main body 1 can be properly adjusted through different tea tree heights through the meshing transmission of the first gear mechanism 15 and the first gear groove 14, so that the picking effect is improved, and the applicability of the device is improved.

Specifically, one end is equipped with stop collar 23 in picking arm 2, and the other end is equipped with connecting rod 22, and connecting rod 22 inserts through spacing groove 16 and establishes in stop collar 23, and the outer pot head of connecting rod 22 is equipped with a spring 6 that is used for the cushioning effect, plays the effect of buffering and inhaling the shake through a spring 6 to the vibrations that the internal plant work produced in this embodiment, and the effectual work precision who picks the arm that improves.

Specifically, be equipped with the spacing groove 16 that is used for connecting rod 22 sliding action in the stop collar 23, be equipped with No. two springs 7 that are used for cushioning effect in the spacing groove 16, through the elastic action of No. two springs 7 in this embodiment, play the effect of buffering and inhaling vibrations, the effectual work precision that improves the picking arm to the vibrations that the internal plant work produced.

Specifically, dead lever 17 one end is equipped with the tooth's socket that is used for the transmission effect, and the other end is equipped with the screw thread that is used for with sliding block 5 joint action, and dead lever 17 evenly distributed is equipped with the rubber pad at sliding block 5's lower extreme in this embodiment, and sliding block 5 and storage device contact department.

Specifically, the main body 1 is internally provided with a controller which is communicated with a visual sensor, the controller is internally provided with a neural network model for realizing the height of tea trees and the identification of tea branches in images fed back by the visual sensor, wherein the model for the tea branches adopts a Dssd-inclusion-v 2-coco model, the model adopts a Dssd target detection algorithm, and an initiation v2 deep neural network is pre-trained by a coco data set. Then, training the model by using a previously prepared data set, finely adjusting various parameters in the deep neural network, and finally obtaining a proper target detection model for detecting the tea branches; the height of the tea tree is identified as a Dssd-BP neural network model, the tea tree is identified based on a Dssd target detection algorithm, and then a corresponding scale is called based on the BP neural network model to realize the height measurement of the tea tree.

The working principle and the using process of the invention are as follows: the invention moves a main body 1 to a proper position through a crawler at the lower end of a base 10, drives a first gear mechanism 15 to rotate through a motor according to the height of tea trees, so that the first gear mechanism 15 is meshed with a first gear groove 14, the telescopic height of a telescopic rod 3 is adjusted, so that the heights of the main body 1 and a picking mechanical arm 2 are adjusted, a connector of a storage device is sleeved at a square ring groove 8, a T-shaped rod 9 is rotated, a tooth groove arranged in the same height of the T-shaped rod 9 is meshed with a tooth groove of a square block 21 for transmission, so that a sliding block 5 is driven to move, when the sliding block 5 moves, a second gear mechanism 19 is meshed with a second gear groove 20 for transmission, so that a rack mechanism 18 is driven to rotate, the rack mechanism 18 drives a fixed rod 17 to rotate through the tooth groove, the other end of the fixed rod 17 is connected with the sliding block 5 through threads, so that the fixed rod 17 moves along with the sliding block 5 to be in a rotating movement state, thereby the interval of continuous change dead lever 17 and square annular 8, after the suitable position, move axis of rotation 12 through the drive belt in the picking arm 2 and rotate, axis of rotation 12 drives blade 13 and rotates, and the branch of tea tree is spacing to slide in picking arm 2 through the rectangular channel of picking arm 2 lower extreme, picks tealeaves through blade 13, collects tealeaves through storage device.

Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that changes may be made in the embodiments and/or equivalents thereof without departing from the spirit and scope of the invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. Equipment is picked to tealeaves intelligence based on degree of deep learning and computer vision recognition, its characterized in that: comprises a main body (1) and a picking mechanical arm (2), wherein the lower end of the main body (1) is connected with a base (10) through a telescopic rod (3), a walking crawler belt for the main body (1) to move is arranged at the lower end of the base (10), the picking mechanical arm (2) for picking is arranged at two ends of the main body (1), a vision sensor is arranged at the position of the picking mechanical arm (2), the main body (1) is connected with the picking mechanical arm (2) through an electric property, the picking mechanical arm (2) is provided with a limit sleeve (23) for the sliding action of a connecting rod (22), a second spring (7) and a first spring (6) for buffering action are arranged in the limit sleeve (23), a square annular groove (8) for the positioning and mounting action of a storage device interface is arranged at the lower end of the picking mechanical arm (2), a sliding groove (11) for the sliding action of a sliding block (5) is arranged at the position of the square annular groove (8), be equipped with No. two gear groove (20) that are used for the transmission effect in sliding tray (11), it is equipped with the rectangular channel that is used for tea tree branch limiting displacement to pick arm (2) lower extreme, it is equipped with blade (13) that are used for picking the effect to pick arm (2) rectangular channel.

2. The tea intelligent picking apparatus based on deep learning and computer vision recognition as claimed in claim 1, wherein: the picking mechanical arm is characterized in that a transmission strip for transmission is arranged in the picking mechanical arm (2), the picking mechanical arm (2) is connected with a rotating shaft (12) through the transmission strip, a blade (13) for picking is sleeved at the position of the rotating shaft (12), and the blade (13) is arc-shaped and is sheet-shaped.

3. The tea intelligent picking apparatus based on deep learning and computer vision recognition as claimed in claim 1, wherein: picking arm (2) side is equipped with T shape pole (9) that are used for the transmission effect, T shape pole (9) are connected through tooth's socket and square piece (21) tooth's socket meshing, square piece (21) other end and sliding block (5) fixed connection, sliding block (5) department is equipped with No. two gear mechanism (19) that are used for with No. two gear groove (20) connection effects, the other end and the rack mechanism (18) meshing of No. two gear mechanism (19) are connected, rack mechanism (18) are connected through tooth's socket and dead lever (17) meshing.

4. The tea intelligent picking apparatus based on deep learning and computer vision recognition as claimed in claim 1, wherein: the motor-driven telescopic rod is characterized in that the base (10) is internally connected with a gear mechanism (15) through a motor, the gear mechanism (15) is connected with the telescopic rod (3) through a square groove (4), and a gear groove (14) for meshing transmission with the gear mechanism (15) is arranged in the telescopic rod (3).

5. The tea intelligent picking apparatus based on deep learning and computer vision recognition as claimed in claim 1, wherein: one end is equipped with stop collar (23) in picking arm (2), and the other end is equipped with connecting rod (22), connecting rod (22) are inserted through spacing groove (16) and are established in stop collar (23), connecting rod (22) outer pot head is equipped with spring (6) that are used for cushioning effect.

6. The tea intelligent picking apparatus based on deep learning and computer vision recognition as claimed in claim 1, wherein: a limiting groove (16) for the sliding action of the connecting rod (22) is formed in the limiting sleeve (23), and a second spring (7) for the buffering action is arranged in the limiting groove (16).

7. The tea intelligent picking apparatus based on deep learning and computer vision recognition as claimed in claim 3, wherein: and one end of the fixed rod (17) is provided with a tooth groove for transmission, and the other end of the fixed rod is provided with a thread for connecting with the sliding block (5).

8. The tea intelligent picking apparatus based on deep learning and computer vision recognition as claimed in claim 3, wherein: the main body (1) is internally provided with a controller which is communicated with the visual sensor, and the controller is internally provided with a neural network model for realizing the height of the tea tree and the identification of the tea branches in the image fed back by the visual sensor.