Intelligent potato seed cutting machine based on visual identification

Abstract

The invention discloses an intelligent potato seed cutting machine based on visual identification, and belongs to the technical field of potato seed cutting machinery, the intelligent potato seed cutting machine comprises a feeding mechanism, the surface of the feeding mechanism is provided with a visual identification device, one side of the feeding mechanism is provided with a sorting mechanism, and the surface of the sorting mechanism is provided with a plurality of transfer mechanisms matched with the sorting mechanism; a workbench is arranged on one side of each transfer mechanism, a plurality of cutting tools matched with the transfer mechanisms are installed on the surface of the workbench, a feeding mechanism matched with the cutting tools is installed on the surface of the workbench, and a pressing mechanism matched with the cutting tools is installed on the top of the workbench. A collecting hopper matched with the cutting tool is installed on the surface of the workbench. On the basis of sorting according to the distribution characteristics of potato bud eyes, the straight knife, the Y-shaped knife, the T-shaped knife and the cross-shaped knife are respectively used for cutting seeds of potatoes so as to meet the requirement of potato seeding.

Description

Technical Field

[0001] The invention relates to the technical field of potato seed cutting machines, and more specifically to an intelligent potato seed cutting machine based on visual recognition. Background Art

[0002] Potatoes are the fourth largest food crop in the world. The seed potato processing has long faced two major technical bottlenecks. First, manual cutting is highly dependent on experience. The efficiency of skilled workers is only 30-50kg / hour, and the qualified rate fluctuates greatly. Field tests show that the rate of intact eyes in manual cutting is only 68.2±7.5%, the degree of standardization is low, and the cutting angle deviation is generally >15°, resulting in a difference of more than 40% in the uniformity of seedling emergence. Second, most existing equipment uses drum sorting, which cannot identify the three-dimensional distribution of eyes, resulting in a waste rate of effective eyes >35%. The fixed tool design results in a single cut shape, which cannot adapt to the distribution characteristics of eyes of different seed potatoes.

[0003] In recent years, some simple mechanical cutting equipment has begun to be used, but these devices usually lack intelligent control and cannot perform precise cutting according to the specific conditions of the buds, which can easily cause waste or damage. During the cutting process, existing equipment cannot ensure that each tuber contains an appropriate number of buds, which may cause some tubers to have too few or too many buds, affecting the planting effect. In addition, there is a lack of data recording and analysis during the cutting process, making it difficult to optimize and improve. Summary of the invention

[0004] 1. Technical problems to be solved

[0005] In view of the problems existing in the prior art, the purpose of the present invention is to provide an intelligent potato seed cutter based on visual recognition, which can sort potatoes according to the distribution characteristics of the potato buds and use a straight knife, a Y-shaped knife, a T-shaped knife, and a cross knife to cut potatoes to meet the needs of potato sowing.

[0006] 2. Technical solution

[0007] To solve the above problems, the present invention adopts the following technical solutions:

[0008] A potato intelligent seed cutter based on visual recognition comprises: a feeding mechanism, a visual recognition device is provided on the surface of the feeding mechanism, a sorting mechanism is installed on one side of the feeding mechanism, a plurality of transfer mechanisms adapted to the sorting mechanism are provided on the surface of the sorting mechanism, a workbench is provided on one side of the transfer mechanism, cutting tools adapted to the plurality of transfer mechanisms are installed on the surface of the workbench, a feeding mechanism adapted to the cutting tools is installed on the surface of the workbench, a pressing mechanism adapted to the cutting tools is installed on the top of the workbench, a collecting bucket adapted to the cutting tools is installed on the surface of the workbench, a collecting box adapted to the collecting bucket is provided on the surface of the workbench, and a control terminal is installed on the surface of one side of the workbench.

[0009] As a preferred solution of the present invention, the loading mechanism includes a placement box, a first support seat is installed on one side of the placement box, a first electric conveying rail for moving potatoes inside the placement box is installed on the surface of the first support seat, loading pallets are equidistantly provided on the surface of the first electric conveying rail, and first baffles are installed on the surfaces of both sides of the first electric conveying rail.

[0010] As a preferred solution of the present invention, the visual recognition device includes an identification frame installed on the surface of the first baffle, the surface of the identification frame is provided with a first guide rail extending longitudinally, the surface of the first guide rail is slidably connected with a first movable part, the surface of the first guide rail is installed with a first electric cylinder for longitudinal adjustment, and the piston rod of the first electric cylinder is fixedly connected to the first movable part, the surface of the first movable part is installed with multiple second electric cylinders, and the surface of the piston rod of the second electric cylinder is installed with a first industrial camera.

[0011] As a preferred embodiment of the present invention, the sorting mechanism includes a sorting frame arranged on one side of the identification frame, a second electric conveying rail is installed on the upper surface of the sorting frame, a guide bucket adapted to the first electric conveying rail is installed on the upper surface of the second electric conveying rail, a plurality of first servo motors are installed on the surface of one side of the sorting frame, a toggle plate is fixedly connected to the surface of the output shaft of the first servo motor, and a sorting bucket adapted to the toggle plate is installed on the surface of the other side of the sorting frame.

[0012] As a preferred embodiment of the present invention, the transfer mechanism includes a transfer frame arranged on one side of the sorting frame, a third electric conveying rail is installed on the upper surface of the transfer frame, second baffles are installed on the surfaces on both sides of the third electric conveying rail, a first mounting frame is installed on the surface of the second baffle, a third electric cylinder for longitudinal adjustment is installed on the surface of the first mounting frame, a second industrial camera is installed on the surface of the piston rod of the third electric cylinder, a fourth electric cylinder is installed on the surface of one side of the third electric conveying rail, a push rod is installed on the surface of the piston rod of the fourth electric cylinder, and a waste box adapted to the push rod is installed on the surface of the transfer frame.

[0013] As a preferred solution of the present invention, the feeding mechanism includes a material moving frame installed on the surface of the workbench, the surface of the material moving frame is provided with a horizontally extending second guide rail, the surface of the second guide rail is slidably connected with a second moving part, the surface of the second moving part is rotatably connected with two driven connecting rods, the outer surface of the driven connecting rod is rotatably connected with a material moving clamping plate, and the inner side of the material moving clamping plate is provided with a rubber plate.

[0014] As a preferred solution of the present invention, the internal rotatable connection of the material moving frame is a threaded rod, and the threaded rod is threadedly connected to the second moving part, a second servo motor is installed on the surface of one side of the material moving frame, and the output shaft of the second servo motor is fixedly connected to one end of the threaded rod, a fifth electric cylinder is installed on the surface of the second moving part, and a transmission gear plate is fixedly connected to the surface of the piston rod of the fifth electric cylinder, the surface of the second moving part is rotatably connected to a toothed connecting rod, and the toothed connecting rod is rotatably connected to the material moving clamp, and the transmission gear plate is meshed with the toothed connecting rods on both sides.

[0015] As a preferred solution of the present invention, the pressing mechanism includes a second support seat and a third servo motor installed on the surface of the workbench, the surface of the second support seat is provided with a third guide rail extending longitudinally, the surface of the third guide rail is slidably connected with a third moving member, the surface of the third moving member is fixedly connected with a moving tooth plate, the bottom of the moving tooth plate is fixedly connected with a pressing plate, the surface of the output shaft of the third servo motor is fixedly connected with a driving gear, and the driving gear is meshed with the moving tooth plate.

[0016] As a preferred solution of the present invention, the control terminal is internally provided with a preprocessing module, a feature extraction module, a bud eye detection and positioning module, a classification module, a decision output module, a post-processing optimization module, a data management and storage module, and a user interaction module;

[0017] The preprocessing module is used to preprocess the original images collected by the first industrial camera and the second industrial camera, and the preprocessing includes Gaussian filtering, contrast enhancement, size normalization and color space conversion. The feature extraction module is used to extract feature information related to the bud eye from the preprocessed image using Canny edge detection. The bud eye detection and positioning module is used to identify and accurately locate the bud eye position on the surface of the potato based on the features extracted by the Hough transform and the YOLO deep learning target detection model. The classification module is used to use a convolutional neural network to classify the bud eyes according to the size and number of the bud eyes. The decision output module is used to generate a final decision recommendation based on the classification result, and output a control signal to the sorting mechanism, the transfer mechanism, the feeding mechanism and the pressing mechanism. The post-processing optimization module is used to evaluate and optimize the performance of the entire system using model hyperparameter adjustment. The data management and storage module is used to manage and store a large amount of image data and classification results. The user interaction module is used to provide a user-friendly interface.

[0018] 3. Beneficial effects

[0019] Compared with the prior art, the advantages of the present invention are:

[0020] (1) The present invention can load potatoes through a loading mechanism, and recognize and classify the buds on the surface of potatoes through a visual recognition device, and then move the potatoes through a sorting mechanism and classify the potatoes. The classified potatoes are moved under the action of a transfer mechanism, and then the moving mechanism moves the potatoes to the inside of a cutting tool, and then the potatoes are cut into pieces under the action of a pressing mechanism, and the cut potatoes fall into the inside of a collection box for collection.

[0021] (2) The present invention uses visual recognition technology to accurately identify the position of the buds on the surface of potatoes, ensuring that each cut potato contains an appropriate amount of buds, thereby improving the germination rate. The entire cutting process is fully automated, reducing human intervention and greatly improving work efficiency. Compared with manual cutting, the intelligent cutting machine can process a large number of potatoes in a short time to meet large-scale planting needs. Automated equipment replaces a large amount of manual labor, reducing labor costs. Through precise cutting, it reduces potato waste and improves resource utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a first-perspective stereoscopic image of an intelligent potato seed cutter based on visual recognition according to the present invention;

[0023] Figure 2 It is a schematic diagram of a feeding mechanism in a potato intelligent seed cutter based on visual recognition according to the present invention;

[0024] Figure 3 It is a schematic diagram of a visual recognition device in a potato intelligent seed cutter based on visual recognition according to the present invention;

[0025] Figure 4 It is a schematic diagram of a sorting mechanism in a potato intelligent seed cutter based on visual recognition according to the present invention;

[0026] Figure 5 It is a schematic diagram of a transport mechanism in a potato intelligent seed cutter based on visual recognition according to the present invention;

[0027] Figure 6 It is a schematic diagram of the local structure of a potato intelligent seed cutter based on visual recognition according to the present invention;

[0028] Figure 7 It is a schematic diagram of a feeding mechanism in a potato intelligent seed cutter based on visual recognition according to the present invention;

[0029] Figure 8 It is a schematic diagram of a material pressing mechanism in a potato intelligent seed cutter based on visual recognition according to the present invention;

[0030] Fig. 9 This is a local structural adaptation diagram of a potato intelligent seed cutter based on visual recognition according to the present invention.

[0031] Description of the numbers in the figure:

[0032] 1. Feeding mechanism; 101. Placement box; 102. First support seat; 103. First electric conveyor rail; 104. Feeding pallet; 105. First baffle; 2. Visual recognition device; 201. Recognition frame; 202. First guide rail; 203. First moving part; 204. First electric cylinder; 205. Second electric cylinder; 206. First industrial camera; 3. Sorting mechanism; 301. Sorting frame; 302. Second electric conveyor rail; 303. Guide bucket; 304. First servo motor; 305. Toggle plate; 306. Sorting bucket; 4. Transfer mechanism; 401. Transfer frame; 402. Third electric conveyor rail; 403. Second baffle; 404. First mounting frame; 405. Third electric cylinder; 406. Second Industrial camera; 407, fourth electric cylinder; 408, push rod; 409, waste box; 5, workbench; 6, feeding mechanism; 601, material transfer frame; 602, second guide rail; 603, second moving part; 604, threaded rod; 605, second servo motor; 606, driven connecting rod; 607, material transfer clamp; 608, rubber plate; 609, fifth electric cylinder; 610, transmission gear plate; 611, tooth connecting rod; 7, cutting tool; 8, pressing mechanism; 801, second support seat; 802, third guide rail; 803, third moving part; 804, moving gear plate; 805, pressing plate; 806, third servo motor; 807, driving gear; 9, collecting bucket; 10, collecting box; 11, control terminal. DETAILED DESCRIPTION

[0033] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] Example:

[0035] See also Figure 1-9A potato intelligent seed cutter based on visual recognition comprises: a feeding mechanism 1, a visual recognition device 2 is arranged on the surface of the feeding mechanism 1, a sorting mechanism 3 is installed on one side of the feeding mechanism 1, a plurality of transfer mechanisms 4 adapted thereto are arranged on the surface of the sorting mechanism 3, a workbench 5 is arranged on one side of the transfer mechanism 4, cutting tools 7 adapted thereto are installed on the surface of the workbench 5, a feeding mechanism 6 adapted thereto is installed on the surface of the workbench 5, a pressing mechanism 8 adapted thereto is installed on the top of the workbench 5, a collecting bucket 9 adapted thereto is installed on the surface of the workbench 5, a collecting box 10 adapted thereto is arranged on the surface of the workbench 5, a control terminal 11 is installed on the surface of one side of the workbench 5, wherein the plurality of cutting tools 7 are respectively a straight knife, a T-shaped knife, a cross-shaped knife and a Y-shaped knife.

[0036] In a specific embodiment of the present invention, potatoes can be loaded through the loading mechanism 1, and the buds on the surface of the potatoes can be identified and classified through the visual recognition device 2, and then the potatoes are moved and classified through the sorting mechanism 3. The classified potatoes are moved under the action of the transfer mechanism 4, and then the feeding mechanism 6 moves the potatoes to the inside of the cutting tool 7, and then the potatoes are cut into pieces under the action of the pressing mechanism 8, and the cut potatoes will fall into the inside of the collection box 10 for collection.

[0037] Specifically, the loading mechanism 1 includes a placement box 101, a first support seat 102 is installed on one side of the placement box 101, a first electric conveying rail 103 for moving potatoes inside the placement box 101 is installed on the surface of the first support seat 102, and loading pallets 104 are equidistantly provided on the surface of the first electric conveying rail 103, and first baffles 105 are installed on the surfaces of both sides of the first electric conveying rail 103.

[0038] In a specific embodiment of the present invention, potatoes are temporarily stored and placed in a placement box 101 , and then the first electric conveyor rail 103 is started and the potatoes are gradually loaded under the action of the loading support plate 104 .

[0039] Specifically, the visual recognition device 2 includes an identification frame 201 installed on the surface of the first baffle 105, the surface of the identification frame 201 is provided with a longitudinally extending first guide rail 202, the surface of the first guide rail 202 is slidably connected with a first movable member 203, the surface of the first guide rail 202 is installed with a longitudinally adjustable first electric cylinder 204, and the piston rod of the first electric cylinder 204 is fixedly connected to the first movable member 203, a plurality of second electric cylinders 205 are installed on the surface of the first movable member 203, and the surface of the piston rod of the second electric cylinder 205 is installed with a first industrial camera 206.

[0040] In a specific embodiment of the present invention, the first electric cylinder 204 is activated by its piston rod to drive the first movable member 203 to move on the surface of the first guide rail 202, and then the second electric cylinder 205 is activated by its piston rod to drive the first industrial camera 206 to move. At this time, potatoes can be identified by multiple first industrial cameras 206.

[0041] Specifically, the sorting mechanism 3 includes a sorting frame 301 arranged on one side of the identification frame 201, a second electric conveying rail 302 is installed on the upper surface of the sorting frame 301, a guide bucket 303 adapted to the first electric conveying rail 103 is installed on the upper surface of the second electric conveying rail 302, a plurality of first servo motors 304 are installed on the surface of one side of the sorting frame 301, a toggle plate 305 is fixedly connected to the surface of the output shaft of the first servo motor 304, and a sorting bucket 306 adapted to the toggle plate 305 is installed on the surface of the other side of the sorting frame 301.

[0042] In a specific embodiment of the present invention, the potatoes are guided by the guide bucket 303, and then the second electric conveyor rail 302 drives the potatoes to move. At this time, the first servo motor 304 adapted to the cut shape of the potatoes starts its output shaft to drive the toggle plate 305 to move, thereby pushing the potatoes into the corresponding sorting bucket 306.

[0043] Specifically, the transfer mechanism 4 includes a transfer frame 401 arranged on one side of the sorting frame 301, a third electric conveying rail 402 is installed on the upper surface of the transfer frame 401, second baffles 403 are installed on the surfaces of both sides of the third electric conveying rail 402, a first mounting frame 404 is installed on the surface of the second baffle 403, a third electric cylinder 405 for longitudinal adjustment is installed on the surface of the first mounting frame 404, a second industrial camera 406 is installed on the surface of the piston rod of the third electric cylinder 405, a fourth electric cylinder 407 is installed on the surface of one side of the third electric conveying rail 402, a push rod 408 is installed on the surface of the piston rod of the fourth electric cylinder 407, and a waste box 409 adapted to the push rod 408 is installed on the surface of the transfer frame 401.

[0044] In a specific embodiment of the present invention, potatoes will move on the surface of the third electric conveyor rail 402 under the action of the second baffle 403, and at the same time, the third electric cylinder 405 starts its piston rod to drive the second industrial camera 406 to move, thereby performing secondary visual recognition on the potatoes, and the fourth electric cylinder 407 starts its piston rod to drive the push rod 408 to move, thereby pushing the unqualified potatoes into the waste box 409.

[0045] Specifically, the feeding mechanism 6 includes a material moving frame 601 installed on the surface of the workbench 5, the surface of the material moving frame 601 is provided with a second guide rail 602 extending horizontally, the surface of the second guide rail 602 is slidably connected to a second moving member 603, the surface of the second moving member 603 is rotatably connected to two driven connecting rods 606, the outer surface of the driven connecting rod 606 is rotatably connected to a material moving clamping plate 607, the inner side of the material moving clamping plate 607 is provided with a rubber plate 608, the interior of the material moving frame 601 is rotatably connected to a threaded rod 604, and the threaded rod 604 is rotatably connected to the second moving member 603. The movable part 603 is threadedly connected, and a second servo motor 605 is installed on the surface of one side of the material transfer frame 601, and the output shaft of the second servo motor 605 is fixedly connected to one end of the threaded rod 604, and a fifth electric cylinder 609 is installed on the surface of the second movable part 603, and a transmission gear plate 610 is fixedly connected to the surface of the piston rod of the fifth electric cylinder 609, and a toothed connecting rod 611 is rotatably connected to the surface of the second movable part 603, and the toothed connecting rod 611 is rotatably connected to the material transfer clamping plate 607, and the transmission gear plate 610 is meshed with the toothed connecting rods 611 on both sides.

[0046] In a specific embodiment of the present invention, the second servo motor 605 is started to drive the threaded rod 604 to rotate through its output shaft, thereby driving the second movable member 603 to move on the surface of the second guide rail 602, and then the fifth electric cylinder 609 is started to drive the transmission gear plate 610 to move. At this time, under the action of the tooth connecting rod 611 and the driven connecting rod 606, the material transfer clamping plate 607 and the rubber plate 608 are driven to rotate, thereby moving the potatoes on the surface of the third electric conveying rail 402 to the surface of the cutting tool 7.

[0047] Specifically, the pressing mechanism 8 includes a second support seat 801 and a third servo motor 806 installed on the surface of the workbench 5. The surface of the second support seat 801 is provided with a third guide rail 802 extending longitudinally. The surface of the third guide rail 802 is slidably connected with a third moving member 803. The surface of the third moving member 803 is fixedly connected with a moving tooth plate 804. The bottom of the moving tooth plate 804 is fixedly connected with a pressing plate 805. The surface of the output shaft of the third servo motor 806 is fixedly connected with a driving gear 807, and the driving gear 807 is meshed with the moving tooth plate 804.

[0048] In a specific embodiment of the present invention, the output shaft of the third servo motor 806 is started to drive the driving gear 807 to rotate. At this time, the movable tooth plate 804 meshing with the driving gear 807 moves. The movement of the movable tooth plate 804 drives the third movable member 803 to move on the surface of the third guide rail 802, thereby driving the pressing plate 805 to descend. At this time, the potatoes are cut into pieces under the action of the cutting tool 7.

[0049] Specifically, the control terminal 11 is internally provided with a preprocessing module, a feature extraction module, a bud eye detection and positioning module, a classification module, a decision output module, a post-processing optimization module, a data management and storage module, and a user interaction module;

[0050] The preprocessing module is used to preprocess the original images collected by the first industrial camera 206 and the second industrial camera 406 to improve the accuracy and efficiency of subsequent processing. The preprocessing includes Gaussian filtering, contrast enhancement, size normalization and color space conversion. The feature extraction module is used to extract feature information related to the buds from the preprocessed image using Canny edge detection. The bud detection and positioning module is used to identify and accurately locate the buds on the surface of the potato based on the features extracted by the Hough transform and the YOLO deep learning target detection model. The classification module is used to use a convolutional neural network to classify the buds according to the size and number of the buds. The decision output module is used to generate a final decision recommendation based on the classification result, and output a control signal to the sorting mechanism 3, the transfer mechanism 4, the feeding mechanism 6 and the pressing mechanism 8. The post-processing optimization module is used to evaluate and optimize the performance of the entire system using model hyperparameter adjustment to ensure the stability and accuracy of the system. The data management and storage module is used to manage and store a large amount of image data and classification results to facilitate subsequent data analysis and model training. The user interaction module is used to provide a user-friendly interface to facilitate user operation and viewing of the system operation status.

[0051] A method for using a potato intelligent seed cutter based on visual recognition comprises the following steps:

[0052] The potatoes are temporarily stored and placed in the placement box 101, and then the first electric conveyor rail 103 is started and the potatoes are gradually loaded under the action of the loading support plate 104, and the first electric cylinder 204 is started and its piston rod drives the first moving member 203 to move on the surface of the first guide rail 202, and then the second electric cylinder 205 is started and its piston rod drives the first industrial camera 206 to move, and at this time, the potatoes can be identified by multiple first industrial cameras 206, and the potatoes are guided by the guide bucket 303. Then, the second electric conveyor rail 302 drives the potatoes to move. At this time, the first servo motor 304 adapted to the cut shape of the potatoes starts its output shaft to drive the toggle plate 305 to move, thereby pushing the potatoes into the corresponding sorting bucket 306. The potatoes will move on the surface of the third electric conveyor rail 402 under the action of the second baffle 403. At the same time, the third electric cylinder 405 starts its piston rod to drive the second industrial camera 406 to move, thereby performing secondary visual recognition of the potatoes. The fourth electric cylinder 407 starts its activity The plug rod drives the push rod 408 to move, thereby pushing the unqualified potatoes into the waste bin 409, and the second servo motor 605 starts its output shaft to drive the threaded rod 604 to rotate, thereby driving the second moving member 603 to move on the surface of the second guide rail 602, and then the fifth electric cylinder 609 starts its piston rod to drive the transmission gear plate 610 to move, at this time, under the action of the tooth connecting rod 611 and the driven connecting rod 606, the material moving clamping plate 607 and the rubber plate 608 are driven to rotate, thereby the third electric cylinder 609 is driven to move the material. The potatoes on the surface of the movable conveying rail 402 move to the surface of the cutting tool 7, and the output shaft of the third servo motor 806 is started to drive the driving gear 807 to rotate. At this time, the movable tooth plate 804 meshed with the driving gear 807 moves, and the movement of the movable tooth plate 804 drives the third movable member 803 to move on the surface of the third guide rail 802, thereby driving the pressing plate 805 to descend. At this time, the potatoes are cut into pieces under the action of the cutting tool 7, and the cut potatoes will fall into the inside of the collecting box 10 under the action of the collecting bucket 9.

[0053] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and improved concepts of the present invention within the technical scope disclosed by the present invention, and they should be covered by the protection scope of the present invention.

Claims

1. A potato intelligent seed cutter based on visual recognition, characterized in that: include: A feeding mechanism (1), wherein a visual recognition device (2) is provided on the surface of the feeding mechanism (1), a sorting mechanism (3) is installed on one side of the feeding mechanism (1), a plurality of transfer mechanisms (4) adapted thereto are provided on the surface of the sorting mechanism (3), a workbench (5) is provided on one side of the transfer mechanism (4), a cutting tool (7) adapted thereto is installed on the surface of the workbench (5), a feeding mechanism (6) adapted thereto is installed on the surface of the workbench (5), a pressing mechanism (8) adapted thereto is installed on the top of the workbench (5), a collecting bucket (9) adapted thereto is installed on the surface of the workbench (5), a collecting box (10) adapted thereto is provided on the surface of the workbench (5), and a control terminal (11) is installed on the surface of one side of the workbench (5).

2. The intelligent potato seed cutter based on visual recognition according to claim 1, characterized in that: The loading mechanism (1) comprises a placement box (101), a first support seat (102) is installed on one side of the placement box (101), a first electric conveying rail (103) for moving potatoes inside the placement box (101) is installed on the surface of the first support seat (102), loading support plates (104) are equidistantly arranged on the surface of the first electric conveying rail (103), and first baffles (105) are installed on the surfaces of both sides of the first electric conveying rail (103).

3. The intelligent potato seed cutter based on visual recognition according to claim 2, characterized in that: The visual recognition device (2) comprises an identification frame (201) mounted on the surface of a first baffle (105), the surface of the identification frame (201) being provided with a first guide rail (202) extending longitudinally, the surface of the first guide rail (202) being slidably connected to a first movable member (203), the surface of the first guide rail (202) being mounted with a first longitudinally adjustable electric cylinder (204), the piston rod of the first electric cylinder (204) being fixedly connected to the first movable member (203), the surface of the first movable member (203) being mounted with a plurality of second electric cylinders (205), the surface of the piston rod of the second electric cylinder (205) being mounted with a first industrial camera (206).

4. The intelligent potato seed cutter based on visual recognition according to claim 3, characterized in that: The sorting mechanism (3) comprises a sorting frame (301) arranged on one side of the identification frame (201); a second electric conveying rail (302) is installed on the upper surface of the sorting frame (301); a guide bucket (303) adapted to the first electric conveying rail (103) is installed on the upper surface of the second electric conveying rail (302); a plurality of first servo motors (304) are installed on the surface of one side of the sorting frame (301); a toggle plate (305) is fixedly connected to the surface of the output shaft of the first servo motor (304); and a sorting bucket (306) adapted to the toggle plate (305) is installed on the surface of the other side of the sorting frame (301).

5. The intelligent potato seed cutter based on visual recognition according to claim 4, characterized in that: The transfer mechanism (4) includes a transfer frame (401) arranged on one side of the sorting frame (301), a third electric conveying rail (402) is installed on the upper surface of the transfer frame (401), second baffles (403) are installed on the surfaces of both sides of the third electric conveying rail (402), a first mounting frame (404) is installed on the surface of the second baffle (403), a third electric cylinder (405) for longitudinal adjustment is installed on the surface of the first mounting frame (404), a second industrial camera (406) is installed on the surface of the piston rod of the third electric cylinder (405), a fourth electric cylinder (407) is installed on the surface of one side of the third electric conveying rail (402), a push rod (408) is installed on the surface of the piston rod of the fourth electric cylinder (407), and a waste box (409) adapted to the push rod (408) is installed on the surface of the transfer frame (401).

6. The intelligent potato seed cutter based on visual recognition according to claim 5, characterized in that: The feeding mechanism (6) includes a material moving frame (601) installed on the surface of the workbench (5), the surface of the material moving frame (601) is provided with a horizontally extending second guide rail (602), the surface of the second guide rail (602) is slidably connected with a second moving member (603), the surface of the second moving member (603) is rotatably connected with two driven connecting rods (606), the outer surface of the driven connecting rod (606) is rotatably connected with a material moving clamping plate (607), and the inner side of the material moving clamping plate (607) is provided with a rubber plate (608).

7. The intelligent potato seed cutter based on visual recognition according to claim 6, characterized in that: The material shifting frame (601) is internally rotatably connected with a threaded rod (604), and the threaded rod (604) is threadedly connected to the second moving member (603); a second servo motor (605) is mounted on the surface of one side of the material shifting frame (601), and the output shaft of the second servo motor (605) is fixedly connected to one end of the threaded rod (604); a fifth electric cylinder (609) is mounted on the surface of the second moving member (603), and a transmission tooth plate (610) is fixedly connected to the surface of the piston rod of the fifth electric cylinder (609); a toothed connecting rod (611) is rotatably connected to the surface of the second moving member (603), and the toothed connecting rod (611) is rotatably connected to the material shifting clamp (607); and the transmission tooth plate (610) is meshed with the toothed connecting rods (611) on both sides.

8. The intelligent potato seed cutter based on visual recognition according to claim 7, characterized in that: The pressing mechanism (8) comprises a second support seat (801) and a third servo motor (806) mounted on the surface of the workbench (5); the surface of the second support seat (801) is provided with a third guide rail (802) extending longitudinally; the surface of the third guide rail (802) is slidably connected with a third moving member (803); the surface of the third moving member (803) is fixedly connected with a moving tooth plate (804); the bottom of the moving tooth plate (804) is fixedly connected with a pressing plate (805); the surface of the output shaft of the third servo motor (806) is fixedly connected with a driving gear (807), and the driving gear (807) is meshed with the moving tooth plate (804).

9. The intelligent potato seed cutter based on visual recognition according to claim 8, characterized in that: The control terminal (11) is internally provided with a preprocessing module, a feature extraction module, a bud eye detection and positioning module, a classification module, a decision output module, a post-processing optimization module, a data management and storage module, and a user interaction module; The preprocessing module is used to preprocess the original images collected by the first industrial camera (206) and the second industrial camera (406), and the preprocessing includes Gaussian filtering, contrast enhancement, size normalization and color space conversion. The feature extraction module is used to extract feature information related to the eye from the preprocessed image using Canny edge detection. The eye detection and positioning module is used to identify and accurately locate the eye position on the surface of the potato based on the features extracted by Hough transform and YOLO deep learning target detection model. The classification module is used to classify the eyes according to the size and number of the eyes using a convolutional neural network. The decision output module is used to generate a final decision suggestion based on the classification result, and output a control signal to the sorting mechanism (3), the transfer mechanism (4), the feeding mechanism (6) and the pressing mechanism (8). The post-processing optimization module is used to evaluate and optimize the performance of the entire system using model hyperparameter adjustment. The data management and storage module is used to manage and store a large amount of image data and classification results. The user interaction module is used to provide a user-friendly interface.

Images