Garlic sorting system based on machine vision

Abstract

The utility model belongs to the technical field of garlic sorting, and particularly relates to a garlic sorting system based on machine vision, which comprises a baffle plate, a conveyor belt and a garlic impurity removal device, the conveyor belt is assembled on the inner side of the baffle plate, a first rack is assembled at the bottom of the baffle plate, and a discharge port platform is arranged on one side of the garlic impurity removal device. A lifting mechanism is arranged in the discharging port platform, a plurality of trays are assembled on the conveying belt, check blocks are assembled on one sides of the trays, a visual recognition module is connected to the position, located in the middle section, of the top of the conveying belt, and a push rod device is arranged on one side of the garlic impurity removing device. And an annular high-angle light and a CMOS camera are arranged above the visual identification module and are used for collecting images and sending the collected images to a local server for processing. Garlic sorting can be achieved through a simple algorithm, and meanwhile the garlic sorting device has the advantages of being simple in structure, small in sorting strength, high in sorting efficiency and the like.

Description

TECHNICAL FIELD

[0001] The utility model belongs to garlic sorting technical field, specifically relates to a kind of garlic sorting system based on machine vision. BACKGROUND

[0002] As an important economic crop, garlic has long relied on manual operation for sorting and grading. However, manual sorting is not only inefficient, but also easily affected by subjective factors, resulting in low sorting accuracy and affecting the overall quality and market competitiveness of garlic. Therefore, there is an urgent need for an automated garlic sorting system to improve sorting efficiency and accuracy.

[0003] Chinese patent with authorization announcement number CN103471514B discloses a garlic grading method based on machine vision and linear regression analysis, which collects and processes images of garlic, uses linear regression analysis for mathematical modeling based on the area and horizontal diameter data of the collected garlic images, and finally grades the pre-graded garlic according to the fitted linear equation.

[0004] Although the above-mentioned patent can perform garlic grading, the original algorithm requires multiple steps of processing on the collected images, which may not achieve real-time sorting, resulting in difficulty in meeting efficiency requirements in high-demand production environments. Moreover, the original algorithm may not perform stably under conditions such as changes in lighting, background environment interference, and presence of damaged garlic, leading to a decrease in sorting efficiency. In addition, traditional sorting equipment has a complex structure and high sorting intensity, which can easily damage the surface of garlic during sorting and discharging, affecting the quality of sorted garlic. CONTENT OF THE UTILITY MODEL

[0005] The utility model aims to provide a garlic sorting system based on machine vision, which can achieve sorting of garlic through a simple algorithm, while having the characteristics of simple structure, low sorting intensity, and high sorting efficiency.

[0006] The technical solution adopted by the utility model is as follows:

[0007] A garlic sorting system based on machine vision includes a baffle, a conveyor belt, and a garlic impurity removal device. The conveyor belt is assembled on the inner side of the baffle. The baffle bottom is equipped with a first rack. The garlic impurity removal device is provided with a discharge port platform on one side. The discharge port platform is provided with a lifting mechanism. The conveyor belt is equipped with multiple trays. The trays are equipped with stop blocks on one side. The conveyor belt top is connected with a vision recognition module at the middle segment position. The garlic impurity removal device is provided with a push rod device on one side.

[0008] The vision recognition module is provided with a ring-shaped high-angle light and a CMOS camera above, for image acquisition. The collected images are sent to a local server for processing.

[0009] The lifting mechanism is at the lowest position and is flush with the discharge port platform, and the lifting mechanism is at the highest position and is aligned with the starting end of the conveying belt.

[0010] Three storage baskets are arranged at positions corresponding to the push rod device on one side of the conveying belt.

[0011] The garlic impurity removing device comprises a second rack for support, an outer shell is connected to the top of the second rack, an inlet pipe is fixedly connected to the side wall of the outer shell near the top, the back of the outer shell is connected to a discharge port platform near the bottom, an air inlet is formed in the side of the outer shell, and the air inlet is used to connect an air compressor.

[0012] A screen is connected between the discharge port platform and the inlet pipe, and the screen is arranged in an inclined manner.

[0013] The lifting mechanism comprises a base and a top inclined platform, two scissor links are movably connected between the base and the top inclined platform, the scissor links on the same side are hingedly connected, a pin shaft is rotatably connected between the scissor links on the two sides, a servo motor is fixedly connected to the top surface of the base, a lead screw is fixedly connected to the output end of the servo motor, a sliding block is threadedly connected to the side wall of the lead screw, and the sliding block is movably connected with the pin shaft and used to push the top inclined platform to ascend or descend.

[0014] The tray comprises a top plate and a bottom plate, at least eight springs are assembled between the top plate and the bottom plate, the top of the top plate is a concave curved surface, and a through hole is formed in the center of the top of the top plate.

[0015] The push rod device comprises a rudder motor fixedly connected to the side wall of the baffle, a connecting rod is fixedly connected to the output end of the rudder motor, a push rod head is hingedly connected to the other end of the connecting rod, and the push rod head corresponds to the storage basket.

[0016] A speed regulating motor is mounted on the side wall of the first rack, a driving wheel is fixedly connected to the output end of the speed regulating motor, a driven wheel is connected to the roller shaft of the conveying belt, and a belt is transmissionally connected between the driving wheel and the driven wheel.

[0017] The utility model achieves the following technical effects:

[0018] The garlic sorting system based on machine vision of the utility model calculates the width and height of each garlic frame, selects the maximum side as the horizontal diameter of the garlic, converts the pixel value into millimeter value by using the known proportional relationship between the pixel and millimeter, and has the characteristics of simple algorithm and high precision.

[0019] The garlic sorting system based on machine vision of the utility model adopts the tray to store and sort garlic, avoids the drawback that the traditional equipment directly pushes out garlic, reduces the damage to the surface of garlic; in addition, the garlic impurity removing device can effectively clean the impurities on the surface of garlic, further improves the quality of sorted garlic. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 The system overall flow chart of the utility model.

[0021] Figure 2 The garlic grading mechanism perspective view of the utility model.

[0022] Figure 3 The side view of Figure 2

[0023] Figure 4 The impurity removing device perspective view of the utility model.

[0024] Figure 5 The screen perspective view of the utility model.

[0025] Figure 6 The lifting mechanism schematic view of the utility model.

[0026] Figure 7 The lifting mechanism descending work schematic view of the utility model.

[0027] Figure 8 The lifting mechanism ascending work schematic view of the utility model.

[0028] Figure 9 The tray overall structure front view schematic view of the utility model.

[0029] Figure 10 The isometric schematic view of Figure 9

[0030] Figure 11 The push rod mechanism schematic view of the utility model.

[0031] In the drawings, the component list represented by each sign is as follows:

[0032] ​​1. Discharge port platform; 2. Feed pipe; 201. Outer shell; 202. Lifting mechanism; 203. Conveyor belt; 204. Stop block; 205. Vision recognition module; 206. Pallet; 207. Push rod device; 208. Storage basket; 209. Baffle; 210. First frame; 211. Driven wheel; 212. Speed ​​regulating motor; 301. Screen; 302. Air compressor; 303. Second frame; 401. Top inclined platform; 402. Scissor linkage; 403. Pin; 404. Servo motor; 405. Lead screw; 406. Slider; 407. Base; 701. Top plate; 702. Bottom plate; 703. Spring; 704. Through hole; 801. Connecting rod; 802. Servo motor; 803. Push rod head. Detailed Implementation

[0033] To make the purpose and advantages of this utility model clearer, the following detailed description is provided in conjunction with embodiments. It should be understood that the following text is merely used to describe one or more specific implementations of this utility model and does not strictly limit the scope of protection specifically claimed by this utility model.

[0034] like Figures 1-11 As shown, a garlic sorting system based on machine vision includes a baffle 209, a conveyor belt 203, and a garlic impurity removal device. The conveyor belt 203 is mounted inside the baffle 209. A first frame 210 is mounted at the bottom of the baffle 209. A discharge port platform 1 is provided on one side of the garlic impurity removal device. A lifting mechanism 202 is provided inside the discharge port platform 1. Multiple trays 206 are mounted on the conveyor belt 203. A stop block 204 is mounted on one side of the tray 206. A vision recognition module 205 is connected to the top of the conveyor belt 203 at the middle position. A push rod device 207 is provided on one side of the garlic impurity removal device.

[0035] The visual recognition module 205 is equipped with a ring-shaped high-angle light and a CMOS camera for image acquisition, and the acquired images are sent to the local server for processing.

[0036] When the lifting mechanism 202 is at its lowest point, it is flush with the discharge port platform 1; when the lifting mechanism 202 is at its highest point, it is aligned with the starting end of the conveyor belt 203.

[0037] Three storage baskets 208 are provided on one side of the conveyor belt 203 at the position corresponding to the push rod device 207.

[0038] In the visual recognition part, a CMOS camera was selected, and the lighting method adopted was uniform diffused ring light. The training framework based on YOLOv10 was used.

[0039] A trained two-class object detection model is used, with class 1 being whole garlic and class 2 being damaged parts of garlic. During the inference stage, the model outputs bounding boxes and their confidence scores for all garlic and damaged parts in the image. To improve the accuracy of the detection, we set the confidence threshold to 0.15 to ensure that as many minor damaged parts as possible are detected.

[0040] Next, we store the garlic boxes and damage boxes separately through a post-processing step, and for each garlic box, we determine whether there is an included damage box. We iterate through each garlic box and check if any damage box is completely contained within the garlic box. If such a containment relationship exists, the garlic is judged to be "damaged garlic", otherwise it is judged to be "normal garlic". Through this spatial containment relationship judgment logic, normal garlic and damaged garlic can be accurately distinguished.

[0041] In addition, in order to accurately measure the cross-diameter size of irregularly shaped garlic, we calculate the width and height of each garlic box, select the maximum side as the cross-diameter of the garlic, and use the known pixel-to-millimeter conversion relationship to convert the pixel value to millimeter value.

[0042] A known size object is used to measure its width in pixels in the image. The following formula is used to calculate the proportion:

[0043] Proportion (mm / px) = Known size (mm) / Known size width in image (px)

[0044] After obtaining the proportion, it is used to convert the detected pixel size to millimeters:

[0045] Diameter (mm) = Diameter (px) x Proportion (mm / px)

[0046] Based on the converted size, we divide the garlic into different categories: those with a cross-diameter of 65 mm or more are classified as category 0, those with a cross-diameter of 60 mm to 65 mm are classified as category 1, and so on, and those with a cross-diameter of 45 mm or less are classified as category 5. For damaged garlic, regardless of its size, it is classified as category 6.

[0047] As shown in Figure 4 and 5 , the garlic impurity removal device includes a second rack 303 for support, and a housing 201 connected to the top of the second rack 303. The side wall of the housing 201 is fixedly connected with an inlet pipe 2 near the top. The back of the housing 201 is connected with an outlet platform 1 near the bottom. An air inlet is formed on the side of the housing 201, which is used to connect an air compressor 302.

[0048] A screen 301 is connected between the outlet platform 1 and the inlet pipe 2, and the screen 301 is inclined.

[0049] Wherein, the garlic from the inlet pipe 2 into the screen 301, under the action of gravity along the screen 301 rolling, in this rolling process, the friction between the garlic and the screen 301 will make the surface gradually loose impurities, air compressor 302 from the side of the high pressure gas injection impact the surface of garlic, so that the skin, soil and other impurities can be quickly peeled from the surface of garlic, then through the screen 301 mesh hole discharge, avoid into the conveyor belt 203 damage the mechanical device, the cleaned garlic is discharged from the discharge port platform 1, provide clean garlic raw material for subsequent sorting process.

[0050] As shown in Figure 6 , Figure 7 and Figure 8 , the lifting mechanism 202 includes a base 407 and a top inclined platform 401, two scissor links 402 are movably connected between the base 407 and the top inclined platform 401, the same side scissor links 402 are hinged, the pin shaft 403 is rotatably connected between the two sides scissor links 402, the servo motor 404 is fixedly connected on the top surface of the base 407, the lead screw 405 is fixedly connected on the output end of the servo motor 404, the sliding block 406 is threadedly connected on the side wall of the lead screw 405, the sliding block 406 is movably connected with the pin shaft 403, for pushing the top inclined platform 401 up and down.

[0051] Specifically, the servo motor 404 at the bottom of the lifting mechanism 202 drives the lead screw 405 to rotate, the rotation of the lead screw 405 will be converted into the translation motion of the sliding block 406, the translation motion of the sliding block 406 further drives the scissor links 402 to increase and decrease the included angle, thereby realizing the lifting function of the top inclined platform 401. Before the garlic enters the conveyor belt 203, the lifting mechanism 202 serves as a transition link between the conveyor belt 203 and the discharge port platform 1, which utilizes the height difference between the conveyor belt 203 and the discharge port platform 1 to avoid the accumulation of garlic on the conveyor belt 203, when the lifting mechanism 202 is lowered, it can receive the garlic, since its top width can only accommodate a single row of garlic, so when the excess garlic slides into the discharge port platform 1 during the lifting process of the lifting mechanism 202. In this way, it can ensure that the garlic sent to the conveyor belt 203 is arranged in a single row, laying a foundation for subsequent stable transportation and accurate identification.

[0052] As shown in Figure 9 and 10 , the tray 206 includes a top plate 701 and a bottom plate 702, at least eight springs 703 are assembled between the top plate 701 and the bottom plate 702, the top of the top plate 701 is a concave curved surface, a through hole 704 is formed in the center of the top of the top plate 701.

[0053] When the garlic is sent to the conveying belt 203 by the lifting mechanism 202, the tray 206 will receive the garlic, and the top plate 701 and the bottom plate 702 of the tray 206 are connected by the spring 703. This connection mode can make the garlic vibrate constantly after sliding into the tray 206, which can effectively slow down the impact speed of the garlic, promote the rolling of the garlic, and ensure that the roots or bulbs of the garlic can pass through the through hole 704, so as to realize the stable fixing of the garlic on the tray 206. From the perspective of the top view, the garlic is in a stable state, and the maximum diameter is clearly displayed, which is beneficial to the accurate feature recognition of the garlic by the subsequent image recognition mechanism.

[0054] As shown in Figure 11 , the push rod device 207 includes a steering engine 802 fixedly connected to the side wall of the baffle 209, and the output end of the steering engine 802 is fixedly connected with a connecting rod 801, and the other end of the connecting rod 801 is hinged with a push rod head 803 corresponding to the storage basket 208. In the last link of the whole sorting process, the push rod device 207 waits for the corresponding garlic to arrive according to the received signal, and the steering engine 802 rotates to drive the connecting rod 801 to move, so that the push rod head 803 moves translationally to impact the garlic and push it into the corresponding specification storage basket 208, thereby completing the sorting work of the garlic. The push rod head 803 is soft to reduce the damage to the garlic.

[0055] As shown in Figure 2 and 3 , the first rack 210 side wall is provided with a speed regulating motor 212, the output end of the speed regulating motor 212 is fixedly connected with a driving wheel, the roller shaft of the conveying belt 203 is connected with a driven wheel 211, and the driving wheel and the driven wheel 211 are transmissionally connected with a belt.

[0056] The first rack 210 provides a support structure for the whole conveying belt 203. The first rack 210 is provided with a speed regulating motor 212 at the edge, the speed regulating motor 212 is connected to drive the driving wheel to rotate, the driving wheel drives the driven wheel 211 through the belt to drive the conveying belt 203, thereby providing a power source for the transportation of the garlic on the conveying belt 203, and adjusting the speed of the conveying belt 203 to adapt to the visual identification speed. The baffle 209 is arranged on one side of the conveying belt 203, which can prevent the garlic from falling off during transportation. Three storage baskets 208 are arranged at the rear end of the visual identification module 205 and close to the end of the conveying belt 203, which provides a classified storage position for the subsequent sorted garlic.

[0057] As shown in Figure 1As shown, it is the system overall flow chart of the utility model. The utility model works as follows: start air compressor 302, conveying belt 203, speed regulation motor 212, lifting mechanism 202, servo motor 404 and other related equipment when using, place the garlic to be sorted at inlet pipe 2, after the garlic is impurity-removed by garlic impurity-removing device, enter discharge port platform 1, by lifting mechanism 202, send the garlic to tray 206 on conveying belt 203 to realize the righting, conveying belt 203 transports the garlic to visual identification module 205 to carry out image acquisition and processing, determine the grade of garlic, and send instruction signal to corresponding push rod device 207, push rod device 207 is on standby, when the garlic is sent to the corresponding position, rudder 802 drives push rod head 803 to push the garlic into storage basket 208, complete sorting.

[0058] The above only describes the preferred embodiments of the utility model, and it should be pointed out that for ordinary skilled persons in the art, without departing from the principle of the utility model, a number of improvements and refinements can be made, and these improvements and refinements should be considered as the protection scope of the utility model. The structures, devices and operation methods not specifically described and explained in the utility model are implemented according to the conventional means in the art without special description and limitation.

Claims

1. A machine vision based garlic sorting system, characterized by: Including baffle (209), conveying belt (203) and garlic impurity removal device, the conveying belt (209) is assembled in the inside of baffle (209), the bottom of baffle (209) is equipped with first rack (210), one side of the garlic impurity removal device is provided with discharge port platform (1), the discharge port platform (1) is provided with lifting mechanism (202), the conveying belt (203) is equipped with a plurality of trays (206), one side of the tray (206) is equipped with a stop block (204), the top of conveying belt (203) is connected with visual identification module (205) in the middle segment position, one side of the garlic impurity removal device is provided with push rod device (207). The visual identification module (205) is provided with annular high-angle light and CMOS camera above, for image acquisition, and the collected image is sent to a local server for processing.

2. The garlic sorting system based on machine vision according to claim 1, characterized in that: When the lifting mechanism (202) stroke is at the lowest, it is flush with the discharge port platform (1), and when the lifting mechanism (202) stroke is at the highest, it is aligned with the starting end of the conveying belt (203).

3. The garlic sorting system based on machine vision according to claim 1, characterized in that: The conveying belt (203) is provided with three storage baskets (208) at the position corresponding to the push rod device (207) on one side.

4. The garlic sorting system based on machine vision according to claim 1, characterized in that: The garlic impurity removal device comprises a second rack (303) for supporting, the top of the second rack (303) is connected with a shell (201), the sidewall of the shell (201) is fixedly connected with an inlet pipe (2) near the top, the back of the shell (201) is connected with the discharge port platform (1) near the bottom, the side of the shell (201) is provided with an air inlet, and the air inlet is used to connect an air compressor (302).

5. A machine vision-based garlic sorting system according to claim 4, characterized in that: The discharge port platform (1) and the inlet pipe (2) are connected with a screen (301), and the screen (301) is inclined.

6. The machine vision-based garlic sorting system according to claim 1, wherein: The lifting mechanism (202) comprises a base (407) and a top inclined surface table (401), two shear type connecting rods (402) are movably connected between the base (407) and the top inclined surface table (401), the shear type connecting rods (402) on the same side are hinged, a pin shaft (403) is rotatably connected between the shear type connecting rods (402) on the two sides, a servo motor (404) is fixedly connected to the top surface of the base (407), a lead screw (405) is fixedly connected to the output end of the servo motor (404), a sliding block (406) is threadedly connected to the sidewall of the lead screw (405), and the sliding block (406) is movably connected with the pin shaft (403) and used to push the top inclined surface table (401) up and down.

7. The machine vision-based garlic sorting system according to claim 1, wherein: The tray (206) comprises a top plate (701) and a bottom plate (702), at least eight springs (703) are assembled between the top plate (701) and the bottom plate (702), the top of the top plate (701) is a concave curved surface, and a through hole (704) is formed in the center of the top of the top plate (701).

8. The machine vision-based garlic sorting system according to claim 1, wherein: The push rod device (207) comprises a steering engine (802) fixedly connected to the side wall of the baffle (209), one end of the steering engine (802) is fixedly connected with a connecting rod (801), and the other end of the connecting rod (801) is hingedly connected with a push rod head (803); the push rod head (803) corresponds to the storage basket (208).

9. The machine vision-based garlic sorting system according to claim 4, wherein: A speed regulating motor (212) is mounted on the side wall of the first rack (210), a driving wheel is fixedly connected to the output end of the speed regulating motor (212), a driven wheel (211) is connected to the roller shaft of the conveying belt (203), and a belt is transmissionally connected between the driving wheel and the driven wheel (211).

Images