Automatic hawthorn sorting system and sorting method thereof

Abstract

The invention belongs to the technical field of automatic fruit sorting, and particularly relates to an automatic hawthorn fruit sorting system and a sorting method.The automatic hawthorn fruit sorting system comprises a feeding optimization mechanism, a conveying mechanism, a sorting mechanism and a sorting mechanism, and single-layer ordered arrangement of hawthorn fruits is achieved through cooperation of a v-shaped feeding disc and a double-brush array; the epidermis detection mechanism comprises a binocular vision system and an annular light supplementing device, analyzes quality indexes such as epidermis damage area and mildew degree based on a Yov11 instance segmentation network, and generates hierarchical codes; according to the conveying mechanism, conveying units are connected with all functional modules through a chain conveying belt composed of flexible conical rollers, and lossless conveying is achieved; and the grading executing mechanism is provided with a pair of cooperative working sliding plates which form different grade channels along a sliding rail cross beam, the fruits are sorted to three grade channels according to grading codes, and a system is integrated with a PLC control module to achieve full-process automation.

Description

Technical Field

[0001] The invention belongs to the technical field of agricultural product sorting, and particularly relates to an automated hawthorn sorting system and a sorting method thereof. Background Art

[0002] As people pursue a higher quality of life under the background of consumption upgrading, the market demand for hawthorn products continues to grow. The traditional manual sorting method relies on experience and judgment, and has problems such as inconsistent sorting standards and decreased work efficiency as working hours increase, which makes it difficult to meet the needs of large-scale production. Although the Chinese invention patent CN119368435A has achieved automated sorting, the traditional image processing process of grayscale-enhancement-segmentation it adopts has essential limitations: the linear processing method based on the two-dimensional pixel matrix can only identify macroscopic defects with a contrast ratio of more than 50% with the background, and cannot identify defects with a color similar to the epidermis (color difference ΔE<5) and an area less than 2mm 2 Minor damage cannot be effectively detected.

[0003] In comparison, visual inspection technology based on deep learning shows significant advantages. The present invention innovatively adopts the Yolov11 instance segmentation network model, and through an end-to-end feature learning mechanism, it independently constructs a multi-dimensional feature space including texture entropy, spectral reflectance, etc. Experimental verification shows that in tests simulating real production environments, the system's comprehensive recognition accuracy for defects such as mildew (recognition rate 99.2%), insect infestation (97.8%), and mechanical damage (96.5%) reached 98.7%, and the processing time for a single fruit was less than 120ms. In particular, by introducing a channel attention module to enhance the characteristic response of the damaged area, and combining near-infrared spectroscopy (900-1700nm) imaging technology to obtain epidermal sub-layer tissue information, early lesions that are difficult to detect with traditional RGB imaging can be effectively identified. Significantly reduce errors caused by manual parameter adjustment, especially when dealing with unstructured or highly variable sorting scenarios, showing stronger robustness and generalization capabilities. Summary of the invention

[0004] In view of the above technical problems, the purpose of the present invention is to propose an automated hawthorn sorting system and a sorting method thereof, which can arrange the piled hawthorn fruits in an orderly manner, realize a scientific evaluation of the surface quality of the hawthorns, safely classify the sorted hawthorns, and prevent the hawthorns from colliding during the sorting process through the coordinated guidance of flexible conical rollers, conveyor belts and sliding plates, thereby improving the efficiency and accuracy of hawthorn sorting.

[0005] In order to achieve the above object, the present invention provides an automated hawthorn sorting system, comprising:

[0006] A feeding optimization mechanism (1), which is installed above the rear end of the entire machine and is used to arrange the disordered hawthorns fed into an orderly single-fruit single-channel;

[0007] A conveying mechanism (3), wherein the conveying mechanism (3) is arranged directly below the feeding optimization mechanism (1), and conveys the hawthorns fed by the feeding optimization mechanism (1) to a subsequent sorting operation, and at the same time drives the hawthorns to rotate, and assists the skin detection mechanism (2) to complete a comprehensive photograph of the hawthorn skin;

[0008] The epidermis detection mechanism (2) is used to photograph the hawthorn fruit in the conveyor mechanism (3) through cameras (23) distributed on the left and right sides of the conveyor belt (41), and transmit the image information to the yolov11 instance segmentation network to detect and identify the surface defects of the hawthorn, and determine the grade of the hawthorn and write it into the database;

[0009] A grading mechanism (4), which is arranged at the end of the conveying mechanism (3), receives the hawthorns conveyed by the conveying mechanism (3), and classifies the hawthorns in order according to the judgment level written into the database by the epidermis detection mechanism (2);

[0010] The feeding optimization mechanism (1) comprises a feeding body (11), a brush shielding cover (12), and a brush (13); the feeding body (11) is a top-to-bottom oblique line type, and is fixedly installed above the right end of the overall system housing above the conveying mechanism (3), so as to guide the hawthorns to be sorted into the conveying mechanism (3); the brush shielding cover (12) is vertically installed above the right end of the conveying mechanism (3) along the left-right symmetrical axis of the machine, and is connected to the feeding body (11); a pair of brushes (13) are fixedly installed on the left and right sides of the brush shielding cover (12).

[0011] The conveying mechanism (3) comprises a support seat (31), a conical roller (32), a driving motor (33), a pulley (34), a transmission sprocket (35), a chain (36), and a rotating shaft (37); the support seat (31) is fixedly mounted on the front and rear ends of the conveying mechanism (3) in two pairs, and is connected to the support mechanism with the rotating shaft (37); the conical roller (32) and the chain (36), the transmission sprocket (35), and the rotating shaft (37) together form a conveyor belt (41) fixedly mounted in the middle of the left and right sides of the whole machine, and its taper is designed along the approximate circular shape of the hawthorn fruit, and the arrangement interval is the radius of the hawthorn fruit, and the hawthorn fruit is driven to rotate in the opposite direction of the conical roller (32) by friction with the hawthorn fruit skin; the driving motor (33) is connected to the pulley (34) and is fixedly mounted below the right end of the conveying mechanism (3) to provide power.

[0012] The epidermis detection mechanism (2) comprises a photographic cover support (21), a photographic cover (22), a camera (23), and a camera fixing screw (24); the photographic cover support (21) is fixedly mounted above the front end and the rear end of the support of the hawthorn sorting system; the photographic cover (22) is fixedly mounted above the front end and the rear end of the photographic cover support (21), and crosses the transverse baffle of the transmission unit; the camera fixing screw (24) is placed horizontally with the ground and vertically with the conveying mechanism (3), and two cameras (23) are respectively fixedly mounted on the front end inner wall and the rear end inner wall of the photographic cover (22). The cameras (23) are divided into two groups, the front end and the rear end, which are respectively fixed below the photographic cover (22) along the front and rear sides of the conveyor belt (41) and facing the conveyor belt (41). At the same time, the cameras (23) are connected to the controller through wires. When in use, the cameras (23) detect the image information of the hawthorn, and the image information is transmitted to the controller. The controller recognizes the defects of the hawthorn in the image through the Yolov11 deep learning algorithm and evaluates the grade at the same time, and writes it into the database.

[0013] The grading mechanism (4) comprises a conveyor belt (41), a collecting channel (42), a slider (43), a slide rail cross beam (44), a sliding plate (45), a separation channel (46), and a photoelectric sensor (47); the conveyor belt (41) is fixedly installed at the front end of the machine and is parallel to the upper surface of the machine, and drives the hawthorn fruit to move forward through friction; the collecting channel (42) is fixedly installed at a position between the conveying mechanism (3) and the grading mechanism (4), and transfers the hawthorn fruit to the grading mechanism (4); the sliding plate (45) is composed of two pieces and is connected to the slider (43), the slide rail cross beam (44), the sliding plate (45), the separation channel (46), and the photoelectric sensor (47); The beam (44) is connected to a photoelectric sensor (47), and the photoelectric sensor (47) is connected to a control system to detect whether hawthorn fruits pass through. The controller calculates the delay time of hawthorns arriving at the sorting port according to the conveyor belt speed, and dynamically adjusts the moving timing of the slider (43). The slider (43) moves left and right along the slide rail beam (44) according to the control system information to drive the sliding plate (45) to form different grade channels; the separation channel (46) is fixedly installed at the left end of the conveyor belt, and cooperates with the sliding plate (45) to guide hawthorns of different grades into the grading channels of corresponding grades.

[0014] An automated hawthorn sorting method comprises the following steps:

[0015] S1, hawthorn fruits enter the machine through the feeding optimization mechanism (1), and the disordered hawthorns are transformed into orderly single fruits and single channels through the optimization device by the brush (13);

[0016] S2, the camera (23) continuously photographs the hawthorns passing through the conveying mechanism (3), the conical roller (32) drives the hawthorns to flip, the auxiliary camera (23) uniformly photographs the hawthorn skin, and the photographed hawthorn pictures are uploaded to the image processing module;

[0017] S3, the image processing module conducts a comprehensive evaluation of the hawthorn epidermis through deep learning technology based on the Yolov11 instance segmentation network, compares the defect types and defect area information detected on the hawthorn epidermis with the pre-training data in the model, and analyzes the types and defect areas of different surface defects;

[0018] S4, segmenting the ratio of the sum of pixel coverage areas of different damage types to the total pixel area of ​​hawthorn according to Yolov11, the image system classifies different hawthorns according to different ratios, and enters the grade information of the tested hawthorns into the database according to the order of passing the epidermis detection mechanism (2);

[0019] S5, when the photoelectric sensor (47) detects that the hawthorn has arrived at the designated position, the controller sequentially controls the slider (43) to drive the sliding plate (45) to move horizontally along the slide rail crossbeam (44) to form a corresponding channel posture according to the hawthorn fruit grade information entered into the database, and cooperates with the operation of the conveyor belt to send hawthorns of different grades into the collection bin of the corresponding grade.

[0020] In step S4, the database stores information about each hawthorn, including the type of hawthorn epidermis defect, the area occupied by different defect types, and the order in which the epidermis detection mechanism (2) detects the hawthorn.

[0021] Compared with the prior art, the present invention has the following advantages:

[0022] 1. The present invention adjusts the disordered hawthorn fruits entering into a single channel by means of a feeding optimization mechanism (1), so that the skin detection mechanism (2) can detect the skin of the hawthorn fruits.

[0023] 2. The present invention uses the conical roller (32) in the conveying mechanism (3) to drive the single-fruit single-channel hawthorn fruits to rotate evenly, so that the camera (23) can capture the complete skin of the hawthorn fruits.

[0024] 3. The present invention detects the complete epidermis of hawthorn fruit by epidermis detection mechanism (2), performs reasonable and scientific evaluation on different defects of hawthorn fruit and different degrees of damage of defects, grades hawthorn fruit one by one according to the evaluation criteria, and writes them into a database so that a grading and classification mechanism can classify hawthorns of different grades according to the database.

[0025] 4. The grading mechanism (4) of the present invention ensures that hawthorn fruits accurately enter the collecting bin through the collecting channel (42), and two sliding plates (45) are installed in the collecting bin; the top slider (43) drives the sliding plate (45) to swing left and right along the slide rail beam (44) to guide hawthorn fruits of different grades to enter the corresponding separation channel (46), so as to efficiently complete the classification task without damaging or spoiling the fruits, and the classification result meets expectations.

[0026] 5. The present invention uses Yolov11 instance segmentation deep learning technology to accurately classify and scientifically evaluate the epidermal characteristics of hawthorn fruits, improve the rationality of grading, and ensure the quality of hawthorn grading. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a schematic diagram of the mechanism composition of the hawthorn sorting system of the present invention;

[0028] Figure 2 It is a lower oblique view of the feeding optimization mechanism in the hawthorn sorting system of the present invention;

[0029] Figure 3 A bottom view of the epidermis detection mechanism in the hawthorn sorting system of the present invention;

[0030] Figure 4 It is an upper oblique view of the conveying mechanism in the hawthorn sorting system of the present invention;

[0031] Figure 5 A top view of the grading mechanism in the hawthorn sorting system of the present invention;

[0032] Figure 6 It is the work flow chart of the present invention.

[0033] Legend: 1. Feeding optimization mechanism; 2. Surface detection mechanism; 3. Conveying mechanism; 4. Grading mechanism; 11. Feeding body; 12. Brush shielding cover; 13. Brush; 21. Photo cover support; 22. Photo cover; 23. Camera; 24. Camera fixing screw; 31. Support seat; 32. Conical roller; 33. Driving motor; 34. Pulley; 35. Drive sprocket; 36. Chain; 37. Rotating shaft; 41. Conveyor belt; 42. Collecting channel; 43. Sliding block; 44. Slide rail beam; 45. Sliding plate; 46. Separation channel; 47. Photoelectric sensor. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the examples of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. The described embodiments are only part of the embodiments of the present invention, not all of them. Based on the examples in the present invention, all other examples obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0035] In this application, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0036] Embodiment 1

[0037] like Figure 1-5 As shown, the present invention is to efficiently, accurately and automatically complete the grading of hawthorn fruits of different grades. This application takes hawthorn as an example for detailed description.

[0038] The machine as a whole includes: a feeding optimization mechanism 2, which is used to arrange the disordered hawthorns fed into an orderly single fruit single channel; please refer to Figure 2 The feeding optimization mechanism 2 includes a feeding body 11, a brush shielding cover 12, and a brush 13; the feeding body 11 is a slash type from top to bottom, and is fixedly installed above the right end of the overall system housing and above the conveying mechanism 3 to guide the hawthorns to be sorted into the conveying mechanism 3; the brush shielding cover 12 is fixedly installed above the right end of the conveying mechanism 3 and is installed perpendicular to the upper surface of the machine along the left-right symmetrical axis of the machine and is connected to the feeding body 11; a pair of brushes 13 are fixedly installed on the left and right sides of the brush shielding cover 12.

[0039] See also Figure 3 , the conveying mechanism 3 conveys the hawthorn fed by the feeding optimization mechanism 1 to the subsequent sorting operation, and at the same time drives the hawthorn fruit to rotate, assisting the skin detection mechanism 2 to complete the comprehensive shooting of the hawthorn fruit skin; the conveying mechanism 3 includes a support seat 31, a conical roller 32, a driving motor 33, a pulley 34, a transmission sprocket 35, a chain 36, and a rotating shaft 37; the support seat 31 is fixedly installed at the front and rear ends of the conveying mechanism 3 in two pairs, and the support mechanism is connected to the rotating shaft 37; the conical roller 32 and the chain 36, the transmission sprocket 35, and the rotating shaft 37 together form a conveyor belt fixedly installed in the middle of the left and right sides of the whole machine, and its taper is designed along the approximate circle of the hawthorn fruit shape, and the arrangement interval is the radius of the hawthorn fruit. The hawthorn fruit is driven to rotate in the opposite direction of the conical roller 32 by friction with the hawthorn fruit skin; the driving motor 33 is connected to the pulley 34 and fixedly installed at the bottom of the right end of the conveying mechanism 3 to provide power.

[0040] See also Figure 4The skin detection mechanism 2 takes pictures of the hawthorn fruits in the conveying mechanism 3 through the cameras 23 distributed on the left and right sides of the conveyor belt 41, and transmits the picture information to the Yolov11 instance segmentation network model to detect and identify the surface defects of the hawthorn, and judge the grade of the hawthorn and write it into the database; the skin detection mechanism 2 includes a photographic cover support 21, a photographic cover 22, a camera 23, and a camera fixing screw 24; the photographic cover support 21 is fixedly installed on the front end and the rear end of the bracket of the hawthorn sorting system; the photographic cover 22 is fixedly installed on the front end and the rear end of the photographic cover support 21, and crosses the conveying mechanism; the camera fixing screw 24 is placed horizontally with the ground and vertically with the conveying mechanism 3, and the two cameras 23 are fixedly installed on the front end inner wall and the rear end inner wall of the photographic cover 22 respectively. The camera 23 is divided into two groups, the front end and the rear end, which are respectively fixed under the photographic cover 22 along the front and rear sides of the conveyor belt 41 facing the conveyor belt 41. At the same time, the camera 23 is connected to the controller through a wire. When in use, the camera 23 detects the image information of the hawthorn, and the image information is transmitted to the controller. The controller recognizes the defects of the hawthorn in the image through the Yolov11 deep learning algorithm and evaluates the grade at the same time, and writes it into the database.

[0041] See also Figure 5 , a grading mechanism 4, which is arranged at the end of the conveying mechanism 3, receives the hawthorns conveyed by the conveying mechanism 3, and classifies the hawthorns in order according to the judgment level of the database written by the epidermis detection mechanism 2; it comprises a conveyor belt 41, a collecting channel 42, a slider 43, a slide rail beam 44, a sliding plate 45, a separation channel 46, and a photoelectric sensor 47; the conveyor belt 41 is fixedly installed at the front end of the machine and is parallel to the upper surface of the machine, and drives the hawthorn fruits forward by friction; the collecting channel is fixedly installed in the middle position between the conveying mechanism 3 and the grading mechanism 4, and the hawthorn fruits are transferred to the grading mechanism 4. Move to the grading mechanism 4; the two sliding plates 45 are connected to the slider 43, the slide rail beam 44, and the photoelectric sensor 47. The photoelectric sensor 47 is connected to the control system to detect whether there are hawthorn fruits passing through. The controller calculates the delay time for the hawthorn to reach the sorting port according to the conveyor belt speed, and dynamically adjusts the movement timing of the slider 43. The slider 43 moves left and right along the slide rail beam 44 according to the control system information, driving the sliding plate 45 to form different grade channels; the separation channel 46 is fixedly installed at the left end of the conveyor belt, and works with the sliding plate 45 to guide different grades of hawthorn into the corresponding graded grading channel 46.

[0042] Embodiment 2

[0043] See also Figure 6 Based on the first embodiment, this embodiment proposes an automated hawthorn sorting system and a sorting method thereof, comprising the following steps:

[0044] S1, hawthorn fruits enter the machine through the feeding optimization mechanism 1, and the disordered hawthorns are transformed into orderly single fruits and single channels through the optimization device by the brush 13.

[0045] S2, the camera 23 continuously photographs the hawthorns passing through the conveying mechanism 3, and the conical roller 32 drives the hawthorns to flip and the auxiliary camera 23 evenly photographs the hawthorn skin, and uploads the photographed hawthorn pictures to the image processing module.

[0046] S3, the image processing module conducts a comprehensive evaluation of the hawthorn epidermis through deep learning technology based on the Yolov11 instance segmentation network, compares the defect type and defect area information detected on the inspected hawthorn epidermis with the pre-training data in the model, and analyzes the type and defect area of ​​different surface defects.

[0047] S4, according to the ratio of the sum of pixel coverage areas of different damage types segmented by Yolov11 to the overall pixel area of ​​hawthorn, the image system classifies different hawthorns according to different ratios, and enters the grade information of the detected hawthorns into the database in the order of passing the epidermis detection mechanism 2.

[0048] S5, when the photoelectric sensor 47 detects that the hawthorn has arrived at the designated position, the controller controls the slider 43 in sequence according to the hawthorn fruit grade information entered into the database to drive the sliding plate 45 to swing left and right to form a corresponding channel posture, and cooperates with the operation of the conveyor belt 41 to send the hawthorns of different grades into the collection bins of the corresponding grades.

[0049] In step S4, the database stores information about each hawthorn, including the type of hawthorn epidermis defect, the area occupied by different defect types, and the order in which the epidermis detection mechanism 2 detects the hawthorn.

[0050] The process of hawthorn segmentation based on YOLOv11 instance segmentation is as follows: Image input and feature extraction: Input hawthorn RGB image, the network first locates and identifies hawthorn individuals in the RGB image one by one, and extracts multi-scale feature resistance through the backbone network (CSPDarknet) of YOLOv11; Instance segmentation is performed on the identified and located hawthorn individuals: Output the target bounding box (BBox), category probability (Class) and mask (Mask) through the detection head (Head); Defect analysis and classification: Calculate the pixel area R of the defect area defect , if R defect ≤5% is first-class fruit, 5%<R defect ≤15% is second-grade fruit, and the rest is defective fruit, triggering the grading mechanism to act. Key logic: The model realizes defect location, classification and mask generation through end-to-end training, combined with dynamic sorting of area thresholds to meet the needs of efficient automation.

[0051] This embodiment uses Yolov11 instance segmentation deep learning technology to accurately identify and evaluate the types of appearance defects of hawthorn fruits, improve the accuracy of recognition, and thus ensure the quality of sorting.

[0052] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An automated hawthorn sorting system, characterized in that: It comprises a feeding optimization mechanism (1), a conveying mechanism (3), a surface detection mechanism (2) and a grading mechanism (4); The feeding optimization mechanism (1) is installed above the rear end of the whole machine and is used to arrange the randomly fed hawthorns into orderly single fruit single channel; The conveying mechanism (3) is arranged directly below the feeding optimization mechanism (1), and conveys the hawthorns fed by the feeding optimization mechanism (1) to the subsequent sorting operation, and at the same time drives the hawthorns to rotate, and assists the skin detection mechanism (2) to complete a comprehensive photograph of the hawthorn fruit skin; The epidermis detection mechanism (2) comprises a photographic cover support (21), a photographic cover (22), a camera (23) and a camera fixing screw (24); the photographic cover support (21) is fixedly mounted above the front end and above the rear end of the support of the hawthorn sorting system; the photographic cover (22) is fixedly mounted above the front end and above the rear end of the photographic cover support (21) and crosses the transverse baffle of the transmission unit; the camera fixing screw (24) is placed horizontally on the ground and vertically on the conveying mechanism (3); two cameras (23) are respectively fixedly mounted on the front end inner wall and the rear end inner wall of the photographic cover (22); the cameras (23) are divided into two groups, the front end and the rear end, which are respectively fixed below the photographic cover (22) along the front and rear sides of the conveyor belt (41) and facing the conveyor belt (41); at the same time, the cameras (23) are connected to the controller via wires; when in use, the cameras (23) detect image information of the hawthorn, and the image information is transmitted to the controller; the controller identifies the defects of the hawthorn in the image through the Yolo v11 deep learning algorithm and simultaneously evaluates the grade and writes it into the database; The grading mechanism (4) is arranged at the end of the conveying mechanism (3), receives the hawthorns conveyed by the conveying mechanism (3), and classifies the hawthorns in order according to the judgment level written into the database by the epidermis detection mechanism (2).

2. The automated hawthorn sorting system according to claim 1, characterized in that: The feeding optimization mechanism (1) comprises a feeding body (11), a brush shielding cover (12), and a brush (13); the feeding body (11) is a top-to-bottom oblique line type, and is fixedly mounted above the right end of the overall system housing and above the conveying mechanism (3), so as to guide the hawthorns to be sorted into the conveying mechanism (3); the brush shielding cover (12) is vertically mounted above the right end of the conveying mechanism (3) along the left-right symmetrical axis of the machine, and is connected to the feeding body (11); a pair of brushes (13) are fixedly mounted on the left and right sides of the brush shielding cover (12).

3. The automated hawthorn sorting system according to claim 1, characterized in that: The conveying mechanism (3) comprises a support seat (31), a conical roller (32), a driving motor (33), a pulley (34), a transmission sprocket (35), a chain (36) and a rotating shaft (37). The support seats (31) are fixedly mounted in two pairs at the front and rear ends of the conveying mechanism (3) and connected to the support mechanism with a rotating shaft (37); the conical roller (32), the chain (36), the transmission sprocket (35) and the rotating shaft (37) together form a conveyor belt (41) fixedly mounted in the middle of the left and right sides of the whole machine, and its taper is designed along the approximate circular shape of the hawthorn fruit, and the arrangement interval is the radius of the hawthorn fruit. The hawthorn fruit is driven to rotate in the opposite direction of the conical roller (32) by friction with the hawthorn fruit skin; the driving motor (33) is connected to the pulley (34) and fixedly mounted below the right end of the conveying mechanism (3) to provide power.

4. The automated hawthorn sorting system according to claim 1, characterized in that: The grading mechanism (4) comprises a conveyor belt (41), a collecting channel (42), a slider (43), a slide rail cross beam (44), a sliding plate (45), a separation channel (46) and a photoelectric sensor (47); the conveyor belt (41) is fixedly installed at the front end of the machine and is parallel to the upper surface of the machine, and drives the hawthorn fruit to move forward through friction; the collecting channel (42) is fixedly installed at a position between the conveying mechanism (3) and the grading mechanism (4), and transfers the hawthorn fruit to the grading mechanism (4); the sliding plate (45) is composed of two pieces and is connected to the slider (43), the slide rail cross beam (44), the sliding plate (45), the separation channel (46) and the photoelectric sensor (47); The beam (44) is connected to a photoelectric sensor (47), and the photoelectric sensor (47) is connected to a control system to detect whether hawthorn fruits pass through. The controller calculates the delay time of hawthorns arriving at the sorting port according to the conveyor belt speed, and dynamically adjusts the moving timing of the slider (43). The slider (43) moves left and right along the slide rail beam (44) according to the control system information to drive the sliding plate (45) to form different grade channels; the separation channel (46) is fixedly installed at the left end of the conveyor belt, and cooperates with the sliding plate (45) to guide hawthorns of different grades into the grading channels of corresponding grades.

5. An automated hawthorn sorting method, characterized in that: The method comprises the following steps: S1, hawthorn fruits are fed into the machine through a feeding optimization mechanism (1), and disordered hawthorn fruits are transformed into orderly single fruit single channel through an optimization device by a brush (13); S2, the camera (23) continuously photographs the hawthorns passing through the conveying mechanism (3), the conical roller (32) drives the hawthorns to flip, the auxiliary camera (23) uniformly photographs the hawthorn skin, and the photographed hawthorn pictures are uploaded to the image processing module; S3, the image processing module conducts a comprehensive evaluation of the hawthorn epidermis through deep learning technology based on the Yolov11 instance segmentation network, compares the defect types and defect area information detected on the hawthorn epidermis with the pre-training data in the model, and analyzes the types and defect areas of different surface defects; S4, segmenting the ratio of the sum of pixel coverage areas of different damage types to the total pixel area of ​​hawthorn according to Yolov11, the image system classifies different hawthorns according to different ratios, and enters the grade information of the tested hawthorns into the database according to the order of passing the epidermis detection mechanism (2); S5, when the photoelectric sensor (47) detects that the hawthorn has arrived at the designated position, the controller sequentially controls the slider (43) to drive the sliding plate (45) to swing left and right to form a corresponding channel posture according to the hawthorn fruit grade information entered into the database, and cooperates with the operation of the conveyor belt (41) to send hawthorns of different grades into the collection bin of the corresponding grade.

6. The automated hawthorn sorting method according to claim 5, characterized in that: In S1, the feeding optimization mechanism (1) uses a brush (13) that coincides with the vertical center plane of the conveying mechanism (3) and is arranged in front and behind to optimize the disordered hawthorns into a single fruit and a single channel.

7. The automated hawthorn sorting method according to claim 5, characterized in that: In S2, the cameras (23) are fixedly installed along both sides of the conveying mechanism (3), and the conical roller (32) drives the hawthorn to rotate in the opposite direction of the conical roller (32). The cameras (23) on both sides take three photos in total, for a total of six photos.

8. The automated hawthorn sorting method according to claim 5, characterized in that: In step S3, the YOLOv11 instance segmentation model locates the hawthorn in the image uploaded by the camera (23) in turn, segments the hawthorn individuals, and simultaneously segments the hawthorn as a whole, the skin defects, the fruit stalks and the fruit pedicle area, and calculates the number of pixels occupied by the segmented hawthorn as a whole, the hawthorn defects, the hawthorn fruit stalks, and the hawthorn pedicles.

9. The automated hawthorn sorting method according to claim 6, characterized in that: In step S5, the photoelectric sensor (47) is connected to the control system to detect whether hawthorn fruits pass through. The controller calculates the delay time of the hawthorns reaching the sorting port according to the speed of the conveyor belt (41) and dynamically adjusts the movement timing of the slider (43). The slider (43) moves left and right along the slide rail beam (44) according to the control system information to drive the sliding plate (45) to form different levels of channels.

10. The automated hawthorn sorting method according to claim 8, characterized in that: Hawthorn defects include at least one of disease and insect pest defects, bruises and rot.

Images