Agricultural product sorting system and method based on artificial intelligence

Abstract

The invention discloses an agricultural product sorting system and method based on artificial intelligence, and relates to the technical field of agricultural product sorting, the agricultural product sorting system comprises a conveying mechanism and an intelligent detection mechanism, the intelligent detection mechanism is arranged above the front end of the conveying mechanism, the appearance of agricultural products is recognized through a machine vision technology, and the quality of the agricultural products is judged; a material guiding mechanism and a defective product removing mechanism; and the sorting mechanism is arranged at the tail end of the conveying mechanism and is used for receiving the good products conveyed by the conveying mechanism. The appearance of the agricultural products is detected and identified through the intelligent detection mechanism, after inferior-quality products in the agricultural products are identified, the inferior-quality products are pushed into the inferior-quality product guiding-out hopper through the inferior-quality product removing mechanism and fall into the inferior-quality product collecting box to be collected, the non-defective products enter the sorting mechanism, and the agricultural products are classified and collected according to the installation diameter sizes through the two sorting rods; subsequent classified sales is facilitated; the system does not need manual operation, automatic sorting is achieved in the whole process, and the sorting efficiency is greatly improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of agricultural product sorting, and in particular relates to an agricultural product sorting system and method based on artificial intelligence. Background Art

[0002] With the acceleration of agricultural modernization, the output and variety of agricultural products are becoming increasingly rich. Traditional agricultural product sorting mainly relies on manual labor, and sorters judge and classify agricultural products based on their experience, such as appearance, size, maturity, etc. This method is inefficient and difficult to meet the needs of large-scale production. Long working hours can easily lead to a decrease in sorting accuracy.

[0003] After searching, a Chinese patent (CN111359887B) discloses a green agricultural product collection, sorting and processing system and a sorting method thereof. The collection, sorting and processing system includes a bottom plate, a storage frame, a conveying pipe, a sorting device and a collection frame. A storage frame is installed at the left end of the bottom plate, a feeding port is opened at the lower end of the storage frame, a conveying pipe is installed at the lower end of the feeding port, the lower end of the conveying pipe is in close contact with the upper end of the sorting device, the sorting device is installed on the bottom plate, and a collection frame is placed in the sorting device. The sorting method includes the following steps: equipment adjustment, feeding, sorting, repeated sorting and discharging. The invention sorts fruits and melons according to their sizes through a sorting device, and the sorted fruits and melons fall into the corresponding collection frame, and the collection frame can play a certain buffering and guiding role on the fruits and melons to prevent the fruits and melons from being damaged due to collisions, and the sorting device can adaptively adjust the size of the sorting port according to the size of the fruits and melons to be sorted, so that the function of sorting a variety of fruits and melons can be achieved using one device.

[0004] However, the invention cannot accurately identify agricultural products, and the sorting misjudgment rate is high when faced with complex agricultural product shapes, surface defects, and different lighting conditions. Therefore, it is of great significance to develop an agricultural product sorting system and method based on artificial intelligence, which can greatly improve sorting efficiency and accuracy, reduce labor costs, and promote the intelligent upgrading of the agricultural product processing industry. Summary of the invention

[0005] The purpose of the present invention is to provide an agricultural product sorting system and method based on artificial intelligence to address the problems of low efficiency and accuracy of existing agricultural product sorting.

[0006] The present invention is achieved through the following technical solutions: The AI-based agricultural product sorting system includes: A conveying mechanism, which is arranged at the upper end of the conveying support and is used to convey the picked agricultural products for sorting operation; An intelligent detection mechanism, which is arranged above the front end of the conveying mechanism and uses machine vision technology to identify the appearance of agricultural products and judge the quality of agricultural products; A material guiding mechanism, which is arranged above the conveying mechanism and is used to arrange the conveyed agricultural products in order for transportation; A defective product rejection mechanism, which is arranged above the conveying mechanism and behind the intelligent detection mechanism, and is used to reject the detected defective products; A sorting mechanism, which is arranged at the end of the conveying mechanism, receives good products conveyed by the conveying mechanism, and sorts the good products according to their installation size; A controller is installed on the conveying bracket and is used to control the operation of each mechanism.

[0007] Preferably, the conveying mechanism includes a conveyor belt, a driving roller, a driving motor and a supporting shaft; both ends of the conveyor belt are rotatably mounted on the upper end of the conveying bracket through the two driving rollers, the driving motor is mounted on the outside of the conveying bracket, and its output end is connected to one of the driving rollers, and the transmission belt is driven to rotate by the driving motor to realize the transmission of agricultural products; a plurality of supporting shafts are also arranged between the two driving rollers, and the supporting shafts are rotatably connected to the inner wall of the upper end of the conveying bracket.

[0008] Preferably, the intelligent detection mechanism includes a detection bracket, a detection camera mounting bracket and a detection camera, the detection bracket is L-shaped, the lower end of the detection bracket is fixedly mounted on the conveying bracket, and the detection camera is fixed to the upper end of the detection bracket through the detection camera mounting bracket; at the same time, the detection camera is connected to the controller through a wire.

[0009] When in use, the detection camera captures the appearance image of the agricultural products and transmits the data to the controller. The controller identifies the quality of the agricultural products through an algorithm. If they are defective, the controller sends instructions to the defective rejection mechanism to make it work and reject the defective products.

[0010] Preferably, there are two groups of defective product rejection mechanisms, which are arranged one in front and one behind above the conveyor belt; the defective product rejection mechanism includes a rejection bracket, a crossbeam, a rejection motor, a roller and a rejection plate, the lower end of the rejection bracket is fixedly installed on the rear side of the conveyor bracket, the crossbeam is fixedly connected to the upper end of the rejection bracket, the rejection motor is fixedly installed on one side of the crossbeam, and the output end of the rejection motor passes through the crossbeam and is connected to the roller, and a plurality of the rejection plates are evenly arranged on the outer wall of the roller; in the present application, there are three rejection plates.

[0011] When in use, the controller sends out a command, and the reject motor rotates to drive the roller and the reject plate to rotate, and the reject plate is used to push the defective products conveyed from the conveyor belt.

[0012] Preferably, there are three groups of material guiding mechanisms, which are respectively arranged at the front end of the conveyor belt, the rear end of the conveyor belt, and the conveyor belt is located between the two groups of defective product rejection mechanisms; The material guiding mechanism comprises a fixed material guiding plate, a movable material guiding plate, a support plate and a U-shaped connecting frame; the outer side wall of the fixed material guiding plate is fixedly connected to the end portion of the support plate, and the support plate is provided with a rectangular through groove near one end portion of the fixed material guiding plate, and a plurality of mounting holes are provided at one end portion away from the fixed material guiding plate. During installation, the outer side wall of the movable material guiding plate is fixedly connected to one end of the U-shaped connecting frame by bolts passing through the mounting holes, and the other end of the U-shaped connecting frame crosses the fixed material guiding plate and is inserted into the rectangular through groove. The support plate is provided with a fixed bottom plate at the lower end of the rectangular through groove, and the fixed bottom plate is connected to the end portion of the U-shaped connecting frame by bolts, so that the U-shaped connecting frame is fixed to the support plate. At the same time, by adjusting the fixed position of the U-shaped connecting frame in the rectangular through groove, the spacing between the movable material guiding plate and the fixed material guiding plate can be changed, thereby meeting the material guiding requirements of agricultural products of different sizes.

[0013] Preferably, it also includes two defective product guide buckets arranged on the side of the conveying bracket, the two defective product guide buckets are flush with the positions of the two defective product rejection mechanisms, and the defective product guide buckets are installed on the side of the conveying bracket by bolts, and a defective product collection box is arranged below the discharge port of the defective product guide bucket; the side panel of the conveying bracket is provided with a notch at the connection with the defective product guide bucket, so that the rejected defective products can roll smoothly into the defective product guide bucket.

[0014] Preferably, the sorting mechanism comprises a sorting bracket and a sorting plate, the sorting plate is arranged obliquely, and the two ends of the sorting plate are fixed to the bottom surface by the sorting bracket, and the height of the sorting plate is located below the end of the conveyor belt; a sorting feed trough is arranged in the middle of the sorting plate, and two sorting rods are symmetrically arranged in the middle of the sorting feed trough, and the spacing between the two sorting rods gradually increases from top to bottom; a sorting guide plate is also arranged on the upper end surface of the sorting plate, and there are two sorting guide plates, which are symmetrically arranged in the middle of the sorting plate; When good agricultural products fall into the sorting plate, they pass through the middle sorting rods in turn under the action of the sorting guide plate. As the spacing between the sorting rods gradually increases, agricultural products with smaller diameters fall from the gaps first and enter the good product collection box, and agricultural products with larger diameters fall from the gaps later and enter the good product collection box, thereby sorting agricultural products of different diameters according to their diameter sizes.

[0015] Preferably, a sorting rod installation groove is provided at the upper part of the sorting material discharge trough of the sorting plate, the upper end of the sorting rod is movably installed in the sorting rod installation groove, a sorting rod gap adjustment groove is provided at the lower part of the sorting material discharge trough, a sorting rod gap adjustment pad is provided at the bottom of the sorting rod gap adjustment groove, and a plurality of arc grooves are provided on the sorting rod gap adjustment pad; When the spacing between the sorting rods needs to be adjusted, the distance between the upper ends of the two sorting rods is adjusted first, and then the lower ends of the sorting rods are placed in different arc grooves, so that the spacing between the two sorting rods can be adjusted to adapt to the sorting of agricultural products of different sizes.

[0016] Preferably, sorting guide plate mounting grooves are symmetrically provided on the upper end of the upper surface of the sorting plate, and a bidirectional threaded rod is penetrated inside the two sorting guide plate mounting grooves, and both ends of the bidirectional threaded rod are rotatably connected to the two side walls of the sorting plate, and the lower end protrusions of the two sorting guide plates respectively extend into the two sorting guide plate mounting grooves and are threadedly connected to the bidirectional threaded rod; one end of the bidirectional threaded rod extends out of the outside of the sorting plate and is connected to an adjustment handle.

[0017] When in use, the adjusting handle is turned to drive the bidirectional threaded rod to rotate, so that the two sorting guide plates are moved closer to or farther away from each other, thereby adjusting the distance between the two sorting guide plates, so that the sorting mechanism can adapt to the sorting of agricultural products of different sizes.

[0018] The sorting method of the agricultural product sorting system based on artificial intelligence comprises the following steps: S1, using machine vision technology to identify the appearance of the agricultural products transmitted, converting the captured target agricultural products into image signals, and transmitting them to the image processing module; S2, the image processing module uses image processing technology to obtain the shape information of the target agricultural product, and converts it into digital signals according to the pixel distribution, brightness and color information; the image system performs various operations on these signals to extract the characteristics of the target agricultural product and compares them with the good agricultural product behaviors in the database; S3, using the analysis module to calculate the surface area, perimeter-to-area ratio, and aspect ratio parameters of the agricultural products, and to evaluate the surface damage of the agricultural products, and to determine whether the agricultural products are defective or good; At the same time, the analysis module can learn and analyze a large amount of agricultural product data through deep learning algorithms, thereby continuously improving the accuracy of identifying various characteristics; S4, the control module controls the rejection motor to work, and the rejection plate is used to push the identified defective products into the defective product outlet bucket; the good products enter the sorting mechanism for size sorting.

[0019] Compared with the prior art, the present invention has the following advantages: 1. The present invention detects and identifies the appearance of agricultural products through an intelligent detection mechanism. After identifying defective agricultural products, the defective products are pushed into a defective product outlet bucket through a defective product rejection mechanism and fall into a defective product collection box for collection. Good products enter a sorting mechanism, and two sorting rods are used to classify and collect agricultural products according to their installation diameters for subsequent classified sales. The system does not require manual operation and is fully automated in sorting, which greatly improves the efficiency of sorting.

[0020] 2. The present invention arranges a material guiding mechanism on the conveyor belt, so that the conveyed agricultural products can be arranged in order for transportation, which is convenient for the intelligent detection mechanism to identify good and defective products, and also convenient for the defective product rejection mechanism to reject defective products; at the same time, the distance between the fixed material guiding plate and the movable material guiding plate in the material guiding mechanism can be adjusted to meet the material guiding needs of agricultural products of different sizes and increase the adaptability of the system.

[0021] 3. The present invention is provided with two groups of defective rejection mechanisms, each group of defective rejection mechanisms is provided with three rejection plates, and the rejection plates are driven by the rotation of the motor to complete the rejection action, which greatly reduces the action time of the rejection plates and increases the rejection efficiency; at the same time, the two groups of defective rejection mechanisms can better prevent rejection omissions, thereby better ensuring the shipment quality of agricultural products.

[0022] 4. The sorting mechanism in the present invention can sort the size of agricultural products through the different spacings between the two sorting rods. It has a simple structure and a very good sorting effect. At the same time, the spacing between the two sorting rods is adjustable. It only needs to adjust the distance between the upper ends of the two sorting rods first, and then put the lower ends of the sorting rods into different arc grooves, so that the spacing between the two sorting rods can be adjusted to adapt to the sorting of agricultural products of different sizes.

[0023] 5. The present invention utilizes machine vision technology and image processing technology to accurately identify the appearance damage of agricultural products, thereby improving the accuracy of identification and ensuring the quality of sorting. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a schematic diagram of the structure of an agricultural product sorting system based on artificial intelligence; Figure 2 It is a left oblique view of the agricultural product sorting system based on artificial intelligence; Figure 3 It is a structural schematic diagram of the material guiding mechanism in the agricultural product sorting system based on artificial intelligence; Figure 4 A downward oblique view of a material guide mechanism in an artificial intelligence-based agricultural product sorting system; Figure 5 It is a structural diagram of the conveying mechanism in the agricultural product sorting system based on artificial intelligence; Figure 6 for Figure 2A is a schematic diagram of the enlarged structure of the middle part; Figure 7 for Figure 2 A schematic diagram of the enlarged structure of B; Figure 8 It is a structural diagram of the sorting mechanism in the agricultural product sorting system based on artificial intelligence; Fig. 9 An upper oblique view of a sorting mechanism in an artificial intelligence-based agricultural product sorting system; Fig.10 for Fig. 9 Schematic diagram of the enlarged structure of C in the middle; Fig.11 It is a structural schematic diagram of the sorting rod in the agricultural product sorting system based on artificial intelligence; Fig.12 It is a structural schematic diagram of the sorting guide plate in the agricultural product sorting system based on artificial intelligence; Fig.13 This is the structural block diagram of the controller in the agricultural product sorting system based on artificial intelligence; Fig.14 The present invention is a flow chart of the sorting method of the agricultural product sorting system based on artificial intelligence.

[0025] Numbers in the figure: 1. Conveying bracket; 2. Conveying mechanism; 3. Intelligent detection mechanism; 4. Material guiding mechanism; 5. Controller; 6. Defective product rejection mechanism; 7. Sorting mechanism; 8. Defective product guide bucket; 9. Defective product collection box; 10. Good product collection box; 21. Conveyor belt; 22. Driving roller; 23. Driving motor; 24. Supporting shaft; 31. Detection bracket; 32. Detection camera mounting frame; 33. Detection camera; 41. Fixed material guide plate; 42. Moving material guide plate; 43. Support plate; 44. U-shaped connecting frame; 45. Rectangular through groove; 46. Fixed bottom plate; 47. Mounting hole; 61. Rejection bracket; 62. Crossbeam; 63. Rejection motor; 64. Roller; 65. Rejection plate; 71. Sorting bracket; 72. Sorting plate; 73. Sorting guide plate; 74. Adjustment handle; 75. Sorting discharge chute; 76. Sorting rod; 77. Sorting guide plate mounting groove; 78. Sorting rod mounting groove; 79. Sorting rod gap adjustment pad; 710. Sorting rod gap adjustment groove; 711. Bidirectional threaded rod. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. 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.

[0027] Embodiment 1 See also Figure 1-Figure 12 , an agricultural product sorting system based on artificial intelligence, this system is mainly used to sort round or quasi-round agricultural products, such as oranges, mandarin oranges, grapefruits, etc. This application takes oranges as an example for detailed description.

[0028] The system includes: a conveying mechanism 2, which is used to convey the picked oranges for sorting operations; see Figure 4 The conveying mechanism 2 includes a conveyor belt 21, a driving roller 22, a driving motor 23 and a supporting shaft 24; both ends of the conveyor belt 21 are rotatably mounted on the upper end of the conveying bracket 1 through two driving rollers 22, the driving motor 23 is mounted on the outer side of the conveying bracket 1, and its output end is connected to one of the driving rollers 22, and the driving motor 23 drives the transmission belt to rotate to realize the conveyance of oranges; a plurality of supporting shafts 24 are also arranged between the two driving rollers 22, and the supporting shafts 24 are rotatably connected to the inner wall of the upper end of the conveying bracket 1.

[0029] See also Figure 2 , Figure 7 The intelligent detection mechanism 3 recognizes the appearance of the oranges through machine vision technology and determines the quality of the oranges; the intelligent detection mechanism 3 includes a detection bracket 31, a detection camera 33 mounting bracket 32 ​​and a detection camera 33, the detection bracket 31 is L-shaped, the lower end of the detection bracket 31 is fixedly mounted on the conveying bracket 1, and the detection camera 33 is fixed to the upper end of the detection bracket 31 through the detection camera 33 mounting bracket 32; at the same time, the detection camera 33 is connected to the controller 5 through a wire.

[0030] When in use, the detection camera 33 captures the appearance image of the orange and transmits the data to the controller 5. The controller 5 identifies the quality of the orange through an algorithm. If the orange is defective, the controller 5 sends a command to the defective rejection mechanism 6 to make it work and reject the defective orange.

[0031] See also Figure 3 , Figure 4 The material guiding mechanism 4 is used to arrange the oranges conveyed in order for transportation; there are three groups of material guiding mechanisms 4, which are respectively arranged at the front end of the conveyor belt 21, the rear end of the conveyor belt 21, and the conveyor belt 21 is located between the two groups of defective rejecting mechanisms 6; The material guide mechanism 4 includes a fixed material guide plate 41, a movable material guide plate 42, a support plate 43, and a U-shaped connecting frame 44; the outer wall of the fixed material guide plate 41 is fixedly connected to the end of the support plate 43, and the support plate 43 is provided with a rectangular through groove 45 at one end close to the fixed material guide plate 41, and a plurality of mounting holes 47 at one end away from the fixed material guide plate 41. When installing, it is fixed to the upper end of the conveying bracket 1 by passing bolts through the mounting holes 47, and the outer wall of the movable material guide plate 42 is fixedly connected to one end of the U-shaped connecting frame 44. The other end of the U-shaped connecting frame 44 crosses the fixed material guide plate 41 and is inserted into the rectangular through slot 45. The support plate 43 is provided with a fixed bottom plate 46 at the lower end of the rectangular through slot 45. The fixed bottom plate 46 is connected to the end of the U-shaped connecting frame 44 by bolts, so that the U-shaped connecting frame 44 is fixed on the support plate 43. At the same time, by adjusting the fixed position of the U-shaped connecting frame 44 in the rectangular through slot 45, the spacing between the movable material guide plate 42 and the fixed material guide plate 41 can be changed, so as to meet the material guide of oranges of different sizes. Among them, the front ends of the fixed material guide plate 41 and the movable material guide plate 42 are both bent, so that the fixed material guide plate 41 and the movable material guide plate 42 are combined to form a Y-shaped material guide channel, which is convenient for the orderly export of oranges.

[0032] See also Figure 2 , Figure 6 The defective rejecting mechanism 6 is arranged above the conveying mechanism 2 and behind the intelligent detection mechanism 3, and is used to reject the detected defective products; the defective rejecting mechanism 6 has two groups, and the two groups are arranged above the conveyor belt 21 one after the other; the defective rejecting mechanism 6 includes a rejecting bracket 61, a beam 62, a rejecting motor 63, a roller 64 and a rejecting plate 65, the lower end of the rejecting bracket 61 is fixedly installed on the rear side of the conveying bracket 1, the beam 62 is fixedly connected to the upper end of the rejecting bracket 61, the rejecting motor 63 is fixedly installed on one side of the beam 62, and the output end of the rejecting motor 63 passes through the beam 62 and is connected to the roller 64, and a plurality of rejecting plates 65 are evenly arranged on the outer wall of the roller 64; in the present application, there are three rejecting plates 65.

[0033] When in use, the controller 5 issues a command to rotate the rejection motor 63 to drive the roller 64 and the rejection plate 65 to rotate, and the defective products transmitted are pushed out of the conveyor belt 21 by the rotating rejection plate 65.

[0034] See also Figure 8-Figure 12, the sorting mechanism 7 is arranged at the end of the conveying mechanism 2, receives the good products conveyed by the conveying mechanism 2, and sorts the good products according to the installation size; it includes a sorting bracket 71 and a sorting plate 72, the sorting plate 72 is arranged obliquely, and the two ends of the sorting plate 72 are fixed to the bottom surface by the sorting bracket 71, and the height of the sorting plate 72 is located below the end of the conveyor belt 21; a sorting feed trough 75 is arranged in the middle of the sorting feed trough 75, and two sorting rods 76 are symmetrically arranged in the middle of the sorting feed trough 75, and the spacing between the two sorting rods 76 gradually increases from top to bottom; a sorting guide plate 73 is also arranged on the upper end surface of the sorting plate 72, and there are two sorting guide plates 73, which are symmetrically arranged in the middle of the sorting plate 72; After the good oranges fall onto the sorting plate 72, they pass through the middle sorting rods 76 in sequence under the action of the sorting guide plate 73. As the spacing between the sorting rods 76 gradually increases, the oranges with smaller diameters fall from the gaps first and enter the good product collection box 10, and the oranges with larger diameters fall from the gaps later and enter the good product collection box 10, thereby sorting out oranges of different diameters according to their diameter sizes.

[0035] The upper part of the sorting trough 75 of the sorting plate 72 is provided with a sorting rod installation groove 78, the upper end of the sorting rod 76 is movably installed in the sorting rod installation groove 78, the lower part of the sorting trough 75 is provided with a sorting rod gap adjustment groove 710, the bottom of the sorting rod gap adjustment groove 710 is provided with a sorting rod gap adjustment pad 79, and the sorting rod gap adjustment pad 79 is provided with a plurality of arc grooves; When the spacing between the sorting rods 76 needs to be adjusted, the distance between the upper ends of the two sorting rods 76 is adjusted first, and then the lower ends of the sorting rods 76 are placed in different arc grooves, so that the spacing between the two sorting rods 76 can be adjusted to adapt to the sorting of oranges of different sizes.

[0036] The upper end of the upper surface of the sorting plate 72 is symmetrically provided with sorting guide plate mounting grooves 77, and a bidirectional threaded rod 711 is penetrated inside the two sorting guide plate mounting grooves 77. Both ends of the bidirectional threaded rod 711 are rotatably connected to the two side walls of the sorting plate 72, and the lower end protrusions of the two sorting guide plates 73 respectively extend into the two sorting guide plate mounting grooves 77 and are threadedly connected with the bidirectional threaded rod 711; one end of the bidirectional threaded rod 711 extends out of the outside of the sorting plate 72 and is connected to an adjustment handle 74.

[0037] When in use, by rotating the adjustment handle 74, the bidirectional threaded rod 711 is driven to rotate, so that the two sorting guide plates 73 are moved closer to or farther away from each other, thereby adjusting the distance between the two sorting guide plates 73, so that the sorting mechanism 7 can adapt to the sorting of oranges of different sizes.

[0038] See also Figure 1 , Figure 2The system also includes two defective product guide buckets 8, which are arranged on the side of the conveying bracket 1. The two defective product guide buckets 8 are flush with the positions of the two defective product rejection mechanisms 6, and the defective product guide buckets 8 are installed on the side of the conveying bracket 1 by bolts. A defective product collection box 9 is arranged below the discharge port of the defective product guide bucket 8; a notch is provided on the side panel of the conveying bracket 1 at the connection with the defective product guide bucket 8, so that the rejected defective products can roll into the defective product guide bucket 8 smoothly.

[0039] See also Fig.13 The controller 5 is installed on the conveying bracket 1, and is respectively connected to the detection camera 33 and the driving motor 23 and the rejection motor 63, and controls the operation; the controller 5 includes an image processing module, an analysis module and a control module.

[0040] The working process of this embodiment: When in use, the sorted oranges are transported to the conveyor belt 21 through the handling equipment, and the driving motor 23 rotates to drive the conveyor belt 21 to transport the oranges from left to right; when the oranges are under the action of the guide mechanism 4, the oranges pass through the channel between the fixed guide plate 41 and the movable guide plate 42 one by one; at this time, the detection camera 33 above the conveyor belt 21 detects the appearance of the oranges and transmits the data to the controller 5, which determines the quality of the oranges; When defective oranges with damaged appearance appear, the controller 5 sends a signal to the rejection motor 63, and the rejection motor 63 drives the rejection plate 65 to rotate, thereby pushing the defective oranges from the conveyor belt 21 into the defective product outlet bucket 8, and falling into the defective product collection box 9 for collection; the good oranges continue to be transported to the right along the conveyor belt 21 and enter the sorting mechanism 7; The good oranges falling on the sorting mechanism 7 move toward the middle of the sorting plate 72 under the action of the sorting guide plate 73, and roll down from between the two sorting rods 76. As the distance between the two sorting rods 76 gradually increases, the oranges with smaller diameters fall from the gap first and enter the good product collection box 10 for small fruits, and the oranges with larger diameters fall from the gap later and enter the good product collection box 10 for medium or large fruits, thereby sorting out oranges of different diameters according to their diameters for subsequent classified sales. This system does not require manual operation and is fully automated in sorting, which greatly improves the efficiency of sorting.

[0041] At the same time, the distance between the fixed material guiding plate 41 and the movable material guiding plate 42 in the material guiding mechanism 4 can be adjusted to meet the material guiding needs of oranges of different sizes and increase the adaptability of the system.

[0042] The system is provided with two groups of defective rejecting mechanisms 6, each group of defective rejecting mechanisms 6 is provided with three rejecting plates 65, and the rejecting plates 65 are driven by the rotation of the motor to complete the rejecting action, which greatly reduces the action time of the rejecting plates 65 and increases the rejecting efficiency; at the same time, the two groups of defective rejecting mechanisms 6 can better prevent rejection omissions, thereby better ensuring the delivery quality of oranges.

[0043] The sorting mechanism 7 in the present system can sort the oranges by size by means of different spacings between the two sorting rods 76. The structure is simple and the sorting effect is very good. At the same time, the spacing between the two sorting rods 76 is adjustable. It is only necessary to adjust the distance between the upper ends of the two sorting rods 76 first, and then put the lower ends of the sorting rods 76 into different arc grooves, so that the spacing between the two sorting rods 76 can be adjusted to adapt to the sorting of oranges of different sizes.

[0044] Embodiment 2 See also Fig.14 Based on the first embodiment, this embodiment proposes a sorting method of an agricultural product sorting system based on artificial intelligence, comprising the following steps: S1, using machine vision technology to identify the appearance of the agricultural products transmitted, converting the captured target agricultural products into image signals, and transmitting them to the image processing module; S2, the image processing module uses image processing technology to obtain the shape information of the target agricultural product, and converts it into digital signals according to the pixel distribution, brightness and color information; the image system performs various operations on these signals to extract the characteristics of the target agricultural product and compares them with the good agricultural product behaviors in the database; S3, using the analysis module to calculate the surface area, perimeter-to-area ratio, and aspect ratio parameters of the agricultural products, and to evaluate the surface damage of the agricultural products, and to determine whether the agricultural products are defective or good; At the same time, the analysis module can learn and analyze large amounts of agricultural product data through deep learning algorithms, thereby continuously improving the accuracy of identifying various features.

[0045] S4, the control module controls the rejection motor to work, and the rejection plate is used to push the identified defective products into the defective product outlet bucket; the good products enter the sorting mechanism for size sorting.

[0046] Among them, in step S2, the image processing module collects the image and performs image preprocessing, removes the background, segments the damaged area, and then calculates the damaged area count, damaged area area and damaged area length.

[0047] Image preprocessing includes image digitization, image grayscale, image enhancement, image segmentation and mathematical morphology processing technology; The background removal steps are as follows: convert the color image into a grayscale image G; use the maximum inter-class variance method to convert the grayscale image into a binary image E, where the background B area is black and the agricultural product L area is white; use the edge tracking algorithm to detect each agricultural product area and mark the agricultural product, and calculate the area SLi of each agricultural product; use area L as a template, and obtain a complete color agricultural product picture from the original image according to the formula to remove the background, where ƒpd(x, y) is the pixel value of the pixel point in the complete agricultural product area; pr(x, y) is the pixel point in the original image, and ƒpr(x, y) is the pixel value of the pr(x, y) pixel point; Convert the damaged area image D after segmentation into a grayscale image DG; then use the maximum inter-class variance method to convert the grayscale image DG into a binary image DE, and then perform pixel calculation on this basis. Damaged area counting: define the inside of each closed edge as a region, and use the edge tracking algorithm to detect the regions in each agricultural product picture marked in DE in turn, and the number of recorded regions is the number of damaged areas on each agricultural product picture; damaged area area: calculate the area of ​​discrete damaged areas on each agricultural product picture, and calculate the total area; damaged area length and width: use the minimum circumscribed rectangle to calculate the length and width of each damaged area on the agricultural product picture, and sort them, and take the longest parameter as the grading standard.

[0048] This embodiment utilizes machine vision technology and image processing technology to accurately identify the appearance damage of agricultural products, improve the accuracy of identification, and thus ensure the quality of sorting.

[0049] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including 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. The sentence "includes an element defined by ... does not exclude the existence of other identical elements in the process, method, article or device including the element".

[0050] 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. The agricultural product sorting system based on artificial intelligence is characterized by: include: A conveying mechanism (2), the conveying mechanism (2) being arranged at the upper end of the conveying support (1) and being used to convey the picked agricultural products for sorting operations; An intelligent detection mechanism (3), the intelligent detection mechanism (3) being arranged above the front end of the conveying mechanism (2), identifying the appearance of the agricultural products through machine vision technology, and judging the quality of the agricultural products; A material guiding mechanism (4), the material guiding mechanism (4) being arranged above the conveying mechanism (2) and used for arranging the conveyed agricultural products in sequence for transportation; A defective product rejection mechanism (6), the defective product rejection mechanism (6) being arranged above the conveying mechanism (2) and located behind the intelligent detection mechanism (3), and being used to reject detected defective products; A sorting mechanism (7), the sorting mechanism (7) being arranged at the end of the conveying mechanism (2), receiving good products conveyed by the conveying mechanism (2), and sorting the good products according to their installation size; A controller (5), the controller (5) being mounted on the conveying support (1) and being used to control the operation of each mechanism.

2. The artificial intelligence-based agricultural product sorting system according to claim 1 is characterized in that: The conveying mechanism (2) comprises a conveyor belt (21), a driving roller (22), a driving motor (23) and a supporting rotating shaft (24); the two ends of the conveyor belt (21) are rotatably mounted on the upper end of the conveying bracket (1) via the two driving rollers (22); the driving motor (23) is mounted outside the conveying bracket (1) and its output end is connected to one of the driving rollers (22); the driving motor (23) drives the transmission belt to rotate, thereby realizing the conveyance of agricultural products; a plurality of supporting rotating shafts (24) are also arranged between the two driving rollers (22); the supporting rotating shafts (24) are rotatably connected to the inner wall of the upper end of the conveying bracket (1).

3. The artificial intelligence-based agricultural product sorting system according to claim 1 is characterized in that: The intelligent detection mechanism (3) comprises a detection bracket (31), a detection camera (33) mounting bracket (32) and a detection camera (33); the detection bracket (31) is L-shaped; the lower end of the detection bracket (31) is fixedly mounted on the conveying bracket (1); the detection camera (33) is fixed to the upper end of the detection bracket (31) via the detection camera (33) mounting bracket (32); and the detection camera (33) is connected to the controller (5) via a wire.

4. The artificial intelligence-based agricultural product sorting system according to claim 2 is characterized in that: The defective product rejection mechanism (6) has two groups, and the two groups are arranged one in front of the other above the conveyor belt (21); the defective product rejection mechanism (6) comprises a rejection bracket (61), a crossbeam (62), a rejection motor (63), a roller (64) and a rejection plate (65); the lower end of the rejection bracket (61) is fixedly mounted on the rear side of the conveyor bracket (1); the crossbeam (62) is fixedly connected to the upper end of the rejection bracket (61); the rejection motor (63) is fixedly mounted on one side of the crossbeam (62); the output end of the rejection motor (63) passes through the crossbeam (62) and is connected to the roller (64); and a plurality of the rejection plates (65) are evenly arranged on the outer side wall of the roller (64).

5. The artificial intelligence-based agricultural product sorting system according to claim 2 is characterized in that: The material guiding mechanism (4) has three groups, and the three groups of material guiding mechanisms (4) are respectively arranged at the front end of the conveyor belt (21), the rear end of the conveyor belt (21), and the conveyor belt (21) is located between the two groups of defective product rejection mechanisms (6); The material guide mechanism (4) comprises a fixed material guide plate (41), a movable material guide plate (42), a support plate (43), and a U-shaped connecting frame (44); the outer side wall of the fixed material guide plate (41) is fixedly connected to the end of the support plate (43), and a rectangular through groove (45) is provided at one end of the support plate (43) close to the fixed material guide plate (41), and a plurality of mounting holes (47) are provided at one end away from the fixed material guide plate (41). When installing, the support plate (43) is fixed to the upper end of the conveying bracket (1) by means of bolts passing through the mounting holes (47); the outer side wall of the movable material guide plate (42) is fixed to one end of the U-shaped connecting frame (44); The U-shaped connecting frame (44) is connected to the fixed material guide plate (41), and the other end of the U-shaped connecting frame (44) crosses the fixed material guide plate (41) and is inserted into the rectangular through groove (45). The support plate (43) is provided with a fixed bottom plate (46) at the lower end of the rectangular through groove (45). The fixed bottom plate (46) is connected to the end of the U-shaped connecting frame (44) by bolts, so that the U-shaped connecting frame (44) is fixed on the support plate (43). At the same time, by adjusting the fixed position of the U-shaped connecting frame (44) in the rectangular through groove (45), the spacing between the movable material guide plate (42) and the fixed material guide plate (41) can be changed, thereby meeting the material guiding requirements of agricultural products of different sizes.

6. The artificial intelligence-based agricultural product sorting system according to any one of claims 1 to 5, characterized in that: It also includes two defective product guide buckets (8) arranged on the side of the conveying bracket (1), the two defective product guide buckets (8) are respectively flush with the positions of the two defective product rejection mechanisms (6), and the defective product guide buckets (8) are installed on the side of the conveying bracket (1) by bolts, and a defective product collection box (9) is arranged below the discharge port of the defective product guide bucket (8); the side panel of the conveying bracket (1) is provided with a notch at the connection position of the defective product guide bucket (8) so that the rejected defective products can smoothly roll into the defective product guide bucket (8).

7. The artificial intelligence-based agricultural product sorting system according to claim 1, characterized in that: The sorting mechanism (7) comprises a sorting bracket (71) and a sorting plate (72), wherein the sorting plate (72) is arranged obliquely, and both ends of the sorting plate (72) are fixed to the bottom surface by means of the sorting bracket (71), and the height of the sorting plate (72) is located below the end of the conveyor belt (21); a sorting feed trough (75) is arranged in the middle of the sorting plate (72), and two sorting rods (76) are symmetrically arranged in the middle of the sorting feed trough (75), and the spacing between the two sorting rods (76) gradually increases from top to bottom; a sorting guide plate (73) is also arranged on the upper end surface of the sorting plate (72), and there are two sorting guide plates (73) symmetrically arranged in the middle of the sorting plate (72); When the good agricultural products fall onto the sorting plate (72), they pass through the middle sorting rods (76) in sequence under the action of the sorting guide plate (73). As the spacing between the sorting rods (76) gradually increases, the agricultural products with smaller diameters fall from the gaps first and enter the good product collection box (10), and the agricultural products with larger diameters fall from the gaps later and enter the good product collection box (10), thereby sorting the agricultural products with different diameters according to their diameter sizes.

8. The artificial intelligence-based agricultural product sorting system according to claim 7, characterized in that: A sorting rod installation groove (78) is provided at the upper part of the sorting material discharge trough (75) of the sorting plate (72), the upper end of the sorting rod (76) is movably installed in the sorting rod installation groove (78), a sorting rod gap adjustment groove (710) is provided at the lower part of the sorting material discharge trough (75), a sorting rod gap adjustment pad (79) is provided at the bottom of the sorting rod gap adjustment groove (710), and a plurality of arc-shaped grooves are provided on the sorting rod gap adjustment pad (79).

9. The artificial intelligence-based agricultural product sorting system according to claim 8, characterized in that: The upper end of the upper surface of the sorting plate (72) is symmetrically provided with sorting guide plate mounting grooves (77), and a bidirectional threaded rod (711) is arranged inside the two sorting guide plate mounting grooves (77). The two ends of the bidirectional threaded rod (711) are rotatably connected to the two side walls of the sorting plate (72), and the lower end protrusions of the two sorting guide plates (73) extend into the two sorting guide plate mounting grooves (77) respectively and are threadedly connected to the bidirectional threaded rod (711); one end of the bidirectional threaded rod (711) extends out of the outside of the sorting plate (72) and is connected to an adjustment handle (74).

10. A sorting method for an agricultural product sorting system based on artificial intelligence, comprising the following steps: S1, using machine vision technology to identify the appearance of the agricultural products transmitted, converting the captured target agricultural products into image signals, and transmitting them to the image processing module; S2, the image processing module uses image processing technology to obtain the shape information of the target agricultural product, and converts it into digital signals according to the pixel distribution, brightness and color information; the image system performs various operations on these signals to extract the characteristics of the target agricultural product and compares them with the good agricultural product behaviors in the database; S3, using the analysis module to calculate the surface area, perimeter-to-area ratio, and aspect ratio parameters of the agricultural products, and to evaluate the surface damage of the agricultural products, and to determine whether the agricultural products are defective or good; At the same time, the analysis module can learn and analyze a large amount of agricultural product data through deep learning algorithms, thereby continuously improving the accuracy of identifying various characteristics; S4, the control module controls the rejection motor to work, and the rejection plate is used to push the identified defective products into the defective product outlet bucket; the good products enter the sorting mechanism for size sorting.

Images