Lemon sorting machine with visual detection and control method

Abstract

The present application relates to a kind of lemon sorting machine with visual detection and control method, belong to sorting machine technical field, including drainage device, lemon machine cover, blanking device, belt feeding assembly and camera, belt feeding assembly is installed on second band bracket water pan, belt feeding assembly includes band hanger flat belt, and according to preset size array installation is carried out on the band hanger flat belt with scraper.The blanking device includes feeding pulp, and feeding pulp is connected with feeding pulp speed reducer through coupling, and then cooperates with blanking hopper to form blanking device, and blanking device is installed on the driven wheel left support plate, driven wheel right support plate on belt feeding assembly;Through advanced visual identification system, fruit is automatically detected, whether fruit is accurately judged to be deteriorated, size shape is suitable, effectively avoid the subjectivity and error of artificial judgment, improve the accuracy and objectivity of fruit quality judgment, provide high-quality raw materials for subsequent fruit tea production.

Description

Technical Field

[0001] This invention relates to the field of technology, and more particularly to a lemon sorting machine with visual detection and its control method. Background Technology

[0002] In the traditional fruit tea making process, operators first need to cut the fruit into pieces, and then count the required number of pieces one by one and place them into a container. This manual method has many drawbacks. On the one hand, manual counting is inefficient and time-consuming, affecting the speed of fruit tea production. On the other hand, manual counting is prone to errors, leading to inaccurate fruit quantities and affecting the quality of the fruit tea. Furthermore, when manually judging whether the fruit is spoiled or whether its size and shape are appropriate, errors can occur due to individual differences and subjective factors, making it impossible to guarantee the consistency of fruit quality. In addition, manual operation increases labor costs and may introduce hygiene risks during the process, which is detrimental to food safety. Therefore, developing a device that can automatically sort fruit quality and accurately count them is of great significance for improving fruit processing efficiency, ensuring fruit quality, and reducing labor intensity and costs. Therefore, to address existing needs, we propose a lemon sorting machine with visual inspection. In existing technologies, such as the fruit inspection system based on computer vision disclosed in patent CN213633221U, although the system can detect the size and quality of fruit through a camera and a vision system, firstly, the structure adopts a vertical feeding method, with a large feeding opening in the storage bin 2 and a small feeding opening in the arrangement tube 4, which easily causes fruit to accumulate and get stuck; secondly, the rack 9 drives the gear 6 to rotate, and the gear can scratch the fruit skin, seriously affecting the quality of the fruit, and the rack 9 and the inner wall of the arrangement tube 4 are prone to jamming, causing the fruit to be crushed and damaged; finally, the rack 9 driving the gear 6 to rotate for a long time will cause wear on the rack and gear, resulting in iron powder and iron slag, which poses a potential health hazard to consumers. Summary of the Invention

[0003] This invention overcomes the shortcomings of the prior art and provides a lemon sorting machine with visual detection and a control method thereon.

[0004] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0005] The first aspect of the present invention provides a lemon sorting machine with visual detection, including a drainage device, a lemon machine cover, a feeding device, a belt feeding assembly and a camera;

[0006] The drainage device includes a first water receiving tray with a bracket and a second water receiving tray with a bracket. The second water receiving tray with a bracket is fixed to the first water receiving tray with a bracket. A water baffle and an electric cylinder mounting support plate are fixed on the second water receiving tray with a bracket. An electric cylinder and a linear guide rail are mounted on the electric cylinder mounting support plate. A slider push plate is mounted on the electric cylinder.

[0007] The lemon machine cover includes a lemon machine lid and a door, and the camera is installed on the upper inner side of the lemon machine cover;

[0008] The feeding device includes a feeding slurry, which is connected to a feeding slurry reduction motor via a coupling, and then cooperates with a feeding funnel to form the feeding device. The feeding device is installed on the left support plate and the right support plate of the drive wheel on the belt feeding assembly.

[0009] The belt feeding assembly is installed on the second belt support water receiving tray. The belt feeding assembly includes a flat belt with hanging plates, and scraper blades are installed on the flat belt with hanging plates in a preset size array.

[0010] Furthermore, in the lemon sorting machine with visual inspection, the first water receiving tray with bracket, the second water receiving tray with bracket, the baffle plate, and the drain connector cooperate to form a drainage trough, which also serves as a support base for the lemon machine with camera.

[0011] Furthermore, in the lemon sorting machine with visual inspection, the camera is installed on the upper inner side of the lemon machine cover, and the lemon machine cover is fixed to the second water receiving tray with bracket of the drainage device and supported by the first water receiving tray with bracket.

[0012] Furthermore, in a lemon sorting machine with visual inspection, the feeding slurry is connected to a feeding slurry reduction motor via a coupling, and then cooperates with a discharge funnel to form the discharge device.

[0013] Furthermore, in the lemon sorting machine with visual inspection, the feeding device is installed on the left support plate and the right support plate of the drive wheel of the belt feeding assembly. The feeding funnel of the feeding device is inclined and an adjustable baffle is provided at the outlet.

[0014] Furthermore, in the lemon sorting machine with vision detection, a feeding slurry control sensor is installed on the left and right support plates of the drive wheel of the belt feeding assembly, and a counting sensor is installed at the lemon hopper of the belt feeding assembly.

[0015] A second aspect of the present invention provides a control method for a lemon sorting machine with vision detection, applied to a lemon sorting machine with vision detection, comprising the following steps:

[0016] Acquire image data information of various historical defect types of lemons, and construct a lemon defect recognition model based on the image data information of various historical defect types of lemons;

[0017] By testing the lemon defect recognition model, the recognition speed of the lemon defect recognition model within a unit time was obtained under different image quantities.

[0018] Based on the recognition speed information of the lemon defect recognition model within a unit time under different image quantity information, the motion parameters of the lemon sorting machine with visual detection are configured.

[0019] Furthermore, in the control method of the lemon sorting machine with vision detection, image data information of each historical defect type of lemon is acquired, and a lemon defect recognition model is constructed based on the image data information of each historical defect type of lemon, specifically including:

[0020] Obtain image data information of various historical defect types of lemons, construct a training set based on the image data information of various historical defect types of lemons, and construct a lemon defect recognition model based on a deep neural network;

[0021] The training set is input into the lemon defect recognition model for training, and the prediction accuracy information of the lemon defect recognition model during the training process is obtained, and a prediction accuracy threshold is set.

[0022] Determine whether the prediction accuracy information of the lemon defect identification model during the training process is greater than the prediction accuracy threshold;

[0023] When the prediction accuracy of the lemon defect recognition model during training is greater than the prediction accuracy threshold, the training of the lemon defect recognition model ends; when the prediction accuracy of the lemon defect recognition model during training is not greater than the prediction accuracy threshold, the training of the lemon defect recognition model continues.

[0024] Furthermore, in the control method of the lemon sorting machine with vision detection, the lemon defect recognition model is tested to obtain the recognition speed information of the lemon defect recognition model within a unit time under different image quantities, specifically including:

[0025] Configure the number of images within a unit time period, set test conditions based on the number of images within the unit time period, and test the lemon defect recognition model based on the test conditions;

[0026] Through testing, the recognition speed information of the lemon defect recognition model within a unit time under different image quantity information is obtained, and the recognition speed information of the lemon defect recognition model within a unit time under different image quantity information is output.

[0027] Furthermore, in the control method of the lemon sorting machine with vision detection, the motion parameters of the lemon sorting machine with vision detection are configured based on the recognition speed information of the lemon defect recognition model within a unit time under different image quantity information, specifically including:

[0028] Randomly select the motion parameters of a lemon sorting machine with vision detection, and obtain the number of images collected per unit time based on the motion parameters of the lemon sorting machine with vision detection.

[0029] Based on the number of images collected within the unit time and the recognition speed of the lemon defect recognition model within the unit time under different image quantity information, the recognition speed of the lemon defect recognition model under the number of images collected within the unit time is estimated.

[0030] Obtain the maximum data processing speed of the data processing system, and determine whether the recognition speed information of the lemon defect recognition model under the number of image data collected within the unit time is greater than the maximum data processing speed of the data processing system;

[0031] When the recognition speed of the lemon defect recognition model exceeds the maximum data processing speed of the data processing system under the number of image data collected within the unit time, the motion parameters of the lemon sorting machine with vision detection are reconfigured.

[0032] This invention addresses the shortcomings of the prior art and has the following beneficial effects:

[0033] 1. Improve the accuracy of fruit quality assessment: Through an advanced visual recognition system, fruits can be automatically detected, accurately determining whether they are spoiled and whether their size and shape are appropriate. This effectively avoids the subjectivity and errors of manual judgment, improving the accuracy and objectivity of fruit quality assessment and providing high-quality raw materials for subsequent fruit tea production.

[0034] 2. This invention utilizes a belt conveyor mechanism and a spiral feeding device to automatically remove problematic fruits, reducing manual intervention and greatly improving the speed and efficiency of fruit processing. It can meet the needs of rapid fruit processing in scenarios such as fruit tea making.

[0035] 3. Enhanced counting accuracy and automation: The control system can accurately count the fruit according to the preset quantity, and automatically drop the fruit when the set quantity is reached by pressing a button. This avoids errors that may occur with manual counting, improves the accuracy and automation of fruit processing, and reduces the labor intensity of operators.

[0036] 4. Reduced costs and improved hygiene standards: Applicable to various scenarios such as fruit tea making, it can significantly reduce labor costs, reduce hygiene risks caused by manual operation, and improve the hygiene standards of the production process, with broad application prospects and market value. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained from these drawings without creative effort.

[0038] Figure 1 A 3D structural diagram of a lemon-making machine with a camera;

[0039] Figure 2 This is a schematic diagram of the front structure of a lemon-making machine with a camera;

[0040] Figure 3 for Figure 2 Schematic diagram of the structure at point AA;

[0041] Figure 4 for Figure 2 Schematic diagram of the structure at point BB;

[0042] Figure 5 for Figure 2 Schematic diagram of the structure at point CC.

[0043] In the diagram: 1. Drainage device; 2. First water receiving tray with support; 3. Second water receiving tray with support; 4. Door; 5. Lemon machine cover; 6. Lemon machine outer cover; 7. Feeding device; 8. Feeding funnel; 9. Feeding slurry; 10. Feeding slurry geared motor; 11. Belt feeding assembly; 12. Drive wheel; 13. Belt support plate; 14. Water baffle; 15. Drainage connector; 16. Sliding pusher plate; 17. Lemon hopper; 18. Main power unit. 19. Wheel; 20. Flat belt with scraper; 21. Upper guide baffle; 22. Camera; 23. Belt drive reducer; 24. Electric cylinder mounting support plate; 25. Electric cylinder; 26. Linear guide rail; 27. Left support plate of driven wheel; 28. Left guide baffle; 29. ​​Left support plate of drive wheel; 30. Counting sensor; 31. Right support plate of drive wheel; 32. Right guide baffle; 33. Feed slurry control sensor; 34. Right support plate of driven wheel; Detailed Implementation

[0044] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0045] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0046] Please see Figures 1 to 5 The present invention provides an embodiment of a lemon machine with a camera, comprising a drainage device 1, a lemon machine cover 6, a feeding device 7, a belt feeding assembly 11 and a camera 21.

[0047] The drainage device 1 includes a first water receiving tray with a bracket 2, a second water receiving tray with a bracket 3, a baffle plate 14, an electric cylinder 24, a linear guide rail 25, an electric cylinder mounting support plate 23, a slider push plate 16, and a drainage connector 15. The second water receiving tray with a bracket is fixed to the first water receiving tray with a bracket. The baffle plate and the electric cylinder mounting support plate are fixed to the second water receiving tray with a bracket. The electric cylinder 24 and the linear guide rail 25 are mounted on the electric cylinder mounting support plate 23. The slider push plate 16 is mounted on the electric cylinder 24.

[0048] The lemon machine outer cover 6 includes a lemon machine cover 5 and a door 4, both of which are mounted on the lemon machine outer cover 6; the camera 21 is mounted on the upper inner side of the lemon machine outer cover 6;

[0049] The feeding device 7 includes a feeding funnel 8, a feeding slurry 9, and a feeding slurry reduction motor 10. The feeding slurry is connected to the feeding slurry reduction motor 10 via a coupling, and then cooperates with the feeding funnel 8 to form the feeding device 7. The feeding device 7 is installed on the left support plate 28 and the right support plate 30 of the drive wheel on the belt feeding assembly 11.

[0050] The belt feeding assembly 11 includes a drive wheel 18, a driven wheel 12, a flat belt with hanging plates 19, a belt support plate 13, a lemon hopper 17, a belt-driven reduction motor 22, a left guide baffle 27, a right guide baffle 31, an upper guide baffle 20, a left support plate 28 for the drive wheel, a right support plate 30 for the drive wheel, a left support plate 26 for the driven wheel, a right support plate 33 for the driven wheel, a feeding slurry control sensor 32, and a counting sensor 29. The belt feeding assembly 11 is mounted on the second belt bracket water receiving tray 3.

[0051] Furthermore, the drainage device 1 includes a first water receiving tray with a support 2, a second water receiving tray with a support 3, a baffle plate 14, an electric cylinder 24, a linear guide rail 25, an electric cylinder mounting support plate 23, a slider push plate 16, and a drainage connector 15; by adopting the above technical solution, the drainage device 1 forms a complete drainage system.

[0052] Furthermore, the lemon machine cover 6 includes a lemon machine cover 5 and a door 4; by adopting the above technical solution, the lemon machine cover 6 plays a safety protection role for the entire lemon machine transmission mechanism.

[0053] The feeding device 7 includes a feeding funnel 8, a feeding slurry 9, and a feeding slurry reduction motor 10; by adopting the above technical solution, the feeding device 7 controls the feeding of a certain quantity of lemon chunks.

[0054] Furthermore, the feeding funnel of the feeding device is inclined, making it easy for manual pouring of the cut lemon pieces into it. An adjustable baffle is provided at the outlet of the feeding funnel. By adjusting the opening of the baffle, the speed and quantity of lemon pieces falling can be controlled, preventing a large number of lemon pieces from falling at once.

[0055] Furthermore, the belt feeding assembly 11 includes a drive wheel 18, a driven wheel 12, a flat belt with hanging plates 19, a belt support plate 13, a lemon hopper 17, a belt-driven reduction motor 22, a left guide baffle 27, a right guide baffle 31, an upper guide baffle 20, a left support plate 28 for the drive wheel, a right support plate 30 for the drive wheel, a left support plate 26 for the driven wheel, a right support plate 33 for the driven wheel, a feeding slurry control sensor 32, and a counting sensor 29. By adopting the above technical solution, the belt feeding assembly 11 performs the functions of conveying and counting lemon blocks.

[0056] Furthermore, the camera 21 is installed on the upper inner side of the lemon machine cover 6; by adopting the above technical solution, the camera 21 collects and monitors the current size and quality of the lemon pieces.

[0057] Furthermore, a pair of feeding slurry control sensors are installed on the left and right support plates of the drive wheel of the belt feeding assembly to control whether there are lemon blocks on the flat belt with hanging plates, thereby controlling the start and stop of the feeding slurry rotation.

[0058] Furthermore, the surface of the flat belt with hanging tabs is equipped with scrapers of a certain size, which play a certain role in dispersing lemon pieces that fall onto the flat belt with hanging tabs.

[0059] Furthermore, a pair of counting sensors are installed at the lemon hopper of the belt feeding assembly to count the number of lemon pieces;

[0060] Furthermore, the camera is installed on the upper inner side of the lemon machine's outer cover to perform image recognition, record, and feedback on the number, shape, size, and quality of lemon pieces that fall onto the belt with hanging tabs.

[0061] Working principle: The operator manually cuts lemon pieces and pours them into the lemon hopper 17, sets the required quantity, and starts the power supply. The feeding slurry 9 pushes the lemon pieces onto the conveyor belt. The flat conveyor belt 19 with hanging tabs continues to move, and the lemons fall into the finished product box. The counting sensor 29 counts the lemons. When the set quantity is reached, pressing a button releases the lemon pieces from the finished product box, which are then collected by the operator using a cup. Note: When the feeding slurry control sensor 32 detects no lemon pieces on the flat conveyor belt 19 with hanging tabs, the feeding slurry 9 in the discharge funnel 8 stirs and discharges the lemons.

[0062] In summary, this invention improves the accuracy of fruit quality assessment: Through an advanced visual recognition system, fruits are automatically detected, accurately determining whether they are spoiled and whether their size and shape are appropriate. This effectively avoids the subjectivity and errors of manual judgment, improving the accuracy and objectivity of fruit quality assessment and providing high-quality raw materials for subsequent fruit tea production. Utilizing a belt conveyor mechanism and a spiral feeding device, this invention automatically removes problematic fruits, reducing manual intervention and significantly increasing the speed and efficiency of fruit processing. This meets the needs of rapid fruit processing in scenarios such as fruit tea production. The control system accurately counts fruits according to a preset quantity and automatically drops them upon reaching the set number, avoiding errors that may occur with manual counting, improving the accuracy and automation level of fruit processing, and reducing the labor intensity of operators. Applicable to various scenarios such as fruit tea production, this invention significantly reduces labor costs, minimizes hygiene risks associated with manual operation, and improves hygiene standards in the production process, possessing broad application prospects and market value.

[0063] A second aspect of the present invention provides a control method for a lemon sorting machine with vision detection, applied to a lemon sorting machine with vision detection, comprising the following steps:

[0064] Acquire image data information of various historical defect types of lemons, and construct a lemon defect recognition model based on the image data information of various historical defect types of lemons;

[0065] By testing the lemon defect recognition model, the recognition speed of the lemon defect recognition model within a unit time was obtained under different image quantities.

[0066] Based on the recognition speed information of the lemon defect recognition model within a unit time under different image quantity information, the motion parameters of the lemon sorting machine with visual detection are configured.

[0067] It should be noted that this invention configures the motion parameters of a lemon sorting machine with visual detection by fusing the lemon defect recognition model's recognition speed per unit time under different image quantity information and the data processing system's recognition speed. This enables deep fusion of the motion parameters of the lemon sorting machine with visual detection, avoids missed detections, and prevents the lemon defect recognition model from crashing during the recognition process.

[0068] Furthermore, in the control method of the lemon sorting machine with vision detection, image data information of each historical defect type of lemon is acquired, and a lemon defect recognition model is constructed based on the image data information of each historical defect type of lemon, specifically including:

[0069] Obtain image data information of various historical defect types of lemons, construct a training set based on the image data information of various historical defect types of lemons, and construct a lemon defect recognition model based on a deep neural network;

[0070] The training set is input into the lemon defect recognition model for training, and the prediction accuracy information of the lemon defect recognition model during the training process is obtained, and a prediction accuracy threshold is set.

[0071] Determine whether the prediction accuracy information of the lemon defect identification model during the training process is greater than the prediction accuracy threshold;

[0072] When the prediction accuracy of the lemon defect recognition model during training is greater than the prediction accuracy threshold, the training of the lemon defect recognition model ends; when the prediction accuracy of the lemon defect recognition model during training is not greater than the prediction accuracy threshold, the training of the lemon defect recognition model continues.

[0073] It should be noted that the defect type image data includes deviations such as size quality, blemish quality, surface damage, color deviation, and maturity. Deep neural networks include recurrent neural networks, convolutional neural networks, and multilayer perceptron neural networks. This method can obtain a lemon defect recognition model to identify lemon defects.

[0074] Furthermore, in the control method of the lemon sorting machine with vision detection, the lemon defect recognition model is tested to obtain the recognition speed information of the lemon defect recognition model within a unit time under different image quantities, specifically including:

[0075] Configure the number of images within a unit time period, set test conditions based on the number of images within the unit time period, and test the lemon defect recognition model based on the test conditions;

[0076] Through testing, the recognition speed information of the lemon defect recognition model within a unit time under different image quantity information is obtained, and the recognition speed information of the lemon defect recognition model within a unit time under different image quantity information is output.

[0077] It should be noted that by setting test conditions based on the number of images within a unit of time, the lemon defect recognition model is tested based on the test conditions. Through testing, the recognition speed information of the lemon defect recognition model within a unit of time under different number of images is obtained, so as to better set the motion parameters of the lemon sorting machine, so as to deeply integrate the motion parameters of the lemon sorting machine with vision detection, avoid missed detection, and avoid the lemon defect recognition model from crashing during the recognition process.

[0078] Furthermore, in the control method of the lemon sorting machine with vision detection, the motion parameters of the lemon sorting machine with vision detection are configured based on the recognition speed information of the lemon defect recognition model within a unit time under different image quantity information, specifically including:

[0079] Randomly select the motion parameters of a lemon sorting machine with vision detection, and obtain the number of images collected per unit time based on the motion parameters of the lemon sorting machine with vision detection.

[0080] Based on the number of images collected within the unit time and the recognition speed of the lemon defect recognition model within the unit time under different image quantity information, the recognition speed of the lemon defect recognition model under the number of images collected within the unit time is estimated.

[0081] Obtain the maximum data processing speed of the data processing system, and determine whether the recognition speed information of the lemon defect recognition model under the number of image data collected within the unit time is greater than the maximum data processing speed of the data processing system;

[0082] When the recognition speed of the lemon defect recognition model exceeds the maximum data processing speed of the data processing system under the number of image data collected within the unit time, the motion parameters of the lemon sorting machine with vision detection are reconfigured.

[0083] It should be noted that, due to the upper limit of the maximum data processing speed of the data processing system, when the recognition speed of the lemon defect recognition model exceeds the maximum data processing speed of the data processing system, given the number of images collected within the specified unit time, the motion parameters of the lemon sorting machine with vision detection are reconfigured. This allows for deep fusion of the motion parameters of the lemon sorting machine with vision detection, avoiding missed detections and preventing the lemon defect recognition model from crashing during the recognition process. The motion parameters of the lemon sorting machine with vision detection include data such as the transmission speed of the belt assembly, the operating parameters of the electric cylinder, and the motion parameters of the motor.

[0084] In addition, this method also includes:

[0085] The main defect characteristics of lemons under various seasonal differences are obtained through big data, and a knowledge graph is constructed. The main defect characteristics of lemons under various seasonal differences are then input into the knowledge graph for storage.

[0086] Obtain the seasonal type of the lemon production area of ​​the current batch to be tested, input the seasonal type of the lemon production area of ​​the current batch to be tested into the knowledge graph for data matching, and obtain the estimated main defect feature types under the seasonal type of the lemon production area of ​​the current batch to be tested.

[0087] The historical identification speed information of the estimated main defect characteristics of the lemons in the current batch to be tested under the seasonal type of the lemon production area is used to identify the defect characteristics using the lemon defect identification model.

[0088] Based on the estimated main defect characteristics of the lemons produced in the current batch to be inspected under the seasonal type, the motion parameters of the lemon sorting machine with visual inspection are readjusted and optimized using the historical recognition speed information of the lemon defect recognition model.

[0089] It should be noted that different seasons bring different lemon defects, such as different lemon diseases. By using the historical recognition speed information of the lemon defect recognition model based on the estimated main defect characteristics under the seasonal type of the lemon production area of ​​the current batch to be tested, the motion parameters of the lemon sorting machine with visual inspection are re-optimized and adjusted. By adjusting the working parameters of the sorting machine according to the main defect characteristics, sorting efficiency can be further improved while reducing the phenomenon of missed detection.

[0090] In addition, this method also includes:

[0091] The characteristics of lemon sorting efficiency variation of sorting machines in the target workshop were statistically analyzed, a time series-based characteristic matrix of lemon sorting efficiency variation of sorting machines was constructed, and a lemon sorting efficiency prediction model was constructed based on deep neural networks.

[0092] The lemon sorting efficiency variation feature matrix of the time series-based sorting machine is input into the lemon machine sorting efficiency prediction model for training, so as to obtain a lemon machine sorting efficiency prediction model that meets the expectations.

[0093] The lemon sorting efficiency change characteristics of each sorting machine in the target workshop within a preset time are obtained. The lemon sorting efficiency change characteristics of each sorting machine in the target workshop within a preset time are used as model input and input into the lemon machine sorting efficiency prediction model that meets the expectations for prediction.

[0094] By predicting, the sorting efficiency characteristic data of lemons in the target workshop at the current timestamp is obtained, and the sorting efficiency requirement information of the current lemon sorting order is obtained. The lemon sorting efficiency characteristic data that is greater than the sorting efficiency requirement information of the current lemon sorting order is recommended as the priority sorting equipment for sorting.

[0095] It should be noted that this method can further improve the sorting efficiency of lemon sorting equipment.

[0096] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0097] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0098] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0099] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0100] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0101] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A control method for a lemon sorting machine with vision detection, characterized in that, The lemon sorting machine with vision detection includes a drainage device, a lemon machine cover, a feeding device, a belt feeding assembly, and a camera. The drainage device includes a first water receiving tray with a bracket and a second water receiving tray with a bracket. The second water receiving tray with a bracket is fixed to the first water receiving tray with a bracket. A water baffle and an electric cylinder mounting support plate are fixed on the second water receiving tray with a bracket. An electric cylinder and a linear guide rail are mounted on the electric cylinder mounting support plate. A slider push plate is mounted on the electric cylinder. The lemon machine cover includes a lemon machine lid and a door, and the camera is installed on the upper inner side of the lemon machine cover; The feeding device includes a feeding slurry, which is connected to a feeding slurry reduction motor via a coupling, and then cooperates with a feeding funnel to form the feeding device. The feeding device is installed on the left support plate and the right support plate of the drive wheel on the belt feeding assembly. The belt feeding assembly is installed on the second belt support water receiving tray. The belt feeding assembly includes a flat belt with hanging plates, and scraper blades are installed on the flat belt with hanging plates in a preset size array. It also includes the following steps: Acquire image data information of various historical defect types of lemons, and construct a lemon defect recognition model based on the image data information of various historical defect types of lemons; By testing the lemon defect recognition model, the recognition speed of the lemon defect recognition model within a unit time was obtained under different image quantities. Based on the recognition speed information of the lemon defect recognition model within a unit time under different image quantity information, the motion parameters of the lemon sorting machine with visual detection are configured. Based on the recognition speed information of the lemon defect recognition model within a unit time under different image quantity information, the motion parameters of the lemon sorting machine with visual detection are configured, specifically including: Randomly select the motion parameters of a lemon sorting machine with vision detection, and obtain the number of images collected per unit time based on the motion parameters of the lemon sorting machine with vision detection. Based on the number of images collected within the unit time and the recognition speed of the lemon defect recognition model within the unit time under different image quantity information, the recognition speed of the lemon defect recognition model under the number of images collected within the unit time is estimated. Obtain the maximum data processing speed of the data processing system, and determine whether the recognition speed information of the lemon defect recognition model under the number of image data collected within the unit time is greater than the maximum data processing speed of the data processing system; When the recognition speed of the lemon defect recognition model exceeds the maximum data processing speed of the data processing system under the number of images collected within the unit time, the motion parameters of the lemon sorting machine with vision detection are reconfigured. In addition, this method also includes: The main defect characteristics of lemons under various seasonal differences are obtained through big data, and a knowledge graph is constructed. The main defect characteristics of lemons under various seasonal differences are then input into the knowledge graph for storage. Obtain the seasonal type of the lemon production area of ​​the current batch to be tested, input the seasonal type of the lemon production area of ​​the current batch to be tested into the knowledge graph for data matching, and obtain the estimated main defect feature types under the seasonal type of the lemon production area of ​​the current batch to be tested. The historical identification speed information of the estimated main defect characteristics of the lemons in the current batch to be tested under the seasonal type of the production area is used to identify the defects through the lemon defect identification model. Based on the estimated main defect characteristics of the lemons produced in the current batch to be inspected under the seasonal type, the motion parameters of the lemon sorting machine with visual inspection are readjusted and optimized using the historical recognition speed information of the lemon defect recognition model. In addition, this method also includes: The characteristics of lemon sorting efficiency variation of sorting machines in the target workshop were statistically analyzed, a time series-based characteristic matrix of lemon sorting efficiency variation of sorting machines was constructed, and a lemon sorting efficiency prediction model was constructed based on deep neural networks. The lemon sorting efficiency variation feature matrix of the time series-based sorting machine is input into the lemon machine sorting efficiency prediction model for training, so as to obtain a lemon machine sorting efficiency prediction model that meets the expectations. The lemon sorting efficiency change characteristics of each sorting machine in the target workshop within a preset time are obtained. The lemon sorting efficiency change characteristics of each sorting machine in the target workshop within a preset time are used as model input and input into the lemon machine sorting efficiency prediction model that meets the expectations for prediction. By predicting, the sorting efficiency characteristic data of lemons in the target workshop at the current timestamp is obtained, and the sorting efficiency requirement information of the current lemon sorting order is obtained. The lemon sorting efficiency characteristic data that is greater than the sorting efficiency requirement information of the current lemon sorting order is recommended as the priority sorting equipment for sorting.

2. The control method for a lemon sorting machine with vision detection according to claim 1, characterized in that, The first water receiving tray with bracket, the second water receiving tray with bracket, the baffle plate, and the drain connector work together to form a drainage channel, which also serves as a support base for the lemon machine with camera.

3. The control method for a lemon sorting machine with vision detection according to claim 1, characterized in that, The camera is installed on the upper inner side of the lemon machine's outer cover. The lemon machine's outer cover is fixed to the second water receiving tray with a bracket of the drainage device and is supported by the first water receiving tray with a bracket.

4. The control method for a lemon sorting machine with vision detection according to claim 1, characterized in that: The feeding slurry is connected to the feeding slurry reduction motor via a coupling, and then cooperates with the feeding funnel to form the feeding device.

5. The control method for a lemon sorting machine with vision detection according to claim 1, characterized in that: The feeding device is installed on the left support plate and the right support plate of the drive wheel of the belt feeding assembly. The feeding funnel of the feeding device is inclined and an adjustable baffle is provided at the outlet.

6. The control method for a lemon sorting machine with vision detection according to claim 1, characterized in that: The belt feeding assembly has feeding slurry control sensors installed on the left and right support plates of the drive wheel, and a counting sensor installed at the lemon hopper of the belt feeding assembly.

7. The control method for a lemon sorting machine with vision detection according to claim 1, characterized in that: Acquire image data information of various historical defect types of lemons, and construct a lemon defect recognition model based on the image data information of various historical defect types of lemons, specifically including: Obtain image data information of various historical defect types of lemons, construct a training set based on the image data information of various historical defect types of lemons, and construct a lemon defect recognition model based on a deep neural network; The training set is input into the lemon defect recognition model for training, and the prediction accuracy information of the lemon defect recognition model during the training process is obtained, and a prediction accuracy threshold is set. Determine whether the prediction accuracy information of the lemon defect identification model during the training process is greater than the prediction accuracy threshold; When the prediction accuracy of the lemon defect recognition model during training is greater than the prediction accuracy threshold, the training of the lemon defect recognition model ends; when the prediction accuracy of the lemon defect recognition model during training is not greater than the prediction accuracy threshold, the training of the lemon defect recognition model continues.

8. The control method for a lemon sorting machine with vision detection according to claim 1, characterized in that: By testing the lemon defect recognition model, the recognition speed of the lemon defect recognition model per unit time was obtained under different image quantities, specifically including: Configure the number of images within a unit time period, set test conditions based on the number of images within the unit time period, and test the lemon defect recognition model based on the test conditions; Through testing, the recognition speed information of the lemon defect recognition model within a unit time under different image quantity information is obtained, and the recognition speed information of the lemon defect recognition model within a unit time under different image quantity information is output.

Images