Torreya grandis sorting device and torreya grandis detection method

Abstract

The invention relates to the technical field of intelligent detection of agricultural products, in particular to a torreya grandis sorting device and a torreya grandis detection method. The feeding mechanism comprises a discharging channel; the conveying mechanism comprises a conveying disc and a driving part, and the driving part is connected with the conveying disc; the discharging channel is arranged above the conveying disc; the detection mechanism comprises a camera and a control system; and the sorting mechanism is used for moving the detected articles out of the upper surface of the conveying disc. According to the torreya grandis sorting device, the rotatable conveying disc is arranged to serve as a conveying mechanism, reasonable position arrangement of the feeding mechanism, the detecting mechanism and the sorting mechanism is facilitated, the different mechanisms do not interfere with one another during cooperative work, and the space utilization rate is increased; and the camera is arranged to perform visual identification on the quality of the Chinese torreya, so that the problem that the shell-broken Chinese torreya cannot be mechanically sorted is solved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of intelligent detection of agricultural products, and particularly relates to a Chinese torreya nut sorting device and a Chinese torreya nut detection method. BACKGROUND

[0002] Chinese torreya nut is a unique and rare dry fruit tree species. After being fried, Chinese torreya nut seeds are crisp and delicious, and are a kind of high-quality dry fruit with high economic value. Classification and hull breaking are key links in the processing of Chinese torreya nut seeds. The fruit of Chinese torreya nut is wrapped by a shell, and is different from other nuts. The shell surface is smooth and has no obvious gap, so it is difficult to peel. There is a lack of hull breaking equipment for Chinese torreya nuts on the market. In the process of mechanical hull breaking, the quality of hull breaking is uneven due to the different sizes of Chinese torreya nuts. In the case of a certain stroke, the mechanical hull breaking cannot meet the needs of different sizes of Chinese torreya nuts, and small-size Chinese torreya nuts may have complete shells, while large-size Chinese torreya nuts may be damaged. How to distinguish Chinese torreya nuts of different qualities is a problem encountered in production.

[0003] CN204653707U provides a Chinese torreya nut vibrating screen cleaning machine, which comprises a rack, a water tank, a double-layer vibrating screen, a feeding hopper, a large Chinese torreya nut outlet, a small Chinese torreya nut outlet, a bucket wheel, and a conveyor. The water tank is installed on the rack. The double-layer vibrating screen is installed on the rack by a spring and is located in the water tank. The feeding hopper is located above the double-layer vibrating screen. The large Chinese torreya nut outlet and the small Chinese torreya nut outlet are respectively located on the left side and the right side of the water tank. The bucket wheel is installed between the double-layer vibrating screen and the large Chinese torreya nut outlet. One end of the conveyor is located at the bottom of the double-layer vibrating screen, and the other end is connected to the small Chinese torreya nut outlet. The Chinese torreya nut vibrating screen cleaning machine has the advantages of simple structure, high automation, sorting and cleaning functions, water saving, and the like.

[0004] CN215189264U provides a new Chinese torreya nut cleaning and sorting device, which comprises a cleaning cavity, a screening cavity, and a first screening roller. Water pipes are installed on both sides of the top end of the cleaning cavity. A water pipe is installed on the top end of the cleaning cavity. A screening cavity is installed at the bottom end of the cleaning cavity. A second screening roller is arranged at the top end of the first screening roller, and the two sides of the second screening roller are fixedly connected to the two sides inside the screening cavity. Connection rollers are arranged inside the first screening roller, and bearings are installed at both ends of the connection rollers. The new Chinese torreya nut cleaning and sorting device has a filtering structure. After cleaning, the water flows into the inside of the screening cavity through the flow port. When the water flows to the bottom end inside the screening cavity, the water passes through the filter plate, which can filter the impurities in the water. By opening the water valve, the treated water can be discharged or recycled, reducing pollution of water resources and increasing energy saving.

[0005] The above patents are all for sorting unbroken Chinese torreya nuts, and cannot solve the problem of sorting broken Chinese torreya nuts. SUMMARY

[0006] The present application aims to overcome the shortcomings of the prior art Chinese torreya nut sorting device that cannot sort broken Chinese torreya nuts.

[0007] In one aspect, the present application provides a Chinese torreya nut sorting device, comprising:

[0008] A rack comprising a mounting table;

[0009] A feeding mechanism, the feeding mechanism comprising a discharge channel;

[0010] A conveying mechanism, the conveying mechanism comprising a conveying disc and a driving member, the driving member being connected to the conveying disc and driving the conveying disc to rotate; the conveying disc being installed on the upper surface of the mounting table; the discharge channel being arranged above the conveying disc;

[0011] A detection mechanism, the detection mechanism comprising a camera and a control system, the camera being electrically connected to the control system; the camera being installed on the surface of the mounting table;

[0012] A sorting mechanism, the sorting mechanism being used to remove the detected articles from the upper surface of the conveying disc.

[0013] According to an embodiment of the present application, the feeding mechanism comprises a vibrating feeding disc, the vibrating feeding disc being arranged on one side of the rack.

[0014] According to an embodiment of the present application, the feeding mechanism further comprises a discharging valve, the discharging valve being installed at the discharge port of the discharge channel.

[0015] According to an embodiment of the present application, the upper surface of the conveying disc is provided with a plurality of positioning grooves, the plurality of positioning grooves being arranged at intervals.

[0016] According to an embodiment of the present application, the upper surface of the conveying disc is provided with an annular positioning groove, the annular positioning groove being concentrically arranged with the conveying disc.

[0017] According to an embodiment of the present application, the disc body of the conveying disc is transparent.

[0018] According to an embodiment of the present application, the camera comprises two, the two cameras being arranged on the upper and lower sides of the conveying disc, respectively.

[0019] According to an embodiment of the present application, the sorting mechanism comprises a first track and a second track, the feed end of the first track and the second track is arranged close to the lower surface of the conveying disc; the blowing assembly comprises a gas nozzle and a solenoid valve for controlling the gas nozzle, and the solenoid valve is electrically connected with the control system.

[0020] The present application also provides a Chinese torreya detection method, which comprises:

[0021] Step S1, the feeding mechanism conveys Chinese torreya to the conveying disc through the discharge channel, and the driving member drives the conveying disc to rotate by a fixed angle to a detection point;

[0022] Step S2, the camera collects Chinese torreya image data, and inputs the image data into the trained YOLO-C2Faster model, and outputs detection data from the YOLO-C2 Faster model and judges the quality of Chinese torreya;

[0023] Step S3, the detection data is sent to the control system, a driving signal is generated from the control system to the sorting mechanism, and the Chinese torreya is transferred to the inferior product track or the superior product track by the sorting mechanism.

[0024] According to an embodiment of the present application, the YOLO-C2Faster model is based on the backbone model in YOLO v8n, replaces the C2f module with the C2f-faster module, adds a CBAM channel attention mechanism module in the backbone model, and uses the SToU function as the loss function.

[0025] Compared with the prior art, the Chinese torreya sorting device provided by the present application guarantees the continuous operation of the sorting work through the feeding mechanism feeding the transmission mechanism, thereby improving the sorting efficiency; the rotatable conveying disc is set as the transmission mechanism, which is beneficial to the reasonable position setting of the feeding mechanism, the detection mechanism and the sorting mechanism, so that different mechanisms do not interfere with each other when working cooperatively, and the space utilization is improved; the camera is set to visually identify the quality of Chinese torreya, so that the Chinese torreya sorting efficiency is high and the precision is high, and the problem that the broken shell Chinese torreya cannot be sorted by machinery is solved; the sorting mechanism is set, so that the sorting mechanism can transfer the identified Chinese torreya to a specific position, and reserve space for the Chinese torreya to be identified.

[0026] Other features and advantages of the present application will be described in detail in the following specific embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0027] The accompanying drawings, which form a part of the present application, are used to provide a further understanding of the present application, and the illustrative embodiments thereof and their descriptions are used to explain the present application, and do not constitute an improper limitation of the present application. In the drawings:

[0028] Figure 1 Structure diagram of the torreya grandis sorting device according to an embodiment of the present application;

[0029] Figure 2 Front view of the torreya grandis sorting device according to an embodiment of the present application;

[0030] Figure 3 Top view of the torreya grandis sorting device according to an embodiment of the present application;

[0031] Figure 4 Top view of the torreya grandis sorting device according to an embodiment of the present application;

[0032] Figure 5 Left view of the torreya grandis sorting device according to an embodiment of the present application;

[0033] Figure 6 YOLO-C2Faster model diagram according to an embodiment of the present application;

[0034] Figure 7 C2f-faster module structure diagram according to an embodiment of the present application;

[0035] Figure 8 FasterNetBlock structure diagram according to an embodiment of the present application;

[0036] Figure 9 PConv working principle diagram according to an embodiment of the present application.

[0037] Legend:

[0038] 1 rack 2 feeding mechanism

[0039] 3 transmission mechanism 4 detection mechanism

[0040] 5 sorting mechanism 21 discharge channel

[0041] 31 conveying disc 41 camera

[0042] 42 control system 51 first track

[0043] 52 second track 53 air blowing assembly

[0044] 311 positioning groove 312 annular positioning groove DETAILED DESCRIPTION

[0045] The specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely intended to illustrate and explain the present application, and are not intended to limit the present application.

[0046] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0047] It should be noted that the terms "first," "second," "third," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of the invention described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0048] To address the above problems, the present invention provides a torreya grandis sorting device, such as... Figures 1 to 5 As shown, it includes:

[0049] The frame 1 includes a mounting platform, the upper and lower surfaces of which can both serve as mounting surfaces, and the lower surface of the mounting platform is provided with multiple support legs.

[0050] The feeding mechanism 2 includes a discharge channel 21, and the feeding mechanism 2 includes a vibrating feeding device, a conveyor belt feeding device, or a screw feeding device.

[0051] The transmission mechanism 3 includes a conveying disk 31 and a driving component. The driving component is connected to the conveying disk 31 and drives the conveying disk 31 to rotate. The driving component includes a motor, preferably a stepper motor. The conveying disk 31 is mounted on the upper surface of the mounting platform. The discharge channel 21 is located above the conveying disk 31, which allows the torreya nuts to fall onto the conveying disk 31 through the discharge channel 21.

[0052] The testing mechanism 4 includes a camera 41 and a control system 42, with the camera 41 electrically connected to the control system 42. The camera 41 is mounted on the surface of the mounting platform; optionally, the camera 41 may be an industrial camera. The camera 41 is used to capture images of the torreya nuts and transmit them to the control system 42. The control system 42 identifies the images and determines the quality of the torreya nuts. After the determination is completed, the control mechanism sends a corresponding signal.

[0053] A sorting mechanism 5 is arranged for moving the detected articles from the upper surface of the conveying disc 31, and the sorting mechanism 5 receives signals from the control mechanism and makes corresponding actions.

[0054] Compared with the prior art, the Chinese torreya sorting device provided by the application has the following advantages: the feeding mechanism 2 feeds the transmission mechanism 3, thereby ensuring the continuous operation of the sorting work and improving the sorting efficiency; the rotatable conveying disc 31 is arranged as the transmission mechanism 3, which is beneficial to the reasonable position arrangement of the feeding mechanism 2, the detection mechanism 4 and the sorting mechanism 5, so that the different mechanisms do not interfere with each other when working cooperatively, and the space utilization is improved; the camera 41 is arranged to visually identify the quality of the Chinese torreya, so that the Chinese torreya sorting efficiency is high and the precision is high, and the problem that the broken-shell Chinese torreya cannot be sorted by a machine is solved; the sorting mechanism 5 is arranged, so that the sorting mechanism 5 can transfer the identified Chinese torreya to a specific position and reserve space for the Chinese torreya to be identified.

[0055] In order to better match the transmission mechanism 3, the feeding mechanism 2 comprises a vibrating feeding disc, which is arranged on one side of the rack 1 and is electrically connected with the control system 42. The vibrating feeding disc arranges the disordered Chinese torreya in order through periodic vibration and outputs the Chinese torreya through the discharge channel 21; further, the vibrating feeding disc can complete feeding at a specific rhythm by starting and pausing, or control the feeding rhythm by controlling the vibration frequency. Such an arrangement makes the vibration of the vibrating feeding disc as little as possible to be transmitted to the conveying mechanism, the transmission mechanism 3 and the sorting mechanism 5, thereby improving the running stability of the whole machine. At the same time, since the detection object of the Chinese torreya detection device is the broken-shell Chinese torreya, the broken-shell Chinese torreya is lighter on the side where the shell is cracked or missing, and under the action of vibration, the heavier side of the Chinese torreya moves downward under the action of gravity, and the lighter side of the Chinese torreya moves upward, so that the Chinese torreya falls onto the upper surface of the conveying disc 31 in this posture, which is beneficial to the image acquisition of the camera 41 on the cracked or missing part of the Chinese torreya, thereby improving the detection accuracy.

[0056] In order to better control the feeding rhythm, the feeding mechanism 2 further comprises a discharging valve, which is installed at the discharge port of the discharge channel 21. The discharging valve controls the feeding rhythm by opening and closing rhythmically.

[0057] In order to improve the detection accuracy, according to one embodiment of the application, as Figure 3As shown in the drawings, the upper surface of the conveying disc 31 is provided with a plurality of positioning grooves 311, the plurality of positioning grooves 311 are arranged at intervals, preferably, the length direction of the positioning grooves 311 is close to perpendicular to the radial direction of the rotating disc. Such arrangement makes the torreya grandis not produce displacement in the radial direction and turn around itself during the rotation of the rotating disc.

[0058] According to another embodiment of the present application, as Figure 4 As shown in the drawings, the upper surface of the conveying disc 31 is provided with an annular positioning groove 312, and the annular positioning groove 312 is concentrically arranged with the conveying disc 31.

[0059] In order to improve the sorting efficiency, the disc body of the conveying disc 31 is transparent. Such arrangement makes the camera 41 can continue image scanning of the torreya grandis from multiple angles, and multiple surfaces of the torreya grandis are detected and identified at the same time, thereby improving the detection efficiency; at the same time, compared with only detecting the upper surface of the torreya grandis, the detection of the lower surface of the torreya grandis is increased, and the detection accuracy is greatly improved.

[0060] In the above, as Figure 2 As shown in the drawings, the camera 41 includes two, and the two cameras 41 are respectively installed on the upper surface and the lower surface of the mounting table, and the two cameras 41 are respectively arranged on the upper and lower sides of the conveying disc 31, preferably, the two cameras 41 are oppositely arranged, such arrangement makes the two cameras 41 can simultaneously acquire the images of the upper surface and the lower surface of the same torreya grandis.

[0061] In order to facilitate the screening of torreya grandis with different qualities, as Figure 4 As shown in the drawings, the sorting mechanism 5 includes a first track 51, a second track 52 and a blowing assembly 53, the feeding end of the first track 51 and the second track 52 is arranged close to the lower surface of the conveying disc 31; the blowing assembly 53 includes a gas nozzle and an electromagnetic valve for controlling the gas nozzle, and the electromagnetic valve is electrically connected with the control system 42. The first track 51 and the second track 52 are respectively inferior product track and superior product track. The electromagnetic valve receives the electrical signal from the control system 42, thereby controlling the opening and closing of the gas nozzle; preferably, the blowing assembly 53 is two, and the two blowing assemblies 53 are respectively arranged corresponding to the first track 51 and the second track 52. When the gas nozzle blows, the airflow blows the torreya grandis away from the surface of the conveying disc 31, and the torreya grandis is selectively fallen into the first track 51 or the second track 52.

[0062] The present application provides a torreya grandis detection method, as Figures 6-7 As shown in the drawings, the torreya grandis detection method includes:

[0063] Step S1, the feeding mechanism 2 transports the torreya grandis to the conveying disc 31 through the discharge channel 21, and the driving member drives the conveying disc 31 to rotate by a fixed angle to a detection point; in the above, whether the torreya grandis in the discharge channel 21 falls above the conveying disc 31 is detected by the laser sensor, and the laser sensor is electrically connected with the control system 42;

[0064] Step S2, the camera 41 collects the image data of the torreya grandis, inputs the image data into the trained YOLO-C2 Faster model, and outputs the detection data from the YOLO-C2 Faster model and judges the quality of the torreya grandis;

[0065] Step S3, the detection data is sent to the control system 42, a driving signal is generated from the control system 42 to the sorting mechanism 5, and the torreya grandis is transferred to the poor quality track or the high quality track by the sorting mechanism 5. Specifically, after the control system 42 receives the monitoring data, a driving signal is generated from the control system 42 to the sorting mechanism 5, and the sorting mechanism 5 controls the opening and closing of the two electromagnetic valves to blow the torreya grandis to the poor quality track or the high quality track.

[0066] The torreya grandis detection method realizes the quality detection and sorting of the torreya grandis through the automatic process of steps S1-S3. The feeding mechanism transports the torreya grandis to the conveying disc through the discharge channel, and the driving member drives the conveying disc to rotate to the detection point. In this process, the laser sensor monitors whether the torreya grandis falls accurately above the conveying disc in real time, and feeds back the signal to the control system; then the camera collects the image data of the torreya grandis, inputs the trained YOLO-C2 Faster model for quality detection, and outputs the detection data; finally, the control system generates a driving signal according to the detection data, controls the opening and closing of the two electromagnetic valves of the sorting mechanism, and accurately blows the torreya grandis to the poor quality track or the high quality track, completing the automatic sorting. In the backbone network, the C2f-faster lightweight module is used to replace the C2f module in the middle, reducing the parameters and calculation amount; the C BAM attention mechanism is added to enhance the learning ability of the model to the feature map; the SIoU loss function is introduced in the detection head to improve the positioning ability of the torreya grandis. The experimental results show that the precision and mAP50 of the lightweight model are improved by 1.2% and 0.2% respectively compared with the basic model, and the recall rate is flat with the basic model. The calculation amount and parameter amount are reduced by 21.3% and 18.8% respectively, and the frame rate is increased by 17.8%. The lightweight algorithm model is tested on the real machine, and the correct rates of the device for sorting four kinds of torreya grandis, i.e. high quality torreya grandis, shell rupture, overcooked, and torreya grandis pulp, are 81.5%, 76.7%, 100%, and 100% respectively. The lightweight algorithm model improves the detection efficiency while maintaining the accuracy of the basic algorithm model, and is suitable for accurate and rapid detection of torreya grandis.

[0067] In YOLO v8, the C2f module is a CSP bottleneck module with two convolutions (Cross Stage Partial Bottleneck with 2 Convolutions).

[0068] According to an embodiment of the present application, the YOLO-C2Faster model is based on the backbone model in YOLO v8n, replaces the C2f module with the C2f-faster module, adds a CBAM channel attention mechanism module (Convolutional Block Attention Module, CBAM) in the middle of the backbone model, and uses the function (Scale-Invariant Intersection over Union, SIoU) as the loss function.

[0069] SIoU is a loss function used in target detection, and SIoU is composed of four loss functions: angle loss, distance loss, shape loss, and IoU loss. It aims to improve the similarity calculation between target boxes. The SIoU loss function introduces the concept of angle loss function, that is, the vector angle size between the real box and the predicted box, so as to avoid the possibility of a worse model due to the position of the predicted box being indefinite during the training process.

[0070] YOLO v8n includes a C2f module in the backbone model Backbone replaced by a C2f-faster module. To further improve the model performance, while improving the model training speed and accuracy, a CBAM channel attention mechanism module is set in the last module of the backbone model Backbone. After the feature map is input into the CBAM attention mechanism, it will first pass through the channel attention module, learn the information at the channel level, and strengthen the key channel features. Generally, if the input image is a grayscale image, it is a 1-channel feature map, if the input image is an RGB image, it is a 3-channel feature map, and if the input image is an intermediate layer feature map, it is a multi-channel feature map. After the multi-channel feature map is adjusted by the channel attention module in CBAM, it will enter the spatial attention module to learn the key spatial information of the feature map, and finally output a feature map with key channel information and key spatial information. After the feature map with channel weight information enters the spatial attention module, it will first be down-sampled by global max pooling and global average pooling. Global max pooling is used to capture key information, and global average pooling is used to capture context information. After global max pooling and global average pooling, two feature maps with a size of H*W*1 are generated. Secondly, the two feature maps with a size of H*W*1 are spliced in the channel dimension to obtain a feature map with a size of H*W*2, and then a convolution operation is performed on it to change the number of channels of the feature map to 1. The feature map with a size of H*W*1 after convolution is activated by a sigmoid function to map the spatial features to the range of 0-1. Finally, the feature map with a size of H*W*1 with spatial weight information is multiplied by F 1 , and the final output is a feature map F 11 with channel weight information and spatial weight information with a size of H*W*C.

[0071] In fact, the C2f module increases the computational complexity and parameter amount of the model while enhancing the feature extraction capability, resulting in an increase in the time cost of training and inference, as well as an increase in the demand for storage space and computing resources.

[0072] Among them, as Figure 7As shown, the C2f-faster includes a first convolution module, at least two layers of FasterNetBlock, and a second convolution module in series. Among them, a split operation is arranged between the first convolution module and the first FasterNetBlock. A part after the split is spliced with the output of each layer of FasterNetBlock through a Concat operation, and the data after the Concat operation is input into the second convolution module. Another part after the split is input into the at least two layers of FasterNetBlock for layer-by-layer calculation. The feature map input into the C2f-faster module generates an intermediate feature map through a convolution layer, and the intermediate feature map is split into two parts, one of which is directly transmitted to the Concat module, and the other is transmitted to the multi-layer FasterNetBlock module for feature extraction. Finally, all branches are spliced in the channel dimension to realize feature fusion. Compared with the Bottleneck module, the FasterNetBlock module cancels the option of whether to use a shortcut connection, and all use this connection mode. From Figure 8 It can be seen that the FasterNetBlock module has a PConv layer followed by two Conv1*1 layers. They are displayed together as an inverted residual block. Among them, the middle layer expands the number of channels and places a Shorcut to reuse the input features. The PConv structure of the FasterNetBlock reduces the number of floating-point operations compared with the traditional Conv structure, thereby improving the operation speed.

[0073] As Figure 9 shown, PConv utilizes the redundancy in the feature map to systematically apply convolution only on a part of the output channels without affecting the remaining channels. The first or last consecutive channel number c1 is used as the channel number of the feature map for calculation of consecutive memory usage, and the number of floating-point operations of PConv is FLOP sPconv ,

[0074] FLOPs = h * w * k 2 * c1 2

[0075] Among them, h is the height of the output feature map, w is the width of the output feature map, k 2 is the area of the convolution kernel; c1 is the number of consecutive channels, and c is the total number of channels.

[0076] FLOP sPConv = 1 / 16 * FLOP sConv

[0077] Among them, FLOP sConvThe floating point operation number of the C2f module is C2f-faster module floating point operation number is only 1 / 16 of the original. Compared with the C2f module, C2f-faster changes the convolution structure to ensure the accuracy, reduces the floating point operation number, and improves the operation speed of the model.

[0078] It should be noted that, for the foregoing method embodiments, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the action sequence described, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the actions and modules involved are not necessarily necessary for the present application.

[0079] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0080] In several embodiments provided by the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only schematic, and the division of units is only a logical function division, and there can be another division way in actual implementation, for example, 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 or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical or other forms.

[0081] The above only describes the preferred embodiments of the present application and is not used to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A torreya grandis sorting device, comprising: a rack (1) comprising a mounting table; a feeding mechanism (2) comprising a discharge channel (21); a conveying mechanism (3) comprising a conveying disc (31) and a driving member connected with the conveying disc (31) and driving the conveying disc (31) to rotate; the conveying disc (31) is installed on the upper surface of the mounting table; the discharge channel (21) is arranged above the conveying disc (31); a detection mechanism (4) comprising a camera (41) and a control system (42), the camera (41) is electrically connected with the control system (42); the camera (41) is installed on the surface of the mounting table; a sorting mechanism (5) for moving the detected articles from the upper surface of the conveying disc (31).

2. The torreya grandis sorting device according to claim 1, characterized in that, The feeding mechanism (2) comprises a vibrating feeding disc arranged on one side of the rack (1).

3. The Japanese nut sorting device according to claim 2, characterized by The feeding mechanism (2) further comprises a discharging valve installed at the discharge port of the discharge channel (21).

4. The Japanese nut sorting apparatus according to claim 1, characterized by A plurality of positioning grooves (311) are arranged on the upper surface of the conveying disc (31).

5. The Japanese nut sorting apparatus according to claim 1, wherein An annular positioning groove (312) is arranged on the upper surface of the conveying disc (31) and is concentrically arranged with the conveying disc (31).

6. The Japanese nut sorting apparatus according to claim 1, wherein The disc body of the conveying disc (31) is transparent.

7. The Japanese nut sorting device according to claim 6, characterized by The camera (41) comprises two cameras (41) arranged on the upper and lower sides of the conveying disc (31) respectively.

8. The Japanese nut sorting apparatus according to claim 1, characterized by The sorting mechanism (5) comprises a first track (51), a second track (52) and a blowing assembly (53), the feeding ends of the first track (51) and the second track (52) are arranged adjacent to the lower surface of the conveying disc (31); the blowing assembly (53) comprises a gas nozzle and a solenoid valve for controlling the gas nozzle, and the solenoid valve is electrically connected with the control system (42).

9. A method for detecting Torreya grandis according to any one of claims 1 to 8, wherein The torreya grandis detection method comprises: Step S1: The feeding mechanism (2) transports the torreya grandis to the conveying disc (31) through the discharge channel (21), and the driving member drives the conveying disc (31) to rotate by a fixed angle to a detection point; Step S2: The camera (41) collects the image data of the torreya grandis, inputs the image data into the trained YOLO-C2Faster model, outputs the detection data from the YOLO-C2Faster model and judges the quality of the torreya grandis; Step S3: The detection data is sent to the control system (42), the control system (42) generates a driving signal to the sorting mechanism (5), and the sorting mechanism (5) transfers the torreya grandis to the inferior product track or the superior product track.

10. The torreya grandis detection method according to claim 9, wherein, The YOLO-C2Faster model is based on the backbone model in YOLOv8n, replaces the C2f module with the C2f-faster module, adds a CBAM channel attention mechanism module in the backbone model, and uses a SIoU function as the loss function.

Images