Tobacco impurity removing method and system

Abstract

The embodiment of the invention discloses a tobacco impurity removing method and system. The method is applied to the tobacco impurity removal system, the tobacco impurity removal system comprises at least one image acquisition device, a controller connected with the image acquisition device and a removal device connected with the controller, and the method comprises the following steps: collecting a tobacco image of to-be-detected tobacco conveyed on a conveyor belt through the image acquisition device, the tobacco leaf image is sent to a controller; inputting the received tobacco leaf image into a pre-trained impurity identification model through the controller, determining the impurity position of the impurity in the to-be-detected tobacco leaf in the tobacco leaf image, generating a rejection control instruction based on the impurity position, and sending the rejection control instruction to a rejection device; and removing impurities in the to-be-detected tobacco leaves through the removing equipment based on the received removing control instruction. According to the technical scheme, the accuracy and efficiency of impurity recognition are improved, the impurities are removed through the removing equipment, manual treatment is not needed, and the production efficiency is improved.

Description

Technical Field

[0001] This invention relates to the field of tobacco debris removal technology, and more particularly to a method and system for removing tobacco debris. Background Technology

[0002] During the process of tobacco harvesting, re-drying, and packing, a lot of impurities are often present. In order to reduce the impurity content during the tobacco collection process, it is necessary to remove impurities from the collected tobacco.

[0003] In existing technologies, tobacco leaves are typically placed on a conveyor belt for transport during tobacco production. Workers manually inspect the leaves for impurities and remove them. However, in developing this invention, it was discovered that existing technologies have at least the following problems: workers spend long periods focusing on the moving tobacco leaves, which can cause visual fatigue and lead to missed inspections; furthermore, manually removing impurities is labor-intensive and time-consuming, increasing production costs and reducing efficiency. Summary of the Invention

[0004] This invention provides a method and system for removing tobacco debris, aiming to improve the efficiency and accuracy of debris identification and reduce production costs.

[0005] According to one aspect of the present invention, a method for removing tobacco debris is provided, applied to a tobacco debris removal system, the tobacco debris removal system comprising at least one image acquisition device, a controller connected to the image acquisition device, and a removal device connected to the controller, the method comprising:

[0006] The image acquisition device acquires images of the tobacco leaves to be inspected being transported on the conveyor belt, and sends the tobacco leaf images to the controller;

[0007] The controller inputs the received tobacco leaf image into a pre-trained debris recognition model to determine the location of the debris in the tobacco leaf image, generates a rejection control command based on the debris location, and sends the rejection control command to the rejection device.

[0008] The rejection device removes impurities from the tobacco leaves to be inspected based on the received rejection control command.

[0009] According to another aspect of the present invention, a tobacco debris removal system is provided, comprising: at least one image acquisition device, a controller connected to the image acquisition device, and a removal device connected to the controller;

[0010] The image acquisition device is used to acquire images of tobacco leaves to be inspected being transported on a conveyor belt, and to send the tobacco leaf images to the controller;

[0011] The controller is used to input the received tobacco leaf image into a pre-trained debris recognition model, determine the location of the debris in the tobacco leaf image, generate a rejection control command based on the debris location, and send the rejection control command to the rejection device.

[0012] The rejection device is used to reject impurities in the tobacco leaf to be tested based on the received rejection control command.

[0013] The technical solution of this invention is applied to a tobacco debris removal system. The system includes at least one image acquisition device, a controller connected to the image acquisition device, and a removal device connected to the controller. The image acquisition device acquires images of tobacco leaves to be inspected transported on a conveyor belt and sends these images to the controller, eliminating the need for manual inspection. The tobacco leaf images reflect information about the tobacco leaves to be inspected. The controller inputs the received tobacco leaf images into a pre-trained debris recognition model to determine the location of debris in the tobacco leaf image. Based on the debris location, a removal control command is generated and sent to the removal device. The removal device then removes the debris from the tobacco leaf based on the received removal control command. This technical solution eliminates the need for manual intervention. The debris recognition model identifies the location of debris in the tobacco leaf image, improving the accuracy and efficiency of debris recognition and reducing missed detections. Furthermore, the removal of debris by the removal device eliminates the need for manual processing, thus reducing production costs and increasing production efficiency.

[0014] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart of a method for removing tobacco debris according to an embodiment of the present invention;

[0017] Figure 2 This is a flowchart of another method for removing tobacco debris provided according to an embodiment of the present invention;

[0018] Figure 3This is a flowchart of another method for removing tobacco debris provided according to an embodiment of the present invention;

[0019] Figure 4 This is a schematic diagram of a rejection device provided according to an embodiment of the present invention;

[0020] Figure 5 This is a flowchart of a vision-based tobacco debris removal process according to an embodiment of the present invention;

[0021] Figure 6 This is a schematic diagram of a tobacco debris removal system provided according to an embodiment of the present invention. Detailed Implementation

[0022] 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.

[0023] It should be noted that the terms "first," "second," 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 so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "etc." and any variations thereof are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises 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.

[0024] It should be noted that the collection, gathering, updating, analysis, processing, use, transmission, and storage of user personal information involved in this disclosed technical solution all comply with relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. Necessary measures are taken to prevent unauthorized access to user personal information data and to safeguard user personal information security and network security.

[0025] Figure 1This is a flowchart of a tobacco impurity removal method according to an embodiment of the present invention. This embodiment is applicable to situations where impurities present in tobacco leaves need to be removed during tobacco production. The method of this embodiment is applied to a tobacco impurity removal system, which includes at least one image acquisition device, a controller connected to the image acquisition device, and a removal device connected to the controller. The system interfaces with the tobacco production line through a standardized interface, ensuring equipment stability and compatibility. Furthermore, the system provides multiple different interfaces to support multiple production line modes, adapting to different post-cut tobacco processing scenarios, facilitating production expansion. Figure 1 As shown, the method may specifically include:

[0026] S110. Acquire images of the tobacco leaves to be inspected on the conveyor belt using an image acquisition device, and send the tobacco leaf images to the controller.

[0027] The image acquisition equipment may include cameras, laser scanning cameras, and industrial cameras. The image acquisition equipment can be installed above the conveyor belt to capture images of items transported on the conveyor belt.

[0028] In this embodiment, the tobacco leaves to be tested can be laid flat on a conveyor belt, which continuously transports the tobacco leaves for impurity removal and shredding. During the transport of the tobacco leaves, the image acquisition device can be aligned with the conveyor belt to capture images of the tobacco leaves being transported on the conveyor belt, which are then sent to the controller.

[0029] Specifically, a periodic acquisition method can be adopted to control the image acquisition device to acquire images of tobacco leaves. The acquisition period can be determined based on the conveyor belt speed to ensure that each portion of tobacco leaves to be inspected laid out on the conveyor belt is captured with a corresponding image.

[0030] S120. The received tobacco leaf image is input into the pre-trained debris recognition model through the controller to determine the location of the debris in the tobacco leaf image, generate a rejection control command based on the debris location, and send the rejection control command to the rejection device.

[0031] In this embodiment, after receiving the tobacco leaf image, the controller can perform image preprocessing operations on the image. For example, image preprocessing operations include image cropping, rotation, magnification, reduction, or resolution adjustment. Through image preprocessing, the format of the acquired tobacco leaf image meets preset format requirements.

[0032] In practice, after preprocessing the tobacco leaf images, the images can be input into a pre-trained debris recognition model. Optionally, the debris recognition model is trained based on a deep learning model and embeds a multi-receptor field feature adaptive fusion module. For example, the debris recognition model can be an object detection model, such as the YOLO8 model.

[0033] In this embodiment, the clutter recognition model employs a multi-receptive-field feature adaptive fusion module, which is a technique used to improve the performance of deep learning models. By using convolutional kernels of different sizes in parallel, feature maps with different receptive fields are generated. These feature maps are then adaptively weighted and fused using an attention feature fusion mechanism, thereby highlighting effective feature information and improving the representation ability of the clutter recognition model.

[0034] For scenarios involving the detection of impurities in tobacco leaves, the size and style of the impurities vary, making it difficult to comprehensively identify smaller impurities using conventional object detection models. In this embodiment, by embedding a multi-receptive-field feature adaptive fusion module into the impurity recognition model, the characteristics of different impurities can be observed, facilitating the effective identification of impurities of different sizes and styles and reducing missed detections.

[0035] It should be noted that the debris recognition model can be pre-trained. For example, a YOLO8 model embedded with a multi-receptive-field feature adaptive fusion module is trained using a series of non-tobacco debris samples to obtain the trained debris recognition model. This model enables accurate identification of different types of debris and effectively analyzes multi-dimensional features such as color, shape, and texture, significantly improving the detection accuracy of small debris. Furthermore, based on historical time periods, when detecting debris in tobacco leaves, the non-tobacco debris samples can be continuously updated and optimized using tobacco leaf images acquired by image acquisition equipment. This allows for updating samples with real production data, achieving adaptive learning and further improving recognition accuracy and removal efficiency.

[0036] Optionally, after the tobacco leaf image is input into the debris recognition model, the output of the debris recognition model is the image position of each debris contained in the tobacco leaf image, i.e., the debris position. In order to accurately remove debris, a removal control command can be generated based on the debris position and sent to the removal device.

[0037] In this embodiment, after receiving the tobacco leaf image, the controller can also generate a stop operation command and send it to the conveyor belt to stop its movement. Correspondingly, the implementation of generating a rejection control command based on the location of the debris may include: determining the debris coordinates on the conveyor belt based on the debris location and a pre-defined image acquisition area on the conveyor belt; generating a rejection control command based on the debris coordinates; and sending the rejection control command to the rejection device.

[0038] S130. The rejecting device removes impurities from the tobacco leaves to be inspected based on the received rejecting control command.

[0039] Miscellaneous items may include plastic sheets, hemp rope, nylon rope, chicken feathers, paper scraps, and metal shavings.

[0040] It should be noted that since the conveyor belt has stopped moving, the debris is also stationary on the conveyor belt. After receiving the rejection control command, the rejection device parses the command to obtain the coordinates of the debris on the conveyor belt. Since the debris is stationary on the conveyor belt, it can be rejected according to these coordinates.

[0041] Optionally, the rejection device may include a robotic arm and clamps mounted on the robotic arm. After the rejection device determines the coordinates of the debris, it controls the robotic arm to move to the location of the debris and uses the clamps to pick up the item at the location of the debris, thereby removing the debris from the tobacco leaves to be inspected.

[0042] The technical solution of this invention is applied to a tobacco debris removal system. The system includes at least one image acquisition device, a controller connected to the image acquisition device, and a removal device connected to the controller. The image acquisition device acquires images of tobacco leaves to be inspected transported on a conveyor belt and sends these images to the controller, eliminating the need for manual inspection. The tobacco leaf images reflect information about the tobacco leaves to be inspected. The controller inputs the received tobacco leaf images into a pre-trained debris recognition model to determine the location of debris in the tobacco leaf image. Based on the debris location, a removal control command is generated and sent to the removal device. The removal device then removes the debris from the tobacco leaf based on the received removal control command. This technical solution eliminates the need for manual intervention. The debris recognition model identifies the location of debris in the tobacco leaf image, improving the accuracy and efficiency of debris recognition and reducing missed detections. Furthermore, the removal of debris by the removal device eliminates the need for manual processing, thus reducing production costs and increasing production efficiency.

[0043] Figure 2This is a flowchart of another tobacco debris removal method provided by an embodiment of the present invention. Based on the above embodiments, this embodiment generates a removal control command based on the debris position as follows: determining the actual movement speed of the debris on the conveyor belt; determining the arrival time of the debris when it is transferred from the image acquisition device to the removal device based on the actual movement speed; and generating a removal control command based on the debris position when the current time is the arrival time. The explanations of terms that are the same as or corresponding to those in the above embodiments are not repeated here. Figure 2 As shown, the method includes:

[0044] S210. Acquire images of the tobacco leaves to be inspected on the conveyor belt using an image acquisition device, and send the tobacco leaf images to the controller.

[0045] S220. The received tobacco leaf image is input into the pre-trained debris recognition model through the controller to determine the location of the debris in the tobacco leaf image and the actual movement speed of the debris on the conveyor belt.

[0046] The controller is a PLC (Programmable Logic Controller), a digital electronic device designed specifically for industrial applications.

[0047] In practical applications, in order to improve the efficiency of tobacco processing and accurately remove impurities from the tobacco leaves to be inspected without stopping the conveyor belt, it is necessary to determine when to control the removal device to perform the impurity removal operation during the movement of the conveyor belt.

[0048] Specifically, one way to determine the actual speed of debris on the conveyor belt is to use the conveyor belt's speed as the actual speed of the debris. However, due to the relative displacement between the conveyor belt and the debris, using the conveyor speed as the actual speed of the debris introduces a certain error, resulting in poor accuracy in removing debris.

[0049] To accurately determine the actual speed of the conveyor belt, the method of acquiring images of the tobacco leaves to be inspected transported on the conveyor belt using an image acquisition device and sending these images to the controller can be as follows: During the conveyor belt transport process, the image acquisition device acquires multiple images of the tobacco leaves to be inspected transported on the conveyor belt according to a fixed acquisition direction, and sends each tobacco leaf image and its corresponding acquisition time to the controller. Correspondingly, the specific implementation method for determining the actual speed of debris on the conveyor belt includes: for each tobacco leaf image, the controller inputs the tobacco leaf image into the debris recognition model to determine the location of the debris in the tobacco leaf image; based on the acquisition time and debris location corresponding to different tobacco leaf images, the actual speed of the debris is determined.

[0050] The fixed acquisition direction can be a direction perpendicular to the strip-shaped plane formed by the conveyor belt.

[0051] In practice, while the conveyor belt is running at a fixed speed, an image acquisition device can capture multiple images of the tobacco leaves to be inspected, and record the acquisition time of each image. For example, multiple tobacco leaf images can be acquired at preset time intervals.

[0052] Furthermore, each acquired tobacco leaf image is sent to the controller. The controller then inputs each tobacco leaf image into the debris recognition model to determine the location of any debris within each image. For example, the preset time interval can be 1 second, and the number of acquired tobacco leaf images can be 5.

[0053] In this embodiment, after determining the location of the debris in each tobacco leaf image, the trajectory of the debris within the tobacco leaf image can be determined. Based on the trajectory and the acquisition time of the tobacco leaf image, the moving speed of the debris within the tobacco leaf image is determined. The moving speed reflects the number of pixels the debris moves within the tobacco leaf image per second. Optionally, the moving speed can be determined as follows:

[0054]

[0055] Where v2 is the moving speed, n represents the nth tobacco leaf image, and n is a positive integer; i is less than n. x(i+1) represents the position of the debris in the (i+1)th tobacco leaf image, x(i) represents the position of the debris in the ith tobacco leaf image, t represents the difference in acquisition time between the ith and (i+1)th tobacco leaf images; v1(i) is the number of pixels the debris moves per second from x(i) to x(i+1).

[0056] Furthermore, based on the moving speed, the actual speed of the debris during its movement on the conveyor belt is determined using the following formula:

[0057] v3 = v2 × k

[0058] Where v3 represents the actual speed of movement, and k is the ratio between the predetermined actual spatial distance and the pixels in the tobacco leaf image, and k is a positive number.

[0059] S230. Based on the installation distance between the image acquisition device and the rejection device, the location of the debris, and the actual movement speed, determine the arrival time of the debris when it is transferred from the image acquisition device to the rejection device.

[0060] The installation distance is the distance between the image acquisition device and the rejection device along the conveyor belt's conveying direction.

[0061] In this embodiment, the ratio of the installation distance to the actual movement speed can be used as the time required for the debris to be transmitted from the image acquisition device to the rejection device. The arrival time of the debris when it is transmitted from the image acquisition device to the rejection device is obtained by adding the acquisition time of the last tobacco leaf image to the required time.

[0062] S240. If the current time is the arrival time, generate a rejection control command based on the location of the debris, and send the rejection control command to the rejection device.

[0063] In practice, in order to accurately remove debris while avoiding pausing the conveyor belt, it is possible to detect whether the current time is the arrival time. If the current time is the arrival time, a control command is generated based on the location of the debris, and the removal command is sent to the removal device.

[0064] S250, The rejecting device removes impurities from the tobacco leaves to be inspected based on the received rejecting control command.

[0065] It should be noted that after receiving the rejection control command, the rejection equipment can determine the position of the debris on the conveyor belt. Furthermore, since the current time is the arrival time, the debris has already been transported to the area where the rejection equipment can perform the rejection operation. The rejection equipment can directly perform the rejection operation on the determined position of the debris on the conveyor belt.

[0066] This embodiment can accurately remove debris while the conveyor belt is running. Through the synergistic effect between different devices, an integrated debris removal system is achieved, ensuring both the accuracy of debris removal and the efficiency of tobacco production.

[0067] Figure 3 This is a flowchart of another method for removing tobacco impurities according to an embodiment of the present invention. Based on the above embodiments, this embodiment optionally includes a removal device comprising a solenoid valve, a cylinder connected to the solenoid valve, and at least one air pipe connected to the cylinder. The method of removing impurities by the removal device can be as follows: the solenoid valve receives a removal control command, causing the valve to open, allowing gas to enter the cylinder. The gas pushes a piston in the cylinder to begin moving, and the air pipe connected to the cylinder delivers gas to remove impurities from the tobacco leaf to be tested. The explanations of terms that are the same as or corresponding to those in the above embodiments are not repeated here. Figure 3 As shown, the method includes:

[0068] S310. Acquire images of the tobacco leaves to be inspected on the conveyor belt using an image acquisition device, and send the tobacco leaf images to the controller.

[0069] S320. The received tobacco leaf image is input into the pre-trained debris recognition model through the controller to determine the location of the debris in the tobacco leaf image, generate a rejection control command based on the debris location, and send the rejection control command to the rejection device; wherein, the rejection device includes a solenoid valve, a cylinder connected to the solenoid valve, and at least one air pipe connected to the cylinder.

[0070] To clearly explain the structure of the rejection equipment, such as Figure 4 As shown. Figure 4 In this configuration, image acquisition device 1 is mounted on the first support 5 to perform image acquisition operations on the tobacco leaves transported on the conveyor belt. A rejection device consists of cylinder 2, solenoid valve 3, and air pipe 4, with the air pipe 4 mounted on the second support 6. For example, there can be six air pipes 4 arranged above the conveyor belt, with each air pipe corresponding to one solenoid valve.

[0071] S330: The solenoid valve receives a rejection control command, causing the valve to open and gas to enter the cylinder. The gas pushes the piston in the cylinder to start moving, and the gas pipe connected to the cylinder delivers gas to remove impurities from the tobacco leaf to be tested.

[0072] In this embodiment, the rejection control command can be an electrical signal. Upon receiving the signal, the solenoid valve becomes energized, opening its valve. This allows external gas to enter the cylinder through an air pipe connected to the solenoid valve. The gas pushes the piston in the cylinder, causing a pressure difference between the inside and outside of the cylinder, which in turn causes air to be ejected from the nozzle of the air pipe connected to the cylinder. This air ejection removes debris from the conveyor belt. This embodiment provides a rejection device capable of rapidly removing debris through gas injection.

[0073] Furthermore, the location of the debris includes the image coordinates of the debris in the tobacco leaf image along the first image direction; the first image direction is the direction perpendicular to the conveyor belt direction in the tobacco leaf image; there are multiple air tubes, and the coverage area of ​​the gas transported by different air tubes on the conveyor belt differs in the first conveyor belt direction; the first conveyor belt direction is the direction perpendicular to the conveyor belt direction.

[0074] It should be noted that because the coverage areas of the gas delivered by different air pipes on the conveyor belt differ along the first conveyor belt direction, different air pipes can remove debris from different positions on the conveyor belt. For example, six air pipes can be arranged along the first conveyor belt direction. The coverage areas of different air pipes along the first conveyor belt direction are interconnected or overlap, and the total coverage area of ​​all air pipes along the first conveyor belt direction can encompass the entire conveyor belt area in the first conveyor direction. That is, when all six air pipes are spraying gas, they can cover the entire position of the conveyor belt in the first conveyor direction, thereby ensuring that debris at any position on the conveyor belt can be removed by the air pipes. Furthermore, compared to using a single air pipe, a removal device with multiple air pipes can select any air pipe for removal based on the location of the debris on the conveyor belt, thereby minimizing the effective range of the gas on the conveyor belt while ensuring accurate removal of debris and reducing the accidental removal of tobacco leaves.

[0075] In this embodiment, generating a rejection control command based on the location of the debris and sending the rejection control command to the rejection device includes: determining the actual coordinates of the debris in the first conveyor belt direction based on image coordinates; determining the target coverage area containing the actual coordinates in each coverage area, and taking the trachea corresponding to the target coverage area as the target trachea; generating a rejection control command for controlling the operation of the target trachea, and sending the rejection control command to the solenoid valve associated with the target trachea in the rejection device.

[0076] Optionally, the image coordinates are multiplied by k to obtain the actual coordinates. Here, k is the ratio between the predetermined actual spatial distance and the pixels in the tobacco leaf image, and k is a positive number. Each air duct corresponds to a coverage area, and the coverage area containing the actual coordinates can be determined as the target coverage area. The air duct that generates the target coverage area is determined as the target air duct. Each air duct corresponds to a solenoid valve. When the solenoid valve is open, differential gas is transmitted through the air duct. After the target air duct is determined, a rejection control command can be generated and sent to the solenoid valve associated with the target air duct to control the valve to open, thereby generating gas through the target air duct to spray towards the actual coordinates, thus rejecting impurities.

[0077] In this embodiment, by setting up multiple air pipes and equipping each air pipe with a solenoid valve, it is possible to control the air jet of each air pipe individually. This ensures accurate removal of debris while minimizing the effective range of the gas on the conveyor belt, thus reducing the accidental removal of tobacco leaves. Furthermore, the coverage area of ​​the multiple air pipes can completely cover the position of the conveyor belt in the first conveyor belt direction, thereby ensuring that debris in any position can be removed.

[0078] In this embodiment, the tobacco debris removal system also includes an alarm device; after determining the location of the debris in the tobacco leaf image, the system further includes: generating an alarm command based on the location of the debris, and sending the alarm command to the alarm device so that the alarm device can perform an alarm operation based on the alarm command.

[0079] Specifically, after detecting the location of the foreign object, it indicates that the tobacco leaf to be tested contains foreign objects. In order to make the staff pay attention to the foreign object situation, an alarm command can be generated and sent to the alarm device. The alarm device can then parse the alarm command and perform the alarm operation.

[0080] For example, the alarm device includes an audible alarm device and / or a visual alarm device; the audible alarm device includes at least one of an electronic buzzer, an alarm horn, and a broadcasting device for playing pre-stored recordings; the visual alarm device includes a light-emitting diode and / or a laser warning device. For the visual alarm device, the location of the obstruction can be indicated by emitting light, and different emitting methods can be set for different areas. For the audible alarm device, the location of the obstruction can be indicated by at least one of different frequencies, sound types, and playback content. Alarms using both audible and visual alarm devices are applicable to various environments. For example, in tobacco manufacturing processes where noise is high, a visual alarm device can be used to detect obstructions. When workers are far from the conveyor belt, an audible alarm device can be used to effectively indicate the presence of obstructions.

[0081] In this embodiment, an alarm device is used to provide an alarm prompt, which makes it easier for staff to promptly detect impurities in the tobacco leaves to be tested, thus meeting the actual work needs of the staff.

[0082] Furthermore, to avoid misidentification of debris, identified debris can be re-confirmed. For details of the confirmation process, please refer to [link / reference needed]. Figure 5 , Figure 5In this process, images of the tobacco leaves to be inspected can be acquired in real time using image acquisition equipment to detect the color and shape of each component of the tobacco leaf. The tobacco leaf images are then input into a debris recognition model. If the debris recognition model determines the location of the debris and an anomaly is found in the tobacco leaf image, the image of the item corresponding to the debris location can be compared with tobacco leaf samples in a pre-built comparison database. If there is no difference, it means that the item corresponding to the debris location is tobacco leaf, and a new tobacco leaf image can be acquired again using the image acquisition equipment to confirm the debris. If there is a difference, it means that the item corresponding to the debris location is debris, and an audible and visual alarm can be triggered. The X and Y coordinates corresponding to the debris location are transmitted to the controller, which determines the energizing time of the solenoid valve and sends a rejection control command to the solenoid valve according to the energizing time to complete the debris rejection operation.

[0083] This embodiment compares the item corresponding to the location of the debris with the tobacco sample to perform a secondary confirmation of the identified debris, which helps to avoid misidentification and improves the accuracy of debris identification.

[0084] Figure 6 This is a schematic diagram of a tobacco debris removal system according to an embodiment of the present invention. This system is used to execute the tobacco debris removal method provided in any of the above embodiments. This system and the tobacco debris removal methods of the above embodiments belong to the same inventive concept. Details not described in detail in the embodiments of the tobacco debris removal system can be found in the embodiments of the tobacco debris removal methods described above. Figure 6 As shown, the system includes: at least one image acquisition device, a controller connected to the image acquisition device, and a rejection device connected to the controller.

[0085] Image acquisition device 10 is used to acquire images of tobacco leaves to be inspected being transported on a conveyor belt and send the tobacco leaf images to the controller;

[0086] The controller 11 is used to input the received tobacco leaf image into the pre-trained debris recognition model, determine the location of the debris in the tobacco leaf image, generate a rejection control command based on the debris location, and send the rejection control command to the rejection device.

[0087] The rejection device 12 is used to reject impurities in the tobacco leaves to be inspected based on the received rejection control command.

[0088] Optionally, based on any of the optional technical solutions in the embodiments of the present invention, the system further includes: an alarm device;

[0089] The controller 10 is also used to generate an alarm command based on the location of the debris in the tobacco leaf image after determining the location of the debris in the tobacco leaf to be detected, and to send the alarm command to the alarm device.

[0090] Alarm devices are used to perform alarm operations based on alarm commands.

[0091] Based on any optional technical solution in the embodiments of the present invention, the alarm device may optionally include a sound alarm device and / or a light alarm device; the sound alarm device includes at least one of an electronic buzzer, an alarm horn, and a broadcasting device for playing pre-stored recordings; the light alarm device includes a light-emitting diode and / or a laser warning device.

[0092] Based on any optional technical solution in the embodiments of the present invention, optionally, the controller 11 includes:

[0093] The speed determination unit is used to determine the actual speed of the debris on the conveyor belt;

[0094] The arrival time determination unit is used to determine the arrival time of the debris when it is transferred from the image acquisition device to the rejection device based on the installation distance between the image acquisition device and the rejection device and the actual movement speed; wherein, the installation distance is the distance between the image acquisition device and the rejection device in the conveying direction of the conveyor belt;

[0095] The instruction sending unit is used to generate a rejection control instruction based on the location of the debris when the current time is not the arrival time, and to send the rejection control instruction to the rejection device.

[0096] Based on any optional technical solution in the embodiments of the present invention, the image acquisition device 10 is optionally used to acquire multiple tobacco leaf images of the tobacco leaf to be detected being transported on the conveyor belt in a fixed acquisition direction during the conveyor belt transmission process, and send each tobacco leaf image and the acquisition time corresponding to the tobacco leaf image to the controller.

[0097] The velocity determination unit includes:

[0098] The velocity determination subunit is used to input the tobacco leaf image into the debris recognition model through the controller for each tobacco leaf image, determine the location of the debris in the tobacco leaf image, and determine the actual movement speed of the debris based on the acquisition time and debris location corresponding to different tobacco leaf images.

[0099] Based on any optional technical solution in the embodiments of the present invention, the rejection device may optionally include a solenoid valve, a cylinder connected to the solenoid valve, and at least one air pipe connected to the cylinder.

[0100] The solenoid valve is used to receive rejection control commands, so that the valve opens and gas enters the cylinder. The gas pushes the piston in the cylinder to start moving, and the gas pipe connected to the cylinder delivers gas to remove impurities from the tobacco leaves to be tested.

[0101] Based on any optional technical solution in the embodiments of the present invention, optionally, the location of the debris includes the image coordinates of the debris in the tobacco leaf image in the first image direction; the first image direction is the direction in the tobacco leaf image that is perpendicular to the conveying direction of the conveyor belt; there are multiple air tubes, and the coverage area of ​​the gas transported by different air tubes on the conveyor belt is different in the first conveyor belt direction; the first conveyor belt direction is the direction on the conveyor belt that is perpendicular to the conveying direction.

[0102] Controller 11 includes:

[0103] The coordinate determination unit is used to determine the actual coordinates of the debris in the first conveyor belt direction based on the image coordinates;

[0104] The trachea determination unit is used to determine the target coverage area containing actual coordinates in each coverage area, and to take the trachea corresponding to the target coverage area as the target trachea.

[0105] The instruction generation unit is used to generate rejection control instructions for controlling the operation of the target trachea, and to send the rejection control instructions to the solenoid valve associated with the target trachea in the rejection device.

[0106] Based on any optional technical solution in the embodiments of the present invention, the debris recognition model is optionally trained based on a deep learning model, and a multi-receptive-field feature adaptive fusion module is embedded in the debris recognition model.

[0107] The technical solution of this invention is applied to a tobacco debris removal system. The system includes at least one image acquisition device, a controller connected to the image acquisition device, and a removal device connected to the controller. The image acquisition device acquires images of tobacco leaves to be inspected transported on a conveyor belt and sends these images to the controller, eliminating the need for manual inspection. The tobacco leaf images reflect information about the tobacco leaves to be inspected. The controller inputs the received tobacco leaf images into a pre-trained debris recognition model to determine the location of debris in the tobacco leaf image. Based on the debris location, a removal control command is generated and sent to the removal device. The removal device then removes the debris from the tobacco leaf based on the received removal control command. This technical solution eliminates the need for manual intervention. The debris recognition model identifies the location of debris in the tobacco leaf image, improving the accuracy and efficiency of debris recognition and reducing missed detections. Furthermore, the removal of debris by the removal device eliminates the need for manual processing, thus reducing production costs and increasing production efficiency.

[0108] It is worth noting that in the embodiments of the tobacco debris removal device described above, the various units and modules are divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the scope of protection of the present invention.

[0109] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0110] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for removing tobacco debris, characterized in that, An application to a tobacco debris removal system, the tobacco debris removal system comprising at least one image acquisition device, a controller connected to the image acquisition device, and a removal device connected to the controller, the method comprising: The image acquisition device acquires images of the tobacco leaves to be inspected being transported on the conveyor belt, and sends the tobacco leaf images to the controller; The controller inputs the received tobacco leaf image into a pre-trained debris recognition model to determine the location of the debris in the tobacco leaf image, generates a rejection control command based on the debris location, and sends the rejection control command to the rejection device. The rejection device removes impurities from the tobacco leaves to be inspected based on the received rejection control command.

2. The method according to claim 1, characterized in that, The step of generating a rejection control command based on the location of the debris and sending the rejection control command to the rejection device includes: Determine the actual speed at which the debris moves on the conveyor belt; Based on the installation distance between the image acquisition device and the rejection device and the actual movement speed, the arrival time of the debris when it is transferred from the image acquisition device to the rejection device is determined; wherein, the installation distance is the distance between the image acquisition device and the rejection device in the conveying direction of the conveyor belt; If the current time is the arrival time, the rejection control command is generated based on the location of the debris, and the rejection control command is sent to the rejection device.

3. The method according to claim 2, characterized in that, The step of acquiring images of the tobacco leaves to be inspected transported on the conveyor belt using an image acquisition device and sending the tobacco leaf images to the controller includes: During the conveyor belt transport process, the image acquisition device acquires multiple images of the tobacco leaves to be inspected on the conveyor belt according to a fixed acquisition direction, and sends each tobacco leaf image and the acquisition time corresponding to the tobacco leaf image to the controller; Determining the actual speed of the debris on the conveyor belt includes: For each tobacco leaf image, the controller inputs the tobacco leaf image into the debris recognition model to determine the location of the debris in the tobacco leaf image; Based on the acquisition time and the location of the debris corresponding to different tobacco leaf images, the actual movement speed of the debris is determined.

4. The method according to claim 1, characterized in that, The rejection device includes a solenoid valve, a cylinder connected to the solenoid valve, and at least one air pipe connected to the cylinder. The process of removing impurities from the tobacco leaves to be inspected by a rejection device based on the received rejection control command includes: The solenoid valve receives the rejection control command to open the valve, allowing gas to enter the cylinder. The gas pushes the piston in the cylinder to start moving, and the gas pipe connected to the cylinder delivers gas to remove impurities from the tobacco leaf to be tested.

5. The method according to claim 4, characterized in that, The location of the debris includes the image coordinates of the debris in the tobacco leaf image along a first image direction; the first image direction is the direction in the tobacco leaf image that is perpendicular to the conveyor belt's conveying direction; there are multiple air tubes, and the coverage area of ​​the gas delivered by different air tubes on the conveyor belt differs along the first conveyor belt direction; the first conveyor belt direction is the direction on the conveyor belt that is perpendicular to the conveying direction. The step of generating a rejection control command based on the location of the debris and sending the rejection control command to the rejection device includes: Based on the image coordinates, determine the actual coordinates of the debris in the first conveyor belt direction of the conveyor belt; In each of the coverage areas, a target coverage area containing the actual coordinates is determined, and the trachea corresponding to the target coverage area is taken as the target trachea; Generate a rejection control command to control the operation of the target trachea, and send the rejection control command to the solenoid valve associated with the target trachea in the rejection device.

6. The method according to claim 1, characterized in that, The tobacco debris removal system also includes an alarm device; After determining the location of the impurities in the tobacco leaf in the tobacco leaf image, the method further includes: An alarm command is generated based on the location of the debris, and the alarm command is sent to the alarm device so that the alarm device can perform an alarm operation based on the alarm command.

7. The method according to claim 1, characterized in that, The debris recognition model is trained based on a deep learning model, and a multi-receptive-field feature adaptive fusion module is embedded in the debris recognition model.

8. A tobacco debris removal system, characterized in that, include: At least one image acquisition device, a controller connected to the image acquisition device, and a rejection device connected to the controller; The image acquisition device is used to acquire images of tobacco leaves to be inspected being transported on a conveyor belt, and to send the tobacco leaf images to the controller; The controller is used to input the received tobacco leaf image into a pre-trained debris recognition model, determine the location of the debris in the tobacco leaf image, generate a rejection control command based on the debris location, and send the rejection control command to the rejection device. The rejection device is used to reject impurities in the tobacco leaf to be tested based on the received rejection control command.

9. The system according to claim 8, characterized in that, Also includes: Alarm equipment; The controller is further configured to, after determining the location of the foreign matter in the tobacco leaf image, generate an alarm command based on the location of the foreign matter, and send the alarm command to the alarm device; The alarm device is used to perform alarm operations based on the alarm command.

10. The system according to claim 9, characterized in that, The alarm device includes an audible alarm device and / or a visual alarm device; the audible alarm device includes at least one of an electronic buzzer, an alarm horn, and a broadcasting device for playing pre-stored recordings; the visual alarm device includes a light-emitting diode and / or a laser warning device.

Images