System for controlling quality of stem slivers after secondary winnowing based on image recognition and weighing feedback

Abstract

The invention discloses a stem quality control system after second-stage winnowing based on image recognition and weighing feedback, which comprises a first-stage air door, a second-stage air door, a third-stage air door, a third-stage air door, a fourth-stage air door, a fifth-stage air door, a sixth-stage air door, a fifth-stage air door and a sixth-stage air door, the material conveying assembly is used for conveying the separated slivers to the downstream of the separation area provided with the weighing detection assembly; the visual acquisition assembly is used for acquiring tobacco shred image data at equal intervals and sending the data to the controller; the weighing detection assembly is used for collecting weight data of the separated slivers and sending the weight data to the controller; the controller is used for receiving the tobacco shred image data, performing feature extraction and calculation, receiving and processing the weight data, fusing the two types of processed data, generating a control signal through a preset control algorithm, and transmitting the control signal to the actuator; and the actuator is used for receiving the control signal and adjusting the opening degree of the corresponding air door, so that the adaptability to working conditions such as upstream incoming material flow fluctuation and transmission speed change is improved, and the stability of separation quality is improved.

Description

Technical Field

[0001] This invention relates to the field of cigarette processing and tobacco impurity removal technology, and in particular to a two-stage air-separated stem quality control system based on image recognition and weighing feedback. Background Technology

[0002] Air separation in tobacco processing is a crucial step in separating tobacco shreds from stems. It works by utilizing the difference in levitation force of airflow to separate tobacco shreds and stems of different densities and shapes into strata within the airflow field, thereby completing the sorting of materials.

[0003] In current tobacco processing, the operation and control of air separation largely rely on a combination of traditional manual operation and basic mechanical adjustment. The damper opening is manually set by on-site operators based on production experience, the material conveying path adopts a fixed mechanical flow guiding structure, and the airflow parameters of air separation also maintain relatively stable preset values, only requiring manual adjustment when switching batches.

[0004] However, relying on manual experience to set the damper opening lacks responsiveness to the real-time status of incoming materials. When the load of incoming materials fluctuates, it reduces the stability of stem and skewer separation quality and affects the uniformity of materials in subsequent processing. When the proportion of stems and skewers in the mixture fluctuates, fixed air separation parameters cannot match the separation characteristics of the materials, resulting in inconsistent stem and skewer separation rates and reducing the purity of the separated tobacco and stem and skewer recovery rate. Manual adjustment is subject to lag and subjectivity. Differences in the experience of different operators can lead to batch-to-batch fluctuations in air separation effects, increasing the risk of quality fluctuations in the production process and making it difficult to adapt to the dynamic working conditions of actual production. Summary of the Invention

[0005] This invention provides a two-stage air-separated stem quality control system based on image recognition and weighing feedback to solve the problems of low stability of stem separation quality after air separation, low tobacco purity, low stem recovery rate, and the resulting difficulty in adapting the air separation stem separation method to the dynamic working conditions of actual production.

[0006] According to one aspect of the present invention, a two-stage air separation stalk quality control system is provided. The system includes a material conveying component, a vision acquisition component, a weighing and detection component, a controller, and actuators. The material conveying component includes a primary air damper, a conveyor belt, a secondary air damper, and a separation zone arranged sequentially along the upstream material inlet to the stalk air conveying position. The vision acquisition component is electrically connected to the controller. The weighing and detection component is electrically connected to the controller. The actuators include a primary air damper actuator and a secondary air damper actuator, which are respectively connected to the primary air damper and the secondary air damper, and are respectively electrically connected to the controller.

[0007] The material conveying assembly sequentially transports incoming materials from upstream to the primary damper for initial diversion, then conveys them via a conveyor belt to the secondary damper. After being guided by the secondary damper, the materials are sent to the separation zone for separation of tobacco shreds and stems. The separated stems are then conveyed downstream of the separation zone, where a weighing and detection assembly is installed. A vision acquisition assembly collects image data of the tobacco shreds on the conveyor belt at preset intervals and sends it to the controller. The weighing and detection assembly collects the weight data of the separated stems and sends it to the controller. The controller receives the tobacco shred image data transmitted by the vision acquisition assembly and performs feature extraction and calculation. It also receives the weight data transmitted by the weighing and detection assembly and performs filtering, calibration, and delay compensation processing. The two types of processed data are fused and a control signal is generated through a preset control algorithm, which is then transmitted to the actuator. The actuator receives the control signal from the controller and adjusts the opening of the corresponding primary and / or secondary dampers according to the control signal.

[0008] According to another aspect of the present invention, a method for quality control of stem labels after secondary wind separation is provided, executed by the controller of the system in any embodiment of the present invention, the method comprising:

[0009] The system receives image data of tobacco shreds on the conveyor belt after the primary damper, which is acquired at preset intervals by an encoder triggered by the vision acquisition component, and extracts the image features of the tobacco shreds. Based on the regression model constructed in advance by coefficients and intercept terms obtained through weighing and sampling calibration, and the image features of the tobacco shreds, the system calculates the amount of tobacco shreds per unit length. The system receives the weight data of the separated stem sticks transmitted by the weighing and detection component, and performs filtering, calibration, and delay compensation processing to obtain the stem stick mass flow rate data aligned with the action time of the secondary damper. Based on the amount of tobacco shreds per unit length and the stem stick mass flow rate data, the system generates a control signal through an internal model control proportional-integral-derivative or Smith control algorithm, and transmits the control signal to the actuator, so that the actuator adjusts the opening of the corresponding primary damper and / or secondary damper according to the control signal.

[0010] According to another aspect of the present invention, a secondary wind-separated stem quality control device is provided, which is configured in the controller of the system in any embodiment of the present invention. The device includes:

[0011] The image extraction module is used to receive image data of tobacco shreds on the conveyor belt after the primary air damper, which is acquired at preset intervals by the encoder triggered by the vision acquisition component, and to extract the image features of the tobacco shreds.

[0012] The feature calculation module is used to calculate the amount of tobacco per unit length based on the regression model constructed in advance by the coefficients and intercept terms obtained through weighing and sampling calibration, and the image features of the tobacco.

[0013] The weighing processing module is used to receive the weight data of the separated stems transmitted by the weighing detection component, and perform filtering, calibration and delay compensation processing to obtain the stem mass flow rate data at the time of action of the second-stage damper.

[0014] The opening adjustment module is used to generate control signals based on the unit length tobacco shred quantity data and stem mass flow rate data, through internal model control proportional-integral-derivative or Smith control algorithms, and transmit the control signals to the actuators, so that the actuators adjust the opening of the corresponding primary damper and / or secondary damper according to the control signals.

[0015] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0016] At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the secondary wind-sorting post-stem label quality control method according to any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions, the computer instructions being configured to cause a processor to execute and implement the secondary wind-sorting post-stem quality control method according to any embodiment of the present invention.

[0018] According to another aspect of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps of the method as described in any embodiment of the present invention.

[0019] The two-stage air-separated stem quality control system of this invention includes: a material conveying component for conveying upstream incoming material to a primary air damper for initial diversion, conveying it to a secondary air damper for guidance, and then sending it to a separation zone for separation; conveying the separated stems to the downstream of the separation zone equipped with a weighing and detection component; a vision acquisition component for equidistantly acquiring tobacco image data and sending it to the controller; a weighing and detection component for acquiring the weight data of the separated stems and sending it to the controller; a controller for receiving tobacco image data and performing feature extraction and calculation, receiving weight data and performing filtering, calibration, and delay compensation processing, fusing the two types of processed data, generating a control signal through a preset control algorithm, and transmitting it to the actuator; and an actuator for receiving the control signal and adjusting the opening of the corresponding air damper. The material conveying path employs a progressive approach: initial diversion via a primary damper, stable conveyor belt transport, and precise guidance via a secondary damper. This is complemented by a visual acquisition component for pre-monitoring of tobacco image data after the primary damper, and a weighing and detection component for post-processing feedback of stem weight data. The controller integrates these two types of data and generates signals through a preset control algorithm to drive the actuators to dynamically adjust the damper opening. Compared to traditional single-stage damper adjustment, this method can anticipate changes in tobacco load and promptly correct separation parameters, improving the stability of stem separation quality, reducing fluctuations in stem flow rate, and ensuring the purity of the separated tobacco. The twig recovery rate remains stable within the preset range; the visual acquisition component triggers equidistant image acquisition; the weighing detection component eliminates interference errors through filtering, calibration, and delay compensation; the controller adopts a control algorithm with time delay compensation capability, enabling the system to cope with complex working conditions such as upstream material flow fluctuations and conveyor belt speed changes, improving control robustness and applicability, and realizing automated closed-loop control of data acquisition, analysis and processing, signal generation, and execution adjustment, reducing manual intervention, lowering labor costs and human error, improving production efficiency, reducing material waste, and reducing system maintenance costs and downtime losses.

[0020] 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

[0021] 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.

[0022] Figure 1 This is a schematic diagram of a two-stage air-separated stem quality control system according to Embodiment 1 of the present invention;

[0023] Figure 2 This is a flowchart of a method for quality control of stem tags after secondary wind separation according to Embodiment 2 of the present invention;

[0024] Figure 3 This is a flowchart of another method for quality control of stem tags after secondary wind separation according to Embodiment 3 of the present invention;

[0025] Figure 4 This is a schematic diagram of a secondary air-separated stem quality control device according to Embodiment 4 of the present invention;

[0026] Figure 5 This is a schematic diagram of the structure of an electronic device for implementing the secondary wind separation and stem quality control method of the present invention. Detailed Implementation

[0027] 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.

[0028] 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 "having," 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.

[0029] Example 1

[0030] Figure 1 This is a schematic diagram of a two-stage air-separation stem quality control system provided in Embodiment 1 of the present invention. This embodiment is applicable to situations where tobacco shreds and stems are separated through air separation. Figure 1 As shown, the system includes:

[0031] The system includes a material conveying assembly 110, a vision acquisition assembly 120, a weighing and detection assembly 130, a controller 140, and an actuator 150. The material conveying assembly 110 includes a primary damper 160, a conveyor belt 170, a secondary damper 180, and a separation zone 190 arranged sequentially along the upstream material inlet to the pneumatic conveying position. The vision acquisition assembly 120 is electrically connected to the controller 140. The weighing and detection assembly 130 is electrically connected to the controller 140. The actuator 150 includes a primary damper actuator 1100 and a secondary damper actuator 1110, which are respectively connected to the primary damper 160 and the secondary damper 180, and are also electrically connected to the controller 140.

[0032] The material conveying assembly 110 is used to sequentially convey the upstream material to the primary damper 160 for initial diversion, and then convey it to the secondary damper 180 via the conveyor belt 170. After being guided by the secondary damper 180, it is sent to the separation zone 190 for separation of tobacco shreds and stems. The separated stems are then conveyed to the downstream of the separation zone 190, where the weighing and detection assembly 130 is installed.

[0033] The visual acquisition component 120 is used to acquire tobacco image data on the conveyor belt 170 at preset intervals and send it to the controller 140.

[0034] The weighing and detection component 130 is used to collect the weight data of the separated twigs and send it to the controller 140;

[0035] The controller 140 is used to receive tobacco image data transmitted by the vision acquisition component 120 and perform feature extraction and calculation, receive weight data transmitted by the weighing detection component 130 and perform filtering, calibration and delay compensation processing, fuse the two types of processed data to generate a control signal through a preset control algorithm, and transmit the control signal to the actuator 150.

[0036] Actuator 150 is used to receive control signals sent by controller 140 and adjust the opening degree of the corresponding primary damper 160 and / or secondary damper 180 according to the control signals.

[0037] In this embodiment of the invention, the material conveying component can be specifically understood as: a carrier used to achieve the orderly conveying of the mixture of tobacco shreds and stems from upstream to the separation zone, and the downstream transfer of the separated stems. The visual acquisition component can be specifically understood as: a device for acquiring images of tobacco shreds on the conveyor belt at a preset frequency, used to obtain visual data on the tobacco shred load and stem content, providing a pre-monitoring signal to the controller. The weighing detection component can be specifically understood as: a device installed downstream of the separation zone, used to acquire weight data of the separated stems, providing feedback data on the separation effect to the controller. The controller can be specifically understood as: a control device with data processing, algorithm calculation, and signal output functions, capable of analyzing the acquired data and generating control commands.

[0038] An actuator can be specifically understood as a device that receives signals from the controller and drives the corresponding damper to adjust its opening. It is the execution carrier of control commands and can include primary and secondary damper actuators. A primary damper can be specifically understood as a device that performs preliminary diversion of upstream material to stabilize the material load entering the secondary air separation stage. A secondary damper can be specifically understood as a device that guides the material after primary diversion, working in conjunction with the separation zone to separate tobacco shreds from stems.

[0039] The separation zone can be specifically understood as the functional area that uses the difference in airflow levitation force to separate tobacco shreds from stems; it is the operating area of ​​the air separation process. The delay compensation processing can be specifically understood as the operation of time calibration of the weight detection data to match the actual operating time of the secondary damper, eliminating the impact of material conveying lag.

[0040] Specifically, the quality control system for tobacco stems after the secondary air separation consists of a material conveying component, a vision acquisition component, a weighing and detection component, a controller, and actuators. The material conveying component is arranged sequentially from the upstream material to the stem conveying position, including a primary air damper, a conveyor belt, a secondary air damper, and a separation zone. The conveyor belt (e.g., a belt conveyor) receives the tobacco output from the primary air damper and transports it to the secondary air damper. The separation zone separates the tobacco from the stems and transports the stems to subsequent stages.

[0041] Both the visual acquisition component and the weighing detection component are electrically connected to the controller. The actuators, including primary and secondary damper actuators, are electrically or pneumatically driven and connected to their respective dampers, and also electrically connected to the controller. The primary and secondary damper actuators receive control signals from the controller and adjust the opening of the corresponding primary or secondary damper accordingly. Both actuators have position feedback and limit functions, enabling the construction of a closed-loop control circuit based on the position feedback signal. They can collect the actual opening position data of the damper in real time and send this data back to the controller as a feedback signal. The controller compares the actual opening position of the damper with the preset target opening position, calculates the deviation, and dynamically adjusts the output control signal based on this deviation to drive the actuator to correct the damper opening. This ensures that the damper opening accurately and stably reaches and maintains the target value, avoiding over- or under-adjustment. Simultaneously, the limit function prevents the damper opening from exceeding the preset limit range, ensuring the safety and accuracy of damper adjustment.

[0042] During system operation, the upstream tobacco shreds and stems mixture is conveyed to the primary damper of the material conveying assembly. The primary damper adjusts its opening according to the controller's instructions to initially divert the mixture, thereby regulating the material load entering subsequent stages and maintaining a relatively stable material quantity. The material after initial diversion is smoothly conveyed onto the conveyor belt, which operates at a set speed, uniformly transporting the material to the secondary damper. The secondary damper adjusts its opening according to the controller's control signal, providing secondary guidance for the material brought in by the conveyor, optimizing the angle and speed of the material entering the separation zone to better suit the airflow environment of the separation zone. After entering the separation zone, under the action of airflow levitation force, tobacco shreds and stems with differences in density and shape are separated into layers. The lighter tobacco shreds are guided by the airflow to a designated area for collection, while the heavier stems fall to the discharge channel of the separation zone. Finally, the separated stems are conveyed along the discharge channel to a specific position downstream of the separation zone, where a weighing and detection assembly is installed. The stems pass through this assembly so that the weighing and detection assembly can collect their weight data. Meanwhile, the visual acquisition component collects image data of tobacco shreds on the conveyor belt at preset intervals and sends it to the controller, while the weighing and detection component collects the weight data of the separated stems and sends it to the controller.

[0043] After receiving the image data of tobacco shreds from the conveyor belt transmitted by the vision acquisition component, the controller performs feature extraction and calculation on the image data using image processing algorithms. For example, it performs grayscale processing on the original image data, converting the image into a grayscale image containing only brightness information to reduce data redundancy. Then, it uses an image segmentation algorithm to divide the image into a foreground region (tobacco shreds and stems) and a background region (conveyor belt surface) by setting pixel thresholds, achieving separation of the target object from the background. Next, it uses a connected component analysis algorithm to identify all independent regions with consistent pixel characteristics within the foreground region. Then, based on preset aspect ratio thresholds and grayscale value thresholds, it selects target connected components that conform to the morphological characteristics of stems (manifested as long, thin, and bright connected components). Finally, it calculates the proportion of the pixel area of ​​the target connected component to the total pixel area of ​​the foreground region to obtain the stem index quantification index used to characterize the stem content of the tobacco shreds. At the same time, combined with the image acquisition interval and the conveyor belt running speed, it calculates the amount of tobacco shreds per unit length of the conveyor belt, completing the feature extraction and calculation of the image data and obtaining the amount of tobacco shreds per unit length.

[0044] After acquiring the tag weight data collected by the weighing and detection component, the controller performs filtering, calibration, and delay compensation on the weight data. For example, filtering algorithms (such as moving average filtering or median filtering) are used to process the raw weight data to remove random noise caused by factors such as equipment vibration or airflow disturbance, making the weight data smoother and more stable. The raw tag weight data collected in real time by the weighing and detection component is compared with the standard weight calibration value stored in the system in advance. The deviation between the real-time acquired data and the standard calibration value is calculated. Then, according to the system's preset calibration coefficient (such as linear correction coefficient), the real-time acquired weight data is compensated and adjusted through the corresponding correction formula (e.g., corrected data = real-time acquired data + deviation value × calibration coefficient). This solves the system error problem of the weighing and detection component itself, such as detection deviation caused by zero-point drift of the equipment, sensor sensitivity decay, and the influence of ambient temperature, ensuring that the calibrated tag weight data can truly reflect the actual tag weight.

[0045] To eliminate the impact of the time lag in material transport from the separation zone to the weighing and detection component on the control effect, the controller employs a control algorithm with time lag compensation capability (such as the Smith predictor algorithm or a PID (Proportional Integral Derivative) algorithm with delay compensation). This algorithm calculates the accurate delay time by combining the conveyor belt speed and the material transport distance. The time point of the calibrated stick weight data is then aligned with the actual action time of the secondary damper adjustment, thus eliminating the mismatch between the detection data and the adjustment effect caused by the time lag. By calculating the weight per unit time, the stick mass flow rate data, which accurately reflects the effect of the secondary damper adjustment, is obtained. In a specific example, based on the weighing data acquisition period (the time interval corresponding to equidistant acquisition) and the cumulative weight of the calibrated sticks within that acquisition period, the stick mass flow rate (unit: kg / s) within that period is calculated using the formula: cumulative stick weight / acquisition period. If continuous weighing is used, the instantaneous mass flow rate can be obtained by differentiating the weight data (or calculating the difference within a sliding window). Simultaneously, by combining the previously calculated transmission delay time (distance from the separation point to the weighing point divided by the conveying speed), the timestamp of the calculated mass flow rate data is shifted backward by the transmission delay time to align with the action time of the secondary damper, thus obtaining the final usable stem mass flow rate data.

[0046] The controller integrates and analyzes the data obtained through feature extraction and calculation (tobacco quantity per unit length and stem mass flow rate quantification index) with the stem mass flow rate data after filtering, calibration, and delay compensation. Using a preset control algorithm (such as a PID algorithm), the controller calculates the control quantity that can adjust the opening of the primary damper and the control quantity of the secondary damper, which includes feedforward and feedback composite adjustment logic. Finally, these two sets of control quantities are converted into electrical signals recognizable by the actuators and transmitted in real time to the corresponding primary and secondary damper actuators via an electrical connection channel.

[0047] In a specific example, the primary damper aims to stabilize the amount of tobacco per unit length. It compares the real-time collected and calculated data of the amount of tobacco per unit length with the preset target value to obtain the deviation value, which is then used as the input variable for the PID algorithm. The secondary damper aims to accurately control the stem mass flow rate. It compares the real-time acquired stem mass flow rate data after time-delay compensation with the preset target value to obtain the corresponding deviation value. At the same time, a stem index quantification index is introduced as a feedforward compensation parameter to predict in advance the impact of changes in the stem content of the incoming material on the separation effect. The controller uses a preset PID algorithm to calculate the deviation variables of the primary damper, the secondary damper, and the feedforward compensation parameters. The PID algorithm for the primary damper outputs the corresponding primary damper opening adjustment control quantity based on the proportional, integral, and derivative terms of the deviation. The PID algorithm for the secondary damper combines deviation feedback regulation with feedforward compensation of the stem index. For example, the quantified stem index obtained by image feature extraction can be multiplied by a preset feedforward weighting coefficient to obtain the feedforward control quantity, thereby adapting in advance to the impact of changes in the stem content of the incoming material on the secondary air separation effect. The controller sums the feedback control quantity and the feedforward control quantity to obtain the initial control quantity of the secondary damper. Then, the initial control quantity is limited to the physical upper and lower limits of the secondary damper opening by a limiting function, and finally outputs the secondary damper opening adjustment control quantity that integrates feedforward and feedback composite regulation logic.

[0048] After receiving the control signal transmitted by the controller, the primary and secondary damper actuators drive their own power mechanisms (electric or pneumatic) to rotate, thereby adjusting the damper opening. During the adjustment process, the actuator's built-in position feedback element collects the actual damper opening data in real time and sends this data back to the controller. The controller compares the actual opening with the target opening, and if there is a deviation, it further fine-tunes the control signal until the damper opening reaches the preset target value. At the same time, the actuator's limit function restricts the damper's rotation range to avoid mechanical damage caused by over-adjustment.

[0049] Optionally, based on the above embodiments, the visual acquisition component may include a camera, a light source, an optics cover, and an encoder. The camera and the light source are mounted above the conveyor belt after the primary damper, the optics cover is located outside the camera and the light source, and the encoder is connected to the conveyor belt for transmission.

[0050] Specifically, the vision acquisition component consists of a camera, a light source, an optics cover, and an encoder. This component can be connected to an IPC (Industrial Personal Computer) to provide the data foundation for vision algorithm operation, primary damper control, and feedforward control quantity generation. The camera and light source are mounted above the conveyor belt after the primary damper. The optics cover is positioned outside the camera and light source. The encoder is connected to the conveyor belt drive, and the optics cover can be used with an air curtain for dust protection. The camera can be either an area scan camera or a line scan camera. The area scan camera can have a pixel range of 5 to 12 megapixels and a frame rate of 30 to 60 frames per second, while the line scan camera can have a resolution of 2,000 to 4,000 pixels and a line frequency of 20 to 60 kHz. The camera also features a global shutter function to avoid image blurring of high-speed moving tobacco. The light source can be a constant-current driven LED strip light, which can be equipped with a polarizer and flat-field correction to ensure clear and stable acquired image data.

[0051] The vision acquisition component triggers equidistant sampling through an encoder. The encoder provides real-time feedback of displacement information as the conveyor belt moves. When the displacement reaches a preset interval, the camera is triggered to take a picture. The camera, in conjunction with a light source, acquires image data of the tobacco shreds on the conveyor belt, providing a basis for subsequent image feature extraction and control quantity calculation.

[0052] By mounting the camera and light source above the conveyor belt after the primary damper, images of tobacco shreds with a relatively stable material load after initial diversion can be directly acquired, avoiding interference from upstream material fluctuations. An optical cover positioned outside the camera and light source reduces the impact of environmental factors such as light, dust, and moisture on imaging, while protecting the camera and light source from collisions and contamination in the industrial environment. The encoder, connected to the conveyor belt drive, acquires real-time conveyor belt displacement information, triggering the camera to perform equidistant sampling. This ensures consistent spacing between adjacent tobacco shred samples, preventing uneven sampling caused by conveyor belt speed fluctuations. The combination of these three elements enables the acquisition of clear, uniform, and consistent tobacco shred image data, providing a high-quality data source for subsequent stem index calculation and unit length tobacco shred quantity statistics, reducing the impact of poor image quality. The encoder's transmission connection with the conveyor belt synchronizes image acquisition with material conveying, ensuring that each frame corresponds to a fixed length of tobacco material on the conveyor belt. This guarantees a one-to-one correspondence between image data and the actual position and conveying status of the material. The controller can then calculate the current material's stem content and load based on the real-time image data, eliminating the problem of image acquisition lagging behind material conveying. This improves the timeliness of feedforward and feedback control calculations, thereby accelerating the response speed of damper adjustment and enhancing the dynamic adjustment capability of the entire air separation control system. Furthermore, the physical protection of the optical cover and the stable transmission connection of the encoder improve the environmental adaptability and durability of the vision acquisition components in industrial settings, reducing downtime for maintenance due to equipment failure and lowering long-term system operation and maintenance costs.

[0053] Furthermore, based on the above embodiments, the controller can also be used for:

[0054] Monitor the operating status of the vision acquisition component. When the vision acquisition component meets the vision acquisition fault conditions, switch the control loop of the primary damper to manual or open-loop state, and simultaneously control the secondary damper to operate independently based on the weight data transmitted by the weighing detection component; and / or monitor the operating status of the weighing detection component. When the weighing detection component meets the weighing detection fault conditions, switch the secondary damper to manual control state, maintain the closed-loop steady flow control of the primary damper, and trigger an alarm signal.

[0055] In this embodiment of the invention, the visual acquisition failure condition can be specifically understood as: the criteria for determining that the visual acquisition component cannot output valid data. Specifically, this may include substandard image clarity or contrast, or a hardware failure in the camera or light source. Accordingly, for each frame of image transmitted by the visual acquisition component, feature values ​​of clarity (e.g., calculated using methods such as Laplacian variance) and contrast (e.g., calculated using indicators such as grayscale histogram dynamic range and variance) can be extracted. These feature values ​​are compared in real time with a preset qualified threshold. If the feature values ​​of a consecutive preset number of frames are lower than the threshold, the image quality is determined to be substandard. Simultaneously, the device's operating status parameters (e.g., camera frame rate, exposure value, temperature, and error code; light source current, voltage, and brightness feedback values) are read in real time through the device communication interface between the camera and the light source. If a hardware abnormality code, parameters exceeding the normal operating range, or communication interruption are read, a hardware failure in the camera or light source is determined.

[0056] The weighing detection fault conditions can be specifically understood as the criteria for determining whether the weighing detection component can collect accurate weighing data, such as no sensor signal or data fluctuation exceeding a threshold. Correspondingly, the output signal of the weighing sensor can be continuously monitored. If no valid electrical signal is received within a preset time (e.g., the signal amplitude is always 0 or in a dead zone with no response), or the communication link is interrupted (e.g., communication timeout), it is determined to be a sensor no signal fault. Simultaneously, for multiple sets of continuously collected weight data, the standard deviation or range of the data fluctuation index is calculated through a sliding window. This fluctuation index is compared with a preset normal fluctuation threshold. If the fluctuation index exceeds the threshold for a consecutive preset number of windows, and interference from normal fluctuations in the material itself is excluded (e.g., combined with operating parameters such as conveyor belt speed and damper opening), it is determined to be a data fluctuation exceeding the threshold fault. Closed-loop steady flow control can be specifically understood as: an automatic adjustment mode of the primary damper based on feedback data of the amount of tobacco per unit length, used to stabilize the material conveying load.

[0057] Specifically, the controller has component operation status monitoring and fault emergency handling functions. On the one hand, it monitors the operation status of the vision acquisition component. When the vision acquisition component meets the vision acquisition fault conditions (such as substandard image quality, camera malfunction, or light source malfunction), the controller switches the automatic closed-loop control mode of the primary damper to manual adjustment mode or open-loop control mode that does not rely on feedback data (maintaining the safe opening of the primary damper to ensure stable material conveying). At the same time, it controls the secondary damper to operate independently based on the weight data transmitted by the weighing detection component (ensuring the continuous separation of the chaff and tumbler). On the other hand, it monitors the operation status of the weighing detection component. When the weighing detection component meets the weighing detection fault conditions, it switches the secondary damper to manual adjustment mode or open-loop control mode that does not rely on feedback data (maintaining the safe opening of the primary damper to ensure stable material conveying). The primary damper is switched to manual control to maintain closed-loop flow control (to stabilize the upstream tobacco conveying state). Simultaneously, an alarm signal is triggered, and a maintenance process can be initiated, prompting personnel to promptly inspect faulty components. This includes recording the time of the fault, fault type, and current system operating parameters (such as damper opening, conveyor belt speed, and material load), generating and storing a fault diagnosis log. Subsequently, a fault alarm window pops up through the industrial HMI, displaying the fault location, cause, and suggested maintenance steps, while simultaneously triggering an audible and visual alarm to alert on-site personnel. If the system is connected to an equipment management system, a fault work order is automatically pushed to the corresponding maintenance personnel's terminal, initiating the maintenance process to ensure timely location and repair of faulty components.

[0058] Understandably, when the visual acquisition component malfunctions, the secondary damper can operate independently based on the weight data from the weighing detection component, and the primary damper can maintain a safe opening to ensure stable material conveying. The separation function of the entire air separation system can still be performed normally, without causing serious consequences such as stem and stick separation failure or material accumulation. However, the weighing detection component is the main data source for the feedback control of the secondary damper. Once the weighing fails, the secondary damper will lose the basis for precise adjustment and can only switch to manual control. At this time, the system cannot guarantee the stem and stick separation effect, which may lead to quality problems such as tobacco loss and excessive stem and stick content, and may even cause material blockage in subsequent processes. Therefore, it is necessary to trigger an alarm signal and start the maintenance process to avoid greater production losses.

[0059] The controller designs different fault response strategies for the vision acquisition component and the weighing detection component. When the vision acquisition component fails, the first-stage air damper is switched to the manual or open-loop state to maintain a safe opening degree, and the second-stage air damper operates independently relying on the weighing data. When the weighing detection component fails, the second-stage air damper is switched to the manual state, while the first-stage air damper maintains closed-loop steady flow control and triggers an alarm signal. Both strategies avoid full-line shutdown caused by a single component failure, improve the system's fault tolerance, and ensure the continuous and stable operation of the tobacco leaf threshing and winnowing process; when there is a vision failure, the second-stage air damper retains the weighing feedback adjustment, which can continue to basically control the separation effect of the stem and chip separation. When there is a weighing failure, the first-stage air damper maintains closed-loop steady flow, which can avoid the deterioration of the separation effect caused by sudden changes in the material flow rate. At the same time, the alarm signal can promptly remind the staff to intervene and handle, reducing the quality problems such as excessive stem content in cut tobacco or stem and chip with silk under fault conditions, reducing the probability of generating non-conforming products, and ensuring the consistency of product quality; the different fault switching strategies clarify the system operation modes when different components fail, enabling the staff to quickly locate the fault source, shortening the fault troubleshooting and repair time, reducing the unplanned downtime of the equipment, and reducing the maintenance cost and production losses caused by long-term downtime or fault expansion; by flexibly switching the control mode, it resists the risk of component failures caused by interference such as dust and vibration in the industrial field, enhances the system's adaptability to the industrial environment, and improves the reliability and stability of the overall operation.

[0060] Optionally, based on the above embodiments, the weighing detection component may include a weighing sensor and a weighing instrument, which are installed on the stem and chip conveying path downstream of the separation area, and the weighing instrument is electrically connected to the controller; the weighing sensor is equipped with an analog-to-digital conversion component, which is calibrated for zero point and span and compensated for temperature drift, and is fixed by an anti-vibration installation method.

[0061] In the embodiments of the present invention, the weighing sensor can be specifically understood as: a detection element for converting the physical signal (such as pressure or deformation) of the weight of the stem and chip into an electrical signal. The weighing instrument can be specifically understood as: a signal processing device supporting the weighing sensor, which can perform processing such as amplification, filtering, and calibration on the electrical signal output by the sensor, and finally output standardized data. The analog-to-digital conversion component can be specifically understood as: a component that converts the analog electrical signal output by the weighing sensor into a digital signal, enabling the data to be recognized and calculated by the controller. The zero point and span calibration can be specifically understood as: the zero point calibration is to eliminate the output deviation of the sensor when it is unloaded, and the span calibration is to calibrate the output linearity of the sensor through a standard weight to ensure the accuracy of the detection data.

[0062] Temperature drift compensation can be understood as using algorithms or hardware circuits to offset the impact of temperature changes on sensor accuracy, thus avoiding measurement errors caused by fluctuations in ambient temperature. Vibration-resistant installation methods can be understood as employing vibration-damping brackets or buffer components to reduce the interference of industrial equipment vibration on sensor detection accuracy.

[0063] Specifically, the weighing detection component consists of two parts: a weighing sensor and a weighing instrument. The entire assembly is installed on the stalk conveying path downstream of the separation zone, and the weighing instrument is electrically connected to the controller. The weighing sensor has a built-in or external high-resolution analog-to-digital converter that can convert the physical signal of the stalk weight into a digital signal. The sensor has also undergone zero-point calibration, span calibration, and temperature drift compensation processing in advance, and is fixed in a vibration-proof installation method, which can effectively resist interference factors such as on-site vibration and temperature fluctuations. Finally, the signal output by the sensor is further amplified and filtered by the weighing instrument and converted into a standard signal or data format that the controller can recognize and receive, providing stalk weight data for the feedback control of the secondary damper.

[0064] By calibrating the load cells at zero point and span, no-load deviation and linearity error can be eliminated. Temperature drift compensation can offset the impact of ambient temperature changes on sensor performance. Vibration-resistant installation reduces the interference of industrial equipment vibration on detection. Combined with an analog-to-digital converter to convert analog signals into digital signals, it reduces signal loss and distortion during transmission, ensuring accurate, stable, and highly interference-resistant output weight data, thus improving the control accuracy of the stalk separation in the air-separation system. The weighing instrument is directly electrically connected to the controller, allowing the processed digital weight data to be directly transmitted to the controller without additional signal conversion. The analog-to-digital converter adapts the sensor output signal to the controller's digital interface, avoiding attenuation and interference problems associated with long-distance analog signal transmission, shortening data transmission and processing time, increasing the feedback rate of weight data, and enhancing the system's dynamic response capability. Vibration-resistant installation reduces the impact and wear on the sensor's internal structure, extending its lifespan. Zero-point and span calibration and temperature drift compensation reduce the probability of control failure due to detection errors, improving the adaptability and operational stability of the weighing detection components under industrial conditions and reducing the long-term operating costs of the system.

[0065] Optionally, based on the above embodiments, the controller may include an industrial computer and a programmable logic controller (PLC); the PLC is used to receive weight data transmitted by the weighing and detection component, and perform filtering, calibration, and delay compensation processing to obtain stem mass flow rate data aligned with the action time of the secondary damper; the industrial computer is used to receive tobacco image data transmitted by the vision acquisition component, extract the image features of the tobacco, and combine them with a regression model constructed in advance by coefficients and intercept terms obtained through weighing and sampling calibration to calculate the amount of tobacco per unit length; and based on the amount of tobacco per unit length and the stem mass flow rate data, generate control signals through internal model control proportional-integral-derivative or Smith control algorithms and transmit them to the actuator.

[0066] In this embodiment of the invention, the Programmable Logic Controller (PLC) can be specifically understood as a digital computing and operating electronic system used for the acquisition, processing, and actuator control of field data. The Industrial Computer (IPC) can be specifically understood as a dedicated computer for industrial scenarios, capable of running complex image algorithms and control strategies. Filtering, calibration, and delay compensation processing can be specifically understood as: filtering is an operation used to eliminate noise interference in the data; calibration is an operation used to correct system errors in sensors; and delay compensation is an operation used to align the time difference between the data acquisition time and the control action time. The regression model can be specifically understood as: a mathematical model constructed based on the correlation between image features and the actual amount of tobacco, used to estimate the amount of tobacco per unit length from image data. Internal Model Control-Proportional Integral Derivative (IMC-PID) and the Smith control algorithm can be specifically understood as: two control algorithms with time delay compensation capabilities, suitable for industrial process control with transmission delays, improving control accuracy and stability.

[0067] Specifically, the controller consists of an industrial computer (configured with a central processing unit and optional small graphics processing unit) and a programmable logic controller (PLC). It exchanges data via communication protocols such as industrial Ethernet, Modbus transmission control protocol, or an open platform communication unified architecture. It also has analog signals (such as 4-20 mA or 0-10 volt analog signals) as a backup communication method and interfaces with data acquisition and monitoring control systems or manufacturing execution systems to achieve tobacco production formula management and data archiving. The PLC receives weight data from the weighing and detection components, performs filtering, calibration, and delay compensation processing sequentially to obtain stem mass flow rate data aligned with the action time of the secondary damper. For example, Kalman filtering or exponential smoothing algorithms are first used to filter out airflow and mechanical vibration interference, followed by online zero-point drift correction. Based on the distance from the separation point to the weighing point and the conveying speed, the transmission delay is calculated as: distance from separation point to weighing point / conveying speed. The weighing data timestamp is aligned with the action time of the secondary damper based on the transmission delay, and finally, the stem mass flow rate data is calculated.

[0068] An industrial computer receives tobacco image data from a vision acquisition unit, extracts image features of the tobacco, and substitutes them into a regression model with coefficients and intercept terms obtained through weighing and sampling calibration to calculate the amount of tobacco per unit length. For example, the image is segmented to extract features such as coverage, texture density (Tidx), and average gray level (Gmean). It can also be extended to extract the stem index quantification index (Sidx, i.e., the proportion of thin, bright connected regions). These are substituted into a linear regression model obtained through weighing and sampling calibration. The linear regression model can be M1est = w1 × Cover + w2 × Tidx + w3 × Gmean + b (where b is the intercept term, used to compensate for optical and calibration deviations, and w1, w2, and w3 are preset weight coefficients corresponding to each feature). A blank band threshold protection is set to avoid invalid calculations when there is no material, thus obtaining the tobacco amount per unit length data M1est.

[0069] Specifically, the linear regression model calibrated by weighing and random sampling can be understood as follows: by combining offline standard sample weighing experiments with manual random sampling verification, the model is obtained by solving the optimal values ​​of the weight coefficients and intercept terms in the regression model, and the mapping relationship between image features and quality is corrected using real physical quantity data. For example, multiple sets of standard tobacco samples with different densities and compositions can be prepared in advance. Under target transmission speed and lighting conditions, the samples are laid on a fixed length area on the conveyor belt. The corresponding tobacco images are acquired using a vision acquisition component, and features such as coverage, texture density, and average gray level are extracted. At the same time, the total mass of the fixed-length sample is weighed and converted into the amount of tobacco per unit length. After completing the acquisition of feature and true mass data of multiple sets of samples, abnormal samples (such as uneven sample stacking or blurred images) are manually sampled and removed. Then, linear regression algorithms such as least squares are used to fit the filtered sample data to solve for the weight coefficients and intercept terms that minimize the error between the model output and the true value. Finally, the accuracy of the model is verified using a reserved validation sample set. If the error exceeds the allowable range, samples are re-acquired or the feature extraction method is optimized until the model accuracy meets the control requirements. The final coefficients and intercept terms are stored in the control program of the industrial computer for online estimation of the amount of tobacco per unit length.

[0070] Based on the two types of data mentioned above, the industrial computer executes a differentiated control strategy. The primary damper control aims to stabilize the tobacco quantity per unit length (M1est) to the corresponding target value (M1set), employing an IMC-PID or Smith algorithm with delay compensation. It can also be combined with limiting, rate limiting, and anti-integral saturation mechanisms, and upstream feedforward can be added to further optimize the control effect, smoothly handling the material load and composition entering the secondary damper. Specifically, the limiting mechanism sets upper and lower boundaries for the primary damper's control output to prevent the damper opening from exceeding the allowable range of the mechanical structure or the safe range required by the process, thus avoiding equipment damage and material blockage. The rate limiting mechanism controls the rate of change of the primary damper opening to prevent drastic fluctuations in tobacco flow or impact wear on mechanical components caused by excessively rapid adjustments. The anti-integral saturation mechanism addresses the problem in IMC-PID controllers where long-term deviations (such as material supply interruptions or sudden changes in target values) cause the integral quantity to accumulate continuously, pushing the output to its limit and preventing timely recovery. This can be achieved using integral separation or reverse integration strategies. Integral separation involves pausing the integral component and retaining only proportional and derivative actions when the deviation between M1est and M1set exceeds a preset threshold; integration is then activated again when the deviation is within the threshold. Reverse integration involves initiating reverse integration to offset the accumulated integral quantity when the damper opening has reached the limit boundary and the deviation direction remains unchanged, ensuring the controller can quickly respond and exit the limit state when the deviation direction changes. Upstream feedforward is a control strategy that intervenes in advance. By introducing the operating parameter signals of the upstream feeding equipment, it offsets the disturbance of upstream material flow fluctuations on M1est, and further improves the flow stabilization effect of the primary damper. A correlation model between upstream feeding parameters (such as feed belt speed or feeder opening) and M1est can be established first. The feedforward weighting coefficient is calibrated through offline experiments. The controller collects the operating parameters of the upstream feeding equipment in real time, substitutes them into the model to calculate the corresponding feedforward adjustment amount, and superimposes the adjustment amount into the output value of IMC-PID or Smith algorithm. Before the material flow changes affect M1est, the opening of the primary damper is adjusted in advance to offset the disturbance.

[0071] The secondary damper control aims to stabilize the stem mass flow rate Mstem to the corresponding target value Mstem_set. A composite logic combining feedforward and feedback can be employed. The feedback part inputs the deviation between Mstem and Mstem_set into the IMC-PID or Smith control algorithm to obtain the feedback control quantity u2fb. The feedforward part calculates the feedforward control quantity u2ff = Kff1 × M1est + Kff2 × Sidx based on the tobacco quantity per unit length data M1est and the stem index quantification index Sidx. Here, Kff1 is the feedforward weighting coefficient for the tobacco quantity per unit length, and Kff2 is the feedforward weighting coefficient for the stem index quantification index. This can be calibrated through offline experiments. By fixing the secondary damper opening and changing the values ​​of M1est and Sidx, the corresponding changes in stem mass flow rate are recorded. The gain coefficient that minimizes the fluctuation of the stem mass flow rate is then solved using the least squares method or gradient descent method.

[0072] After the two are combined (u2ff+u2fb), the lower and upper limits of the final control quantity are obtained through upper and lower limit clamping. At the same time, a set value coordination mechanism is set so that Mstem_set is slowly scheduled and filtered along with the recipe or M1set.

[0073] For example, a mapping table is first established between formula parameters, M1set, and Mstem_set (the optimal stem mass flow rate target value is determined through offline process experiments for different tobacco formulas and different target unit lengths of tobacco). When the production formula is switched or M1set is adjusted, the controller does not directly jump Mstem_set to the new target value. Instead, it starts a slow scheduling logic and gradually approaches the new target value by setting a fixed rate limit (such as allowing Mstem_set to change no more than a preset process threshold per minute). At the same time, a first-order low-pass filter or moving average filter module is added to the scheduling path to further smooth the slowly changing Mstem_set setpoint, filter out the small fluctuations in the setpoint scheduling process, ensure that the change of Mstem_set is smooth and without sharp jumps, avoid large adjustments of the secondary damper caused by sudden changes in the setpoint, and ensure the consistency between the stem separation strategy and the primary flow stability target.

[0074] Furthermore, when the primary damper is in a large deviation state, the control gain of the secondary damper is reduced to avoid amplifying disturbances. For example, a deviation threshold for the primary damper is preset (e.g., when the absolute value of the deviation between M1est and M1set exceeds the preset proportional value of M1set, and the duration exceeds the preset judgment period). The controller calculates and monitors this deviation value in real time. If the deviation does not exceed the threshold, the secondary damper maintains normal IMC-PID control gain to ensure accurate adjustment of the flow rate of the stalk. If the deviation exceeds the threshold and meets the duration requirement, the primary damper is determined to be in a large deviation state. At this time, the controller automatically triggers the secondary gain attenuation logic, reducing the control gain of the secondary damper through a preset gain attenuation coefficient (e.g., multiplying the normal gain by an attenuation factor of 0.5 to 0.7), while recording the attenuation trigger time and current operating parameters. After the primary damper deviation falls back to within the threshold and remains stable, the normal control gain of the secondary damper is restored through a slow recovery logic (e.g., gradually restoring the gain to the normal value at a fixed rate). This avoids the adjustment action of the secondary damper amplifying the disturbances caused by the fluctuations of the primary material, ensuring the control stability of the entire air separation system.

[0075] Finally, the industrial computer transmits the generated control signals to the PLC, which then sends them to the primary and secondary damper actuators to complete the opening adjustment.

[0076] By assigning the real-time-critical weighing data processing task to a programmable logic controller (PLC), filtering, calibration, and delay compensation are used to eliminate on-site interferences such as airflow and equipment vibration, and to align control timing, thus obtaining stalk mass flow rate data that accurately reflects the adjustment effect of the secondary damper. Complex, computationally intensive tasks such as image feature extraction and regression model calculations are performed by an industrial computer. Based on visual image features and a pre-calibrated regression model, the amount of tobacco shreds per unit length is calculated, providing a basis for the primary damper's flow stabilization control. Furthermore, by combining the time delay compensation capability of the IMC-PID or Smith algorithm with the transmission delay characteristics of the air separation system, the primary and secondary damper flow control is achieved. The coordinated control of the gates reduces the impact of tobacco load fluctuations on stem-to-stick separation, decreases the probability of excessive stem content in tobacco or stem-to-stick mix, and ensures consistent product quality. The division of labor between the two avoids computational resource conflicts, shortens the time spent on data processing and control signal generation, and improves the system's response speed to material changes, meeting the real-time control requirements of continuous production in the tobacco processing air separation process. The visual regression model, built based on offline calibration data, can offset the effects of optical system errors and lighting changes, reducing interference from complex on-site conditions on control data, decreasing the probability of controller misadjustment, and improving the stability and reliability of the air separation system during long-term operation.

[0077] The two-stage air-separated stem quality control system of this invention includes: a material conveying component for conveying upstream incoming material to a primary air damper for initial diversion, conveying it to a secondary air damper for guidance, and then sending it to a separation zone for separation; conveying the separated stems to the downstream of the separation zone equipped with a weighing and detection component; a vision acquisition component for equidistantly acquiring tobacco image data and sending it to the controller; a weighing and detection component for acquiring the weight data of the separated stems and sending it to the controller; a controller for receiving tobacco image data and performing feature extraction and calculation, receiving weight data and performing filtering, calibration, and delay compensation processing, fusing the two types of processed data, generating a control signal through a preset control algorithm, and transmitting it to the actuator; and an actuator for receiving the control signal and adjusting the opening of the corresponding air damper. The material conveying path employs a progressive approach: initial diversion via a primary damper, stable conveyor belt transport, and precise guidance via a secondary damper. This is complemented by a visual acquisition component for pre-monitoring of tobacco image data after the primary damper, and a weighing and detection component for post-processing feedback of stem weight data. The controller integrates these two types of data and generates signals through a preset control algorithm to drive the actuators to dynamically adjust the damper opening. Compared to traditional single-stage damper adjustment, this method can anticipate changes in tobacco load and promptly correct separation parameters, improving the stability of stem separation quality and the uniformity of materials in subsequent processing. It also reduces fluctuations in stem mass flow rate, ensuring the purity of the separated tobacco. The twig recovery rate remains stable within the preset range; the visual acquisition component triggers equidistant image acquisition, and the weighing detection component eliminates interference errors through filtering, calibration, and delay compensation; the controller adopts a control algorithm with time delay compensation capability, enabling the system to cope with complex working conditions such as upstream material flow fluctuations and conveyor belt speed changes, improving control robustness and applicability, and realizing automated closed-loop control of data acquisition, analysis and processing, signal generation, and execution adjustment, reducing manual intervention, lowering labor costs and human error, improving production efficiency and reducing material waste, and reducing system maintenance costs, downtime losses, and the risk of quality fluctuations in the production process.

[0078] Example 2

[0079] Figure 2 This is a flowchart of a method for quality control of tobacco stems after secondary air separation, provided in Embodiment 2 of the present invention. This embodiment is applicable to situations where tobacco shreds and stems are separated by air separation. This method can be executed by a secondary air separation stem quality control device, which can be implemented in hardware and / or software and is generally configured in the controller of the secondary air separation stem quality control system. Figure 2 As shown, the method includes:

[0080] S210: Receive image data of tobacco shreds on the conveyor belt after the primary damper, which is acquired at preset intervals by the encoder triggered by the vision acquisition component, and extract the image features of the tobacco shreds.

[0081] S220. Based on the regression model constructed in advance by the coefficients and intercept terms obtained through weighing and sampling calibration, and the image characteristics of the tobacco shreds, the amount of tobacco shreds per unit length is calculated.

[0082] S230: Receive the weight data of the separated stem tag transmitted by the weighing and detection component, and perform filtering, calibration and delay compensation processing to obtain the stem tag mass flow rate data at the time of action of the second-stage damper.

[0083] S240. Based on the data of tobacco shreds per unit length and the mass flow rate of the stem, a control signal is generated through the proportional-integral-derivative or Smith control algorithm of the internal model control, and the control signal is transmitted to the actuator, so that the actuator adjusts the opening of the corresponding primary damper and / or secondary damper according to the control signal.

[0084] Specifically, the controller receives the image data of the first-stage air damper, which is triggered by the encoder and collected at preset intervals, and then transmits the image data of the tobacco shreds. The controller segments the image and extracts image features such as the coverage, texture density, and average grayscale characteristics of the tobacco shreds. These features are then substituted into a regression model constructed in advance by obtaining coefficients and intercept terms through weighing and sampling calibration to calculate the amount of tobacco shreds per unit length.

[0085] Simultaneously, it receives the weight data of the separated stem tag transmitted by the weighing and detection component, filters out noise caused by airflow and mechanical vibration through filtering methods such as low-pass filtering, Kalman filtering or exponential smoothing algorithm, performs zero-point drift online correction to offset sensor system errors, calculates the transmission delay based on the distance from the separation point to the weighing point and the conveying speed, aligns the timestamp of the weighing data with the action time of the secondary damper to complete the delay compensation process, and finally calculates the stem tag mass flow rate data aligned with the action time of the secondary damper by the weight per unit time.

[0086] Based on the data on the amount of tobacco shreds per unit length and the mass flow rate of the stems, control signals are generated using internal model control proportional-integral-derivative (PID) or Smith control algorithms. The primary damper control employs an IMC-PID or Smith algorithm with delay compensation, coupled with amplitude limiting, rate limiting, and anti-integral saturation mechanisms. Upstream feedforward can be added to smooth the material load and composition entering the secondary damper. The secondary damper control uses a composite logic combining feedforward and feedback. The feedback part inputs the deviation into the IMC-PID algorithm to obtain the feedback control quantity. The feedforward part calculates the feedforward control quantity based on the amount of tobacco shreds per unit length using the formula: Feedforward control quantity = Feedforward weighting coefficient of tobacco shreds per unit length × Tobacco shreds per unit length. The two are then combined and clamped to the mechanical and technological safety range of the secondary damper using a clamp function (amplitude limiting constraint function) to obtain the final control quantity.

[0087] The controller transmits control signals to the actuator, which then adjusts the opening of the primary or secondary damper to stabilize the amount of tobacco per unit length to the set value and ensure that the stem mass flow rate meets the preset requirements.

[0088] The technical solution of this invention uses an encoder to trigger a vision acquisition component to acquire images of tobacco shreds on the conveyor belt after the primary air damper at preset intervals. Combined with a regression model constructed from weighing and sampling calibration, the image features are converted into data on the amount of tobacco shreds per unit length. This allows for real-time and accurate capture of changes in tobacco shred quantity, dynamically adjusting the opening of the primary air damper to offset fluctuations in the incoming material load, thus maintaining a stable amount of tobacco shreds per unit length entering the secondary air classifier and reducing processing deviations in subsequent processes. Simultaneously, this method filters, calibrates, and compensates for the weight data from the weighing detection component, eliminating interference and aligning the action time of the secondary air damper. Combined with a control algorithm with time-delay compensation capabilities, a control signal is generated, which can stably control the stem mass flow rate within a preset range. Within the designated area, the system avoids issues such as tobacco residue on stems or stems carrying tobacco, ensuring consistent stem separation across different batches. By integrating data on the amount of tobacco per unit length with stem mass flow rate data, the system achieves coordinated adjustment of the primary and secondary air dampers. This combines the goal of stable material load in the primary stage with precise stem separation control in the secondary stage, enhancing the synergy between the two stages of air separation. Compared to independent adjustment modes, this significantly improves the overall control accuracy of the system, further stabilizing the purity of the separated tobacco and the stem recovery rate. It also improves the stability of stem separation quality and the uniformity of materials in subsequent processing, enabling the system to cope with complex operating conditions such as fluctuations in upstream feed flow and changes in conveyor belt speed. This reduces system maintenance costs, downtime losses, and the risk of quality fluctuations during production.

[0089] Example 3

[0090] Figure 3 This is a flowchart of another method for controlling the quality of stems after secondary air separation, provided in Embodiment 3 of the present invention. This embodiment refines the steps of "extracting image features of tobacco shreds" and "generating control signals based on the unit length tobacco shred quantity data and stem mass flow rate data using an internal model control proportional-integral-derivative or Smith control algorithm, and transmitting the control signals to the actuators so that the actuators adjust the opening degree of the corresponding primary and / or secondary air dampers according to the control signals" in the above embodiments. Figure 3 As shown, the method includes:

[0091] S310: Receives image data of tobacco shreds on the conveyor belt after the primary damper, which is acquired at preset intervals by the encoder triggered by the visual acquisition component. Separates the foreground region of the tobacco shred image data through grayscale processing and image segmentation algorithms, and identifies the connected components within the foreground region through connected component analysis algorithms. Target connected components with aspect ratios greater than a preset threshold and grayscale values ​​higher than a set grayscale threshold are selected.

[0092] In this embodiment of the invention, the connected component analysis algorithm can be specifically understood as an algorithm for identifying regions where pixels are connected in an image, used to distinguish the contour features of tobacco shreds and stems. Correspondingly, the target connected component can be specifically understood as a connected component whose aspect ratio and gray value both meet a preset threshold, i.e., a slender and bright connected component, representing the visual features of the stem.

[0093] S320. Calculate the target proportion of the pixel area of ​​the target connected region to the total pixel area of ​​the foreground region, and use it as a quantitative indicator of the tag index.

[0094] In this embodiment of the invention, the stem index quantification index can be specifically understood as: the proportion of the target connected region pixel area to the total area of ​​the foreground region, used to quantify the content and distribution of stems in tobacco shreds, and is an important feature of the secondary damper feedforward control.

[0095] Specifically, the controller receives image data of tobacco shreds from the first-stage air damper, which is triggered by the encoder and collected at preset intervals. The controller then performs grayscale conversion, adaptive threshold segmentation, and morphological opening operations on the image to separate the foreground region of the tobacco shreds. The connected component analysis algorithm is then used to identify the connected components within the foreground region. Slender, bright target connected components with an aspect ratio greater than a preset threshold and a grayscale value higher than the set grayscale threshold are selected. The proportion of the pixel area of ​​the target connected component to the total pixel area of ​​the foreground region is calculated and used as a quantification index for the stem content of the tobacco shreds.

[0096] S330. Based on the regression model constructed in advance by the coefficients and intercept terms obtained through weighing and sampling calibration, and the image features of the tobacco shreds, the amount of tobacco shreds per unit length is calculated.

[0097] S340: Receive the weight data of the separated stem tag transmitted by the weighing and detection component, and perform filtering, calibration and delay compensation processing to obtain the stem tag mass flow rate data at the time of action of the second-stage damper.

[0098] S350: Calculate the deviation between the unit length tobacco shred quantity data and the preset unit length tobacco shred quantity target value, calculate the primary damper adjustment amount through internal model control proportional integral derivative or Smith control algorithm, and generate a matching control signal based on the primary damper adjustment amount and transmit it to the primary damper actuator.

[0099] S360. Multiply the quantified index of the stem and the data of the amount of tobacco per unit length by a preset feedforward weighting coefficient and then sum them to obtain the feedforward control quantity of the secondary damper.

[0100] S370. Calculate the deviation between the preset target value of the stem mass flow rate and the stem mass flow rate data, and calculate the secondary damper feedback control quantity through internal model control proportional-integral-derivative or Smith control algorithms.

[0101] S380. The initial control quantity of the secondary damper is obtained by summing the feedforward control quantity and the feedback control quantity of the secondary damper. The initial control quantity of the secondary damper is then limited to the physical lower limit and upper limit of the opening of the secondary damper by the limiting function to obtain the adjustment quantity of the secondary damper.

[0102] S390: Generate a matching control signal based on the adjustment amount of the secondary damper and transmit it to the secondary damper actuator.

[0103] Specifically, the image is segmented and image features such as tobacco coverage, texture density, and average grayscale are extracted. These features are then substituted into a pre-weighed or sampled and calibrated regression model to calculate the amount of tobacco per unit length.

[0104] For primary damper control, the controller calculates the deviation between the unit length tobacco shred quantity data and the preset target value. It then calculates the primary damper adjustment amount using an IMC-PID or Smith algorithm with delay compensation. During the calculation process, it performs amplitude limiting, rate limiting, and anti-integral saturation processing. It can also superimpose upstream feedforward signals. The generated control signal is transmitted to the primary damper actuator to adjust the opening to stabilize the unit length tobacco shred quantity and smoothly introduce material load and composition into the secondary damper.

[0105] For the composite control of the secondary damper, the received stem weight data is subjected to low-pass, Kalman, or exponential smoothing filtering, and zero-point drift is corrected online. Then, the transmission delay is calculated based on the distance from the separation point to the weighing point and the conveying speed, and the action time of the secondary damper is aligned. Finally, the stem mass flow rate data is obtained through calculation.

[0106] The feedforward control quantity of the secondary damper is obtained by multiplying the quantified stem index and the tobacco shred quantity per unit length by their respective preset feedforward weighting coefficients and summing the results. Based on the deviation between the target and actual stem mass flow rates, the feedback control quantity of the secondary damper is calculated using either the IMC-PID algorithm or the Smith control algorithm. After summing the two to obtain the initial control quantity, it is limited by the clamp function to the physical upper and lower limits of the secondary damper opening, thus obtaining the secondary damper adjustment quantity.

[0107] Understandably, if the deviation of the primary damper control is detected to exceed the preset threshold, the gain coefficient of the secondary feedback control quantity (such as proportional gain or feedback gain) is reduced before summing. In addition, the controller can perform setpoint coordinated control, so that the target value of the stem mass flow rate is slowly scheduled and filtered according to the formula or the target value of the tobacco quantity per unit length. Finally, a control signal is generated according to the adjustment amount of the secondary damper and transmitted to the secondary damper actuator to achieve precise adjustment of the opening of the secondary damper.

[0108] Understandably, the controller can also monitor the operating status of the vision acquisition component and the weighing detection component in real time. If the image quality of the vision component is substandard or the camera or light source fails, the primary damper will be switched to manual or open-loop mode and maintained at a safe opening. The secondary damper will operate independently based on the weighing data. If the weighing component fails, the secondary damper will be switched to manual control, the primary damper will be kept under closed-loop steady flow control, and an alarm signal will be triggered to start the maintenance process, ensuring that the system can still operate safely under single-point failure.

[0109] The technical solution of this invention involves triggering a visual acquisition component via an encoder to acquire images of tobacco shreds on the conveyor belt behind the primary air damper at preset intervals. The foreground region of the tobacco shred images is separated using grayscale processing and image segmentation algorithms. Target connected regions meeting both aspect ratio and grayscale value thresholds are selected through connected component analysis. The percentage of pixel area in the target connected region is used as a quantitative indicator for the stem index. This avoids the subjectivity of traditional manual visual judgment and achieves a digital and quantitative representation of the stem-containing state of the tobacco shreds. Compared to vague qualitative judgments, the quantified stem index more accurately reflects the mixing ratio of tobacco shreds and stems, providing a secondary... The air separation adjustment provides an objective and reliable preliminary judgment basis, avoiding adjustment deviations caused by misjudgment of the stem-containing state; the regression model constructed by combining weighing and sampling calibration converts image features into tobacco shred quantity data per unit length; the weight data of the weighing detection component is filtered, calibrated, and time-delay compensated to obtain stem mass flow rate data aligned with the action time of the secondary air damper; the deviation between the tobacco shred quantity data per unit length and the preset target value of tobacco shred quantity per unit length is calculated, and the adjustment amount of the primary air damper is calculated through internal model control proportional-integral-derivative or Smith control algorithms, and a matching control signal is generated based on the adjustment amount of the primary air damper and transmitted to the primary air damper. The primary damper actuator implements adjustment logic based on the deviation of tobacco shred quantity per unit length to stabilize material load. It weights and fuses the quantified stem index with the tobacco shred quantity per unit length to generate the feedforward control quantity for the secondary damper. This is then combined with the feedback control quantity calculated from the stem mass flow rate deviation for coordinated adjustment. Finally, a limiting function constrains the damper opening within the physical range. Compared to the adjustment lag problem of single feedback control, the coordinated control mode of feedforward and feedback can respond to changes in incoming material earlier, shortening the adjustment response time. Simultaneously, feedback correction can compensate for errors in the feedforward model, and the limiting constraint avoids mechanical damage. Ultimately, precise and stable adjustment of the secondary damper is achieved, ensuring the consistency of the stem-to-stick separation rate. A feedforward and feedback composite adjustment logic is designed for the secondary damper to precisely control the stem-to-stick separation effect. The feedforward signal for the secondary adjustment comes from the tobacco state data after the primary damper, realizing the linkage between the two levels of control. The graded control avoids mutual interference between the primary and secondary damper adjustments. The stable material load of the primary damper lays a uniform material foundation for the precise control of stem-to-stick separation in the secondary damper, while the precise adjustment in the secondary damper ensures the final separation quality. The two work together to improve the control stability of the air separation system and reduce the impact of incoming material fluctuations on the final product quality.

[0110] Example 4

[0111] Figure 4 This is a schematic diagram of a secondary air-separated stem quality control device provided in Embodiment 4 of the present invention. Figure 4 As shown, the device includes: an image extraction module 410, a feature calculation module 420, a weighing processing module 430, and an opening adjustment module 440, wherein:

[0112] The image extraction module 410 is used to receive the image data of tobacco shreds on the conveyor belt after the primary air damper, which is acquired at preset intervals by the encoder and transmitted by the vision acquisition component, and to extract the image features of the tobacco shreds.

[0113] The feature calculation module 420 is used to calculate the amount of tobacco per unit length based on the regression model constructed in advance by the coefficients and intercept terms obtained through weighing and sampling calibration, and the image features of the tobacco.

[0114] The weighing processing module 430 is used to receive the weight data of the separated stems transmitted by the weighing detection component, and to perform filtering, calibration and delay compensation processing to obtain the stem mass flow rate data at the time of action of the second-stage damper.

[0115] The opening adjustment module 440 is used to generate a control signal based on the unit length tobacco shred quantity data and stem mass flow rate data, through an internal model control proportional-integral-derivative or Smith control algorithm, and transmit the control signal to the actuator, so that the actuator adjusts the opening of the corresponding primary damper and / or secondary damper according to the control signal.

[0116] The technical solution of this invention uses an encoder to trigger a vision acquisition component to acquire images of tobacco shreds on the conveyor belt after the primary air damper at preset intervals. Combined with a regression model constructed from weighing and sampling calibration, the image features are converted into data on the amount of tobacco shreds per unit length. This allows for real-time and accurate capture of changes in tobacco shred quantity, dynamically adjusting the opening of the primary air damper to offset fluctuations in the incoming material load, thus maintaining a stable amount of tobacco shreds per unit length entering the secondary air classifier and reducing processing deviations in subsequent processes. Simultaneously, this method filters, calibrates, and compensates for the weight data from the weighing detection component, eliminating interference and aligning the action time of the secondary air damper. Combined with a control algorithm possessing time-delay compensation capabilities, a control signal is generated, which can stably control the stem mass flow rate within a preset range. Within the specified range, this system avoids issues such as tobacco residue on stems or stems carrying tobacco, ensuring consistent stem separation across different batches. By integrating data on the amount of tobacco per unit length with stem mass flow rate data, it achieves coordinated adjustment of the primary and secondary air dampers. This combines the goal of stable material load in the primary stage with precise stem separation control in the secondary stage, strengthening the synergy between the two stages of air separation. Compared to independent adjustment modes, this improves the overall control accuracy of the system, further stabilizing the purity of the separated tobacco and the stem recovery rate. It also enhances the stability of stem separation quality and the uniformity of materials in subsequent processing, enabling the system to cope with complex operating conditions such as fluctuations in upstream material flow and changes in conveyor belt speed. This reduces system maintenance costs, downtime losses, and the risk of quality fluctuations during production.

[0117] Based on the above embodiments, the image extraction module 410 is specifically used for:

[0118] The foreground region of the tobacco image data is separated by grayscale processing and image segmentation algorithms, and the connected components within the foreground region are identified by connected component analysis algorithms. Target connected components with aspect ratios greater than a preset threshold and grayscale values ​​higher than a set grayscale threshold are selected.

[0119] The target proportion of the pixel area of ​​the target connected region to the total pixel area of ​​the foreground region is calculated and used as a quantitative indicator of the tag index.

[0120] Accordingly, based on the above embodiments, the opening adjustment module 440 is specifically used for:

[0121] The deviation between the unit length tobacco quantity data and the preset target value is calculated. The primary damper adjustment is calculated using an internal model control proportional-integral-derivative (PID) or Smith control algorithm. A matching control signal is generated based on the primary damper adjustment and transmitted to the primary damper actuator. The stem index quantification index and the unit length tobacco quantity data are multiplied by preset feedforward weighting coefficients and summed to obtain the secondary damper feedforward control quantity. The deviation between the preset stem mass flow rate target value and the stem mass flow rate data is calculated. The secondary damper feedback control quantity is calculated using an internal model control PID or Smith control algorithm. The secondary damper feedforward control quantity and the secondary damper feedback control quantity are summed to obtain the initial control quantity of the secondary damper. The initial control quantity of the secondary damper is then limited to the physical lower and upper limits of the secondary damper opening using a limiting function to obtain the secondary damper adjustment quantity. A matching control signal is generated based on the secondary damper adjustment quantity and transmitted to the secondary damper actuator.

[0122] The secondary wind-separated stem label quality control device provided in this embodiment of the invention can execute the secondary wind-separated stem label quality control method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0123] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0124] Example 5

[0125] Figure 5 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0126] like Figure 5As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0127] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0128] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, central processing unit (CPU), graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as the secondary wind selection post-straining quality control method, i.e.:

[0129] The system receives image data of tobacco shreds on the conveyor belt after the primary damper, which is acquired at preset intervals by an encoder triggered by the vision acquisition component, and extracts the image features of the tobacco shreds. Based on the regression model constructed in advance by coefficients and intercept terms obtained through weighing and sampling calibration, and the image features of the tobacco shreds, the system calculates the amount of tobacco shreds per unit length. The system receives the weight data of the separated stem sticks transmitted by the weighing and detection component, and performs filtering, calibration, and delay compensation processing to obtain the stem stick mass flow rate data aligned with the action time of the secondary damper. Based on the amount of tobacco shreds per unit length and the stem stick mass flow rate data, the system generates a control signal through an internal model control proportional-integral-derivative or Smith control algorithm, and transmits the control signal to the actuator, so that the actuator adjusts the opening of the corresponding primary damper and / or secondary damper according to the control signal.

[0130] In some embodiments, the secondary wind-sorting post-stem tag quality control method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the secondary wind-sorting post-stem tag quality control method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the secondary wind-sorting post-stem tag quality control method by any other suitable means (e.g., by means of firmware).

[0131] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0132] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0133] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0134] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0135] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0136] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through a communication network. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0137] 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.

[0138] 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 quality control system for stem labels after a two-stage air separation, characterized in that, The system includes a material conveying component, a vision acquisition component, a weighing and detection component, a controller, and actuators. The material conveying component includes a primary damper, a conveyor belt, a secondary damper, and a separation zone arranged sequentially along the upstream material inlet to the pneumatic conveying position. The vision acquisition component is electrically connected to the controller. The weighing and detection component is electrically connected to the controller. The actuators include a primary damper actuator and a secondary damper actuator, which are respectively connected to the primary damper and the secondary damper, and are also electrically connected to the controller. The material conveying assembly is used to sequentially convey the upstream material to the primary air damper for initial diversion, then convey it to the secondary air damper via a conveyor belt, and after being guided by the secondary air damper, it is sent to the separation zone for separation of tobacco shreds and stems. The separated stems are then conveyed to the downstream of the separation zone, where a weighing and detection assembly is installed. The visual acquisition component is used to acquire image data of tobacco shreds on the conveyor belt at preset intervals and send it to the controller. The weighing and detection component is used to collect the weight data of the separated twigs and send it to the controller; The controller is used to receive tobacco image data transmitted by the vision acquisition component and perform feature extraction and calculation, receive weight data transmitted by the weighing detection component and perform filtering, calibration and delay compensation processing, fuse the two types of processed data to generate a control signal through a preset control algorithm, and transmit the control signal to the actuator. The actuator is used to receive control signals sent by the controller and adjust the opening degree of the corresponding primary damper and / or secondary damper according to the control signals.

2. The system according to claim 1, characterized in that, The visual acquisition component includes a camera, a light source, an optics cover, and an encoder. The camera and light source are installed above the conveyor belt after the primary damper, the optics cover is located outside the camera and light source, and the encoder is connected to the conveyor belt for transmission.

3. The system according to claim 1, characterized in that, The controller is also used for: Monitor the operating status of the vision acquisition component. When the vision acquisition component meets the vision acquisition fault conditions, switch the control loop of the primary damper to manual or open-loop mode, and simultaneously control the secondary damper to operate independently based on the weight data transmitted by the weighing detection component; and / or Monitor the operating status of the weighing detection component. When the weighing detection component meets the weighing detection fault conditions, switch the secondary damper to manual control mode, maintain the closed-loop steady flow control of the primary damper, and trigger an alarm signal.

4. The system according to claim 1, characterized in that, The weighing detection component includes a weighing sensor and a weighing instrument, which are installed on the skewer conveying path downstream of the separation zone, and the weighing instrument is electrically connected to the controller; the weighing sensor is equipped with an analog-to-digital converter, and has undergone zero-point and span calibration and temperature drift compensation, and is fixed by a vibration-proof installation method.

5. The system according to claim 1, characterized in that, The controller includes an industrial computer and a programmable logic controller; The programmable logic controller is used to receive the weight data transmitted by the weighing and detection component, and to perform filtering, calibration and delay compensation processing to obtain the sludge mass flow rate data at the moment of action of the secondary damper. An industrial computer is used to receive tobacco image data transmitted by a vision acquisition component, extract the image features of the tobacco, and calculate the amount of tobacco per unit length by combining a regression model constructed in advance with coefficients and intercept terms obtained through weighing and sampling calibration. Based on the amount of tobacco per unit length and the stem mass flow rate data, a control signal is generated and transmitted to the actuator through an internal model control proportional-integral-derivative or Smith control algorithm.

6. A method for quality control of stem labels after secondary air separation, executed by the controller of the system described in any one of claims 1-5, characterized in that, The method includes: The system receives image data of tobacco shreds on the conveyor belt after the primary air damper, which is acquired at preset intervals by the encoder triggered by the visual acquisition component, and extracts the image features of the tobacco shreds. Based on the regression model constructed in advance by obtaining coefficients and intercept terms through weighing and sampling calibration, and the image features of tobacco shreds, the amount of tobacco shreds per unit length is calculated. The system receives the weight data of the separated stem from the weighing and detection component, and performs filtering, calibration and delay compensation processing to obtain the stem mass flow rate data at the moment of action of the second-stage damper. Based on the data of tobacco shreds per unit length and the mass flow rate of the stem, a control signal is generated by the internal model control proportional-integral-derivative or Smith control algorithm, and the control signal is transmitted to the actuator, so that the actuator adjusts the opening of the corresponding primary damper and / or secondary damper according to the control signal.

7. The method according to claim 6, characterized in that, Extracting image features of tobacco shreds, including: The foreground region of the tobacco image data is separated by grayscale processing and image segmentation algorithms, and the connected components within the foreground region are identified by connected component analysis algorithms. Target connected components with aspect ratios greater than a preset threshold and grayscale values ​​higher than a set grayscale threshold are selected. The target proportion of the pixel area of ​​the target connected region to the total pixel area of ​​the foreground region is calculated and used as a quantitative indicator of the tag index. Accordingly, based on the data of tobacco shred quantity per unit length and stem mass flow rate, a control signal is generated using an internal model control proportional-integral-derivative or Smith control algorithm. This control signal is then transmitted to the actuator, causing the actuator to adjust the opening of the corresponding primary and / or secondary dampers according to the control signal. This includes: The deviation between the unit length tobacco shred amount data and the preset unit length tobacco shred amount target value is calculated. The adjustment amount of the first-level damper is calculated by the internal model control proportional integral derivative or Smith control algorithm. The matching control signal is generated according to the adjustment amount of the first-level damper and transmitted to the first-level damper actuator. The feedforward control quantity of the secondary damper is obtained by multiplying the quantified index of the stem and the data of the amount of tobacco per unit length by a preset feedforward weighting coefficient and then summing them. Calculate the deviation between the preset target value of the stem mass flow rate and the stem mass flow rate data, and calculate the secondary damper feedback control quantity through internal model control proportional-integral-derivative or Smith control algorithms; The initial control quantity of the secondary damper is obtained by summing the feedforward control quantity and the feedback control quantity of the secondary damper. The initial control quantity of the secondary damper is then limited to the physical lower limit and upper limit of the opening of the secondary damper by the limiting function to obtain the adjustment quantity of the secondary damper. A matching control signal is generated based on the adjustment amount of the secondary damper and transmitted to the secondary damper actuator.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the secondary wind-separated post-stem label quality control method according to any one of claims 6-7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the secondary wind-sorting post-stem quality control method as described in any one of claims 6-7.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the secondary wind-separated stem label quality control method according to any one of claims 6-7.

Images