Cigarette capsule linear detection system and method based on machine vision

Abstract

The invention provides a cigarette capsule linear detection system and method based on machine vision, the cigarette capsule linear detection system comprises a linear conveying module, a visual imaging module and a classification module, a feeding module is connected with the linear conveying module, the linear conveying module is provided with a fixing groove, a cigarette spherical capsule is placed in the fixing groove, and the visual imaging module is connected with the classification module; the fixing groove is used for rolling and turning over the spherical capsules for cigarettes in the process of linearly conveying the spherical capsules for cigarettes; the visual imaging module is arranged above the linear conveying module and is used for carrying out image acquisition on the spherical capsule for the cigarette to obtain a spherical capsule image; and the classification module classifies and screens the spherical capsules for cigarettes according to the spherical capsule images. According to the invention, the inherent defects of low detection flux, complicated structure, inconsistent capsule attitude control and the like of the existing equipment are overcome, and the detection system which is compact in structure, stable in operation and capable of realizing ultrahigh detection flux is provided by adopting the design of the linear conveying conveyor belt.

Description

Technical Field

[0001] This application belongs to the field of tobacco product testing, specifically relating to a linear detection system and method for tobacco capsules based on machine vision. Background Technology

[0002] During the industrial production of tobacco capsules (also known as spherical tobacco capsules), various factors such as production processes and equipment precision can easily lead to a variety of appearance and internal quality defects, including irregular roundness, oil seepage from cuts, internal bubbling, and color differences. These defects directly affect the product performance and usage effect of tobacco capsules, therefore, strict quality testing of tobacco capsules is necessary.

[0003] Traditional testing methods rely on manual inspection, which is not only inefficient and labor-intensive, but also susceptible to the influence of subjective judgment and operational proficiency of the inspectors, resulting in poor consistency. Furthermore, manual contact testing does not meet the requirements of hygiene and quality management standards (such as GMP standards) during capsule production, and cannot meet the quality control needs of large-scale and standardized production.

[0004] To address the drawbacks of manual inspection, automated capsule inspection equipment has emerged in existing technologies, with the mainstream structures being rotary or vibratory disc types. The rotary inspection equipment works by placing cigarette capsules into pre-set slots on the edge of a rotating disc. The capsules rotate synchronously with the disc, passing through multiple functional stations such as a visual inspection station and a sorting station to complete defect detection and sorting. However, this type of rotary inspection equipment has the following inherent technical limitations, making it difficult to meet the high-efficiency, high-precision inspection requirements for large-scale production of cigarette capsules: 1. Limited throughput: Due to the upper limit of the diameter of the turntable and the fact that the turntable speed is limited by factors such as mechanical strength and capsule stability, it cannot be increased indefinitely. This results in a significant bottleneck in the number of capsules that can be detected per unit time. For example, the maximum detection efficiency of the current mainstream turntable pellet sorting machine is only 800,000 capsules / hour. Its detection rate has reached the performance limit of this structure and equipment, and it is difficult to further improve it to meet the needs of larger-scale production.

[0005] 2. Complex structure and large footprint: In order to achieve simultaneous operation of multiple processes, the rotary testing equipment needs to be set up with multiple independent functional stations around the rotary table, which results in a complex overall mechanical structure and a large number of parts, thus making the equipment bulky and occupying a large area, increasing the site occupancy cost of the production workshop.

[0006] 3. Difficulty in controlling capsule posture: The capsules used in cigarettes have a spherical structure. During the high-speed rotation of the turntable, the capsules located in the inner and outer ring slots of the turntable have different linear velocities due to the inconsistent rotation radii. This leads to posture deviation and wobbling of the capsules, which disrupts the posture stability of the capsules during imaging, affects the imaging quality of the visual inspection module, and ultimately reduces the detection accuracy, making it easy for defects to be missed or falsely detected.

[0007] 4. Inconvenient equipment maintenance: The turntable body, edge slots, and multiple independent workstations of the turntable structure are all areas where dust and capsule debris are easily accumulated, making cleaning difficult. At the same time, the multi-workstation collaborative design makes maintenance operations such as equipment debugging and parts replacement cumbersome, resulting in low maintenance efficiency and increased equipment maintenance costs and downtime losses. Summary of the Invention

[0008] In view of this, the purpose of this application is to provide a linear detection system and method for tobacco capsules based on machine vision, so as to solve the above problems.

[0009] To solve the above-mentioned technical problems, this application adopts the following technical solution: In a first aspect, this application provides a linear detection system for cigarette capsules based on machine vision. The linear detection system for cigarette capsules includes: a linear conveying module, a visual imaging module, and a classification module. The linear conveying module is provided with a fixing groove in which spherical cigarette capsules are placed. The fixing groove is used to roll and flip the spherical cigarette capsules during the linear conveying process. The visual imaging module is located above the linear conveying module and is used to acquire images of the spherical cigarette capsules to obtain images of the spherical cigarette capsules. The classification module classifies and filters the spherical cigarette capsules according to the images of the spherical cigarette capsules.

[0010] Furthermore, the linear detection system for cigarette capsules also includes a feeding module, which comprises a hopper, a spiral directional track, and a vibrating plate. The hopper is a storage cavity with an open top and a narrowing bottom, used to store the spherical cigarette capsules to be tested. The spiral directional track is arranged spirally upward along its inner wall and is connected to the upper opening of the hopper and the linear conveying module, respectively, to transfer the spherical cigarette capsules to the linear conveying module. The spiral directional track is equipped with a vibrating plate, which drives the spiral directional track to generate spiral lift and radial micro-vibration in the circumferential direction, causing the spherical cigarette capsules to automatically climb, disperse, and gradually adjust their posture along the inner wall of the spiral directional track under the action of vibration.

[0011] Furthermore, the width, height, and sidewall inclination of the spiral directional track are designed to match the external dimensions of the spherical tobacco capsules. Only the spherical tobacco capsules are allowed to pass through with their axial direction horizontal and their long axis along the conveying direction. This is to screen out tobacco capsules that do not conform to the orientation, which then fall back to the bottom of the spiral directional track under the action of vibration and gravity for reordering.

[0012] Furthermore, the linear conveying module also includes a drive unit and a conveyor belt. The conveyor belt is an array-type conveyor belt, and the surface of the array-type conveyor belt is provided with a fixing groove that matches the shape of the spherical capsule for cigarettes. When the spherical capsule for cigarettes runs along the conveyor belt, the spherical capsule for cigarettes will roll in the fixing groove with the contact surface. The contact surface has a certain coefficient of friction, which can meet the rolling movement of the spherical capsule for cigarettes.

[0013] Furthermore, the visual imaging module includes a support and an image acquisition device. The support is bridging the top of the linear conveyor module. The image acquisition device includes a camera and an illumination unit. The illumination unit is installed at the bottom of the linear conveyor module to provide a parallel backlight for the camera. The camera is mounted on the support and is used to acquire images of the spherical capsules used for tobacco, thereby obtaining images of the spherical capsules.

[0014] Furthermore, the classification module includes sensors and controllers, and sorting actuators. The sensors are used to count and mark the spherical capsules for tobacco use. The controller processes the images of the spherical capsules according to the image processing algorithm and classifies the spherical capsules for tobacco use into defective capsules and non-defective capsules. The controller controls the sorting actuators to sort the defective capsules to the waste recycling bin and sort the non-defective capsules to the qualified product channel.

[0015] Furthermore, image processing algorithms include OpenCV or deep learning models.

[0016] Secondly, this application provides a linear detection method for cigarette capsules based on machine vision. This method is applied to the aforementioned linear detection system for cigarette capsules and includes the following steps: Step S1: A linear conveying module fixes the spherical cigarette capsule, ensuring that the spherical cigarette capsule is smoothly transported to the target position along a preset trajectory at high speed; Step S2: A visual imaging module captures an image of the spherical cigarette capsule located on the linear conveying module; Step S3: A classification module performs image processing and analysis on the spherical capsule image and classifies the spherical cigarette capsules into defective capsules and non-defective capsules.

[0017] Furthermore, step S3 includes: step S31: the classification module performs image denoising and image enhancement processing on the spherical capsule image; step S32: the classification module extracts features from the processed spherical capsule image to obtain capsule feature data; step S33: the classification module determines whether the capsule feature data belongs to the preset qualified product parameter range: if yes, the spherical capsule for tobacco is determined to be a non-defective capsule; if no, the spherical capsule for tobacco is determined to be a defective capsule.

[0018] Furthermore, the capsule feature data includes at least one of the following: capsule size, capsule center position, capsule outline roundness, and capsule color HSV value.

[0019] The spherical capsule for cigarettes disclosed in this application is primarily a round flavor capsule (also known as a flavor capsule, flavoring capsule, or brittle capsule), a mainstream category of cigarette capsules. It is mainly used in cigarette filters, forming the core functional component of flavor capsule cigarettes. The capsule contains different types of flavoring liquids, which smokers can break during smoking to release, enriching the cigarette's flavor and enhancing its aroma. It also features low tar content and minimal odor buildup, catering to the needs of different smokers and is widely used in large-scale cigarette production. The quality of this round flavor capsule directly determines the taste experience and product quality of the flavor capsule cigarette; therefore, accurate detection of its appearance and internal defects is crucial, placing higher demands on the adaptability and accuracy of the detection system.

[0020] As can be seen from the above technical solution, this application overcomes the inherent defects of existing rotary detection equipment, such as low detection throughput, complex structure, and inconsistent capsule posture control. By adopting a linear conveyor belt design, it provides a compact, stable, and ultra-high detection throughput linear detection system for cigarette capsules. This system specifically possesses the following significant technical features: First, it has high throughput. Its linear motion mode has no inertial limitations. By increasing the conveying speed and camera line frequency, it achieves extremely high detection speed, with a detection rate of 3,000-10,000 particles / minute, far exceeding existing rotary table detection equipment. Secondly, it features high precision and high stability. The tobacco capsules are placed in a square groove and move in a straight line. All the spherical capsules maintain the same posture and are not affected by centrifugal force, resulting in clear and stable imaging, which can significantly improve the accuracy of defect detection. Third, it has a compact structure and is easy to maintain. It adopts a modular linear structure design, which makes the equipment occupy a small area. The conveyor belt and vision inspection module are easy to disassemble and clean, which can meet high standards of production hygiene management requirements. Fourth, it is flexible and expandable, making it easy to add or reduce testing stations on a linear conveyor line to adjust testing efficiency. It can also be adapted to different specifications of spherical tobacco capsules by changing square grooves of different particle sizes. Fifth, it has a high level of intelligence, with the system integrating deep learning algorithms, which can continuously optimize the defect identification model, effectively adapt to new defect types, and further improve detection efficiency and reliability. Attached Figure Description

[0021] The above description of this application and the following detailed embodiments will be better understood when read in conjunction with the accompanying drawings. It should be noted that the drawings are merely examples of the claimed technical solutions.

[0022] Figure 1 This is an overall structural diagram of a linear detection system for tobacco capsules; Figure 2 This is an internal structure diagram of a linear detection system for tobacco capsules; Figure 3 This is a magnified view of the internal structure of a linear detection system for tobacco capsules. Figure 4 This is a schematic diagram of the image acquisition by the visual imaging module; Figure 5 This is a diagram showing the distribution of cameras; Figure 6 This is a flowchart of the linear detection method for tobacco capsules.

[0023] The reference numerals in the attached figures are explained as follows: Feeding module: 1; Linear conveyor module: 2; Drive unit: 21; Conveyor belt: 22; Visual imaging module: 3; Bracket: 31; Camera: 32; Shots: 33; Lighting units: 34; Category modules: 4; Spherical capsules for tobacco: 5. Detailed Implementation

[0024] The following detailed description of the features and advantages of this application is sufficient to enable any person skilled in the art to understand the technical content of this application and implement it accordingly. Based on the specification, claims and drawings disclosed in this specification, those skilled in the art can easily understand the related objectives and advantages of this application.

[0025] The invention will now be described with reference to the accompanying drawings, in which similar reference numerals denote similar elements. While specific structures and arrangements are discussed, it should be understood that this is done merely for illustrative purposes. Those skilled in the art will recognize that other structures and arrangements can be used without departing from the spirit and scope of the invention. It will be apparent to those skilled in the art that the invention can also be used in a variety of other applications.

[0026] In this specification and claims, several terms will be used, and unless otherwise indicated, these terms will be defined to have the following meanings: The singular forms “a” and “the” include their corresponding plural forms. “At least one” means one or more, and “more” means two or more. “At least one of the following” or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can be expressed as: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0027] All figures used to represent component amounts, properties (e.g., molecular weight), reaction conditions, etc., should be considered to be modified in all cases by the terms "within the unavoidable margin of error" or "about". Therefore, the numerical values ​​set forth herein are approximate and may vary depending on the desired properties sought to be obtained by the present invention. The principle of equivalents, which is applied to a minimum and not intended to limit the scope of the claims, should be applied, for example, each value should be interpreted at least according to the specified significant digits and by applying conventional rounding techniques.

[0028] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0029] In the description of this embodiment, it should be noted that the terms "upper", "lower", "inner", "bottom", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship in which the product is usually placed during use. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0030] Unless otherwise indicated, the following abbreviations have the following meanings, and any other abbreviations used herein but not defined have their generally accepted standard meanings: All other terms used herein that are specifically defined herein shall have the general meaning understood by one of ordinary skill in the art, in particular meaning that, upon reading the claims, specification and drawings of this patent, one can directly and without doubt determine how the technical solution of this patent can be implemented.

[0031] Even if there are incomplete descriptions, omissions, or ambiguities in the grammar, words, punctuation, graphics, symbols, etc. of the claims, specification, and drawings of this patent, a person skilled in the art can still arrive at the only correct understanding by reading the claims, specification, and drawings as a whole without extensive reasoning or experimentation, and effectively exclude various incorrect interpretations that are not aimed at achieving the purpose of this patent.

[0032] Those skilled in the art would first choose to read the claims, specification, and drawings of this patent to reasonably interpret the terms; secondly, they would choose to refer to the relevant definitions in other documents published by the applicant before the filing date to reasonably interpret the terms; thirdly, they would choose the references cited in this patent to reasonably interpret the terms; and finally, they would choose to combine the technical dictionaries, technical manuals, reference books, textbooks, national or industry technical standards, etc., commonly used by those skilled in the art to reasonably interpret the terms.

[0033] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0034] Please refer to Figure 1 This application provides a linear detection system for cigarette capsules based on machine vision. The linear detection system for cigarette capsules includes: a feeding module 1, a linear conveying module 2, a vision imaging module 3, and a classification module 4.

[0035] The feeding module 1 is mainly composed of three parts: a hopper, a spiral directional track, and a vibratory feeder. The components work together to achieve orderly, stable, and efficient feeding of the spherical capsules 5 to be tested, providing a reliable material supply guarantee for the subsequent linear testing process.

[0036] The hopper adopts a funnel-shaped storage cavity structure with an open top and a constricted bottom. Its inner wall is polished to effectively prevent jamming, residue, or surface scratches of the tobacco spherical capsules 5 during storage and feeding. The volume of the hopper can be flexibly designed according to the actual production and testing throughput to meet the storage needs of batches of tobacco spherical capsules 5 in large-scale production, and facilitate continuous feeding operations without frequent replenishment.

[0037] The spiral directional track serves as the core channel for material transfer. It is arranged in a spiral upward shape along the inner wall of the silo. Its lower end is smoothly connected to the lower contraction area of ​​the silo, while its upper end is precisely connected to the feeding end of the linear conveyor module 2, forming a complete material transfer link. This ensures that the tobacco spherical capsule 5 can smoothly transition from the silo to the linear conveyor module 2, avoiding problems such as falling or disordered posture during material transfer.

[0038] To achieve automatic climbing, dispersion, and attitude calibration of the spherical tobacco capsules 5, a vibratory feeder is installed on the spiral directional track. This vibratory feeder is fixedly connected to the bottom of the spiral directional track and, during operation, drives the spiral directional track to generate a spiral lift force along the circumferential direction and a slight radial vibration. These two forces work synergistically on the spherical tobacco capsules 5 to be tested. Under the vibration, the spherical tobacco capsules 5 can overcome their own weight and track friction, automatically climbing along the inner wall of the spiral directional track. Simultaneously, the radial micro-vibration allows stacked spherical tobacco capsules 5 to quickly disperse, preventing clumping and congestion.

[0039] Preferably, in order to further improve the attitude consistency of the spherical capsule 5 for tobacco and ensure the accuracy of subsequent testing, the key structural parameters of the spiral directional track are precisely matched and designed according to the specific external dimensions (including particle size, spherical diameter tolerance, etc.) of the spherical capsule 5 to be tested. Among them, the core track width, track height and side wall inclination angle work together to form the key structure for attitude screening and calibration.

[0040] Specifically, the track width is designed to be slightly larger than the maximum diameter of the spherical tobacco capsule 5, while strictly controlling the width margin to prevent the capsule from swaying left and right within the track; the track height matches the capsule diameter setting to ensure that the capsule is placed horizontally within the track and does not tip over or stack; the side wall inclination angle has been adjusted and optimized multiple times, adopting a gentle and reasonable tilt angle, which can provide stable support and guidance for the capsule, and also provide a fall channel for capsules with unqualified posture.

[0041] Through precise matching of the above-mentioned dimensional parameters, the spiral directional track can achieve forced attitude screening of the spherical tobacco capsules 5, allowing only the capsules to pass smoothly in a standard attitude with the axial direction horizontal and the long axis along the conveying direction of the track. For the spherical tobacco capsules 5 that do not conform to the standard attitude (such as tilting, inverting, or tumbling laterally), under the combined action of the track vibration force and their own gravity, they will automatically slide off the track side wall and fall back to the bottom area of ​​the spiral directional track, re-participate in the sorting, climbing and attitude calibration process in the track, until they reach the standard attitude before entering the subsequent conveying stage.

[0042] This attitude screening structure eliminates the need for additional complex screening mechanisms. It can achieve attitude calibration and automatic return of unqualified attitude capsules by relying on the design of the track itself. This not only simplifies the overall structure of the feeding module 1 and reduces the equipment manufacturing cost, but also significantly improves the attitude consistency of capsules entering the linear conveyor module 2. It reduces detection errors caused by attitude disorder from the source, further ensuring the detection accuracy and stability of the entire detection system. At the same time, it avoids the waste of detection resources caused by unqualified attitude capsules entering the detection process.

[0043] It is understandable that the feeding module 1 has a compact structure, strong compatibility with the linear conveyor module 2 and the detection module, stable operation, low noise, easy disassembly and cleaning, and complies with GMP hygiene management standards. It can effectively adapt to the feeding needs of tobacco spherical capsules 5 of different specifications and particle sizes.

[0044] The linear conveying module 2 also includes a drive unit 21 and a conveyor belt 22, which work together to provide stable, uniform, and controllable conveying power for the spherical capsules 5 used in cigarettes, ensuring the smooth progress of subsequent visual inspection processes.

[0045] The drive unit 21 adopts a high-precision servo drive structure, which has the advantages of adjustable speed, stable operation, low noise, and fast response speed. It can flexibly adjust the conveying speed of the conveyor belt 22 according to the actual detection throughput requirements, so as to achieve precise matching with the feeding rate of the feeding module 1 and the detection rate of the vision inspection module, avoiding material congestion or conveying interruption. At the same time, the high-precision control characteristics of the servo drive can ensure the consistency of the running speed of the conveyor belt 22, providing power guarantee for stable capsule rolling and clear imaging. The conveyor belt 22 adopts an array structure design. Compared with the traditional integrated conveyor belt 22, its structure is more targeted and stable. The surface of the array conveyor belt 22 is evenly distributed with several fixed grooves. The number and spacing of the fixed grooves can be flexibly designed according to the detection efficiency and capsule specifications. The inner contour of each fixed groove is precisely matched with the outer dimensions of the tobacco spherical capsule 5. The arc-shaped groove structure can not only reliably limit the capsule, preventing the capsule from shifting, shaking or falling during the conveying process, but also provide sufficient rolling space for the capsule.

[0046] When the spherical tobacco capsule 5 enters the array conveyor belt 22 from the spiral directional track of the feeding module 1, it will fall precisely into the corresponding fixed groove and move in a straight line synchronously with the conveyor belt 22. During this process, the spherical tobacco capsule 5 will roll naturally along the contact surface inside the groove under the limiting action of the fixed groove. This contact surface has undergone special surface treatment and has a preset reasonable coefficient of friction. The range of the coefficient of friction has been adjusted and optimized many times. It can provide sufficient friction to drive the capsule to move synchronously with the conveyor belt 22, meet the requirement of smooth rolling movement of the spherical tobacco capsule 5, and effectively improve the detection probability of bubbling in the defects of spherical capsules.

[0047] This also avoids problems such as scratches and wear on the capsule surface due to excessive friction, or slippage and misalignment of the capsule due to insufficient friction. The rolling motion of the capsule within the fixed groove ensures that all areas of the capsule surface are evenly exposed within the detection field of the visual inspection module, effectively avoiding blind spots caused by stationary capsules and further improving the comprehensiveness and accuracy of defect detection.

[0048] Please refer to Figure 2 and Figure 3 The visual imaging module 3 mainly consists of two parts: a support 31 and an image acquisition device. The structural design of each component is tailored to the detection requirements, working together to achieve efficient and accurate image acquisition. The support 31 is made of high-strength, lightweight alloy material and has a gate-shaped cross-connection structure. It is securely mounted above the linear conveyor module 2. The span of the support 31 precisely matches the width of the linear conveyor module 2, and the height of the support 31 can be flexibly adjusted according to image acquisition needs. The bottom of the support 31 is fixed to the ground or equipment mounting platform with bolts, ensuring that the support 31 is firmly installed and does not shake during operation, preventing the image acquisition device from shifting due to vibration of the support 31 and affecting imaging accuracy. A limit installation position is preset on the crossbeam of the support 31 for precisely fixing the image acquisition device. The installation position can be adjusted horizontally along the crossbeam, facilitating the adjustment of the image acquisition device's acquisition position according to the width of the conveyor belt 22 and the spacing of the fixing slots. This ensures accurate alignment with the spherical tobacco capsules 5 in each fixing slot, achieving full coverage and no omissions in image acquisition.

[0049] Please refer to Figure 4 The image acquisition device, as the core execution component of the visual imaging module 3, mainly includes a camera 32 and an illumination unit 34. Working together, through a reasonable installation layout and parameter settings, they ensure that the acquired images of the spherical tobacco capsules 5 are clear, complete, and have high contrast. This allows for the accurate presentation of surface defects such as color differences and oil seepage from cuts, as well as internal defects such as blistering. Specifically, the illumination unit 34 is embedded in the bottom of the linear conveyor module 2, and its installation position corresponds one-to-one with the acquisition position of the camera 32 above. The illumination unit 34 uses a high-brightness, low-power parallel light source, which emits a uniform and stable parallel backlight during operation. This parallel backlight can penetrate the shell of the spherical tobacco capsules 5 while avoiding interference factors such as reflections and shadows. It can clearly present the internal structure and defects of the capsules, and also highlight the surface features such as color differences and minor scratches, providing reliable illumination for the camera 32 to acquire high-quality images.

[0050] Camera 32 is an industrial-grade high-definition camera, boasting advantages such as high imaging resolution, fast acquisition speed, and strong anti-interference capabilities. It is compatible with the high-speed conveying requirements of the linear conveyor module 2, ensuring accurate capture of complete images of each spherical tobacco capsule 5 during its rolling process. Camera 32 is equipped with a replaceable lens 33. The focal length and magnification of lens 33 can be flexibly selected according to the specifications (particle size) of the spherical tobacco capsule 5. By replacing lenses 33 with different parameters, accurate imaging of capsules of different sizes can be achieved, significantly improving the adaptability of the visual imaging module 3. Diverse detection needs can be met without replacing the entire camera 32, reducing equipment upgrade and maintenance costs.

[0051] Please refer to Figure 5 The camera 32 is fixed to the crossbeam of the portal frame 31 via an adjustable bracket 31. Its installation angle and height are flexibly adjustable, ensuring that the lens 33 is vertically aligned with the fixed slot on the linear conveyor module 2. Furthermore, the field of view of the lens 33 completely covers the spherical tobacco capsule 5 within the fixed slot. As the capsule rolls within the slot, the camera 32 continuously acquires images of various areas on the capsule surface at a preset acquisition frequency. These images are then integrated to form a complete image of the spherical tobacco capsule 5, which is transmitted to the subsequent defect identification module for analysis and processing, laying a solid foundation for accurate defect identification. In addition, the camera 32 is linked to the drive unit 21 of the linear conveyor module 2, automatically adjusting the image acquisition frequency according to the conveyor belt 22's speed. This avoids problems such as image blurring and missed acquisitions, further improving the reliability and consistency of image acquisition.

[0052] The classification module 4 is mainly composed of three parts: sensors, controllers, and sorting actuators. The three parts are linked by signals to realize the accurate counting, trigger control, defect judgment, and automatic sorting of tobacco spherical capsules 5. This ensures that qualified capsules can smoothly enter the next process, while defective capsules are promptly recycled, thus ensuring product quality and avoiding material waste.

[0053] Among them, high-precision photoelectric sensors are selected and are installed in a symmetrical embedded manner. They are respectively deployed at the starting end of the linear conveyor belt 22 and at the corresponding position of the image acquisition by the visual imaging module 3. The sensors at the two installation positions have clear division of labor and work together to realize the dual functions of counting and photo triggering, providing basic data support for the accurate operation of the classification module 4.

[0054] The photoelectric sensor installed at the starting end of the linear conveyor belt 22 corresponds precisely to the fixed slot of the linear conveyor module 2. When the spherical capsule 5 for tobacco falls from the feeding module 1 into the fixed slot of the conveyor belt 22 and enters the starting end of the conveyor line with the conveyor belt 22, the capsule blocks the detection light of the photoelectric sensor. The sensor immediately generates a trigger signal and transmits it to the controller. At the same time, it completes the counting and statistics of single capsules. The controller records the total number of capsules, the number of capsules detected, and the number of capsules to be detected in real time, which makes it easy for staff to monitor the production and testing progress in real time. The detection pass rate can be calculated based on the counting data, providing data reference for production process optimization.

[0055] The photoelectric sensor deployed at the vision station is installed next to the camera 32 of the vision imaging module 3. It is linked with the camera 32 and the drive unit 21 of the linear conveyor module 2. When the capsule moves to the vision inspection station with the conveyor belt 22 and enters the field of view of the camera 32, the photoelectric sensor quickly detects the capsule and sends a trigger signal to precisely control the camera 32 to start taking pictures. This ensures that each capsule can be imaged at the imaging position, avoiding missed or incorrect images. At the same time, the trigger signal is synchronously transmitted to the controller, providing a timing synchronization guarantee for the controller to receive images and start the defect identification algorithm, ensuring the continuity of the inspection process.

[0056] The controller receives trigger signals and counting signals from the sensors, as well as spherical capsule image data transmitted by the vision imaging module 3. It then deeply integrates with preset deep learning defect recognition and image processing algorithms to achieve rapid analysis and defect determination of the capsule images. During operation, after receiving a complete image of the tobacco spherical capsule 5 transmitted by the vision imaging module 3, the controller immediately activates the image processing algorithm to perform preprocessing such as noise reduction, enhancement, and contour extraction. Then, using OpenCV algorithms or deep learning models, it can accurately identify external defects (such as irregular roundness, oil seepage from cuts, and color differences) and internal defects (such as internal blistering) in the capsule image. By comparing these defects with preset qualified capsule standard parameters and various defect feature templates, the controller quickly classifies the tobacco spherical capsule 5 into two categories: defective capsules and non-defective capsules. Simultaneously, it records the defect type and quantity of defective capsules, forming a detection data ledger for easy traceability and process improvement.

[0057] In addition, the controller can also achieve linkage control with the drive unit 21 of the linear conveyor module 2 and the camera 32 of the vision imaging module 3. According to the conveying speed of the conveyor belt 22 and the spacing between capsules, the sensor triggering frequency, the camera 32 shooting frequency and the action sequence of the sorting actuator can be flexibly adjusted to ensure that the entire detection and sorting process is synchronized and coordinated, and to avoid problems such as jamming and sorting misalignment.

[0058] The sorting actuator, as the execution component of the sorting module 4, is connected to the controller signal and installed at the end of the linear conveyor module 2. Its installation position corresponds one-to-one with the fixed groove of the conveyor belt 22, and it is equipped with a qualified product channel and a waste product recycling bin, which are used to receive non-defective capsules and defective capsules, respectively.

[0059] The sorting actuator uses an array-type high-speed solenoid valve air nozzle, with the nozzle position corresponding one-to-one with the capsule position on the conveyor belt. The response time is less than 15 milliseconds, which has the advantages of fast response speed, precise action, stable operation and no damage. It can be adapted to the high-speed conveying requirements of the linear conveyor module 2, ensuring that each capsule can be accurately sorted, while avoiding scratches and wear on the capsule surface during the sorting process, and ensuring the integrity of qualified capsules.

[0060] Once the controller completes the capsule classification and sends a sorting signal, the sorting actuator responds immediately, executing the corresponding sorting action according to the controller's instructions: if a capsule is determined to be defective, the sorting actuator moves quickly, smoothly pushing the defective capsule from the fixed groove of the conveyor belt 22 to the adjacent waste recycling bin. If a capsule is determined to be non-defective, the sorting actuator does not move, and the capsule continues to run with the conveyor belt 22, smoothly falling into the qualified product channel, and then being transported to the subsequent packaging and storage processes.

[0061] To further improve sorting accuracy, the sorting actuator can also be equipped with a position calibration sensor, which can provide real-time feedback on its own position. The controller fine-tunes the actuator's movement based on the feedback signal to ensure accurate sorting and effectively prevent defective capsules from being mixed into the qualified product channel or qualified capsules from being mistakenly sorted into the waste recycling bin.

[0062] It should be noted that the linear detection system for tobacco capsules in this application is specifically designed for tobacco capsules (round tobacco capsules). It is fundamentally different from the detection equipment on the market used for other types of capsules such as pharmaceutical capsules and food capsules. It is designed to be adapted to the structural characteristics, quality requirements and production scenarios of tobacco capsules, thus avoiding the problems of poor compatibility and insufficient detection accuracy of general-purpose detection equipment.

[0063] The specific differences are mainly reflected in the following aspects: Most pharmaceutical capsules on the market are long and non-spherical in shape, and the focus of testing is on performance indicators such as disintegration time and dissolution rate. The requirements for the accuracy of testing on appearance roundness and oil leakage from the cut are relatively low, and the testing throughput is much lower than that of the large-scale production scenario of tobacco capsules. Food capsules are mostly large-diameter, thick-shell structures with no brittleness requirements. There is no need to focus on testing internal foaming and slight color differences on the surface. However, tobacco capsules are small-diameter, thin and brittle shells (with flavoring liquid inside). The requirements for the accuracy of testing defects such as irregular roundness, oil leakage from the cut, and internal foaming are extremely high. It is also necessary to avoid damaging the shell during the testing process, and at the same time, it needs to be adapted to the needs of high-speed large-scale production.

[0064] Based on the specific testing requirements for tobacco capsules, the core relevant parameters (exemplary) of the testing system in this application are as follows, further demonstrating its differences from other capsule testing equipment and its specific compatibility: 1. Compatible Specifications: Specifically designed for cigarette capsules, it can accommodate particle sizes ranging from 0.8 to 3.0 mm (most pharmaceutical capsule testing equipment on the market accommodates particle sizes of 5-10 mm, and food capsules typically accommodate 4-8 mm). By changing the fixing slots of different sizes, it can be compatible with the three mainstream cigarette capsule particle sizes of 1.0 mm, 1.5 mm, and 2.0 mm, providing precise fit to the small particle size characteristics of cigarette capsules. 2. Detection accuracy parameters: Roundness and irregularity detection accuracy ≤ 0.01mm, which can accurately identify minute deformations of the capsule shell (the roundness detection accuracy of other capsule testing equipment on the market is mostly 0.05-0.1mm, which does not require such high precision); Color difference detection accuracy ΔE≤1.0, which can capture minute color differences on the surface of the capsule, meeting the requirements for consistent appearance of tobacco products; Internal bubble detection can identify bubbles with a minimum diameter ≥ 0.05mm, avoiding flavor leakage after the capsule is crushed due to internal bubbles; Cutting oil seepage detection sensitivity ≥ 0.01mg, which can accurately identify minute oil seepage traces and ensure the sealing of the capsule; 3. Detection throughput parameters: The detection rate is adjustable from 3,000 to 10,000 capsules / minute (corresponding to an hourly throughput of 180,000 to 600,000 capsules), which is suitable for the large-scale production needs of tobacco capsules (the throughput of pharmaceutical capsule testing equipment on the market is mostly 500-2,000 capsules / minute, and the throughput of food capsule testing equipment is mostly 1,000-3,000 capsules / minute, both of which are lower than this system). 4. Core Component Adaptation Parameters: The vision imaging module uses a 12-megapixel industrial-grade high-definition camera with adjustable lens magnification (10-50x), specifically adapted for imaging small-diameter capsules used in cigarette testing (other capsule testing equipment on the market typically uses cameras with 5-8 megapixels and magnification ≤20x); the illumination unit uses a parallel backlight with a wavelength of 550-650nm, which can penetrate the thin shell of the capsule without damaging it, clearly revealing internal defects (other capsule testing equipment on the market often uses direct light sources, which are prone to reflection and cannot be adapted for thin-shell testing). 5. Attitude control parameters: The inclination angle of the spiral directional track sidewall is adjustable from 5° to 15°, and the track width is 0.1-0.2mm larger than the size of the compatible burst beads, so as to accurately control the attitude of the burst beads; the friction coefficient of the contact surface of the fixed groove is controlled at 0.2-0.3, which not only ensures the smooth rolling of the burst beads, but also avoids excessive friction from damaging the brittle shell of the burst beads. 6. Sorting adaptation parameters: Sorting actuator response time ≤10ms, sorting accuracy ≤0.02mm, adopting a flexible pushing structure to avoid squeezing and breaking the beads during the sorting process.

[0065] The above parameters are specifically designed for the structural characteristics and quality requirements of tobacco capsules, clearly distinguishing them from other capsule testing equipment on the market. This ensures that the system can accurately, efficiently, and non-destructively complete the quality testing of tobacco capsules, meeting the high-standard control requirements for the large-scale production of tobacco capsules.

[0066] Please refer to Figure 6 Based on the same inventive concept, this application also provides a machine vision-based linear detection method for cigarette capsules, the specific steps of which are as follows: Step S1: The linear conveyor module 2 fixes the spherical tobacco capsule 5, ensuring that the spherical tobacco capsule 5 is smoothly transported to the target position along the preset trajectory under high-speed movement.

[0067] Specifically, the spherical capsules 5 to be tested are first stored in the hopper of the feeding module 1. Under the driving action of the vibrating plate, the spiral directional track generates spiral lift and radial micro-vibration, which makes the capsules automatically climb, disperse and complete the attitude calibration. They fall into the fixed groove of the array-type conveyor belt 22 of the linear conveying module 2 in a standard posture with the axial horizontal and the long axis along the conveying direction. The fixed groove can be a square groove to ensure that the spherical capsules can be transported smoothly to the target position in an orderly and efficient manner along the preset trajectory under high-speed movement.

[0068] The conveying speed of the linear conveying module 2 is adjustable. When the servo drive unit 21 of the linear conveying module 2 is started, the conveyor belt 22 is driven to run at a preset speed. The speed of the conveyor belt 22 is precisely matched with the feeding rate of the feeding module 1 and the shooting frequency of the subsequent visual imaging module 3, so as to avoid capsule congestion, conveying discontinuity or posture disorder. The maximum speed can reach more than 3,000 capsules per minute.

[0069] Step S2: The visual imaging module 3 takes a picture of the spherical capsule 5 for tobacco in the linear conveying module 2 to obtain an image of the spherical capsule.

[0070] When the spherical tobacco capsule 5 moves to the vision inspection station along the conveyor belt 22, the photoelectric sensor deployed at the vision station quickly detects the capsule and immediately sends a trigger signal to simultaneously trigger the camera 32 and the lighting unit 34 to start. The lighting unit 34 (installed at the bottom of the linear conveyor module 2) emits a uniform and stable parallel backlight that penetrates the capsule shell and avoids interference such as reflection and shadow, highlighting the defect features on the surface and inside of the capsule.

[0071] At the same time, the industrial-grade high-definition camera 32 installed on the gantry bracket 31 is activated, and the lens 33 is vertically aligned with the capsule in the fixed groove. The shooting frequency is automatically adjusted according to the conveying speed of the conveyor belt 22. As the capsule rolls naturally under the action of friction on the contact surface in the fixed groove, the camera 32 continuously shoots according to the preset frequency, capturing images of various areas on the surface of the capsule. Finally, the multiple frames of images are integrated and processed to form a spherical capsule image that can fully present the appearance (color difference, oil seepage from the cut, irregular roundness, etc.) and internal (internal bubbling, etc.) characteristics of the capsule. Then, the collected image data is transmitted to the controller of the classification module 4 in real time.

[0072] Step S3: Classification module 4 performs image processing and image analysis on the spherical capsule images, and classifies and filters the spherical capsules 5 for tobacco use into defective capsules and non-defective capsules.

[0073] Step S3 includes: Step S31: Classification module 4 performs image denoising and image enhancement processing on the spherical capsule image.

[0074] For example, the image denoising process employs an adaptive median filtering algorithm to precisely filter out random noise, light interference noise, and electronic noise that may occur during visual imaging and transmission. This algorithm can adaptively adjust the size of the filtering window according to the intensity and distribution characteristics of noise in the image. While effectively removing various noise impurities and avoiding noise interference with subsequent processing, it preserves the detailed features of the capsule image to the greatest extent (such as minor cuts, oil seepage marks, slight color differences, and internal microbubbles), preventing the blurring or loss of defect features due to excessive denoising.

[0075] Image enhancement processing employs a combination of histogram equalization and adaptive contrast adjustment. For capsule images acquired under different lighting conditions, it automatically optimizes the brightness, contrast, and grayscale levels of the images, enhancing the discernibility of the differences between defective and normal areas of the capsule. For example, enhancement processing can clearly highlight the color difference boundaries on the capsule surface, the color distinction between oil seepage from the cut and the capsule body, and improve the grayscale contrast between the bubbles inside the capsule and the capsule shell. This ensures that subsequent feature extraction can accurately capture the image features corresponding to various defects, avoiding misjudgments or omissions of defects due to insufficient image contrast or blurred details.

[0076] Step S32: Classification module 4 extracts features from the processed spherical capsule image to obtain capsule feature data.

[0077] Specifically, the controller of the classification module 4 calls a preset image feature extraction algorithm to perform precise feature extraction on the spherical capsule image after denoising and enhancement preprocessing, and finally obtains capsule feature data that can comprehensively reflect the quality status of the tobacco spherical capsule 5.

[0078] The feature extraction algorithm is precisely adapted to the defect type of the spherical tobacco capsule 5, and can realize the synchronous extraction of multi-dimensional features, ensuring that the extracted capsule feature data is comprehensive, accurate and representative. Among them, the capsule feature data includes at least one of the following: capsule size, capsule center position, capsule outline roundness and capsule color HSV value.

[0079] The extraction logic for each feature data is as follows: Capsule size is obtained by capturing the contour boundary of the capsule image through an edge detection algorithm, and by combining the image pixels with the calibrated ratio of the actual size, the actual particle size, diameter and other size parameters of the capsule are accurately calculated.

[0080] The capsule's center position is determined using the Hough circle detection algorithm to locate the geometric center coordinates of the capsule image, which is used to determine whether the capsule is offset or has abnormal posture.

[0081] Capsule outline roundness is quantified by calculating the deviation between the actual capsule outline and the standard circular outline, and is used to determine whether the capsule has any roundness defects.

[0082] The capsule color HSV value is formed by extracting three parameters—hue (H), saturation (S), and lightness (V)—from different regions of the capsule image to create a color feature vector. This vector is used to accurately identify color-related defects on the capsule surface, such as color differences and oil seepage from cuts.

[0083] In addition, based on actual testing needs, additional feature data such as the number, size, and location of air bubbles inside the capsule can be extracted to further enrich the feature dimensions and improve the comprehensiveness and accuracy of defect judgment.

[0084] Step S33: Classification module 4 determines whether the capsule feature data belongs to the preset qualified product parameter range: if yes, then the spherical tobacco capsule 5 is determined to be a non-defective capsule; if no, then the spherical tobacco capsule 5 is determined to be a defective capsule.

[0085] When the defective capsules reach the end of the conveyor line, the sorting module 4 issues an instruction to use a high-pressure gas nozzle to precisely blow or push them into the waste collection box, while the non-defective capsules continue to fall into the qualified product channel.

[0086] It is understood that the linear detection method for tobacco capsules based on machine vision provided in this application corresponds to the linear detection system for tobacco capsules based on machine vision provided in this application. For the sake of brevity, the same or similar parts can be referred to the content of the linear detection system for tobacco capsules section, and will not be repeated here.

[0087] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0088] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0089] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0090] In this specification, references to "an embodiment" or "a specific implementation" mean that a particular feature, structure, or characteristic described in connection with that embodiment / specific implementation is included in at least one embodiment / specific implementation of the invention. Therefore, the phrase "in one embodiment / specific implementation" appearing in various places in this specification does not necessarily refer to the same embodiment / setting, but rather to potentially different embodiments. Furthermore, specific features, structures, or characteristics may be combined in one or more embodiments / settings in any suitable manner, as will be apparent to those skilled in the art from this disclosure.

[0091] Similarly, it should be understood that in the above description of exemplary embodiments / specific implementations of the invention, various features of the invention are sometimes combined in a single embodiment / specific implementation or its figures and description, with the aim of simplifying the disclosure and aiding in the understanding of one or more of the various aspects of the invention. However, the method of description in this patent should not be construed as reflecting an intention that the claimed features of the invention are more than those expressly stated in each claim, except where explicitly stated otherwise or in obvious technical contradiction or exclusion. Rather, the inventive aspect reflected in the claims lies in not all the features of a single foregoing disclosed embodiment / specific implementation. Therefore, the claims following the detailed description are expressly incorporated herein by reference, each claim existing independently as a separate embodiment / specific implementation of the invention.

[0092] Furthermore, while some embodiments / specific implementations described herein include, but are not limited to, other features included in other embodiments / specific implementations, combinations of features from different embodiments / specific implementations are intended to be within the scope of the invention and form different embodiments / specific implementations, as will be understood by those skilled in the art. For example, in the following claims, embodiments / specific implementations of any claim can be used in any combination.

[0093] The terms and expressions used in this specification are for illustrative purposes and not for limitation. In using these terms and expressions, it is not intended to exclude any equivalents of the features or portions thereof shown and described, but rather to recognize that various modifications may be possible within the scope of the invention.

[0094] Therefore, it should be understood that although the invention has been specifically disclosed through preferred embodiments, exemplary embodiments and optional features, those skilled in the art may take variations or modifications of the concepts disclosed herein, and such variations and modifications are therefore considered to be within the scope of the invention as defined by the appended claims.

[0095] The specific embodiments given in this specification are examples of useful implementations of the present invention. It will be apparent to those skilled in the art that the present invention can be implemented using many variations of the devices, device components, and method steps disclosed in this specification.

[0096] The foregoing description of specific embodiments fully discloses the general features of the present invention, enabling others to easily modify and / or adapt such specific embodiments for various applications by applying knowledge within the scope of the art, without conducting excessive experimentation and without departing from the general concept of the present invention.

[0097] Therefore, based on the teachings and guidance provided herein, it is intended that such modifications and alterations be included within the meaning and scope of equivalents of the disclosed embodiments. It should be understood that the wording or terminology used herein is for descriptive purposes and is not intended to be limiting; thus, the wording or terminology in this specification will be interpreted by those skilled in the art based on the foregoing teachings and guidance.

[0098] Furthermore, the scope of the invention should not be limited to any of the exemplary embodiments described above, but only to the appended claims and their equivalents.

Claims

1. A linear detection system for cigarette capsules based on machine vision, characterized in that, The linear detection system for tobacco capsules includes: a linear conveying module, a visual imaging module, and a classification module. The linear conveying module is provided with a fixing groove, in which a spherical capsule for cigarettes is placed. The fixing groove is used to roll and flip the spherical capsule for cigarettes during the linear conveying process. The visual imaging module is positioned above the linear conveying module and is used to acquire images of the spherical tobacco capsule to obtain an image of the spherical capsule. The classification module classifies and filters the spherical capsules for tobacco based on the image of the spherical capsule.

2. The linear detection system for tobacco capsules according to claim 1, characterized in that, The linear detection system for cigarette capsules also includes a feeding module, which comprises a hopper, a spiral directional track, and a vibratory feeder. The hopper is a storage cavity that is open at the top and narrows at the bottom, used to store spherical tobacco capsules to be tested; The spiral directional track is arranged spirally upward along the inner wall and is connected to the upper opening of the hopper and the linear conveying module, respectively, for transferring the tobacco spherical capsule to the linear conveying module; The spiral directional track is equipped with a vibrating plate, which drives the spiral directional track to generate spiral lift and radial micro-vibration along the circumferential direction. Under the action of vibration, the spherical capsule for tobacco automatically climbs, disperses and gradually adjusts its posture along the inner wall of the spiral directional track.

3. The linear detection system for tobacco capsules according to claim 2, characterized in that, The width, height, and sidewall inclination of the spiral directional track are designed to match the external dimensions of the spherical tobacco capsules. The spherical tobacco capsules are only allowed to pass through with their axial direction horizontal and their long axis along the conveying direction. This is to screen out tobacco capsules that do not conform to the posture, which will fall back to the bottom of the spiral directional track under the action of vibration and gravity for reordering.

4. The linear detection system for tobacco capsules according to claim 1, characterized in that, The linear conveying module also includes a drive unit and a conveyor belt. The conveyor belt is an array-type conveyor belt, and the surface of the array-type conveyor belt is provided with the fixing groove that matches the shape of the spherical capsule for cigarettes. When the spherical tobacco capsule runs along the conveyor belt, it rolls within the fixed groove along with the contact surface. The contact surface has a certain coefficient of friction, which satisfies the rolling movement of the spherical tobacco capsule.

5. The linear detection system for tobacco capsules according to claim 1, characterized in that, The visual imaging module includes a support and an image acquisition device. The bracket is positioned above the linear conveying module; The image acquisition device includes a camera and an illumination unit. The illumination unit is installed at the bottom of the linear conveyor module and is used to provide a parallel backlight for the camera. The camera is mounted on the bracket and is used to acquire images of the spherical tobacco capsule to obtain an image of the spherical capsule.

6. The linear detection system for tobacco capsules according to claim 1, characterized in that, The sorting module includes sensors and controllers, and sorting actuators. The sensor is used to count and mark the spherical tobacco capsules; The controller processes the image of the spherical capsule according to the image processing algorithm and classifies the spherical capsules for tobacco into defective capsules and non-defective capsules; The controller controls the sorting actuator to sort the defective capsules to the waste recycling bin and the non-defective capsules to the qualified product channel.

7. The linear detection system for tobacco capsules according to claim 6, characterized in that, The image processing algorithms include OpenCV or deep learning models.

8. A linear detection method for cigarette capsules based on machine vision, characterized in that, The linear detection method for tobacco capsules is applied to the linear detection system for tobacco capsules according to claim 1, and the linear detection method for tobacco capsules includes: Step S1: The linear conveyor module fixes the spherical capsule for cigarettes, ensuring that the spherical capsule for cigarettes is transported smoothly to the target position along a preset trajectory under high-speed movement; Step S2: The visual imaging module takes a picture of the spherical capsule for tobacco in the linear conveying module to obtain an image of the spherical capsule; Step S3: The classification module performs image processing and image analysis on the spherical capsule image, and classifies and filters the spherical capsules for tobacco use into defective capsules and non-defective capsules.

9. The linear detection method for tobacco capsules according to claim 8, characterized in that, Step S3 includes: Step S31: The classification module performs image denoising and image enhancement processing on the spherical capsule image; Step S32: The classification module extracts features from the processed spherical capsule image to obtain capsule feature data; Step S33: The classification module determines whether the capsule feature data belongs to the preset qualified product parameter range: If so, then the spherical capsule for tobacco is determined to be the non-defective capsule; If not, then the spherical capsule for tobacco is determined to be the defective capsule.

10. The linear detection method for tobacco capsules according to claim 9, characterized in that, The capsule feature data includes at least one of the following: Capsule size, capsule center position, capsule outline roundness, and capsule color HSV value.

Images