Cigarette tipping paper gluing position and gluing thickness in-situ detection system and method

Abstract

The invention provides an in-situ detection system and method for the gluing position and the gluing thickness of tipping paper for cigarettes, the system comprises a detection module, a control module and an operation module, and the position and the thickness of latex are obtained according to the measurement result of the water content in the latex through near infrared rays. The detection system and method are not influenced by ambient light and working distance fluctuation. And for cigarettes of different brands, the detection module does not need to be adjusted, so that the use is more convenient.

Description

Technical Field

[0001] This invention relates to the field of tobacco cork splicing technology, specifically to a system and method for online detection of the glue application position and thickness of tobacco cork during the splicing process. Background Technology

[0002] Ensuring and improving the quality of cigarette splicing during production has always been a research hotspot in the industry. Cigarette splicing is one of the core processes in cigarette production by cigarette rolling machines. Cigarette splicing uses latex as an adhesive, and the cigarette stick, filter, and tipping paper are joined together by a rubbing plate and a rubbing wheel.

[0003] In standard cigarette tipping paper, there are sections in the middle and on both sides where no glue is supplied. There are two reasons for this: First, to prevent glue residue from adhering to the cutter, which would dull the cutter and cause rough edges on the cigarette cut. Second, to prevent latex from overflowing onto the surfaces of the conveyor rollers, paper cutter rollers, combining rollers, cigarette brushes, and other mechanisms on the packing machine, which would create glue residue and affect the operation of the equipment.

[0004] Therefore, the wetting effect (adhesive amount) and application position of the latex on the bonding surface of the tipping paper directly affect the cigarette's bonding effect, easily leading to quality problems such as rough edges, weak adhesion, air leakage, glue overflow, and tipping paper curling. Insufficient adhesive on the tipping paper or misalignment of the adhesive will result in weak adhesion and air leakage; excessive adhesive will cause glue overflow during bonding, and in severe cases, may cause machine shutdowns due to clogging of the corrugated board. Therefore, detecting and monitoring the adhesive quality (adhesive amount and application position) of the tipping paper can effectively improve cigarette quality.

[0005] To address the issue of online detection of the glue coating quality in cigarette tipping paper, Guangxi Tobacco Industry Co., Ltd.'s patent CN17080284A discloses "An apparatus and method for automatic detection and adjustment of the glue coating position in cigarette tipping paper." This system uses a top light source and a high-speed camera to identify the glue position in the tipping paper of the cigarette from a grayscale image. However, this system cannot detect the glue coating thickness. The Baumer GDS 300-M6 glue quantity sensor from Germany uses infrared measurement technology to detect the glue coating quality in tipping paper. This sensor integrates six glue quantity measuring heads, capable of simultaneously collecting data from six glue tracks. Four of these heads are adjustable based on the glue edge position and the location of the excess glue, while the other two are fixed glue quantity measuring heads whose detection positions cannot be moved. These heads are primarily used to measure the glue quantity in the middle area of ​​the tipping paper. However, this type of glue quantity sensor has the following shortcomings: the sensor relies on glue quantity information from six single points to replace the glue quantity information of the entire glued cross section of the tipping paper. Glue quantity information outside the detection points cannot be obtained, which means that the information collected by the sensor cannot fully reflect the glued quality of the tipping paper. Summary of the Invention

[0006] To address the shortcomings of existing technologies, an in-situ detection system for the glue application position and thickness of tobacco tipping paper was developed, and its detection method was studied to achieve a full evaluation of the glue application quality of tipping paper, improve the rejection rate of defective glue-coated products, and reduce the false detection rate of glue application on tipping paper.

[0007] The present invention provides a system and method for detecting the position and thickness of adhesive coating on tobacco water-based cork paper, which uses near-infrared spectroscopy to measure the water content in latex to obtain the position and thickness of the latex.

[0008] Furthermore, the present invention provides a method for detecting the glue application position and thickness of tobacco water-based cork paper, which includes the following steps:

[0009] S1. Measurement: That is, to obtain the physical size corresponding to a single pixel of the target cork paper in the infrared spectrum;

[0010] The specific method is as follows:

[0011] S1.1. Adjust the direction of the near-infrared laser so that it is perpendicular to the direction of movement of the tipping paper;

[0012] S1.2. Use a near-infrared array camera to capture images of the cross-section of the cork paper;

[0013] S1.3. Measure the width b of the tipping paper;

[0014] S1.4. Count the number of pixels n on the tipping paper in the image taken in S1.2;

[0015] S1.5. Calculate the physical size corresponding to a single pixel, i.e., b / n.

[0016] S2. Calibration: That is, after replacing the cork paper with a calibration plate, obtain the homography matrix between the camera image plane and the plane calibration plate;

[0017] The specific method is as follows:

[0018] S2.1. Place the calibration plate on the tipping paper and take a picture of the calibration plate with a camera;

[0019] S2.2. The homography matrix is ​​used to characterize the coordinate transformation relationship between the camera image plane and the cork paper plane (replaced by a plane calibration plate), and its expression is:

[0020]

[0021] Where W represents the coordinate value in the cork paper plane coordinate system, H represents the homography matrix, and C represents the coordinate value of the corresponding point in the image plane coordinate system;

[0022] The homography matrix H is expressed as:

[0023]

[0024] The specific geometric meaning of each element is explained in four functional blocks below;

[0025] (1) Linear transformation part ( These four elements are primarily responsible for the linear transformation of the image within the plane;

[0026] Scaling ( ); change the size of the image.

[0027] Rotation ( ); rotate the image.

[0028] • Miscutting ( ); This causes the image to be distorted into a parallelogram shape.

[0029] (2) Translation part ( , These two elements are responsible for image translation;

[0030] · ; represents the displacement in the horizontal direction (x-axis).

[0031] · ; represents the displacement in the vertical direction (y-axis).

[0032] (3) Perspective / projection section ( , This is the core part of the homography matrix, responsible for perspective distortion;

[0033] They determine whether lines that were originally parallel will remain parallel after being transformed.

[0034] The existence of these two parameters allows for the simulation of a perspective effect where objects appear smaller in the distance and larger in the near.

[0035] (4) Scale factor ( )

[0036] • Typically, the scaling factor is normalized to 1.

[0037] • Homography matrices have “homogeneity”, meaning that if all nine elements of the matrix are multiplied by a non-zero constant, the transformation it describes is still the same; therefore, it actually has only eight degrees of freedom.

[0038] S2.3. A pair of matching points on the image is based on Expanding on the relationship, we get:

[0039]

[0040] Use h９ =1. Normalizing the homography matrix yields:

[0041] ;

[0042] S2.4. A pair of matching points provides two constraint equations, which, due to the homogeneous coordinates, expand as shown below:

[0043]

[0044] The four pairs of matching points provide eight equations as shown below:

[0045]

[0046] Solving this system of linear equations yields the homography matrix H.

[0047] S3. Demonstration: Take a set of high-quality, neatly joined cigarettes (no glue residue, wrinkles, curling edges, uniform circumference, no glue separation, no displacement). The glue coating thickness of this set should be as thick as possible (cigarettes joined within the upper limit of the specified thickness). Take another set of neatly joined cigarettes (also meeting the above conditions), with the glue coating thickness as thin as possible (cigarettes joined within the lower limit of the specified thickness). Use a near-infrared laser to photograph the near-infrared light reflected from the latex on the tipping paper, and record the grayscale value of the line laser. This grayscale value reflects the glue coating thickness. Take the lowest grayscale value in the first set as the upper limit r1, and the highest grayscale value in the second set as the lower limit r2. Also, record a large number of grayscale values ​​r3 of the reflected line laser when the tipping paper is not coated with glue, and calculate the mean μ and standard deviation σ of these samples (for subsequent judgment). In this case, r3 > r2 > r1. (The sample size for the demonstration step can be determined based on specific production requirements and production plans.)

[0048] S4. Inspection: After the near-infrared array camera captures an image triggered by the start signal of the paper cutting roller, the image is analyzed to identify and remove tobacco paper with glue coating defects.

[0049] S4.1. When the paper cutting wheel drives the coated cork paper to move, the near-infrared laser is turned on, and the near-infrared array camera enters the ready state. The trigger signal of the near-infrared array camera is connected to the encoder signal line of the paper cutting wheel motor. When the encoder sends a pulse signal, the near-infrared array camera is activated to capture a line of images.

[0050] S4.2. Remove any tobacco paper with adhesive coating defects.

[0051] The aforementioned defects include: the glue application area not meeting the specifications, and the glue application thickness not meeting the specifications. The specific analysis process is as follows:

[0052] S4.2.1. Based on the homography matrix H, analyze each frame of the row image in real time and identify the gluing area in the row image; use image processing techniques (such as gray threshold segmentation), combined with the significant gray difference between the glued and unglued areas in the near-infrared image, to determine which pixels belong to the gluing area.

[0053] S4.2.2. For each gluing point in the gluing area identified in the row image (a gluing point refers to the position of each tiny pixel representing glue in the captured image, which is a region from a process perspective and a "set of points" in algorithm processing, and coordinate transformation is performed point by point to achieve high-precision positioning.), based on the homography matrix H, perform coordinate transformation on the gluing point to obtain the true position of the gluing point in the physical coordinate system of the tipping paper, and compare it with the physical boundaries of all pre-defined non-glued areas to determine whether the gluing point falls within the non-glued area;

[0054] When the result is "yes", execute "rejection";

[0055] When the result is "no", proceed to S4.2.3;

[0056] S4.2.3. For the gray value r of each pixel in the row image, compare it with the gray value in S3 to determine the gluing thickness of the pixel of the tipping paper for cigarettes;

[0057] When r≈r3 (taking μ±3σ, about 99.73% confidence interval), it is considered that this pixel is a non-glued area, and execute "rejection";

[0058] When r>r2, it is considered that the gluing thickness of this pixel is less than the set gluing thickness, and execute "rejection";

[0059] When r<r1, it is considered that the gluing thickness of this pixel is greater than the set gluing thickness, and execute "rejection".

[0060] Regarding the above rejection time control, the specific method is as follows:

[0061] According to the physical size corresponding to a single pixel in S2, convert the row image into gluing quality data in physical size; the number of pulses for one rotation of the encoder is n, and the diameter of the cutting wheel roller is d mm, then the distance between adjacent two row images obtained by the near-infrared linear array camera is (π*d) / n, calculate the time delay t for the tipping paper for cigarettes with gluing defects to reach the rejection valve through the distance between the near-infrared linear array laser and the rejection valve and the rotation speed of the cutting wheel; when detecting the tipping paper for cigarettes with gluing defects and a rejection instruction occurs, send a rejection signal to the rejection valve after a delay of time t, and at this time, exactly reject the spliced cigarette with gluing defects.

[0062] In addition, the present invention also provides an in-situ detection system for the gluing position and gluing thickness of tipping paper for cigarettes, including:

[0063] The detection module, used for image acquisition, includes a near-infrared laser, a near-infrared array camera, and a reflector;

[0064] The near-infrared laser has a wavelength of 1460nm;

[0065] The near-infrared array camera can accept wavelengths from 900 to 1700 nm and has a maximum line frequency of 20 kHz. The near-infrared array camera is connected to the encoder of the motor in the winding unit to achieve line triggering synchronization.

[0066] This mirror is used to deflect light and avoid structural interference;

[0067] The control module is used to analyze the acquired images to obtain evaluation results on the glue application position and thickness. It is connected to the rejection valve of the roll splicing machine and sends a rejection signal to the rejection valve when a defective product is found, so as to reject the defective product.

[0068] Operation module: Used for functions such as setting system parameters and displaying data curves.

[0069] Furthermore, the present invention provides a detection system for the position and thickness of adhesive coating on tobacco tipping paper, characterized in that: a linear laser emitted by a near-infrared laser is guided to the moving surface of the tipping paper by a laser reflector, and the reflected light is then guided to the near-infrared linear array camera by a reflector, with the optical path deflected by 90°.

[0070] Compared with the prior art, the present invention has the following advantages:

[0071] This invention indirectly measures the thickness of latex on tobacco cork paper based on the absorption curve of moisture to near-infrared light. Moreover, the detection system is an active light measurement system, so the detection accuracy of the system is not affected by ambient light or fluctuations in working distance.

[0072] This invention uses a near-infrared array camera to acquire adhesive quality data from a cross-section of the cigarette paper, providing more comprehensive detection information. Furthermore, it eliminates the need to adjust the detection module for different brands of cigarettes, making it more convenient to use. Attached Figure Description

[0073] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0074] Figure 1 Schematic diagram of an in-situ detection system for the location and thickness of adhesive coating on cigarettes;

[0075] Figure 2 This is a schematic diagram of the absorption curve of water to infrared light;

[0076] Figure 3This is a schematic diagram of the optical path design of the detection system in this invention;

[0077] Figure 4 This is a schematic diagram of the infrared reflection thickness measurement principle in this invention;

[0078] Figure 5 This is a schematic diagram of the principle of adhesive coating profile measurement based on a line scan camera in this invention;

[0079] Figure 6 This is a schematic diagram of the cork paper photographed in this invention;

[0080] Figure 7 This is a schematic diagram of measuring cork paper with vernier calipers in this invention;

[0081] Figure 8 This is a schematic diagram illustrating the relationship between adhesive coating thickness and grayscale value in this invention.

[0082] Figure 9 This is a schematic diagram of the adhesive thickness detection principle in this invention (the upper and lower limits of adhesive thickness in the figure are assumed values, and the actual values ​​are determined by measurement).

[0083] Figure 10 This is a flowchart of the in-situ detection system for the glue application position and thickness of tobacco water cork paper according to the present invention.

[0084] 3-1 Camera reflector, 3-2 Laser reflector, 3-3 Laser, 3-4 Lens, 3-5 Line scan camera, 4-1 Infrared laser, 4-2 Photoelectric sensor, 5-1 Infrared laser, 5-2 Line scan camera, 5-3 Tipping paper, 6-1 Laser, 6-2 Tipping paper, 7-1 Vernier caliper, 7-2 Tipping paper Detailed Implementation

[0085] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0086] This embodiment provides a system for detecting the glue application position and thickness of tobacco water-based cork paper. The system hardware configuration is as follows:

[0087] like Figure 1 As shown, this tobacco paper coating position and coating thickness detection system mainly includes: a detection module, a control module, and an operation module.

[0088] The detection module, whose core components include: a near-infrared laser: This invention selects a semiconductor laser with a wavelength of approximately 1450nm or 1940nm because this wavelength is near the near-infrared absorption peak of water molecules (see...). Figure 2 (Moisture absorption spectrum curve), which is highly sensitive to moisture and has high measurement sensitivity.

[0089] Near-infrared linear array camera: A 1024-pixel linear array camera is selected, whose spectral response range covers the selected laser wavelength.

[0090] Optical components include a laser reflector, a camera reflector, and a beam splitter. As shown in Figure 3, to accommodate the compact installation space inside the winding machine, especially to reduce the dimension in the height direction, this invention designs a 90° deflected optical path. The 90° deflection refers to a right-angle turn in the overall optical path layout, converting the vertical (height) installation space to the horizontal (length) direction. The line laser emitted by laser 3-3 is first deflected by approximately 80° by laser reflector 3-2 before illuminating the cork paper surface. The reflected light is collected by camera reflector 3-1 and finally converged onto the line scan camera 3-5 through lens 3-4. This optical path design effectively transfers the system's height dimension to the length direction, enhancing the compatibility of different models. Without this design, without these reflectors and lasers, the camera and the object being measured need to be aligned in the same straight line in the height direction, which would occupy a significant amount of vertical installation height.

[0091] The control module uses an industrial computer (IPC), which connects to the line scan camera via a Camera Link interface and to the encoder of the winding machine motor and the rejection valve via a digital I / O card.

[0092] The operation module is a touchscreen human-machine interface (HMI) installed on the industrial control computer, used for parameter setting, real-time image display, and data curve display.

[0093] The specific method for detecting the glue application location and thickness of tobacco water-cooled cork paper using the above-mentioned equipment is as follows:

[0094] S1. Measurement and calibration, used to determine spatial scale;

[0095] After the system is installed, measurements are taken first. The direction of the near-infrared laser is adjusted to ensure its light band is perpendicular to the direction of movement of the tipping paper. Then, the uncoated tipping paper is passed through the detection area, and the line scan camera captures one frame. The actual physical width *b* of the tipping paper is precisely measured using calipers, and the number of pixels *n* occupied by the tipping paper area in the camera image is counted. The physical size corresponding to each pixel is then calculated as *b / n* pixels. This parameter is used to convert the pixel position in the image to its actual physical position.

[0096] Then, calibration was performed by placing a calibration plate on the tipping paper, taking pictures of the calibration plate with the camera, and calculating the homography matrix between the camera image plane and the plane calibration plate.

[0097] The specific method is as follows:

[0098] S2.1. Place the calibration plate on the tipping paper and take a picture of the calibration plate with a camera;

[0099] S2.2. The homography matrix is ​​used to characterize the coordinate transformation relationship between the camera image plane and the cork paper plane (replaced by a plane calibration plate), and its expression is:

[0100]

[0101] Where W represents the coordinate value in the cork paper plane coordinate system, H represents the homography matrix, and C represents the coordinate value of the corresponding point in the image plane coordinate system;

[0102] The homography matrix H is expressed as:

[0103]

[0104] The specific geometric meaning of each element is explained in four functional blocks below;

[0105] (5) Linear transformation part ( These four elements are primarily responsible for the linear transformation of the image within the plane;

[0106] Scaling ( ); change the size of the image.

[0107] Rotation ( ); rotate the image.

[0108] • Miscutting ( ); This causes the image to be distorted into a parallelogram shape.

[0109] (6) Translation part ( , These two elements are responsible for image translation;

[0110] · ; represents the displacement in the horizontal direction (x-axis).

[0111] · ; represents the displacement in the vertical direction (y-axis).

[0112] (7) Perspective / Projection Section , This is the core part of the homography matrix, responsible for perspective distortion;

[0113] They determine whether lines that were originally parallel will remain parallel after being transformed.

[0114] The existence of these two parameters allows for the simulation of a perspective effect where objects appear smaller in the distance and larger in the near.

[0115] (8) Scale factor ( )

[0116] • Typically, the scaling factor is normalized to 1.

[0117] • Homography matrices have “homogeneity”, meaning that if all nine elements of the matrix are multiplied by a non-zero constant, the transformation it describes is still the same; therefore, it actually has only eight degrees of freedom.

[0118] S2.3. A pair of matching points on the image is based on Expanding on the relationship, we get:

[0119]

[0120] Use h ９ =1. Normalizing the homography matrix yields:

[0121] ;

[0122] S2.4. A pair of matching points provides two constraint equations, which, due to the homogeneous coordinates, expand as shown below:

[0123]

[0124] The four pairs of matching points provide eight equations as shown below:

[0125]

[0126] Solving this system of linear equations yields the homography matrix H.

[0127] S3. Thickness-Grayscale Relationship Teaching: This step aims to establish the correspondence between image grayscale values ​​and adhesive coating thickness. The principle is as follows: Figure 4 and Figure 5 As shown, the specific method is as follows:

[0128] To adjust the glue application amount, first, adjust the gap of the glue application rollers to the upper limit h1 required by the process. That is, take a set of high-quality, properly joined cigarettes (no glue stains, wrinkles, curling edges, uniform circumference, no glue separation, and no displacement), and ensure the glue thickness of this set is as thick as possible (cigarettes joined within the specified upper limit thickness). After production stabilizes, confirm the glue thickness meets the standard through quality feedback from the joined cigarettes. At this point, record the average grayscale value R1 of the glued area in the image captured by the line scan camera.

[0129] Similarly, adjust the adhesive thickness to the lower limit h2, that is, take another group of twisted cigarettes (which also meet the above conditions), and make the adhesive thickness of this group as thin as possible (the cigarettes are twisted within the specified lower limit thickness), and record the average gray value of the adhesive area at this time as R2.

[0130] Finally, a large amount of data was recorded when the tipping paper was not coated with glue, and the gray value of the tipping paper surface at this time was recorded as R3.

[0131] Because water molecules in latex have a strong absorption effect on near-infrared light (see principle below). Figure 2 The thicker the adhesive coating, the more moisture it contains, resulting in weaker light intensity reflected back to the camera and lower image grayscale values. Therefore, the relationship R3 > R2 > R1 must be satisfied. This set of grayscale values ​​(R1, R2, R3) serves as the benchmark for subsequent real-time detection.

[0132] S4. Synchronous detection and data acquisition, that is, during the production process, the system enters real-time detection mode.

[0133] The trigger signal of the line scan camera is connected to the encoder signal of the roller motor that drives the tipping paper. The encoder emits a pulse every time it rotates a certain angle. For each pulse received, the line scan camera acquires one line of images. In this way, each line of images corresponds precisely to a fixed position of the tipping paper along its length. Let the number of pulses in one revolution of the encoder be N, and the diameter of the driving roller be d, then the length distance of the tipping paper represented by two adjacent lines of images is (π*d) / N. By stitching together a series of lines of images in this way, a two-dimensional grayscale image reflecting the glue coating status of the entire section of tipping paper is obtained.

[0134] S5. Real-time calculation, judgment, and rejection control

[0135] The industrial control computer (control module) processes each row of images in real time. Based on the analysis results of all rows, it determines whether the glue coating quality of this section of tipping paper meets the preset standard pixel-level judgment: comparing the real-time grayscale value r of each pixel in the image with the taught value. This includes the following steps:

[0136] S5.1. Based on the homography matrix H, analyze each frame of the line image in real time and identify the glue-coated areas in the line image; using image processing techniques (such as gray-scale threshold segmentation) and combining the significant gray-scale difference between glue-coated and uncoated areas in the near-infrared image, determine which pixels belong to the glue-coated areas.

[0137] S5.2. For each glue application point identified in the row image within the glue application area (a glue application point refers to each tiny pixel position representing glue in the captured image, which is a region from a process perspective and a "set of points" in algorithm processing. Coordinate transformation is performed point by point to achieve high-precision positioning), based on the homography matrix H, perform coordinate transformation on the glue application point to obtain its true position in the physical coordinate system of the tipping paper, and compare it with the physical boundaries of all pre-defined glue-free areas to determine whether the glue application point falls within the glue-free area;

[0138] When the result is "yes", perform "rejection";

[0139] When the result is "no", proceed to S5.3;

[0140] S5.3. For the gray value r of each pixel in the row image, compare it with the gray value in S3 to determine the glue application thickness of the pixel on the tipping paper for cigarettes;

[0141] When r ≈ r3 (taking μ ± 3σ, approximately 99.73% confidence interval), it is considered that this pixel is a non-glue application area, and perform "rejection";

[0142] When r > r2, it is considered that the glue application thickness of this pixel is less than the set glue application thickness, and perform "rejection";

[0143] When r < r1, it is considered that the glue application thickness of this pixel is greater than the set glue application thickness, and perform "rejection".

[0144] Refer to Figure 9 , in a test case, perform identification and judgment based on the following logic:

[0145] For each section of tipping paper:

[0146] 1. Gray scale dimension judgment → Does [large area too thin / no glue] or [large area too thick] appear?

[0147] Yes → Trigger rejection (major defect)

[0148] No → Proceed to the next level of judgment

[0149] 2. Position dimension judgment → Is [center position] or [glue width] out of tolerance?

[0150] [[ID=4l]]Yes → Trigger rejection (positioning defect)

[0151] No → Proceed to the next level of judgment

[0152] 3. Morphology dimension judgment → Do local defects such as [broken glue] and [glue hole] exist?

[0153] Yes (and the defect size exceeds the standard) → Trigger rejection (appearance / continuous defects)

[0154] No → Determined as [Qualified Product]

[0155] Defect Location and Delay Calculation: Combining the pixel physical dimensions (b / n) obtained from the calibration process, the precise location (distance from the edge) of any defect point on the tipping paper can be calculated. Simultaneously, based on the encoder pulse spacing (πd / N) and the fixed mechanical distance L between the detection point and the rejection valve, the time required for the defect point to move from detection to the rejection valve can be calculated, i.e., the delay time t = L / v, where v is the running speed of the tipping paper (calculated from the encoder pulse frequency). Precise Rejection: Once a glue defect (such as broken glue, uneven glue strip width, glue too thick / too thin) is identified, the control module immediately starts a delay timer. After the delay time t, a high-speed pulse signal is precisely sent to the rejection valve, rejecting the corresponding problematic cigarette.

[0156] Advantages of this embodiment: Through the above implementation, the present invention achieves online detection and precise control of the glue coating quality and position of cork paper. The system is an active optical measurement system, unaffected by ambient light interference; it utilizes near-infrared absorption characteristics, resulting in stable and reliable measurement results; it employs line scanning imaging, obtaining two-dimensional glue coating information for the entire cross-section, which is more comprehensive than single-point sensors; and its optical path design is compact and highly adaptable.

[0157] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for detecting the glue application position and thickness of tobacco water-based cork paper, characterized in that: The location and thickness of the latex can be obtained by measuring the water content in the latex using near-infrared spectroscopy.

2. A method for detecting the glue application location and thickness of tobacco water-based cork paper, characterized in that, Includes the following steps: S1. Obtain the physical dimensions of a single pixel in the infrared image of the target cork paper; S2. Determine the spatial scale; S3. Establish the correspondence between image grayscale values ​​and adhesive coating thickness; S4. After the near-infrared array camera captures an image triggered by the start signal of the paper cutting roller, it analyzes the image and finds and removes tobacco paper with glue coating defects.

3. The method for detecting the glue application position and thickness of tobacco water-based cork paper as described in claim 1, characterized in that: The specific method of S1 is as follows: S1.

1. Adjust the direction of the near-infrared laser so that it is perpendicular to the direction of movement of the tipping paper; S1.

2. Use a near-infrared array camera to capture images of the cross-section of the cork paper; S1.

3. Measure the width b of the tipping paper; S1.

4. Count the number of pixels n on the tipping paper in the image taken in S1.2; S1.

5. Calculate the physical size corresponding to a single pixel, i.e., b / n.

4. The method for detecting the glue application position and thickness of tobacco water-based cork paper as described in claim 1, characterized in that: The specific method of S2 is as follows: S2.

1. Place the calibration plate on the tipping paper and take a picture of the calibration plate with a camera; S2.

2. The homography matrix is ​​used to characterize the coordinate transformation relationship between the camera image plane and the cork paper plane. Its expression is: Where W represents the coordinate value in the cork paper plane coordinate system, H represents the homography matrix, and C represents the coordinate value of the corresponding point in the image plane coordinate system; The homography matrix H is expressed as: S2.

3. A pair of matching points on the image is based on Expanding on the relationship, we get: Use h ９ =1. Normalizing the homography matrix yields: ； S2.

4. A pair of matching points provides two constraint equations, which, due to the homogeneous coordinates, expand as shown below: The four pairs of matching points provide eight equations as shown below: Solving this system of linear equations yields the homography matrix H.

5. The method for detecting the glue application position and thickness of tobacco water-based cork paper as described in claim 1, characterized in that: The specific method of S3 is as follows: S3.

1. Take and record the image. When the adhesive coating thickness reaches the upper limit h1, record the grayscale value r1 of the reflected linear laser light captured by the near-infrared array camera at this time. S3.

2. Take and record the image. When the adhesive coating thickness reaches the lower limit h2, record the gray value r2 of the reflected linear laser light captured by the near-infrared array camera at this time. S3.

3. Take and record the grayscale value r3 of the reflected linear laser light captured by the near-infrared array camera when the tipping paper is not coated with glue.

6. The method for detecting the glue application position and thickness of tobacco water-based cork paper as described in claim 1, characterized in that: The specific method of S4 is as follows: S4.

1. When the paper cutting roller drives the glued tipping paper to move, the near-infrared laser is turned on, and the near-infrared array camera enters the ready state. The trigger signal of the near-infrared array camera is connected to the encoder signal line of the paper cutting roller motor. When the encoder sends a pulse signal, the near-infrared array camera is activated to capture a line of images. When the number of pulses in one revolution of the encoder is N and the diameter of the driving roller is d, the length spacing of the tipping paper represented by two adjacent lines of images is (π*d) / N. Based on the length spacing of the tipping paper, a series of lines of images are stitched together to obtain a two-dimensional grayscale image reflecting the glue coating of the entire section of tipping paper. S4.

2. Identify and remove tobacco paper with adhesive coating defects.

7. The method for detecting the glue application position and thickness of tobacco water-based cork paper as described in claim 5, characterized in that: The method for identifying adhesive coating defects is as follows: S4.2.

1. Analyze each frame of the row image in real time based on the homography matrix H, and identify the glue application area in the row image; S4.2.

2. For each glue application point in the identified glue application area in the row image, perform coordinate transformation on the glue application point based on the homography matrix H to obtain the real position of the glue application point in the physical coordinate system of the tipping paper, and compare it with the physical boundaries of all pre-defined glue-free areas to determine whether the glue application point falls within the glue-free area; When the result is "yes", execute "rejection"; When the result is "no", proceed to S4.2.3; S4.2.

3. For the gray value r of each pixel in the row image, compare it with the gray value in S3 to determine the glue application thickness of the pixel of the tipping paper for cigarettes; When r≈r3, it is considered that the pixel is a non-glue application area, and execute "rejection"; When r2 < r ≤ r3, it is considered that the glue application thickness of the pixel is less than the set glue application thickness, and execute "rejection"; When r < r1, it is considered that the glue application thickness of the pixel is greater than the set glue application thickness, and execute "rejection".

8. A method for in-situ detection of the glue application position and glue application thickness of tipping paper for cigarettes as claimed in claim 1, wherein: The time control method for the rejection is as follows: According to the physical size corresponding to a single pixel in S2, convert the row image into glue application quality data in the physical size; the number of pulses for one rotation of the encoder is n, and the diameter of the roller of the cutting wheel is d mm, then the distance between adjacent row images obtained by the near-infrared linear array camera is (π * d) / n, and calculate the time delay t for the tipping paper for cigarettes with glue application defects to reach the rejection valve through the distance between the near-infrared linear array laser and the rejection valve and the rotation speed of the cutting wheel; When a rejection instruction is generated for the tipping paper for cigarettes with glue application defects detected, send a rejection signal to the rejection valve after a delay of time t.

9. A system for detecting the glue application position and thickness of tobacco water-based cork paper, characterized in that, Including: A detection module for image acquisition, including a near-infrared laser, a near-infrared linear array camera, and a reflector; The near-infrared linear array camera is connected to the encoder of the motor in the cigarette making and tipping unit to achieve trigger synchronization; The reflector is used for deflecting light; A control module for analyzing the acquired image, obtaining an evaluation result of the glue application position and glue application thickness, and being connected to the rejection valve of the cigarette making and tipping machine, and sending a rejection signal to the rejection valve when defective products appear to reject the defective products.

10. A system for in-situ detection of the glue application position and glue application thickness of tipping paper for cigarettes as claimed in claim 9, wherein: The line laser emitted by the near-infrared laser is guided to the surface of the moving tipping paper by a laser reflector, and its reflected light is then guided to the near-infrared linear array camera by the reflector, and the deflected optical path is 90°.

Images