Cigarette end face defect detection system

By using industrial cameras and image processing modules in the cigarette support detection system for image acquisition and analysis of cigarette support end faces, the problem of cigarette support detection in the special-shaped packaging equipment is solved, and automatic identification and early warning of all categories of quality defects in the cigarette support end faces is realized, improving the Product quality and production efficiency.

CN118947958APending Publication Date: 2024-11-15HONGYUN HONGHE TOBACCO (GRP) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411094100.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-09
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

The existing cigarette support end face detection system cannot conduct multiple categories of full inspections on cigarette support in mold boxes in special-shaped packaging equipment, resulting in failure to detect defective cigarette packages in time, product quality problems occur, and rely on manual sampling and poor timeliness and missed inspection rate. high.

Method used

Design a cigarette support end face defect detection system, including industrial cameras, image processing and operation workstation computers, databases and early warning platforms, collect cigarette support end face images through industrial cameras, use image processing modules to identify and classify defects, generate early warning information, and realize Automated detection and positioning.

Benefits of technology

It realizes the detection of all-category quality defects of the end face of the cigarette support, automatically identify defects, and pushes early warning information, improves the quality of cigarette products and production automation level, and reduces costs and manual sampling work intensity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118947958A_ABST
    Figure CN118947958A_ABST
Patent Text Reader

Abstract

The invention discloses a cigarette end face defect detection system, which comprises an industrial personal computer, the input end of the industrial personal computer is connected with an industrial camera, the output end of the industrial personal computer is connected with an image processing and operation workstation computer, the image processing and operation workstation computer is connected with a database, the database is connected with an early warning platform, and the industrial camera is used for collecting cigarette end face images; the industrial personal computer is used for storing cigarette end face images; the image processing and operation workstation computer is used for performing cigarette end face defect identification according to the cigarette end face image to obtain a cigarette end face defect type result and a defective cigarette position; the database is used for storing defect type results and defect positions and sending the results and the positions to the early warning platform. According to the cigarette end face defect detection system, the image of the cigarette end face is collected through the industrial camera, and the end face of the cigarette is analyzed through the image processing and operation workstation computer, so that full-category quality defect detection of the end face of the cigarette is achieved, and online detection of the quality of the end face of the cigarette in the packaging process of the cigarette is achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of cigarette detection, and more specifically, to a cigarette end face defect detection system. Background Art

[0002] While the emergence of special-shaped packaging equipment in the tobacco industry has provided cigarette manufacturers with more options, it has also created new challenges for cigarette end-face quality inspection. The mold boxes used in special-shaped packaging equipment differ from those used in conventional packaging equipment, making current cigarette end-face inspection systems incompatible. Due to the mold box design and the cigarettes' transport in cigarette boxes, special-shaped packaging equipment can easily cause wrinkles and scratches on the ignition end of cigarettes. Furthermore, brands produced using special-shaped packaging equipment often have higher quality requirements, placing even higher demands on quality inspection systems.

[0003] The current inspection system is unable to conduct comprehensive, multi-category inspections of cigarettes in mold boxes used in special-shaped packaging equipment, nor can it provide timely warnings of quality defects. As a result, defective cigarette packages are not discovered in a timely manner, leading to serious product quality issues. Manual sampling methods, on the other hand, can lead to problems such as a high rate of missed detection of cigarette defects and poor timeliness.

[0004] Therefore, there is an urgent need for a cigarette end face defect detection system. Summary of the Invention

[0005] The purpose of the present invention is to provide a cigarette end face defect detection system to solve the above-mentioned problems in the prior art and to achieve comprehensive automatic detection of cigarette end face defects.

[0006] The present invention provides a cigarette end face defect detection system, which includes: an industrial computer, an input end of the industrial computer is connected to an industrial camera, an output end of the industrial computer is connected to an image processing and operation workstation computer, the image processing and operation workstation computer is connected to a database, and the database is connected to an early warning platform, wherein:

[0007] The industrial camera is used to capture images of cigarette end faces;

[0008] The industrial computer is used to store the cigarette end face image and transmit the cigarette end face image to the image processing and operation workstation computer;

[0009] The image processing and operation workstation computer is used to identify cigarette end face defects based on the cigarette end face image, and obtain the cigarette end face defect type result and the defective cigarette position;

[0010] The database is used to store the cigarette end surface defect type results and the defective cigarette position, and send the cigarette end surface defect type results and the defective cigarette position to the early warning platform;

[0011] The early warning platform is used to generate early warning information based on the cigarette end face defect type result and the defective cigarette position.

[0012] In the cigarette end face defect detection system as described above, preferably, the industrial camera is arranged on the side of the lower smoke channel of the cigarette bin of the mold box conveying chain plate, and is used to take pictures of the end faces of the cigarettes in the mold box when the mold box transported on the mold box conveying chain plate is transported to the lower smoke channel of the cigarette bin, so as to obtain the end face image of the cigarette.

[0013] As described above, the cigarette end face defect detection system, wherein preferably, the cigarette end face defect detection system also includes a light source connected to the industrial computer for providing lighting for the industrial camera; the light source is arranged at the front end of the industrial camera.

[0014] The cigarette end face defect detection system as described above, wherein, preferably, the cigarette end face defect detection system also includes a functional personnel mobile terminal, a functional personnel PC terminal and an operator slave computer connected to the early warning platform, and the early warning platform is also used to send the early warning information to at least one of the functional personnel mobile terminal, the functional personnel PC terminal and the operator slave computer.

[0015] In the cigarette end face defect detection system as described above, preferably, the database is connected to the input end of the early warning platform, and the functional personnel mobile terminal, the functional personnel PC terminal and the operator's lower computer are connected to the output end of the early warning platform.

[0016] In the above-mentioned cigarette end face defect detection system, preferably, the image processing and operation workstation computer includes a brightness coefficient calculation module, a brightness anomaly judgment module, an image preprocessing module, a cigarette end face image segmentation module, and a cigarette end face image defect recognition module, wherein:

[0017] The brightness coefficient calculation module is used to calculate the brightness coefficient of the cigarette end face image;

[0018] The brightness abnormality judgment module is used to judge whether the brightness of the cigarette end face image is normal according to the brightness coefficient;

[0019] The image preprocessing module is used to preprocess the cigarette end face image with normal brightness;

[0020] The cigarette end face image segmentation module is used to locate the center of each cigarette end face, fit the radius of the circle, and use the fitted circle corresponding to each cigarette to intercept the end face image of each cigarette;

[0021] The cigarette end face image defect recognition module is used to classify and judge the end face image of each cut cigarette using a trained deep neural network model to determine whether the end face of each cigarette has defects.

[0022] In the cigarette end face defect detection system as described above, preferably, the brightness coefficient calculation module includes a grayscale image processing unit and a brightness coefficient calculation unit, wherein:

[0023] The grayscale image processing unit is used to perform grayscale image processing on the cigarette end face image, obtain the shape and line width of the cigarette end face image, and calculate the mean and deviation of the grayscale image pixels;

[0024] The brightness coefficient calculation unit is used to calculate the brightness coefficient of the cigarette end face image according to the mean and deviation of the grayscale image pixels using the following formula:

[0025] (1)

[0026] (2)

[0027] MA = ∑ | ( x i − 128 ) − DA | * Hist [ i ] ∑ Hist [ i ] (3)

[0028] in, represents the image brightness coefficient, Represents the mean of (image grayscale value -128), represents the mean deviation, represents the grayscale value of the image, Indicates the number of image pixels, Hist [ i ] Indicates the number of pixels at each gray level in the grayscale histogram.

[0029] In the above-mentioned cigarette end face defect detection system, preferably, the brightness abnormality judgment module is specifically used to: if the image brightness coefficient is less than or equal to a preset brightness coefficient threshold, judge that the brightness of the cigarette end face image is normal and within a reasonable range; if the image brightness coefficient is greater than the preset brightness coefficient threshold, judge that the brightness of the cigarette end face image is abnormal.

[0030] The image preprocessing module is specifically used to perform mean shift filtering and noise reduction processing on the cigarette end face image through a filter.

[0031] In the above-mentioned cigarette end face defect detection system, preferably, the cigarette end face image cutting module includes a cigarette end face image cutting unit and a cigarette end face image defect recognition unit, wherein:

[0032] The cigarette end face image segmentation unit is used to locate the center of each cigarette end face using the Hough circle detection algorithm and fit the radius of the circle using the following formula:

[0033] (4)

[0034] (5)

[0035] (6)

[0036] Where (x, y) represents the coordinates of the cigarette end point, (a, b) represents the coordinates of the center of the fitted circle, and r represents the radius of the fitted circle.

[0037] The cigarette end face image segmentation unit is further configured to intercept a square image centered on the circle, with equal length and width, both greater than the diameter, based on the circle located using the Hough circle method and using the known coordinates of the center and radius of the fitted circle for each cigarette, to intercept the end face image of each cigarette. The boundary position of the interception range of a single cigarette end face image is calculated using the following formula:

[0038] (7)

[0039] (8)

[0040] (9)

[0041] (10)

[0042] in, Indicates the The horizontal coordinate of the center of the cigarette circle, Indicates the The vertical coordinate of the center of the cigarette circle, Indicates the side length of the rectangle to be intercepted, obtained by calculation 、 、 、 Determine the cigarette and capture the image of a single cigarette.

[0043] The cigarette end face defect detection system as described above, wherein preferably, the cigarette end face image defect recognition module is specifically used to: send the segmented image to a trained and tuned deep neural network model for classification to determine whether it is defective; in the deep neural network model, the segmented image is sent to the neural network, passes through a convolution block, and then enters three dense blocks for feature extraction, wherein each dense block contains two layers of convolution blocks, each convolution block is composed of a convolution layer, a Batch Normalization layer and a ReLU function activation layer, and then passes through an average pooling layer, and finally a fully connected layer maps the result into two values ​​to indicate whether the cigarette in the image is defective.

[0044] The present invention provides a cigarette end face defect detection system, which collects cigarette end face images through an industrial camera, analyzes the cigarette end faces through image processing and an operation workstation computer, so as to realize full-category quality defect detection of cigarette end faces and realize online detection of cigarette end face quality during the cigarette packaging process; it can automatically identify cigarette end face quality defects, push early warning information to functional personnel and operators through an early warning platform, and locate the mold box where the defect occurs; it effectively improves the quality of cigarette products, improves the automated production level of cigarette rolling and splicing equipment, improves work efficiency, reduces costs, and reduces the intensity of manual random inspections. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be further described below with reference to the accompanying drawings, in which:

[0046] Figure 1 This is a structural block diagram of an embodiment of a cigarette end face defect detection system provided by the present invention;

[0047] Figure 2 A schematic top view of the installation positions of the industrial camera and light source of an embodiment of the cigarette end face defect detection system provided by the present invention;

[0048] Figure 3 A schematic diagram of the working process of an embodiment of a cigarette end face defect detection system provided by the present invention;

[0049] Figure 4 A schematic diagram of the Hough circle positioning of a cigarette end face according to an embodiment of the cigarette end face defect detection system provided by the present invention;

[0050] Figure 5 This is a schematic diagram of the cigarette end face cutting results of an embodiment of the cigarette end face defect detection system provided by the present invention.

[0051] Explanation of the accompanying numbers: 1-industrial camera, 2-light source, 3-industrial computer, 4-image processing and operation workstation computer, 5-database, 6-early warning platform, 7-functional personnel mobile terminal, 8-functional personnel PC terminal, 9-operator lower computer, 10-smoke unloading channel of cigarette warehouse, 11-mold box, 12-mold box conveying chain plate. DETAILED DESCRIPTION

[0052] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. The description of the exemplary embodiments is merely illustrative and is in no way intended to limit the present disclosure, its application, or use. The present disclosure can be implemented in many different forms and is not limited to the embodiments described herein. These embodiments are provided to make the present disclosure thorough and complete and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that unless otherwise specifically stated, the relative arrangement of parts and steps, the composition of materials, numerical expressions, and numerical values ​​set forth in these embodiments should be interpreted as being merely exemplary and not as limiting.

[0053] The terms "first," "second," and similar terms used in this disclosure do not indicate any order, quantity, or importance, but are simply used to distinguish different parts. Terms such as "include" or "comprising" mean that the elements preceding the term include the elements listed after the term, and do not exclude the possibility of also including other elements. Terms such as "upper," "lower," and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0054] In the present disclosure, when a specific component is described as being located between a first component and a second component, there may or may not be an intervening component between the specific component and the first component or the second component. When a specific component is described as being connected to another component, the specific component may be directly connected to the other component without an intervening component, or may not be directly connected to the other component but have an intervening component.

[0055] All terms (including technical or scientific terms) used in this disclosure have the same meaning as those understood by one of ordinary skill in the art to which this disclosure belongs, unless otherwise specifically defined. It should also be understood that terms defined in, for example, general dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an idealized or highly formal sense, unless explicitly defined herein.

[0056] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0057] Current cigarette end-face defect detection systems compare the tobacco pixel intervals and grayscale intervals corresponding to hollow and loose cigarettes to determine whether the cigarette end face contains these quality defects. This system cannot fully detect all types of cigarette end faces. Current technical solutions for measuring cigarette end-face defects have the following shortcomings: They can only detect hollow and loose end defects, failing to detect all types of quality defects; they can only eliminate hollow and loose end defects, failing to provide full-platform early warnings for personnel on mobile devices, PCs, and operator computers; and they cannot identify the specific quality defect causing the model box, thereby providing guidance for maintenance personnel during troubleshooting.

[0058] like Figure 1 As shown, an embodiment of the present invention provides a cigarette end face defect detection system, which includes: an industrial computer 1, an input end of the industrial computer 1 is connected to an industrial camera 1, an output end of the industrial computer 1 is connected to an image processing and operation workstation computer 4, the image processing and operation workstation computer 4 is connected to a database 5, and the database 5 is connected to an early warning platform 6, wherein:

[0059] The industrial camera 1 is used to capture images of cigarette end faces;

[0060] The industrial computer 3 is used to store the cigarette end face image and transmit the cigarette end face image to the image processing and operation workstation computer 4;

[0061] The image processing and operation workstation computer 4 is used to identify cigarette end face defects based on the cigarette end face image, and obtain the cigarette end face defect type result and the defective cigarette location;

[0062] The database 5 is used to store the cigarette end surface defect type results and the defective cigarette positions, and send the cigarette end surface defect type results and the defective cigarette positions to the early warning platform 6;

[0063] The early warning platform 6 is used to generate early warning information based on the cigarette end face defect type result and the defective cigarette position.

[0064] Among them, such as Figure 2 As shown, the industrial camera 1 is positioned on the side of the mold box conveyor chain 12 on the cigarette storage downflow channel 10. It is used to photograph the end faces of cigarettes within the mold box 11 as the mold box 11, transported along the mold box conveyor chain 12, reaches the cigarette storage downflow channel 10, thereby obtaining images of the cigarette end faces. In a specific implementation, an image transmission system can be used to transmit the captured cigarette end face images to the image processing and operation workstation computer 4 for image processing.

[0065] During operation, the end face image of the cigarette is collected by the industrial camera 1, the end face image of the cigarette is obtained by the industrial computer 3, and then the end face defect of the cigarette is identified by the image processing and operation workstation computer to obtain the cigarette end face defect type result and the defective cigarette position. When the image processing and operation workstation computer 4 detects a defective image, the defect result will be uploaded to the database 5, and the defect type and position will be compared and analyzed in the database 5, and the early warning platform 6 will be reported based on the defect statistical results.

[0066] Furthermore, the cigarette end face defect detection system further includes a light source 2 connected to the industrial computer 3 for providing lighting for the industrial camera 1. Figure 2 As shown, the light source 2 is arranged at the front end of the industrial camera 1 .

[0067] Further, if Figure 1 As shown, the cigarette end surface defect detection system also includes a functional personnel mobile terminal 7, a functional personnel PC terminal 8, and an operator slave computer 9 connected to the early warning platform 6. The early warning platform 6 is further configured to send the early warning information to at least one of the functional personnel mobile terminal 7, the functional personnel PC terminal 8, and the operator slave computer 9. Specifically, the database 5 is connected to the input terminal of the early warning platform 6, and the functional personnel mobile terminal 7, the functional personnel PC terminal 8, and the operator slave computer 9 are connected to the output terminal of the early warning platform 6. In some embodiments of the present invention, after the database 5 determines the same defect type and defect location, if defect images of the same defect type and location appear continuously within a recent period (e.g., one hour), the relevant images are extracted for comparison. After confirming the similarity, the database 5 pushes the results to the early warning platform 6, which then pushes the early warning information to the functional personnel mobile terminal 7, the functional personnel PC terminal 8, and the operator slave computer 9. In addition, based on statistical information, it can determine which mold box is responsible for the quality defect and generate a prompt message for maintenance personnel. Moreover, all data are stored in the database to facilitate subsequent retrieval of early warning information for analysis and assessment of cigarette quality and safety.

[0068] Furthermore, the image processing and operation workstation computer 4 includes a brightness coefficient calculation module, a brightness anomaly judgment module, an image preprocessing module, a cigarette end face image segmentation module, and a cigarette end face image defect recognition module, wherein:

[0069] The brightness coefficient calculation module is used to calculate the brightness coefficient of the cigarette end face image;

[0070] The brightness abnormality judgment module is used to judge whether the brightness of the cigarette end face image is normal according to the brightness coefficient;

[0071] The image preprocessing module is used to preprocess the cigarette end face image with normal brightness;

[0072] The cigarette end face image segmentation module is used to locate the center of each cigarette end face, fit the radius of the circle, and use the fitted circle corresponding to each cigarette to intercept the end face image of each cigarette;

[0073] The cigarette end face image defect recognition module is used to classify and judge the end face image of each cut cigarette using a trained deep neural network model to determine whether the end face of each cigarette has defects.

[0074] Specifically, the brightness coefficient calculation module includes a grayscale image processing unit and a brightness coefficient calculation unit, wherein:

[0075] The grayscale image processing unit is used to perform grayscale image processing on the cigarette end face image, obtain the shape and line width of the cigarette end face image, and calculate the mean and deviation of the grayscale image pixels;

[0076] The brightness coefficient calculation unit is used to calculate the brightness coefficient of the cigarette end face image according to the mean and deviation of the grayscale image pixels using the following formula:

[0077] (1)

[0078] (2)

[0079] MA = ∑ | ( x i − 128 ) − DA | * Hist [ i ] ∑ Hist [ i ] (3)

[0080] in, represents the image brightness coefficient, Represents the mean of (image grayscale value -128), represents the mean deviation, represents the grayscale value of the image, Indicates the number of image pixels, Hist [ i ] Indicates the number of pixels at each gray level in the grayscale histogram.

[0081] Furthermore, the brightness abnormality judgment module is specifically configured to: if the image brightness coefficient is less than or equal to a preset brightness coefficient threshold, determine that the brightness of the cigarette end face image is normal and within a reasonable range; if the image brightness coefficient is greater than the preset brightness coefficient threshold, determine that the brightness of the cigarette end face image is abnormal. Abnormal brightness indicates that the packet lacks inner frame paper, resulting in an abnormal cigarette packet, which is rejected. In some embodiments of the present invention, the preset brightness coefficient threshold is 5. When k < 5, the image is considered normal and the brightness is within a reasonable range. When k > 5, the image is considered abnormal. It should be noted that the present invention does not specifically limit the preset brightness coefficient threshold.

[0082] Furthermore, the image preprocessing module is specifically configured to perform mean shift filtering and noise reduction processing on the cigarette end face image through a filter. Mean filtering can effectively eliminate high-frequency noise in the image and achieve image smoothing and blurring functions.

[0083] Furthermore, the cigarette end face image cutting module includes a cigarette end face image cutting unit and a cigarette end face image defect recognition unit, wherein:

[0084] like Figure 4 As shown, the cigarette end face image segmentation unit is used to locate the center of each cigarette end face using the Hough circle detection algorithm and fit the radius of the circle using the following formula:

[0085] (4)

[0086] (5)

[0087] (6)

[0088] Where (x, y) represents the coordinates of the cigarette end point, (a, b) represents the coordinates of the center of the fitted circle, and r represents the radius of the fitted circle.

[0089] If the center and radius of the circle are known, a circular diagram of each cigarette can be drawn, and the position of each cigarette can be located through the center of the circular diagram.

[0090] The cigarette end face image segmentation unit is further configured to intercept a square image with the center of the circle as the center, with equal length and width, both greater than the diameter, based on the circle located by the Hough circle method and using the known coordinates of the center and radius of the fitted circle of each cigarette, and to intercept the end face image of each cigarette (the segmentation result is shown in FIG. Figure 5 As shown), the boundary position of the interception range of a single cigarette image of the cigarette end face is calculated by the following formula:

[0091] (7)

[0092] (8)

[0093] (9)

[0094] (10)

[0095] in, Indicates the The horizontal coordinate of the center of the cigarette circle, Indicates the The vertical coordinate of the center of the cigarette circle, Indicates the side length of the rectangle to be intercepted, obtained by calculation 、 、 、 Determine the cigarette and capture the image of a single cigarette.

[0096] By segmenting the cigarette end face image, the entire image of the cigarette end face can be segmented and the end face image of each cigarette can be extracted, which is convenient for feature extraction of each cigarette image. This can effectively improve the accuracy of cigarette end face recognition. By sorting the positions of cigarettes, it is helpful to better determine the location of defective cigarettes.

[0097] Furthermore, the cigarette end face image defect recognition module is specifically used to: send the segmented image to a trained and tuned deep neural network model for classification to determine whether it has defects. In the deep neural network model, the segmented image is sent to the neural network, passes through a convolution block, and then enters three dense blocks for feature extraction, where each dense block contains two layers of convolution blocks. Each convolution block is composed of a convolution layer, a BatchNormalization layer and a ReLU function activation layer. After passing through an average pooling layer, a fully connected layer finally maps the result into two values ​​to indicate whether the cigarette in the image has defects.

[0098] Accordingly, the cigarette end face defect detection process is as follows: Figure 3 As shown, the end face image of the cigarette is obtained by the industrial computer of the existing cigarette package visual inspection system, and then the image brightness coefficient is calculated. The image brightness is judged to be normal based on the image brightness coefficient, and the cigarette end face images with abnormal brightness are eliminated. If the image brightness is normal, the image is further deeply analyzed. Specifically, after image preprocessing, the cigarette end face image with normal brightness is segmented, and the cigarette end face image defect identification is performed based on the image segmentation result. The cigarettes without defects are normal, and the defect results of the defective ones are uploaded to the database, and the defect type and position are compared and analyzed in the database, and the early warning platform is reported based on the defect statistical results.

[0099] The cigarette end face defect detection system provided by the embodiment of the present invention collects cigarette end face images through an industrial camera, and analyzes the cigarette end faces through image processing and operation workstation computers to realize full-category quality defect detection of cigarette end faces and realize online detection of cigarette end face quality during the packaging process; it can automatically identify cigarette end face quality defects, push early warning information to functional personnel and operators through an early warning platform, and locate the mold box where the defect occurs; it effectively improves the quality of cigarette products, improves the automation production level of cigarette rolling equipment, improves work efficiency, reduces costs, and reduces the intensity of manual random inspections.

[0100] Thus far, various embodiments of the present disclosure have been described in detail. To avoid obscuring the concept of the present disclosure, some details known in the art have not been described. Based on the above description, those skilled in the art can fully understand how to implement the technical solutions disclosed herein.

[0101] Although some specific embodiments of the present disclosure have been described in detail through examples, those skilled in the art will understand that the above examples are for illustration only and are not intended to limit the scope of the present disclosure. Those skilled in the art will understand that the above embodiments may be modified or some technical features may be replaced with equivalents without departing from the scope and spirit of the present disclosure. The scope of the present disclosure is defined by the appended claims.

Claims

1. A cigarette end surface defect detection system, characterized in that: include: An industrial computer, the input end of which is connected to an industrial camera, the output end of which is connected to an image processing and operation workstation computer, the image processing and operation workstation computer is connected to a database, and the database is connected to an early warning platform, wherein: The industrial camera is used to collect cigarette end face images; The industrial computer is used to store the cigarette end face image and transmit the cigarette end face image to the image processing and operation workstation computer; The image processing and operation workstation computer is used to identify cigarette end face defects based on the cigarette end face image, and obtain the cigarette end face defect type result and defective cigarette position; The database is used to store the cigarette end surface defect type results and the defective cigarette position, and send the cigarette end surface defect type results and the defective cigarette position to the early warning platform; The early warning platform is used to generate early warning information according to the cigarette end surface defect type result and the defective cigarette position.

2. The cigarette end surface defect detection system according to claim 1, characterized in that: The industrial camera is arranged on one side of the lower smoke channel of the cigarette bin of the mold box conveying chain plate, and is used to take pictures of the end faces of the cigarettes in the mold box when the mold box conveyed on the mold box conveying chain plate is transferred to the lower smoke channel of the cigarette bin, so as to obtain the end face image of the cigarette.

3. The cigarette end surface defect detection system according to claim 1, characterized in that: The cigarette end surface defect detection system also includes a light source connected to the industrial computer and used to provide lighting for the industrial camera; the light source is arranged at the front end of the industrial camera.

4. The cigarette end surface defect detection system according to claim 1, characterized in that: The cigarette end face defect detection system also includes a functional personnel mobile terminal, a functional personnel PC terminal and an operator slave computer connected to the early warning platform. The early warning platform is also used to send the early warning information to at least one of the functional personnel mobile terminal, the functional personnel PC terminal and the operator slave computer.

5. The cigarette end surface defect detection system according to claim 4, characterized in that: The database is connected to the input end of the early warning platform, and the functional personnel mobile terminal, the functional personnel PC terminal and the operator's lower computer are connected to the output end of the early warning platform.

6. The cigarette end surface defect detection system according to claim 1, characterized in that: The image processing and operation workstation computer includes a brightness coefficient calculation module, a brightness abnormality judgment module, an image preprocessing module, a cigarette end face image cutting module and a cigarette end face image defect recognition module, wherein: The brightness coefficient calculation module is used to calculate the brightness coefficient of the cigarette end face image; The brightness abnormality judgment module is used to judge whether the brightness of the cigarette end face image is normal according to the brightness coefficient; The image preprocessing module is used to preprocess the cigarette end face image with normal brightness; The cigarette end face image segmentation module is used to locate the center of each cigarette end face, fit the radius of the circle, and intercept the end face image of each cigarette using the fitting circle corresponding to each cigarette; The cigarette end face image defect recognition module is used to classify and judge the end face image of each cut cigarette using a trained deep neural network model to determine whether the end face of each cigarette has defects.

7. The cigarette end surface defect detection system according to claim 6, characterized in that: The brightness coefficient calculation module includes a grayscale image processing unit and a brightness coefficient calculation unit, wherein: The grayscale image processing unit is used to perform grayscale image processing on the cigarette end face image, obtain the shape and line width of the cigarette end face image, and calculate the mean and deviation of the grayscale image pixel points; The brightness coefficient calculation unit is used to calculate the brightness coefficient of the cigarette end face image according to the mean and deviation of the grayscale image pixels by the following formula: (1) (2) (3) in, represents the image brightness coefficient, Represents the mean of (image gray value - 128), represents the mean deviation, represents the gray value of the image, Indicates the number of image pixels. Represents the number of pixels at each gray level in the grayscale histogram.

8. The cigarette end surface defect detection system according to claim 7, characterized in that: The brightness abnormality judgment module is specifically used to: if the image brightness coefficient is less than or equal to the preset brightness coefficient threshold, judge that the brightness of the cigarette end face image is normal and within a reasonable range; if the image brightness coefficient is greater than the preset brightness coefficient threshold, judge that the brightness of the cigarette end face image is abnormal. The image preprocessing module is specifically used to: perform mean shift filtering and noise reduction processing on the cigarette end face image through a filter.

9. The cigarette end surface defect detection system according to claim 7, characterized in that: The cigarette end face image cutting module comprises a cigarette end face image cutting unit and a cigarette end face image defect recognition unit, wherein: The cigarette end face image segmentation unit is used to locate the center of each cigarette end face through the Hough circle detection algorithm and fit the radius of the circle through the following formula: (4) (5) (6) Among them, (x, y) represents the coordinates of the cigarette end point, (a, b) represents the coordinates of the center of the fitting circle, and r represents the radius of the fitting circle; The cigarette end face image segmentation unit is further used to intercept a square image with the center of the circle as the center, the length and width being equal and both greater than the diameter, based on the circle positioned by the Hough circle method and using the known coordinates of the center of the fitted circle of each cigarette, and intercept the end face image of each cigarette, wherein the boundary position of the interception range of a single cigarette end face image is calculated by the following formula: (7) (8) (9) (10) in, Indicates The horizontal coordinate of the center of the cigarette circle, Indicates The vertical coordinate of the center of the cigarette circle, Indicates the side length of the rectangle to be intercepted, obtained by calculation , , , Determine the cigarette and capture the image of a single cigarette.

10. The cigarette end surface defect detection system according to claim 7, characterized in that: The cigarette end face image defect recognition module is specifically used to: send the segmented image to a trained and tuned deep neural network model for classification to determine whether it is defective. In the deep neural network model, the segmented image is sent to the neural network, passes through a convolution block, and then enters three dense blocks for feature extraction, each of which contains two layers of convolution blocks. Each convolution block is composed of a convolution layer, a Batch Normalization layer and a ReLU function activation layer. After passing through an average pooling layer, a fully connected layer finally maps the result into two values ​​to indicate whether the cigarette in the image is defective.