A method, device, equipment and medium for optimizing process parameters in cigarette package production

Abstract

The present application discloses a method, device, equipment and medium for optimizing process parameters in cigarette pack production, and relates to the field of data processing technology, and determines key process parameters that affect appearance quality; obtains real-time process parameter data including key process parameters, and associates them with identification codes of cigarette packs to obtain a process parameter set; collects appearance detection data of cigarette packs to determine whether the appearance of the cigarette packs is abnormal; if the appearance is abnormal, screens the corresponding identification code to obtain abnormal process parameter data; analyzes the abnormal process parameter data to obtain the cause of the abnormal appearance, and traces historical cigarette packs according to the abnormal process parameter data to obtain abnormal related parameters corresponding to the historical cigarette packs; combines all abnormal related parameters with the abnormal process parameter data to obtain combined abnormal parameters, and optimizes the combined abnormal parameters according to the cause of the abnormal appearance to obtain optimized parameters, thereby achieving accurate association and tracing of cigarette packs, optimizing the production process of cigarette packs, and improving production efficiency and product quality.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, device, equipment and medium for optimizing process parameters in cigarette package production. Background Art

[0002] As a consumer product, the production process and packaging quality of cigarettes have a significant impact on consumers and corporate brand image. In the existing technology, there are deficiencies in the monitoring and tracing of key process parameters in the cigarette production process, especially in the product appearance quality inspection and tracing, which lacks efficient and accurate means. For example, in the gilding beauty process, when the appearance inspection of the cigarette package fails, it is impossible to locate the problem station, resulting in quality risks. In addition, the cigarette package detector in the existing technology can only perform detection, which limits the optimization and quality control of the production process.

[0003] As can be seen from the above, how to achieve accurate association and traceability of cigarette packages, optimize the production process of cigarette packages, and improve production efficiency and product quality are issues to be solved in this field. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a method, device, equipment and medium for optimizing process parameters in cigarette pack production, which can achieve accurate association and traceability of cigarette packs, optimize the production process of cigarette packs, and improve production efficiency and product quality. The specific scheme is as follows:

[0005] In a first aspect, the present application discloses a method for optimizing process parameters in cigarette pack production, comprising:

[0006] Determine the key process parameters that affect the appearance quality of cigarette packages during production;

[0007] Acquire real-time process parameter data including the key process parameters of cigarette pack production, and associate the real-time process parameter data with the identification code of the cigarette pack to obtain a process parameter set;

[0008] collecting appearance detection data of the cigarette package, and judging whether the appearance of the cigarette package is abnormal according to the appearance detection data;

[0009] If the appearance of the cigarette package is abnormal, selecting an identification code corresponding to the cigarette package with abnormal appearance from the process parameter set, and acquiring corresponding abnormal process parameter data;

[0010] Analyze the abnormal process parameter data to obtain the cause of the abnormal appearance, and trace back the historical cigarette packs according to the abnormal process parameter data to obtain the abnormal related parameters corresponding to the historical cigarette packs;

[0011] All the abnormality-related parameters are combined with the abnormal process parameter data to obtain combined abnormal parameters, and the combined abnormal parameters are optimized according to the cause of the appearance abnormality to obtain optimized parameters.

[0012] Optionally, the key process parameters affecting the appearance quality of cigarette packages during production are determined, including:

[0013] The data of cigarette packs at each station on the production line are saved in a preset cigarette pack database; the data includes temperature, speed, pressure, and humidity;

[0014] Analytical and statistical methods are used to perform correlation analysis on appearance quality and identify key parameters of the data in the cigarette package database, so as to obtain key process parameters that affect the appearance quality of the cigarette packages during the production process.

[0015] Optionally, the acquiring of real-time process parameter data including the key process parameters of cigarette pack production, and associating the real-time process parameter data with the identification code of the cigarette pack, includes:

[0016] The pre-installed high-speed data acquisition device and visual inspection system are used to respectively obtain the real-time process parameter data of cigarette pack production including the key process parameters and the identification code of the cigarette pack, and the real-time process parameter data is associated with the identification code of the cigarette pack.

[0017] Optionally, collecting appearance detection data of the cigarette package and judging whether the appearance of the cigarette package is abnormal according to the appearance detection data includes:

[0018] Use the pre-installed visual inspection system to collect appearance images of cigarette packages from different directions;

[0019] The cigarette package appearance images in different directions are merged to obtain a merged image, and whether the appearance of the cigarette package is abnormal is judged according to the synthesized image; the appearance detection data includes the synthesized image.

[0020] Optionally, merging the cigarette package appearance images in different directions to obtain a combined image, and judging whether the appearance of the cigarette package is abnormal according to the combined image includes:

[0021] Using an image processing algorithm to preprocess, match and align feature points, transform and splice the cigarette package appearance images in different directions to obtain a merged image;

[0022] Inputting the synthesized image into a preset deep learning model for image analysis, and extracting features using a convolutional neural network to obtain a feature vector;

[0023] The classifier is used to perform defect detection and classification on the feature vector to complete the abnormal appearance judgment of the cigarette package.

[0024] Optionally, analyzing the abnormal process parameter data to obtain the cause of the abnormal appearance includes:

[0025] Performing statistical analysis on the abnormal process parameter data to obtain analysis results;

[0026] A prediction model is constructed, and the analysis results are input into the prediction model so that the prediction model can identify and predict the cause of the appearance abnormality based on the analysis results, and output the predicted cause of the appearance abnormality.

[0027] Optionally, the performing statistical analysis on the abnormal process parameter data includes:

[0028] The abnormal process parameter data are cleaned and formatted to obtain the pre-processed abnormal process parameter data, and then correlation analysis and regression analysis are performed on the pre-processed abnormal process parameter data.

[0029] In a second aspect, the present application discloses a device for optimizing process parameters in cigarette pack production, comprising:

[0030] Key process parameter determination module, used to determine the key process parameters that affect the appearance quality of cigarette packages during the production process;

[0031] An association module, used for acquiring real-time process parameter data including the key process parameters of cigarette pack production, and associating the real-time process parameter data with the identification code of the cigarette pack to obtain a process parameter set;

[0032] a judgment module, used for collecting appearance detection data of the cigarette package, and judging whether the appearance of the cigarette package is abnormal according to the appearance detection data;

[0033] an abnormal process parameter data acquisition module, for selecting an identification code corresponding to the cigarette package with abnormal appearance from the process parameter set if the cigarette package has abnormal appearance, and acquiring corresponding abnormal process parameter data;

[0034] A tracing module, used to analyze the abnormal process parameter data to obtain the cause of the abnormal appearance, and trace the historical cigarette packs according to the abnormal process parameter data to obtain the abnormal related parameters corresponding to the historical cigarette packs;

[0035] The optimization module is used to combine all the abnormality-related parameters with the abnormal process parameter data to obtain combined abnormal parameters, and optimize the combined abnormal parameters according to the cause of the appearance abnormality to obtain optimized parameters.

[0036] In a third aspect, the present application discloses an electronic device, comprising:

[0037] Memory, used to store computer programs;

[0038] A processor is used to execute the computer program to implement the aforementioned process parameter optimization method in cigarette package production.

[0039] In a fourth aspect, the present application discloses a computer storage medium for storing a computer program; wherein, when the computer program is executed by a processor, the steps of the process parameter optimization method in cigarette pack production disclosed above are implemented.

[0040] It can be seen that the present application provides a method for optimizing process parameters in cigarette pack production, including determining key process parameters that affect the appearance quality of cigarette packs during the production process; obtaining real-time process parameter data of cigarette pack production including the key process parameters, and associating the real-time process parameter data with the identification code of the cigarette pack to obtain a process parameter set; collecting appearance detection data of the cigarette pack, and judging whether the appearance of the cigarette pack is abnormal based on the appearance detection data; if the appearance of the cigarette pack is abnormal, filtering out the identification code corresponding to the cigarette pack with abnormal appearance from the process parameter set, and obtaining the corresponding abnormal process parameter data; analyzing the abnormal process parameter data to obtain the cause of the abnormal appearance, and tracing back the historical cigarette packs based on the abnormal process parameter data to obtain the abnormality-related parameters corresponding to the historical cigarette packs; combining all the abnormality-related parameters with the abnormal process parameter data to obtain the combined abnormal parameters, and optimizing the combined abnormal parameters based on the cause of the abnormal appearance to obtain the optimized parameters. The present application comprehensively records and monitors the key process parameters that affect the appearance quality in the cigarette production process, associates the real-time process parameter data of cigarette pack production including the key process parameters with the identification code of the cigarette pack, collects the appearance detection data of the cigarette pack, and determines whether the appearance of the cigarette pack is abnormal. If the appearance is abnormal, the identification code corresponding to the cigarette pack with abnormal appearance is screened out from the process parameter set, and the corresponding abnormal process parameter data is obtained, so as to realize the accurate tracing and positioning of the cigarette pack with abnormal appearance quality or abnormal process parameters. The tracing process is fast and accurate, which can greatly reduce the workload of manual investigation of production personnel and improve production efficiency. The abnormal process parameter data is analyzed to obtain the cause of the abnormal appearance, and the historical cigarette packs are traced to obtain the abnormal related parameters corresponding to the historical cigarette packs. All the abnormal related parameters are combined with the abnormal process parameter data, and the combined abnormal parameters are optimized according to the cause of the abnormal appearance. Based on the optimal process parameter combination, the producer is guided to optimize the process parameters, improve the quality stability and consistency of cigarettes, and improve production efficiency and product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0042] Figure 1 A flow chart of a method for optimizing process parameters in cigarette pack production disclosed in the present application;

[0043] Figure 2 A schematic diagram of the structure of a process parameter optimization device in cigarette pack production disclosed in the present application;

[0044] Figure 3 A structural diagram of an electronic device provided for this application. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] As a consumer product, the production process and packaging quality of cigarettes have a significant impact on consumers and corporate brand image. In the prior art, there are deficiencies in the monitoring and tracing of key process parameters in the cigarette production process, especially in the detection and tracing of product appearance quality, which lacks efficient and accurate means. For example, in the gilding beauty process, when the appearance inspection of the cigarette package fails, the problem station cannot be located, resulting in quality risks. In addition, the cigarette package detector in the prior art can only perform detection, which limits the optimization and quality control of the production process. As can be seen from the above, how to achieve accurate association and tracing of cigarette packages, optimize the production process of cigarette packages, and improve production efficiency and product quality are problems to be solved in this field.

[0047] See also Figure 1 As shown, the embodiment of the present invention discloses a method for optimizing process parameters in cigarette pack production, which may specifically include:

[0048] Step S11: Determine the key process parameters that affect the appearance quality of the cigarette pack during the production process.

[0049] In this embodiment, the data of cigarette packs at each workstation on the production line are saved in a preset cigarette pack database; the data include temperature, speed, pressure, and humidity; analytical and statistical methods are used to perform correlation analysis on the appearance quality and identify key parameters of the data in the cigarette pack database to obtain the key process parameters that affect the appearance quality of the cigarette packs during the production process.

[0050] Specifically, the existing data in the cigarette package database is used to conduct a correlation analysis with the appearance quality of the cigarette packages, and parameters with higher correlation are selected as candidate parameters. The candidate parameters are combined with the influencing strength to determine the key process parameters.

[0051] In this embodiment, the identification and analysis of key process parameters is the first step, which involves an in-depth analysis of the correlation between the process parameters of the cigarette package in the entire production process and the product quality. It mainly involves careful observation and research of each link in the cigarette package production process to identify the key process parameters that have a significant impact on the product appearance quality. These parameters may include but are not limited to temperature, speed, pressure, humidity, etc. The detailed steps are as follows:

[0052] First, collect data on cigarette packs at each station on the production line, including but not limited to process parameters such as temperature, speed, pressure, and humidity;

[0053] Correlation analysis: Then, use statistical methods to analyze the correlation between these process parameters and the appearance quality of cigarette packages, and find out which parameter changes have a significant impact on the appearance quality of cigarette packages;

[0054] Key parameter identification: Through the above analysis, a set of key process parameters can be identified, which are the key factors affecting the appearance quality of cigarette packages;

[0055] Finally, all the key process parameters are obtained. This application will monitor each workstation individually, record the real-time data of the process parameters, and form the core key parameters, especially the process parameters that have a greater impact on the gilding beauty effect.

[0056] Step S12: acquiring real-time process parameter data including the key process parameters of cigarette pack production, and associating the real-time process parameter data with the identification code of the cigarette pack to obtain a process parameter set.

[0057] In this embodiment, a pre-installed high-speed data acquisition device and a visual inspection system are used to respectively obtain real-time process parameter data of cigarette pack production including the key process parameters and the identification code of the cigarette pack, and the real-time process parameter data is associated with the identification code of the cigarette pack.

[0058] Specifically, the setting of the visual inspection system and its association with the identification code are specifically located at each workstation in the beauty production line. A visual acquisition mechanism connected to the production control system is assembled, and after the cigarette pack enters the beauty production line, an identification code associated with the visual acquisition mechanism is obtained.

[0059] This application uses the following technical means to ensure that the key process parameters of each cigarette pack at each workstation during the production process can be captured in real time and accurately associated with the unique identifier of the cigarette pack - the QR code information:

[0060] Real-time data capture: A high-speed data acquisition device is installed at each workstation, which is directly connected to the controller of the production line. The controller captures key process parameters in real time according to the preset sampling frequency, as shown below:

[0061] ;

[0062] In this embodiment, Represents the process parameter data captured at a specific time point. When a cigarette pack passes through a certain workstation, the visual inspection system of the workstation will read the QR code information on the cigarette pack. The QR code information is associated with the process parameter data captured in real time through the data processing system. In actual production, when a cigarette pack enters the production line, the system will assign it a unique QR code, which will accompany the cigarette pack throughout the entire production process. At each key workstation, a dedicated data acquisition device is installed, which can quickly read the QR code of the cigarette pack when it passes through, and capture the key process parameters of the workstation in real time.

[0063] In this embodiment, data collection of cigarette packs at each workstation is completed automatically, reducing human intervention and improving data accuracy and timeliness. Through the database management system, the production record of any cigarette pack can be quickly retrieved, achieving transparency and traceability of the production process.

[0064] Step S13: collecting appearance detection data of the cigarette package, and judging whether the appearance of the cigarette package is abnormal according to the appearance detection data.

[0065] In this embodiment, a pre-installed visual inspection system is used to collect appearance images of cigarette packages in different directions; the appearance images of cigarette packages in different directions are merged to obtain a merged image, and whether the appearance of the cigarette package is abnormal is judged based on the composite image; the appearance inspection data includes the composite image.

[0066] Specifically, the process of image merging and appearance abnormality judgment is as follows: using an image processing algorithm to preprocess, match and align feature points, transform and splice the appearance images of the cigarette package in different directions to obtain a merged image; inputting the composite image into a preset deep learning model for image analysis, and using a convolutional neural network to extract features to obtain a feature vector; using a classifier to perform defect detection and classification on the feature vector to complete the appearance abnormality judgment of the cigarette package.

[0067] A cigarette package appearance detection model is constructed, and the cigarette package appearance detection model is used to complete the judgment of the appearance abnormality of the cigarette package. Specifically, features are extracted from the cigarette package image through a convolutional neural network and input into a classifier. The classifier determines whether there is a defect based on the feature vector and classifies the type of defect.

[0068] Among them, the appearance images of the cigarette package in different directions are collected as a six-sided image of the cigarette package, and a three-dimensional image of the cigarette package is formed through image processing technology, and a cigarette package appearance detection model is constructed to detect the appearance of the cigarette package; the image processing technology is specifically: image preprocessing; obtaining and matching key feature points of each surface; using matching points to calculate the transformation matrix between images and align the images; according to the transformation matrix, the image is geometrically transformed so that the images of the six surfaces are aligned in space, and then the adjacent surfaces are seamlessly spliced ​​using multi-band mixing technology.

[0069] In this embodiment, the appearance inspection data is associated with the QR code, and the appearance inspection data of the cigarette package, including pictures, inspection information, etc., is collected and associated with the QR code information. In this step, in order to ensure that the appearance quality of the cigarette package is effectively monitored and accurate traceability can be achieved, the present invention takes the following steps to collect and associate the appearance inspection data with the QR code information: Appearance inspection data collection: At the inspection station after the cigarette package passes through the gilding beauty device, a high-precision visual inspection system is used to collect images of the six sides of the cigarette package. At the same time, the inspection system will analyze the image data to determine whether the appearance of the cigarette package is qualified.

[0070] Among them, the visual inspection system collects image data according to the following formula:

[0071] ;

[0072] in, The image data represent the six surfaces of the cigarette package respectively. At this time, the appearance of each surface can be inspected separately, or multiple surfaces can be fused together to form an integrated structure for inspection.

[0073] When a cigarette pack stops at the inspection station, the QR code reader will scan the unique QR code on the pack and obtain the QR code information. The collected image data and inspection information will then be associated with the read QR code information for identification; the subsequent data processing system will store the image data, inspection results and QR code information together in the database to ensure that the inspection records of each pack of cigarettes can be traced.

[0074] Assuming that a cigarette pack is inspected at the inspection station, the system will collect images of its six sides and analyze these images to determine the appearance quality. If the appearance of the cigarette pack is qualified, the system will associate these images and the "qualified" inspection results with the QR code information of the cigarette pack and store them. Through the above method, when it is necessary to trace the appearance quality of the cigarette pack, the corresponding appearance inspection data and images can be quickly found through the QR code of the cigarette pack for in-depth analysis. If problems are found in subsequent quality inspections, the specific cigarette pack can be traced back through the QR code, and its historical inspection data and images can be viewed to analyze the cause of the problem. This association method not only improves the efficiency of traceability, but also enhances the accuracy of quality control, and provides real-time feedback for the adjustment of process parameters.

[0075] This embodiment ensures that the appearance quality of the cigarette packs is effectively monitored throughout the entire production process, and that problems can be quickly located and resolved when discovered.

[0076] Step S14: If the appearance of the cigarette package is abnormal, an identification code corresponding to the cigarette package with abnormal appearance is selected from the process parameter set, and corresponding abnormal process parameter data is obtained.

[0077] It can be understood that the association code is used to obtain the process parameter data at the workstation associated with the association code and the detection image in the appearance inspection data; the process parameter data and the detection image are analyzed to determine the cause of the abnormality; based on the cause of the abnormality, the possible abnormal workstation is located and the workstation parameter data at the workstation is analyzed; the workstation parameters at different workstations are compared to obtain the parameters with the highest correlation with the impact on the abnormal cigarette packs.

[0078] Step S15: Analyze the abnormal process parameter data to obtain the cause of the abnormal appearance, and trace back the historical cigarette packs according to the abnormal process parameter data to obtain the abnormality-related parameters corresponding to the historical cigarette packs.

[0079] In this embodiment, the abnormal process parameter data is statistically analyzed to obtain analysis results; a prediction model is constructed, and the analysis results are input into the prediction model so that the prediction model can identify and predict the cause of the abnormal appearance based on the analysis results, output the predicted cause of the abnormal appearance, and trace back the historical cigarette packs according to the abnormal process parameter data to obtain the abnormality-related parameters corresponding to the historical cigarette packs.

[0080] The statistical analysis process is as follows: cleaning and formatting the abnormal process parameter data to obtain the pre-processed abnormal process parameter data, and then performing correlation analysis and regression analysis on the pre-processed abnormal process parameter data.

[0081] In the precise tracing and positioning step, the present invention uses the following method to achieve rapid and accurate tracing and positioning of cigarette packages with abnormal appearance quality. First, when a cigarette package is judged to be unqualified in appearance at the inspection station, the system will immediately start the tracing mechanism. The following are the specific implementation steps:

[0082] Traceability start: When the inspection system finds a cigarette pack with abnormal appearance quality, it will trigger the traceability process. First, the system will extract the QR code information of the cigarette pack;

[0083] Data query: Using the extracted QR code information, the system queries the database for all process parameter data and test images associated with the cigarette package;

[0084] Abnormality analysis: Through in-depth analysis of the queried data, including but not limited to statistical analysis of process parameters and visual recognition of test images, to determine the cause of abnormal appearance quality;

[0085] Based on the analysis results, the system will locate the specific workstation that may cause abnormal appearance quality and retrieve the relevant process parameter records of the workstation for further analysis and adjustment;

[0086] Assuming that a cigarette pack is found to have one side with unqualified appearance during inspection, the system will trace back to all the stations it has passed through in the gilded beauty device through the QR code of the cigarette pack, and analyze the process parameter data of each station. By comparing the process parameters of different stations, such as temperature, pressure, etc., the system can find the parameters with the highest correlation with abnormal appearance quality. Once the parameters of a certain station are found to be abnormal, the system will mark it as a potential problem station for subsequent parameter adjustment and quality re-inspection. Through this method, the present invention can not only quickly trace and locate the specific station through which the cigarette pack with abnormal appearance quality has passed, but also provide data support for the optimization of process parameters, thereby improving the overall quality level of the production line. If the process parameters traced back to a certain station are often associated with abnormal appearance quality, the production team can focus on the inspection and maintenance of the station to prevent similar quality problems from occurring in the future. Such a tracing and positioning mechanism significantly improves the efficiency of troubleshooting problem cigarette packs, reduces quality risks in the production process, and helps to continuously improve the production process.

[0087] Step S16: combining all the abnormality-related parameters with the abnormal process parameter data to obtain combined abnormal parameters, and optimizing the combined abnormal parameters according to the cause of the appearance abnormality to obtain optimized parameters.

[0088] In this embodiment, the parameter data of each workstation and the appearance inspection structure of qualified or abnormal cigarette packages are collected, and data preprocessing is performed; the Pearson correlation coefficient method is used to obtain the correlation between different key process parameters to form a preliminary parameter combination, and the principal component analysis method is used to obtain the parameter combination that has the greatest impact on the appearance quality of the cigarette package; the support vector method is used to construct a model between the parameter combination and the appearance quality of the cigarette package, and training is performed to form a quality prediction model with optimized parameters, and the optimized parameters are output.

[0089] In order to further improve production efficiency and product quality, the present invention uses statistical methods and data mining technology to analyze the correlation between process parameters to guide producers to optimize and adjust process parameters. The following are specific implementation steps:

[0090] Data collection and preprocessing: First, collect a large amount of production data, including but not limited to the process parameters of each workstation, the appearance inspection results of cigarette packages, etc. Clean and preprocess the collected data to ensure the quality and availability of the data;

[0091] Correlation analysis: Analyze the correlation between different process parameters using statistical methods, such as correlation coefficient analysis and principal component analysis, and identify the parameter combinations that have a significant impact on product quality;

[0092] Constructing a prediction model: Using data mining techniques, such as support vector machine (SVM), random forest, etc., to construct a prediction model between process parameters and product quality. This model can predict the quality of products under different process parameter combinations;

[0093] Process parameter optimization: According to the results of correlation analysis, adjust the process parameters to achieve the purpose of optimizing product quality. For example, if the analysis finds that the temperature of a certain station is closely related to product quality, the temperature range of the station can be appropriately adjusted;

[0094] Quality prediction: Use the constructed prediction model to predict quality problems that may occur in the production process, so as to take preventive measures in advance.

[0095] Assume that in the gilding beauty process, through data analysis, it is found that temperature and speed are two key parameters that affect the appearance quality of cigarette packs. By adjusting these two parameters, the beauty effect of cigarette packs can be improved. First, collect temperature and speed data for a certain period of time, as well as the appearance quality test results of the corresponding cigarette packs. Secondly, use correlation coefficient analysis to determine the relationship between temperature and speed and appearance quality. Based on the collected data, build an SVM model with temperature and speed as input and appearance quality prediction results as output. According to the model prediction results, adjust the temperature and speed settings, for example, increase the temperature and reduce the speed, to improve the appearance quality of cigarette packs. Use the adjusted parameters for production and collect data again to verify the accuracy of the prediction model and the effectiveness of the process parameter adjustment.

[0096] In this way, the present invention can not only optimize process parameters through data analysis, but also predict potential quality problems, thereby improving production stability and consistency of product quality.

[0097] A detailed introduction is given by taking cigarette package traceability as an example. It can comprehensively record and monitor the key process parameters in the cigarette production process, and realize accurate association and traceability with the product appearance inspection pictures. By setting up a specific visual system on the production line, the images of each surface of the cigarette package are collected and merged to form a complete cigarette package image, so as to judge whether the appearance of the cigarette package is qualified, and the position of the production line can be traced based on the images of each surface of the cigarette package to improve production efficiency and product quality. The entire data collection is mainly carried out with the visual mechanism composed of cameras, etc., and the computer monitoring system with models is used for traceability. The specific steps are as follows:

[0098] The first step is image acquisition: At the core of the production line, a series of high-resolution visual acquisition systems are used to capture image data from all sides of the cigarette pack. Specifically, multiple visual acquisition devices are set up at key locations in the flow of cigarette packs, such as the packaging machine exit, inspection station, and in front of the palletizer. Each device is precisely aimed at a specific side of the cigarette pack to ensure that images of the front, back, two sides, top, and bottom surfaces can be clearly obtained;

[0099] Six visual mechanisms are set up in different sections of the production line, and the settings are as follows: two visual mechanisms are set up at the exit of the packaging machine to collect the front and back images of the cigarette packs respectively; two visual mechanisms are set up at the inspection station to collect the two side images of the cigarette packs respectively; two visual mechanisms are set up in front of the palletizer to collect the top and bottom images of the cigarette packs respectively. In this embodiment, through the high-precision image acquisition equipment, it can be ensured that the required image data can be captured quickly and accurately during the continuous movement of the cigarette packs. Through this arrangement, a complete six-sided image of the cigarette pack can be obtained, providing detailed information for subsequent appearance quality inspection and production location traceability;

[0100] Each visual system is equipped with appropriate lighting equipment to maintain consistency in image brightness and contrast. The image acquisition device has an autofocus function that can adapt to small changes in the size and position of cigarette packages. The collected image data is transmitted to the central processing system via Ethernet for further analysis and processing. In this way, not only can the quality of the image be ensured, but also reliable data support can be provided for the appearance inspection and traceability of cigarette packages;

[0101] The second step is to merge the complete 3D image of the cigarette package: the independent images collected by each visual system are synthesized through image processing algorithms. The specific steps include:

[0102] 1. Preprocess each single-sided image, including denoising, brightness and contrast adjustment, to ensure consistency of image quality.

[0103] 2. Feature point matching and alignment: Use image processing algorithms such as SIFT (Scale-invariant feature transform) or SURF (Speeded Up Robust Features) to detect key feature points in the images of each surface and match them. Through these matching points, the transformation matrix between the images is calculated and used to align the images. In this embodiment, the scale-invariant feature transform is mainly used.

[0104] 3. Image transformation and stitching: Specifically, the image is geometrically transformed according to the calculated transformation matrix so that the images of each surface are aligned in space. Then, an image stitching algorithm, such as Multi-Band Blending, is used to seamlessly stitch the images of adjacent surfaces together.

[0105] In this embodiment, image synthesis is performed based on the principle of the formula:

[0106] ;

[0107] In the above formula, represents the synthesized complete cigarette pack image, i.e., the result, and represents the image of the cigarette pack surface collected by the i-th visual unit, where i is 1, 2, 3, 4, 5 or 6; and It represents the weight of the i-th image in the synthesis process, which is dynamically calculated according to the overlapping area and image quality between images. Then, according to the above algorithm, the images of the six surfaces are synthesized into a three-dimensional model, which can be regarded as the complete six-sided image of the cigarette package.

[0108] In this step, we usually start with a front image of a cigarette pack and use it as the reference image. Then, the back, side, top, and bottom images are matched and aligned with the reference image one by one. In this way, we finally get a complete six-sided image of the cigarette pack, which can be used for subsequent appearance quality inspection and analysis. For example, if there is a deviation between the front image and the back image, the position of the back image can be adjusted by the algorithm until the two images are perfectly aligned.

[0109] Step 3: Determine whether the appearance quality of the cigarette package is qualified: Use advanced image recognition technology to analyze the synthesized complete cigarette package image. Specifically, the following steps are included:

[0110] 1. Image analysis preparation: The synthesized complete cigarette pack image is input into the image analysis system. The system is pre-loaded with a trained deep learning model that can identify and classify various appearance defects of cigarette packs;

[0111] 2. Feature extraction: Use CNN (Convolutional Neural Networks) to extract features from the cigarette pack image. CNN is a deep learning architecture that is particularly suitable for image recognition and can effectively identify complex patterns and structures in images. In this embodiment, feature extraction is specifically performed based on vector features, and the principle of extraction is shown in the formula;

[0112] ;

[0113] In the above formula, F represents the image of the complete cigarette pack. The feature vector extracted from ; f represents the convolutional neural network function; Represents the parameters of the network, which are learned through a large amount of sample data during the training process.

[0114] 3. Defect detection and classification: The extracted feature vector is input into the classifier, which will determine whether the cigarette package has defects based on the feature vector and classify the type of defect (such as damage, stains, printing errors, etc.). The classification process is as follows;

[0115] ;

[0116] In the above formula, C represents the classification result, g represents the classifier function, Represents the parameters of the classifier.

[0117] In this embodiment, model training is also included, specifically using a large amount of cigarette pack image data to train the deep learning model. The data includes various cigarette packs with qualified appearance and cigarette packs with different types of defects.

[0118] During the implementation, the synthesized cigarette pack image is first fed into the image analysis system. The CNN model in the system extracts features from the image, and then the classifier detects and classifies defects based on the feature vector. If any defects are detected, the system automatically marks the cigarette pack as unqualified and associates the unqualified information with the production line position data of the cigarette pack. Suppose that after a cigarette pack image is subjected to feature extraction and classification, the system finds a small damaged area. This damage is identified as a defect by the classifier, and the cigarette pack is therefore judged as unqualified.

[0119] Step 4: trace the location of the cigarette package based on the image information of each side of the cigarette package.

[0120] In this embodiment, when each visual system collects the image of the cigarette package, it will mark the image, and this mark includes the unique identification code of the cigarette package and the acquisition timestamp. This identification code is the "identity card" of the cigarette package on the production line, and follows the cigarette package in every link of the entire production process. The identification code here can be a QR code, etc., and then in the whole process, the appearance inspection data formed by the image is associated with the QR code information, so as to achieve accurate matching between the cigarette package and the QR code.

[0121] In this embodiment, the image marking process formula is as follows:

[0122] ;

[0123] in, represents the image tag of the i-th face, ID represents the unique identification code of the cigarette pack, Indicates the acquisition timestamp of the image.

[0124] The collected image data is transmitted to the central database together with its corresponding identification code and time stamp. In the database, this information is associated with the position data of the production line to form a complete data chain. By analyzing the image marks on each side of the cigarette package, the specific location of the cigarette package on the production line can be tracked.

[0125] The specific steps of tracing in this embodiment are: analyzing the movement trajectory of the cigarette pack on the production line according to the timestamp of image acquisition; matching the timestamp with the timeline of each position of the production line to determine the specific position of the cigarette pack at each time point. When a cigarette pack is detected to be defective at a certain position on the production line, the system will immediately retrieve all images and timestamp data associated with the cigarette pack ID. By analyzing this data, the system can trace back and determine the specific location of the cigarette pack on the production line, such as "packaging machine exit", "inspection station" or "in front of the palletizer". This location information will then be used for production line fault analysis, as well as traceability and recall of unqualified products. Assuming that a cigarette pack is detected to be defective at the inspection station, the system analyzes the image mark and timestamp of the cigarette pack and finds that it is abnormal at the exit of the packaging machine. In this way, the production team can check the packaging machine in a targeted manner to find out the root cause of the problem.

[0126] Step five, build a model to judge unqualified cigarette packs, associate the production line location of unqualified products, and complete the traceability of unqualified products. In order to accurately judge whether the cigarette packs are qualified and effectively associate unqualified products with the production line location, a deep learning model is used to detect appearance defects, and combined with data traceability technology to complete the traceability of unqualified products. A deep learning model based on convolutional neural network (CNN) is constructed to identify and classify the appearance defects of cigarette packs. The specific construction process is: collect a large amount of cigarette pack image data, including normal cigarette packs and cigarette packs with various appearance defects. These data are marked as qualified or unqualified and used to train the model. Design a multi-layer CNN model, which includes multiple convolutional layers, pooling layers, and fully connected layers. Use labeled image data to train the CNN model and adjust the model parameters through the back propagation algorithm , to minimize the classification error. The simplified formula for the specific model design is as follows:

[0127] ;

[0128] Among them, Output represents the prediction result of the model for the input image. Represents the parameters of the model.

[0129] Determination of unqualified cigarette packages:

[0130] Real-time detection: On the production line, images of cigarette packages are collected in real time and input into the trained CNN model. The model will output the result of whether the cigarette package is qualified.

[0131] Defect classification: If the model determines that the cigarette package is unqualified, it will also output the type and location information of the defect. The specific defect judgment formula is as follows:

[0132] ;

[0133] In the above formula, C represents the qualified / unqualified classification result of the cigarette pack; F represents the feature extracted from the cigarette pack image; Represents the parameters of the classifier.

[0134] In this embodiment, unqualified may include damage, stains, etc.

[0135] Correlation with production line location: Data recording: When a cigarette pack is judged as unqualified, the unique identification code, detection timestamp and defect information of the cigarette pack are recorded. Location traceability: The unique identification code and timestamp of the cigarette pack are matched with the production line database to find the specific production location of the cigarette pack.

[0136] The tracing process is as follows:

[0137] ;

[0138] In the above formula, P represents the production line position of the cigarette pack; ID represents the unique identification code of the cigarette pack; t represents the timestamp when the unqualified product is detected; and Database represents the database of the production line.

[0139] In this embodiment, the above scheme can also be used to trace the position of the cigarette pack production line when it is qualified.

[0140] For example, when a cigarette pack is judged as unqualified at the inspection station, the system will immediately extract its unique identification code and query its historical location information on the production line through the database. By analyzing this information, the system can determine at which link the problem occurred in the cigarette pack, thereby tracing the unqualified products. Assuming that a cigarette pack is detected to be damaged at the inspection station, the system analyzes its identification code and timestamp and traces back to the fact that the cigarette pack was damaged at the exit of the packaging machine, thereby locating the source of the problem.

[0141] In this way, not only can unqualified cigarette packs be identified in a timely manner, but the production line location of unqualified products can also be accurately traced, providing an important basis for production line improvement and product quality control.

[0142] This application uses precise cigarette package image acquisition and processing technology combined with efficient data analysis and traceability algorithms to achieve appearance quality inspection and accurate traceability during the cigarette package production process, thereby improving production efficiency and quality control levels.

[0143] Based on the above example of cigarette package traceability, using the existing data in the cigarette package database can significantly improve the quality management level of cigarette production, reduce quality risks, improve production efficiency, and provide strong guarantees for cigarette production. The specific steps are as follows:

[0144] First, data will be collected on the process parameters of the entire cigarette pack production process, including temperature, speed, pressure, humidity, etc. Systematic data collection will be carried out on various process parameters throughout the cigarette pack production process. This includes real-time monitoring and data recording of key process parameters such as temperature, speed, pressure and humidity at each link in the production line. Specifically, sensors and monitoring equipment will be deployed to ensure that the values ​​of these parameters can be accurately captured and the data will be stored in a database for subsequent in-depth analysis. The following is a detailed description of this process:

[0145] Deploy sensors: Install temperature sensors, speed sensors, pressure sensors, and humidity sensors at key nodes of the production line.

[0146] Real-time monitoring: Sensors will monitor changes in process parameters in real time and transmit data to the central data processing system.

[0147] Data recording: The central data processing system records the received data according to the timestamp to form a complete data set.

[0148] On the production line, temperature and pressure data are recorded every minute. For example, if the temperature parameter T represents the temperature value at a certain moment and is recorded, a large amount of data can be accumulated for subsequent analysis of the relationship between process parameters and cigarette pack quality.

[0149] Secondly, use statistical methods to analyze the correlation between the collected process parameter data and the appearance quality of cigarette packages. After completing the data collection step, enter the second step of the detailed technical solution, which is to conduct statistical analysis on the collected process parameter data. The core purpose of this step is to explore the intrinsic relationship between these process parameters and the appearance quality of cigarette packages through statistical methods. Specifically, statistical methods such as correlation analysis and regression analysis will be used to identify which changes in process parameters are significantly correlated with the appearance quality of cigarette packages. The following is a detailed description of this analysis process:

[0150] Statistical analysis steps: 1) Data preprocessing: First, the collected data is cleaned and formatted to ensure data quality and consistency. 2) Correlation analysis: Then, the Pearson correlation coefficient or Spearman rank correlation coefficient is used to evaluate the linear relationship between each process parameter and the appearance quality of cigarette packs. 3) Regression analysis: Then, the quantitative effect of process parameters on the appearance quality of cigarette packs is determined by multiple linear regression or logistic regression models.

[0151] Assuming that temperature T and pressure P are found to be potential correlation parameters, the Pearson correlation coefficient between temperature and cigarette pack appearance quality is calculated using the following formula:

[0152] and represents the correlation coefficient between temperature and appearance quality, and quality i represent the temperature and appearance quality of the i-th sample, respectively. , Represents their average value. Again, based on the analysis results, the key process parameters that have a significant impact on the appearance quality of cigarette packages are identified.

[0153] In this embodiment, after completing the statistical analysis of the correlation between process parameters and cigarette package appearance quality, the core step of the detailed technical solution is to identify the key process parameters that have a significant impact on the cigarette package appearance quality. The following are the detailed steps and methods of this identification process:

[0154] Parameter screening: Based on the results of statistical analysis, the process parameters with a high correlation with the appearance quality of cigarette packages are first screened out. These parameters will be used as candidate key process parameters.

[0155] Impact assessment: Further impact assessment is conducted on the selected candidate parameters.

[0156] Key parameter determination: Based on the impact assessment results, determine the set of key process parameters whose impact exceeds the preset threshold.

[0157] Assume that through statistical analysis, it is found that temperature (T) and pressure (P) are two parameters that are highly correlated with the appearance quality of cigarette packages. Use the following steps to determine whether they are critical process parameters:

[0158] Calculate the effects of temperature and pressure:

[0159] ;

[0160] if and If both exceed the set threshold, temperature and pressure are identified as key process parameters. In this way, it is possible to accurately identify which process parameters are critical to the appearance quality of cigarette packages, thus providing a scientific basis for subsequent process optimization and quality control. Finally, an identification model for the key process parameter set is constructed to guide quality control and process optimization in the production process.

[0161] In order to apply the identified key process parameter set to the actual production process, an identification model was designed and constructed, which aims to provide real-time decision support for quality control and process optimization of the production line. The following are the detailed steps and methods for building the model:

[0162] Data preparation: First, the data set consisting of the previously collected and preprocessed existing data is divided into a training set and a test set. The training set is used to train the model, while the test set is used to verify the accuracy of the model.

[0163] Model selection: Based on the characteristics of the data and the nature of the problem, choose an appropriate machine learning algorithm to build the model. This may include decision trees, random forests, support vector machines, or neural networks, etc.

[0164] Model training: Use the training set to train the selected model and adjust the model parameters to obtain the best performance.

[0165] Model validation: Use the test set to verify the model's predictive ability and ensure that the model has good generalization ability.

[0166] In this embodiment, a support vector machine (SVM) is used as a model, and its basic formula is as follows:

[0167] ;

[0168] In the above, f(x) is the prediction function, is the Lagrange multiplier, is the label of the i-th sample, is the kernel function and b is the bias term.

[0169] It also includes training the model by minimizing the following objective function, and evaluating the model performance by calculating the accuracy, recall and other indicators on the test set. Through the above steps, a model that can accurately identify the key process parameter set is constructed. The model can monitor the process parameters in real time during the production process and adjust the production process in time to ensure that the appearance quality of the cigarette package meets the predetermined standards.

[0170] The present invention also includes verifying the accuracy and effectiveness of the model through experiments to ensure the application value of the model in actual production. In order to test the performance of the model, a series of experiments were designed, which simulated different working conditions in the actual production environment, including normal working conditions and abnormal working conditions. Experimental setup: According to the actual production conditions, a variety of different experimental scenarios were set up to ensure that the model can be tested in a variety of environments. Data collection: During the experiment, the corresponding process parameter data and cigarette package appearance quality results were collected, and these data will be used to compare with the predicted results of the model. Model testing: Input the experimental data into the model, let the model predict the data, and record the predicted results. Result comparison: Compare the predicted results of the model with the actual collected cigarette package appearance quality results, and use the following indicators to evaluate the accuracy of the model:

[0171] ;

[0172] In addition to the accuracy, performance indicators such as recall and F1 score are also calculated to comprehensively evaluate the effectiveness of the model. Experimental setup: Three different temperatures and two different pressure conditions may be set to simulate different production environments. Data collection: Under each condition, a certain number of cigarette packs were produced and their actual appearance quality was recorded. Model testing: The model predicts the process parameters under each experimental condition and outputs the predicted appearance quality results. Result comparison: For example, if the model predicts the appearance quality of 100 cigarette packs, 90 of which are consistent with the actual results, the accuracy is 0.9; through these experimental verifications, it can be ensured that the model has high accuracy and effectiveness in actual production, thereby providing reliable quality control and process optimization guidance for cigarette pack production. By analyzing the correlation between process parameters and appearance quality in the cigarette pack production process through statistical methods, the key process parameter set was successfully identified, providing effective quality control and process optimization means for cigarette pack production.

[0173] Explain the parameter combination and subsequent quality prediction:

[0174] 1) Collect the original data of various process parameters and cigarette package appearance quality test results from the production line. After the data is collected, preprocessing is immediately performed, including data cleaning, data filling and data standardization. For example, for missing values ​​in temperature data, the average value of adjacent records is used to fill in to ensure the continuity and integrity of the data.

[0175] 2) Correlation analysis. By calculating the Pearson correlation coefficient for the process parameter data, a correlation coefficient matrix is ​​obtained, which reveals the degree of interdependence between the parameters. For example, if the correlation coefficient between temperature and humidity is found to be high, it indicates that the two parameters may have a strong interaction, and special attention should be paid to their impact in the quality prediction model. In this embodiment, principal component analysis is used to extract the main features in the process parameter data. This step not only identifies the key parameter combinations that affect product quality, but also reduces the dimensionality of the data set and simplifies the model building process. Through principal component analysis, it is possible to determine which parameters are the main driving factors, so that these key parameters can be focused on when building a prediction model.

[0176] 3) Prediction model construction. In order to quantify the relationship between process parameters and product quality, the present invention uses data mining technology to construct an accurate prediction model. The following is a detailed description of the construction process: Model selection and configuration: Based on the characteristics of the data and the prediction target, the support vector machine (SVM) is selected as the prediction model. SVM can effectively handle nonlinear problems and maintain good generalization ability when the data set is not large. In terms of model configuration, the optimal penalty parameter and kernel function are determined by cross-validation method to improve the prediction performance of the model. Feature engineering: Based on the aforementioned correlation analysis, the process parameters that have the most significant impact on product quality are selected as features. This step not only reduces the complexity of the model, but also helps to improve the interpretability of the model. Model training and verification: The SVM model is trained using the standardized data set. During the training process, we continuously adjust the model parameters and evaluate the accuracy of the model through the verification set. Finally, an SVM model that can accurately predict product quality is obtained. The model can predict the quality grade of the product based on the input process parameter combination. Through this model, producers can intuitively understand the impact of different process parameter settings on product quality, so as to make more accurate parameter adjustments in actual production.

[0177] 4) Process parameter optimization, adjust the process parameters according to the results of correlation analysis. Genetic Algorithm is used because it is an efficient global optimization algorithm suitable for finding the optimal solution in a complex search space. The following is the adjustment method of the genetic algorithm: Encoding: Encode each process parameter into a binary string to form the genetic code of the individual. Initial population: Randomly generate a certain number of individuals as the initial population. Fitness function: Define a fitness function to evaluate the fitness of each individual, that is, the prediction accuracy of the prediction model for product quality. Selection: Select individuals with high fitness for reproduction through roulette and other methods. Crossover and mutation: Generate a new generation of individuals through crossover and mutation operations. Termination condition: When the predetermined number of iterations or fitness threshold is reached, the algorithm terminates. Adjustment implementation: According to the optimal parameter combination output by the genetic algorithm, the process parameters on the production line are actually adjusted. Effect verification: After adjustment, collect new production data, and use the prediction model to verify the quality changes after parameter adjustment to ensure that the optimization effect is as expected. Through the above steps, the present invention can not only provide a scientific process parameter adjustment method, but also ensure the stability and consistency of product quality during the production process. Assume that when optimizing temperature and humidity parameters, the optimal parameter combination obtained by genetic algorithm is temperature X°C and humidity Y%. Apply this parameter combination to the production line and monitor the changes in product quality. By comparing the product quality data before and after optimization, you can intuitively see the quality improvement effect brought about by process parameter optimization.

[0178] 5) Quality prediction, using the prediction model to predict the quality problems that may occur during the production process. In order to ensure the continuity and stability of product quality, the present invention proposes the following quality prediction process: Real-time data input: During the production process, the process parameter data collected by sensors and detection equipment in real time are input into the constructed prediction model. Model prediction: The prediction model calculates the predicted value of product quality based on the input process parameter data and predicts possible abnormal situations. Early warning system: When the prediction model detects a potential quality problem, the early warning system is immediately activated. The early warning system will provide corresponding suggestions to production management personnel or automatically adjust the process parameters according to the severity of the problem to avoid the occurrence of quality problems. Result feedback and adjustment: The prediction results are fed back to the production control system to realize the dynamic adjustment of the process parameters. At the same time, the accuracy of the prediction model is regularly evaluated, and the model is updated and optimized according to the actual production situation. Through this process, the present invention can monitor product quality in real time during the production process, discover and solve potential quality problems in advance, thereby significantly improving production efficiency and product quality. The present invention achieves optimization of process parameters and accurate prediction of product quality by analyzing the correlation between process parameters and constructing a prediction model, thereby improving production efficiency and product quality, and has significant economic and social benefits.

[0179] The innovations and advantages of this application are as follows: Through precise cigarette package image acquisition and processing technology, combined with efficient data analysis and traceability algorithms, the appearance quality detection and accurate traceability of cigarette packages in the production process are realized, and production efficiency and quality control level are improved; the visual mechanism is associated with process parameters, etc., and the identification code composed of the associated process parameters and QR code information is used to realize the accurate traceability and positioning of cigarette packages with abnormal appearance quality or abnormal process parameters. The traceability process is fast and accurate, which can greatly reduce the workload of manual investigation by production personnel and improve production efficiency; using statistical methods and data mining technology, the correlation and influence weights between process parameters are deeply explored to explore the optimal combination of process parameters. Based on the optimal combination of process parameters, guide producers to optimize and adjust process parameters to improve the quality stability and consistency of cigarettes. Further construct a cigarette quality prediction model and establish an early warning mechanism for process parameter changes so that timely intervention and adjustment can be made when parameters deviate from the normal range to ensure product quality.

[0180] In this embodiment, key process parameters that affect the appearance quality of cigarette packages during the production process are determined; real-time process parameter data including the key process parameters of cigarette package production are obtained, and the real-time process parameter data are associated with the identification code of the cigarette package to obtain a process parameter set; appearance detection data of the cigarette package is collected, and whether the appearance of the cigarette package is abnormal is determined based on the appearance detection data; if the appearance of the cigarette package is abnormal, the identification code corresponding to the cigarette package with abnormal appearance is screened out from the process parameter set, and the corresponding abnormal process parameter data is obtained; the abnormal process parameter data is analyzed to obtain the cause of the abnormal appearance, and the historical cigarette packages are traced back according to the abnormal process parameter data to obtain the abnormality-related parameters corresponding to the historical cigarette packages; all the abnormality-related parameters are combined with the abnormal process parameter data to obtain the combined abnormal parameters, and the combined abnormal parameters are optimized according to the cause of the abnormal appearance to obtain the optimized parameters. The present application comprehensively records and monitors the key process parameters that affect the appearance quality in the cigarette production process, associates the real-time process parameter data of cigarette pack production including the key process parameters with the identification code of the cigarette pack, collects the appearance detection data of the cigarette pack, and determines whether the appearance of the cigarette pack is abnormal. If the appearance is abnormal, the identification code corresponding to the cigarette pack with abnormal appearance is screened out from the process parameter set, and the corresponding abnormal process parameter data is obtained, so as to realize the accurate tracing and positioning of the cigarette pack with abnormal appearance quality or abnormal process parameters. The tracing process is fast and accurate, which can greatly reduce the workload of manual investigation of production personnel and improve production efficiency. The abnormal process parameter data is analyzed to obtain the cause of the abnormal appearance, and the historical cigarette packs are traced to obtain the abnormal related parameters corresponding to the historical cigarette packs. All the abnormal related parameters are combined with the abnormal process parameter data, and the combined abnormal parameters are optimized according to the cause of the abnormal appearance. Based on the optimal process parameter combination, the producer is guided to optimize the process parameters, improve the quality stability and consistency of cigarettes, and improve production efficiency and product quality.

[0181] See also Figure 2 As shown, the embodiment of the present invention discloses a process parameter optimization device in cigarette pack production, which may specifically include:

[0182] A key process parameter determination module 11 is used to determine key process parameters that affect the appearance quality of cigarette packages during the production process;

[0183] An association module 12, used to obtain real-time process parameter data including the key process parameters of cigarette pack production, and associate the real-time process parameter data with the identification code of the cigarette pack to obtain a process parameter set;

[0184] A judgment module 13, used for collecting appearance detection data of the cigarette package, and judging whether the appearance of the cigarette package is abnormal according to the appearance detection data;

[0185] The abnormal process parameter data acquisition module 14 is used for selecting the identification code corresponding to the cigarette package with abnormal appearance from the process parameter set if the appearance of the cigarette package is abnormal, and acquiring the corresponding abnormal process parameter data;

[0186] The tracing module 15 is used to analyze the abnormal process parameter data to obtain the cause of the abnormal appearance, and trace the historical cigarette packs according to the abnormal process parameter data to obtain the abnormal related parameters corresponding to the historical cigarette packs;

[0187] The optimization module 16 is used to combine all the abnormality-related parameters with the abnormal process parameter data to obtain combined abnormal parameters, and optimize the combined abnormal parameters according to the cause of the appearance abnormality to obtain optimized parameters.

[0188] In some specific embodiments, the key process parameter determination module 11 may specifically include:

[0189] A data storage module is used to store the data of cigarette packs at each station on the production line into a preset cigarette pack database; the data includes temperature, speed, pressure, and humidity;

[0190] The analysis and identification module is used to use analytical and statistical methods to perform appearance quality correlation analysis and key parameter identification on the data in the cigarette package database, so as to obtain the key process parameters that affect the appearance quality of the cigarette package during the production process.

[0191] In some specific embodiments, the association module 12 may specifically include:

[0192] The identification code association module is used to use a pre-installed high-speed data acquisition device and a visual inspection system to respectively obtain real-time process parameter data of cigarette pack production including the key process parameters and the identification code of the cigarette pack, and associate the real-time process parameter data with the identification code of the cigarette pack.

[0193] In some specific embodiments, the determination module 13 may specifically include:

[0194] An appearance image acquisition module, used to acquire appearance images of cigarette packages from different directions using a pre-installed visual inspection system;

[0195] The appearance abnormality judgment module is used to merge the appearance images of the cigarette package in different directions to obtain a merged image, and judge whether the appearance of the cigarette package is abnormal based on the synthesized image; the appearance detection data includes the synthesized image.

[0196] In some specific embodiments, the determination module 13 may specifically include:

[0197] An image processing module, used for preprocessing, matching and aligning feature points, transforming and splicing the cigarette package appearance images in different directions using an image processing algorithm to obtain a merged image;

[0198] A feature extraction module, used to input the synthesized image into a preset deep learning model for image analysis, and to extract features using a convolutional neural network to obtain a feature vector;

[0199] The defect detection and classification module is used to use a classifier to perform defect detection and classification on the feature vector to complete the judgment of the appearance abnormality of the cigarette package.

[0200] In some specific embodiments, the tracing module 15 may specifically include:

[0201] A statistical analysis module, used for performing statistical analysis on the abnormal process parameter data to obtain analysis results;

[0202] The input-output module is used to construct a prediction model and input the analysis results into the prediction model so that the prediction model can identify and predict the cause of the appearance abnormality based on the analysis results and output the predicted cause of the appearance abnormality.

[0203] In some specific embodiments, the tracing module 15 may specifically include:

[0204] The cleaning and formatting module is used to clean and format the abnormal process parameter data to obtain the pre-processed abnormal process parameter data, and then perform correlation analysis and regression analysis on the pre-processed abnormal process parameter data.

[0205] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 is used to store a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the process parameter optimization method for cigarette pack production performed by the electronic device disclosed in any of the aforementioned embodiments.

[0206] In this embodiment, the power supply 23 is used to provide working voltage for each hardware device on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and the external device, and the communication protocol it follows is any communication protocol that can be applied to the technical solution of the present application, and is not specifically limited here; the input and output interface 25 is used to obtain external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs and is not specifically limited here.

[0207] In addition, the memory 22, as a carrier for storing resources, can be a read-only memory, a random access memory, a disk or an optical disk, etc. The resources stored thereon include an operating system 221, a computer program 222 and data 223, etc. The storage method can be temporary storage or permanent storage.

[0208] Among them, the operating system 221 is used to manage and control the hardware devices and computer programs 222 on the electronic device 20 to realize the operation and processing of the data 223 in the memory 22 by the processor 21, which can be Windows, Unix, Linux, etc. In addition to including a computer program that can be used to complete the process parameter optimization method in cigarette pack production performed by the electronic device 20 disclosed in any of the aforementioned embodiments, the computer program 222 can also further include a computer program that can be used to complete other specific tasks. In addition to including data transmitted from an external device received by the process parameter optimization device in cigarette pack production, the data 223 can also include data collected by its own input and output interface 25, etc.

[0209] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0210] Furthermore, an embodiment of the present application also discloses a computer-readable storage medium, in which a computer program is stored. When the computer program is loaded and executed by a processor, the process parameter optimization method steps in cigarette pack production disclosed in any of the aforementioned embodiments are implemented.

[0211] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0212] The above is a detailed introduction to the process parameter optimization method, device, equipment and storage medium in the cigarette pack production provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for general technical personnel in this field, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for optimizing process parameters in cigarette pack production, characterized in that: include: Determine the key process parameters that affect the appearance quality of cigarette packages during production; Acquire real-time process parameter data including the key process parameters of cigarette pack production, and associate the real-time process parameter data with the identification code of the cigarette pack to obtain a process parameter set; collecting appearance detection data of the cigarette package, and judging whether the appearance of the cigarette package is abnormal according to the appearance detection data; If the appearance of the cigarette package is abnormal, selecting an identification code corresponding to the cigarette package with abnormal appearance from the process parameter set, and acquiring corresponding abnormal process parameter data; Analyze the abnormal process parameter data to obtain the cause of the abnormal appearance, and trace back the historical cigarette packs according to the abnormal process parameter data to obtain the abnormal related parameters corresponding to the historical cigarette packs; All the abnormality-related parameters are combined with the abnormal process parameter data to obtain combined abnormal parameters, and the combined abnormal parameters are optimized according to the cause of the appearance abnormality to obtain optimized parameters.

2. The process parameter optimization method in cigarette pack production according to claim 1, characterized in that: The key process parameters that affect the appearance quality of cigarette packages during production include: The data of cigarette packs at each station on the production line are saved in a preset cigarette pack database; the data includes temperature, speed, pressure, and humidity; Analytical and statistical methods are used to perform correlation analysis on appearance quality and identify key parameters of the data in the cigarette package database, so as to obtain key process parameters that affect the appearance quality of the cigarette packages during the production process.

3. The process parameter optimization method in cigarette pack production according to claim 1, characterized in that: The step of obtaining real-time process parameter data including the key process parameters of cigarette pack production and associating the real-time process parameter data with the identification code of the cigarette pack includes: The pre-installed high-speed data acquisition device and visual inspection system are used to respectively obtain the real-time process parameter data of cigarette pack production including the key process parameters and the identification code of the cigarette pack, and the real-time process parameter data is associated with the identification code of the cigarette pack.

4. The process parameter optimization method in cigarette pack production according to claim 1, characterized in that: The collecting the appearance detection data of the cigarette package and judging whether the appearance of the cigarette package is abnormal according to the appearance detection data includes: Use the pre-installed visual inspection system to collect appearance images of cigarette packages from different directions; The cigarette package appearance images in different directions are merged to obtain a merged image, and whether the appearance of the cigarette package is abnormal is judged according to the synthesized image; the appearance detection data includes the synthesized image.

5. The process parameter optimization method in cigarette pack production according to claim 4, characterized in that: The step of merging the cigarette package appearance images in different directions to obtain a combined image, and judging whether the appearance of the cigarette package is abnormal according to the combined image includes: Using an image processing algorithm to preprocess, match and align feature points, transform and splice the cigarette package appearance images in different directions to obtain a merged image; Inputting the synthesized image into a preset deep learning model for image analysis, and extracting features using a convolutional neural network to obtain a feature vector; The classifier is used to perform defect detection and classification on the feature vector to complete the abnormal appearance judgment of the cigarette package.

6. The process parameter optimization method in cigarette pack production according to any one of claims 1 to 5, characterized in that: The analyzing the abnormal process parameter data to obtain the cause of the abnormal appearance includes: Performing statistical analysis on the abnormal process parameter data to obtain analysis results; A prediction model is constructed, and the analysis results are input into the prediction model so that the prediction model can identify and predict the cause of the appearance abnormality based on the analysis results, and output the predicted cause of the appearance abnormality.

7. The process parameter optimization method in cigarette pack production according to claim 6, characterized in that: The statistical analysis of the abnormal process parameter data includes: The abnormal process parameter data are cleaned and formatted to obtain the pre-processed abnormal process parameter data, and then correlation analysis and regression analysis are performed on the pre-processed abnormal process parameter data.

8. A device for optimizing process parameters in cigarette pack production, characterized in that: include: Key process parameter determination module, used to determine the key process parameters that affect the appearance quality of cigarette packages during the production process; An association module, used for acquiring real-time process parameter data including the key process parameters of cigarette pack production, and associating the real-time process parameter data with the identification code of the cigarette pack to obtain a process parameter set; a judgment module, used for collecting appearance detection data of the cigarette package, and judging whether the appearance of the cigarette package is abnormal according to the appearance detection data; an abnormal process parameter data acquisition module, for selecting an identification code corresponding to the cigarette package with abnormal appearance from the process parameter set if the cigarette package has abnormal appearance, and acquiring corresponding abnormal process parameter data; A tracing module, used to analyze the abnormal process parameter data to obtain the cause of the abnormal appearance, and trace the historical cigarette packs according to the abnormal process parameter data to obtain the abnormal related parameters corresponding to the historical cigarette packs; The optimization module is used to combine all the abnormality-related parameters with the abnormal process parameter data to obtain combined abnormal parameters, and optimize the combined abnormal parameters according to the cause of the appearance abnormality to obtain optimized parameters.

9. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor, used to execute the computer program to implement the process parameter optimization method in cigarette package production as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: Used to store computer programs; wherein, when the computer program is executed by a processor, the method for optimizing process parameters in cigarette package production according to any one of claims 1 to 7 is implemented.

Images