A pressure feedback analysis method and system for visual printing presses

Abstract

The present invention discloses a pressure feedback analysis method and system applied to a visual printer, comprising: obtaining historical printing examples of a target model visual printer, extracting features from each historical printing example and evaluating the printing quality of each historical printing example, and constructing a historical printing example data set; analyzing the relationship between printing pressure, printing quality and printing control parameters based on the historical printing example data set, and constructing a fusion topological structure diagram through the relationship between printing pressure, printing quality and printing control parameters; constructing a printing quality prediction model, performing model training through the historical printing example data set and the fusion topological structure diagram, obtaining real-time printing pressure monitoring information, predicting real-time printing quality and judging whether it meets the printing quality standard; if the printing quality standard is not met, optimizing the current printing control parameters, generating a printing parameter control scheme for control, thereby improving printing quality and reducing scrap rate.

Description

Technical Field

[0001] The present invention relates to the technical field of printing machines, and in particular to a pressure feedback analysis method and system applied to a visual printing machine. Background Art

[0002] In modern printing technology, in order to ensure that the printing quality meets the high standards of image clarity and color reproduction, the pressure control of the printing press is a key factor. During the printing process, the precise adjustment of the roller pressure directly affects the contact effect between the printing medium and the ink, thereby determining the uniformity, color consistency and detail performance of the finished product. However, due to slight differences in the thickness, flexibility and ink absorption of paper or other printing media, the real-time change of the roller pressure may cause local printing quality to be substandard, especially when dealing with high-speed continuous printing and complex patterns. The traditional fixed parameter pressure setting cannot fully adapt to dynamic needs. At the same time, during the printing process, the printing pressure reflects the printing quality of the printed product to a certain extent. If the printing pressure is not properly controlled, the quality of the final printed product will be greatly affected.

[0003] Therefore, how to effectively control the printing pressure during printing, thereby improving the printing quality and reducing the scrap rate is an urgent problem to be solved. Summary of the invention

[0004] The present invention overcomes the defects of the prior art and provides a pressure feedback analysis method and system applied to a visual printing press, the important purpose of which is to improve the printing quality and reduce the scrap rate.

[0005] To achieve the above-mentioned purpose, the present invention provides a pressure feedback analysis method applied to a visual printing press in a first aspect, comprising:

[0006] Obtain historical printing instances of the target model visual printing machine, extract features from each historical printing instance, evaluate the printing quality of each historical printing instance, and construct a historical printing instance dataset;

[0007] Analyzing the relationship between printing pressure, printing quality and printing control parameters based on the historical printing instance data set, and constructing a fusion topology structure diagram through the relationship between printing pressure, printing quality and printing control parameters;

[0008] Constructing a printing quality prediction model, training the model through the historical printing instance data set and the fusion topological structure diagram, obtaining real-time printing pressure monitoring information, predicting real-time printing quality and judging whether it meets the printing quality standard;

[0009] If the printing quality standard is not met, the current printing control parameters are optimized and a printing parameter control scheme is generated for control.

[0010] In this solution, the historical printing examples of the target model visual printing machine are obtained, features are extracted for each historical printing example, and the printing quality of each historical printing example is evaluated to construct a historical printing example data set, which specifically includes:

[0011] Based on the big data network, historical printing instances of the target model visual printing machine are obtained, including the printed product images, printing control parameters and printing pressure data, to form a historical printing instance data set;

[0012] Building a feature pyramid network, extracting the printed product image corresponding to each instance according to the acquired historical printing instances, inputting the image into the feature pyramid network, and performing a dilated convolution on the input printed product image according to a preset dilated rate;

[0013] The feature maps outputted by each branch after the dilated convolution are obtained and horizontally connected to obtain the fused feature maps corresponding to each printed product image, and the image feature time series corresponding to each printed product image is generated according to the preset timestamp;

[0014] The printing image quality evaluation index is set, and the weight of each printing image quality evaluation index is assigned using the intuitive fuzzy analysis method. The printing quality evaluation model is constructed by combining the weight assignment results and the printing image quality evaluation index.

[0015] Inputting the image feature time series corresponding to each printed product into the printing quality evaluation model, analyzing the printing image quality when each historical printing instance performs printing work at different printing time stages, and obtaining printing quality evaluation information;

[0016] The historical printing instance dataset is updated by using the printing quality evaluation information as an attached feature corresponding to the historical printing instance.

[0017] In this solution, the relationship between printing pressure, printing quality and printing control parameters is analyzed based on the historical printing instance data set, and a fusion topology structure diagram is constructed through the relationship between printing pressure, printing quality and printing control parameters, specifically including:

[0018] Acquire a historical printing instance data set, extract printing pressure data corresponding to each historical printing instance according to the historical printing instance data set, and perform time series processing to obtain a printing pressure data sequence;

[0019] Acquire printing quality evaluation information, and generate a printing image quality score sequence when each historical printing instance performs printing work at different printing time stages according to the printing quality evaluation information;

[0020] Perform time-series matching on the printing pressure data sequence and the printing image quality score sequence, introduce a decision tree algorithm, take the printing pressure as the decision attribute and the printing image quality as the target attribute, and initialize the root node;

[0021] According to the initialized root node, a plurality of branches are obtained by splitting based on the range of different printing pressure data, and a hierarchical structure of the decision tree is gradually constructed, and the best splitting node is selected according to the information gain, and a first topological structure diagram is generated according to the constructed decision tree structure;

[0022] Subsequently, the printing control parameters corresponding to each historical printing instance are extracted through the historical printing instance data set, and the printing control parameters are used as decision attributes and the printing pressure is used as the target attribute. A decision tree is constructed to analyze the association between the printing control parameters and the printing pressure data, and a second topological structure diagram is generated;

[0023] The first topology structure graph and the second topology structure graph are associated and fused to generate a fused topology structure graph.

[0024] In this solution, the printing quality prediction model is constructed, the model is trained through the historical printing instance data set and the fusion topology structure diagram, the real-time printing pressure monitoring information is obtained, the real-time printing quality is predicted and whether it meets the printing quality standard is determined, specifically including:

[0025] Acquire a fused topological structure diagram, and extract a relationship matrix between printing image quality, printing pressure data, and printing control parameters based on the fused topological structure diagram;

[0026] A printing quality prediction model is constructed using a deep neural network, the relationship matrix is ​​input into the prediction model for coding learning, and model training and parameter adjustment are performed in combination with a historical printing instance data set to obtain a printing quality prediction model that meets expectations;

[0027] Monitor the printing pressure of the target visual printer when performing a printing task, obtain real-time printing pressure monitoring information, perform data preprocessing on the real-time printing pressure monitoring information, and perform time series processing on the preprocessed real-time printing pressure data;

[0028] Acquire a real-time printing pressure data sequence, input it into the printing quality prediction model, analyze the printing quality in a future time period under the current printing pressure, and obtain printing quality prediction information;

[0029] Acquire the printing quality standard of the current printing task, judge the printing quality prediction information with the printing quality standard of the current printing task, and analyze whether it is necessary to adjust the printing control parameters;

[0030] If the printing quality prediction information is greater than the printing quality standard, it means that the current printing control parameters do not need to be adjusted; if the printing quality prediction information is less than the printing quality standard, it means that the printing control parameters need to be adjusted.

[0031] In this solution, if the printing quality standard is not met, the current printing control parameters are optimized and a printing parameter control scheme is generated for control, which specifically includes:

[0032] Acquire printing quality prediction information, and search in a historical printing instance data set based on the printing quality prediction information to acquire historical printing control parameters corresponding to the historical printing instance under the current printing quality;

[0033] Acquire the real-time printing control parameters of the target visual printer, and calculate the real-time printing control parameters with the retrieved historical printing control parameters to obtain the printing control parameter deviation;

[0034] Add the printing control parameter deviation to the real-time printing control parameter to obtain the added printing control parameter, and construct a restriction interval with the real-time printing control parameter as the lower limit and the added printing control parameter as the upper limit;

[0035] According to the restriction interval, historical printing instances whose historical printing control parameters are within the restriction interval are retrieved from the historical printing instance data set and a search space is formed, a sparrow search algorithm is introduced, and an initial sparrow population is randomly generated using the search space;

[0036] Calculating the fitness value of each individual sparrow in the initial sparrow population, performing fitness sorting according to the fitness value to obtain a fitness sorting table, selecting a corresponding number of individuals as discoverers based on a preset discoverer ratio, and taking the remaining individuals as followers to update their positions;

[0037] The position of the optimal sparrow individual is obtained by repeated iterative updating, the current optimal solution is output, and a number of candidate printing control parameters are generated according to the output current optimal solution.

[0038] In this solution, if the printing quality standard is not met, the current printing control parameters are optimized, and a printing parameter control scheme is generated for control, which also includes:

[0039] Acquire candidate printing control parameters, and analyze printing pressure data when the candidate printing control parameters are used by simulation software based on the candidate printing control parameters to obtain printing pressure simulation data;

[0040] Inputting the printing pressure simulation data into a printing quality prediction model, predicting the quality of the printed product under the corresponding printing pressure, and judging it against the printing quality standard of the current printing task;

[0041] If it is less than the printing quality standard of the current printing task, it means that the corresponding candidate printing control parameter is an unavailable parameter, and a restriction condition is generated based on the corresponding candidate printing control parameter, and the next candidate printing control parameter is obtained by re-iteration;

[0042] If it is greater than the printing quality standard of the current printing task, it means that the corresponding candidate printing control parameter is an available parameter, and the corresponding printing quality prediction result is calculated with the printing quality prediction result of the real-time printing task;

[0043] The calculated difference is used as an evaluation index to sort the candidate printing control parameters, and the candidate printing control parameter with the largest difference is selected as the final printing control parameter according to the sorting result;

[0044] The final printing control parameters are subtracted from the real-time printing control parameters, and a printing parameter control scheme for the target visual printing machine is generated based on the calculation results.

[0045] A second aspect of the present invention provides a pressure feedback analysis system applied to a visual printer, the system comprising: a memory, a processor, the memory containing a pressure feedback analysis method program applied to a visual printer, the pressure feedback analysis method program applied to a visual printer being executed by the processor to implement the following steps:

[0046] Obtain historical printing instances of the target model visual printing machine, extract features from each historical printing instance, evaluate the printing quality of each historical printing instance, and construct a historical printing instance dataset;

[0047] Analyzing the relationship between printing pressure, printing quality and printing control parameters based on the historical printing instance data set, and constructing a fusion topology structure diagram through the relationship between printing pressure, printing quality and printing control parameters;

[0048] Constructing a printing quality prediction model, training the model through the historical printing instance data set and the fusion topological structure diagram, obtaining real-time printing pressure monitoring information, predicting real-time printing quality and judging whether it meets the printing quality standard;

[0049] If the printing quality standard is not met, the current printing control parameters are optimized and a printing parameter control scheme is generated for control.

[0050] The present invention discloses a pressure feedback analysis method and system applied to a visual printer, comprising: obtaining historical printing examples of a target model visual printer, extracting features from each historical printing example and evaluating the printing quality of each historical printing example, and constructing a historical printing example data set; analyzing the relationship between printing pressure, printing quality and printing control parameters based on the historical printing example data set, and constructing a fusion topological structure diagram through the relationship between printing pressure, printing quality and printing control parameters; constructing a printing quality prediction model, performing model training through the historical printing example data set and the fusion topological structure diagram, obtaining real-time printing pressure monitoring information, predicting real-time printing quality and judging whether it meets the printing quality standard; if the printing quality standard is not met, optimizing the current printing control parameters, generating a printing parameter control scheme for control, thereby improving printing quality and reducing scrap rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the drawings required for use in the embodiments or exemplary descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to the drawings without paying creative work.

[0052] Figure 1 A flow chart of a pressure feedback analysis method applied to a visual printing machine provided by one embodiment of the present invention;

[0053] Figure 2 A flow chart of a method for optimizing printing control parameters of a visual printing press based on pressure feedback analysis provided by one embodiment of the present invention;

[0054] Figure 3 A block diagram of a pressure feedback analysis system applied to a visual printing machine provided by one embodiment of the present invention;

[0055] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0056] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0057] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0058] Figure 1 A flow chart of a pressure feedback analysis method applied to a visual printing machine provided by one embodiment of the present invention;

[0059] like Figure 1 As shown, the present invention provides a flow chart of a pressure feedback analysis method applied to a visual printing machine, comprising:

[0060] S102, obtaining historical printing instances of the target model visual printing machine, extracting features from each historical printing instance and evaluating the printing quality of each historical printing instance, and constructing a historical printing instance data set;

[0061] S104, analyzing the relationship between printing pressure, printing quality and printing control parameters based on the historical printing instance data set, and constructing a fusion topology structure diagram through the relationship between printing pressure, printing quality and printing control parameters;

[0062] S106, constructing a printing quality prediction model, training the model through the historical printing instance data set and the fusion topology structure diagram, obtaining real-time printing pressure monitoring information, predicting the real-time printing quality and judging whether it meets the printing quality standard;

[0063] S108, if the printing quality standard is not met, the current printing control parameters are optimized and a printing parameter control scheme is generated for control.

[0064] It should be noted that the present invention provides a pressure feedback analysis method and system for a visual printer. First, a large number of historical printing instances of a target model of a visual printer are obtained by data retrieval, covering different printing conditions, media types and specific pattern features. By extracting features from each historical instance, a variety of key parameters affecting printing quality, such as printing pressure, speed, inkjet intensity, paper material, etc., are extracted. At the same time, image processing technology is used to evaluate the quality of the printed product to form corresponding quality evaluation data, thereby forming a historical printing instance data set. Subsequently, the complex correlation between printing pressure, printing quality and printing control parameters is further analyzed based on the historical printing instance data set. To this end, a method of fusion topological structure graph is introduced to form an explanatory and predictive graph model by integrating the multi-dimensional feature relationship in the printing process. A printing quality prediction model is constructed based on the above-mentioned fusion topological structure graph and the historical printing instance data set, and the nonlinear mapping relationship between printing parameters, pressure changes and quality results is learned. Next, when the real-time printing pressure monitoring information is obtained, the model quickly predicts the real-time printing quality based on the current parameter conditions and determines whether it meets the preset quality standard. If the prediction result shows that the current printing quality fails to meet the standard, the control parameter optimization mechanism will be triggered to generate a specific printing parameter adjustment plan, which will be used for real-time control, allowing the printing machine to gradually improve the current printing quality while operating uninterruptedly, greatly improving the autonomous operation capability and printing quality stability of the visual printing machine.

[0065] Further, in a preferred embodiment of the present invention, the process of obtaining historical printing instances of the target model visual printing machine, extracting features of each historical printing instance and evaluating the printing quality of each historical printing instance, and constructing a historical printing instance data set specifically includes:

[0066] Based on the big data network, historical printing instances of the target model visual printing machine are obtained, including the printed product images, printing control parameters and printing pressure data, to form a historical printing instance data set;

[0067] Building a feature pyramid network, extracting the printed product image corresponding to each instance according to the acquired historical printing instances, inputting the image into the feature pyramid network, and performing a dilated convolution on the input printed product image according to a preset dilated rate;

[0068] The feature maps outputted by each branch after the dilated convolution are obtained and horizontally connected to obtain the fused feature maps corresponding to each printed product image, and the image feature time series corresponding to each printed product image is generated according to the preset timestamp;

[0069] The printing image quality evaluation index is set, and the weight of each printing image quality evaluation index is assigned using the intuitive fuzzy analysis method. The printing quality evaluation model is constructed by combining the weight assignment results and the printing image quality evaluation index.

[0070] Inputting the image feature time series corresponding to each printed product into the printing quality evaluation model, analyzing the printing image quality when each historical printing instance performs printing work at different printing time stages, and obtaining printing quality evaluation information;

[0071] The historical printing instance dataset is updated by using the printing quality evaluation information as an attached feature corresponding to the historical printing instance.

[0072] It should be noted that historical printing instances are collected through the big data network, including the corresponding finished printing images, printing control parameters, and pressure data recorded during the printing process. The feature pyramid network is used to extract deep features from the collected historical instances, and the finished printing images of each historical instance are input into the feature pyramid network. The network performs hole convolution processing on the input image according to different hole rates to capture multi-scale image information. This convolution operation adjusts the receptive field to improve the ability to obtain global features while retaining local details of the image. Subsequently, feature maps are extracted from the convolution results of each branch in the network and horizontally connected to generate fused feature maps. Through the fusion operation, the features of each image can represent its internal pattern in a more comprehensive way. Combined with the printing timestamp information, these fused feature maps are further generated into corresponding image feature time series sequences to characterize the dynamic changes of the image in the time dimension. In order to quantify the quality of the printed product, multiple printing image quality evaluation indicators are set, including color reproduction, pattern clarity, contrast, etc. The importance of the evaluation indicators is weighted by the intuitive fuzzy analysis method, and a printing quality evaluation model is established. Next, the image feature time series extracted from the feature pyramid network is input into the quality assessment model to analyze the image quality performance of each historical instance in different printing time periods, and obtain the printing quality assessment information, which specifically reflects the quality fluctuation of each instance and the staged printing performance, providing data basis for subsequent optimization. Finally, the generated printing quality assessment information is used as a new auxiliary feature and updated to the original historical printing instance data set, so that the data set not only contains the initial printed products and related parameters, but also integrates the quality assessment information.

[0073] Furthermore, in a preferred embodiment of the present invention, the relationship between printing pressure, printing quality and printing control parameters is analyzed based on the historical printing instance data set, and a fusion topology structure diagram is constructed through the relationship between printing pressure, printing quality and printing control parameters, specifically including:

[0074] Acquire a historical printing instance data set, extract printing pressure data corresponding to each historical printing instance according to the historical printing instance data set, and perform time series processing to obtain a printing pressure data sequence;

[0075] Acquire printing quality evaluation information, and generate a printing image quality score sequence when each historical printing instance performs printing work at different printing time stages according to the printing quality evaluation information;

[0076] Perform time-series matching on the printing pressure data sequence and the printing image quality score sequence, introduce a decision tree algorithm, take the printing pressure as the decision attribute and the printing image quality as the target attribute, and initialize the root node;

[0077] According to the initialized root node, a plurality of branches are obtained by splitting based on the range of different printing pressure data, and a hierarchical structure of the decision tree is gradually constructed, and the best splitting node is selected according to the information gain, and a first topological structure diagram is generated according to the constructed decision tree structure;

[0078] Subsequently, the printing control parameters corresponding to each historical printing instance are extracted through the historical printing instance data set, and the printing control parameters are used as decision attributes and the printing pressure is used as the target attribute. A decision tree is constructed to analyze the association between the printing control parameters and the printing pressure data, and a second topological structure diagram is generated;

[0079] The first topology structure graph and the second topology structure graph are associated and fused to generate a fused topology structure graph.

[0080] It should be noted that the printing pressure data corresponding to each historical instance is extracted through the historical printing instance data set. A time series processing method is used to generate a printing pressure data sequence with time continuity, accurately describing the pressure changes of each historical instance at different time stages. At the same time, the printing quality evaluation information of each instance is obtained from the data set, and the printing image quality score corresponding to each time stage is calculated according to the evaluation model, thereby generating an image quality score sequence for each historical printing instance. The generated printing pressure data sequence is time-matched with the printing image quality score sequence, and a decision tree algorithm is introduced to analyze the matching data, setting the printing pressure as the decision attribute and the printing image quality as the target attribute, and initializing the root node of the decision tree. According to different printing pressure data ranges, the root node is split into several branches, and the best split node is selected based on the information gain as the criterion. The hierarchical structure of the decision tree is constructed through multiple iterations, and finally a complete decision tree is formed, which is further converted into the first topological structure diagram. Subsequently, the printing control parameter data corresponding to each instance is extracted from the historical printing instance data set and set as the decision attribute. At the same time, the printing pressure data is set as the target attribute. The decision tree algorithm is used again for analysis to construct the association decision tree between the control parameters and the pressure to generate the second topological structure diagram, which is used to represent the dynamic mechanism of how the control parameters affect the printing pressure. Finally, the first topological structure diagram and the second topological structure diagram are associated and fused to generate a fused topological structure diagram.

[0081] Furthermore, in a preferred embodiment of the present invention, the construction of the printing quality prediction model, the model training is performed through the historical printing instance data set and the fusion topological structure diagram, the real-time printing pressure monitoring information is obtained, the real-time printing quality is predicted and whether it meets the printing quality standard is determined, specifically including:

[0082] Acquire a fused topological structure diagram, and extract a relationship matrix between printing image quality, printing pressure data, and printing control parameters based on the fused topological structure diagram;

[0083] A printing quality prediction model is constructed using a deep neural network, the relationship matrix is ​​input into the prediction model for coding learning, and model training and parameter adjustment are performed in combination with a historical printing instance data set to obtain a printing quality prediction model that meets expectations;

[0084] Monitor the printing pressure of the target visual printer when performing a printing task, obtain real-time printing pressure monitoring information, perform data preprocessing on the real-time printing pressure monitoring information, and perform time series processing on the preprocessed real-time printing pressure data;

[0085] Acquire a real-time printing pressure data sequence, input it into the printing quality prediction model, analyze the printing quality in a future time period under the current printing pressure, and obtain printing quality prediction information;

[0086] Acquire the printing quality standard of the current printing task, judge the printing quality prediction information with the printing quality standard of the current printing task, and analyze whether it is necessary to adjust the printing control parameters;

[0087] If the printing quality prediction information is greater than the printing quality standard, it means that the current printing control parameters do not need to be adjusted; if the printing quality prediction information is less than the printing quality standard, it means that the printing control parameters need to be adjusted.

[0088] It should be noted that in order to achieve accurate prediction and dynamic control of printing quality, the relationship matrix between printing image quality, printing pressure data and printing control parameters is first extracted by fusion topological structure diagram. A printing quality prediction model is built using a deep neural network, and the relationship matrix is ​​used as input to encode and learn the model. The network weights and parameters are optimized in combination with the historical printing instance data set to ensure that the model has reliable prediction ability and high accuracy. Subsequently, the printing pressure of the target visual printing machine is monitored in real time to generate a pressure data sequence reflecting the real-time pressure change trend. The generated real-time printing pressure data sequence is input into the trained printing quality prediction model to predict the printing quality in the future time period under the current printing pressure conditions to obtain printing quality prediction information. Then, the corresponding printing quality standard is obtained, and the predicted printing quality information is compared with the quality standard. If the prediction information shows that the printing quality is higher than the preset standard, it means that the current printing control parameters are reasonably configured and no adjustment is required; on the contrary, if the prediction information shows that the printing quality is lower than the standard, it means that the existing control parameters fail to meet the task requirements and need to be dynamically adjusted to optimize the printing results.

[0089] Furthermore, in a preferred embodiment of the present invention, if the printing quality standard is not met, the current printing control parameters are optimized, and a printing parameter control scheme is generated for control, which specifically includes:

[0090] Acquire printing quality prediction information, and search in a historical printing instance data set based on the printing quality prediction information to acquire historical printing control parameters corresponding to the historical printing instance under the current printing quality;

[0091] Acquire the real-time printing control parameters of the target visual printer, and calculate the real-time printing control parameters with the retrieved historical printing control parameters to obtain the printing control parameter deviation;

[0092] Add the printing control parameter deviation to the real-time printing control parameter to obtain the added printing control parameter, and construct a restriction interval with the real-time printing control parameter as the lower limit and the added printing control parameter as the upper limit;

[0093] According to the restriction interval, historical printing instances whose historical printing control parameters are within the restriction interval are retrieved from the historical printing instance data set and a search space is formed, a sparrow search algorithm is introduced, and an initial sparrow population is randomly generated using the search space;

[0094] Calculating the fitness value of each individual sparrow in the initial sparrow population, performing fitness sorting according to the fitness value to obtain a fitness sorting table, selecting a corresponding number of individuals as discoverers based on a preset discoverer ratio, and taking the remaining individuals as followers to update their positions;

[0095] The position of the optimal sparrow individual is obtained by repeated iterative updating, the current optimal solution is output, and a number of candidate printing control parameters are generated according to the output current optimal solution.

[0096] It should be noted that when the predicted real-time printing quality does not meet the standards, the printing control parameters need to be adjusted to obtain a printed product that meets the standards. First, based on the obtained printing quality prediction information, a search is performed in the historical printing instance data set to determine the historical printing instances corresponding to the current quality prediction information, and the historical printing control parameters in these historical instances are extracted to find reference parameters that can match the current quality performance. Next, by monitoring the real-time printing control parameters of the target visual printer, the real-time printing control parameters are calculated with the retrieved historical printing control parameters to obtain the deviation between the current control settings and the historical reference parameters, and generate the printing control parameter deviation. In order to guide the optimization direction, the deviation value is added to the real-time printing control parameters to calculate the added printing control parameters. On this basis, the limit interval of the printing control parameters is constructed with the real-time printing control parameters as the lower limit and the added printing control parameters as the upper limit. Using the above-mentioned restricted interval, further search is conducted in the historical printing instance data set to screen out all historical instances whose historical printing control parameters fall within this interval, and the search space is constructed based on this. The sparrow search algorithm is introduced, and the constructed search space is used as the input of the algorithm to randomly generate a population consisting of multiple sparrow individuals, which represent different parameter combination possibilities. By calculating the fitness value of each sparrow individual, the potential optimization effect of each parameter combination is evaluated, and the fitness is sorted according to the fitness to generate a fitness sorting table. In the optimization process, according to the rules of the sparrow search algorithm, individuals with high fitness rankings are selected as discoverers based on the preset discoverer ratio. These discoverers adjust their positions through exploratory update rules to expand the diversity of the search area. At the same time, the remaining individuals with lower fitness are used as followers, and their positions are adjusted using the follow-up update rule to follow a better solution. In each iteration, all individuals continuously update their positions according to the dynamic changes of fitness, and gradually converge to the global optimal solution, and finally output the position of the current optimal sparrow individual, that is, the optimal printing control parameter combination in the current search space, and generate several candidate printing control parameters.

[0097] Furthermore, in a preferred embodiment of the present invention, if the printing quality standard is not met, the current printing control parameters are optimized, and a printing parameter control scheme is generated for control, which also includes:

[0098] Acquire candidate printing control parameters, and analyze printing pressure data when the candidate printing control parameters are used by simulation software based on the candidate printing control parameters to obtain printing pressure simulation data;

[0099] Inputting the printing pressure simulation data into a printing quality prediction model, predicting the quality of the printed product under the corresponding printing pressure, and judging it against the printing quality standard of the current printing task;

[0100] If it is less than the printing quality standard of the current printing task, it means that the corresponding candidate printing control parameter is an unavailable parameter, and a restriction condition is generated based on the corresponding candidate printing control parameter, and the next candidate printing control parameter is obtained by re-iteration;

[0101] If it is greater than the printing quality standard of the current printing task, it means that the corresponding candidate printing control parameter is an available parameter, and the corresponding printing quality prediction result is calculated with the printing quality prediction result of the real-time printing task;

[0102] The calculated difference is used as an evaluation index to sort the candidate printing control parameters, and the candidate printing control parameter with the largest difference is selected as the final printing control parameter according to the sorting result;

[0103] The final printing control parameters are subtracted from the real-time printing control parameters, and a printing parameter control scheme for the target visual printing machine is generated based on the calculation results.

[0104] It should be noted that after determining the candidate printing control parameters, firstly, the simulation software is used to analyze these candidate parameters, and the printing pressure performance when applied to the target visual printing machine is simulated, so as to generate the corresponding printing pressure simulation data. The printing pressure data obtained by simulation is input into the previously constructed printing quality prediction model for prediction, and the predicted value of the quality of the printed product when each set of candidate parameters is used is calculated. Next, the quality of the printed product obtained by each set of predictions is compared and judged with the quality standard of the current printing task. If the prediction result of a set of candidate parameters is less than the quality standard of the current task, it indicates that the set of candidate parameters cannot meet the task requirements and is an unavailable parameter. In this case, new constraints are generated according to the corresponding candidate parameters, the search space is expanded or corrected, and the next round of iteration is triggered to re-acquire new candidate printing control parameters. If the prediction result of a set of candidate parameters is higher than the printing quality standard, it means that the parameter can meet the current task requirements and is an available parameter. In order to further screen the optimal control scheme, the difference calculation is performed between the prediction results of these available parameters and the quality prediction results of the real-time printing task, so as to quantify the specific performance of each candidate parameter in improving printing quality. The difference of all candidate parameters is used as an evaluation index and sorted to intuitively show the advantages and disadvantages of each group of parameters. Finally, the candidate printing control parameter with the largest difference is selected according to the sorting result and regarded as the optimal parameter, that is, the final printing control parameter. In order to accurately adjust the parameters, the difference between the final parameter and the current real-time printing control parameter is calculated to clarify the amplitude and direction of the parameter adjustment and generate the printing parameter control plan of the target visual printing machine. This ensures that the printing machine is always in the optimal control state, improves printing quality and optimizes production efficiency.

[0105] Figure 2 A flow chart of a method for optimizing printing control parameters of a visual printing press based on pressure feedback analysis provided by one embodiment of the present invention;

[0106] like Figure 2 As shown, the present invention provides a flow chart of a method for optimizing printing control parameters of a visual printing press based on pressure feedback analysis, comprising:

[0107] S202, acquiring real-time printing pressure monitoring information, performing data preprocessing and time sequence processing on the real-time printing pressure monitoring information, and obtaining a real-time printing pressure data sequence;

[0108] S204, inputting into the printing quality prediction model, predicting the printing quality in the future time period under the current printing pressure, and judging with the printing quality standard of the current printing task, analyzing whether the printing control parameters need to be adjusted;

[0109] S206, if the printing quality prediction result is less than the printing quality standard, it means that the printing control parameters need to be adjusted, then based on the printing quality prediction information, a search is performed in the historical printing instance data set to obtain the historical printing control parameters corresponding to the historical printing instance under the current printing quality;

[0110] S208, calculating the real-time printing control parameter and the retrieved historical printing control parameter to obtain a printing control parameter deviation, and constructing a restriction interval according to the printing control parameter deviation;

[0111] S210, searching the historical printing instances whose historical printing control parameters are within the restricted interval in the historical printing instance data set according to the restricted interval and forming a search space, optimizing the printing control parameters by using a sparrow search algorithm, and obtaining candidate printing control parameters;

[0112] S212, predicting the quality of the printed product of the candidate printing control parameters, and selecting the candidate printing control parameters that meet the printing standards to formulate a printing parameter control plan.

[0113] It should be noted that when the visual printer performs a printing task, the printing pressure is monitored to obtain the corresponding real-time printing pressure monitoring information, and the collected real-time printing pressure monitoring information is preprocessed and time-series processed to obtain a real-time printing pressure data sequence that can reflect the trend of printing pressure changes. Subsequently, it is input into the printing quality prediction model to predict the printing quality performance in the future time period under the current printing pressure conditions. Based on the prediction results, it is compared and judged with the printing quality standard of the current printing task to determine whether the printing control parameters need to be adjusted. When the predicted printing quality result is lower than the quality standard, it means that the existing control parameters cannot meet the printing requirements and the printing control parameter optimization needs to be performed. At this time, the historical printing instance data set is retrieved through the printing quality prediction information to locate the historical printing instance matching the current predicted quality result, and the corresponding historical printing control parameters are extracted. The parameter restriction interval is further constructed to define the adjustment range. On this basis, the historical printing instance data set is used again for retrieval to filter out the instances whose historical printing control parameters fall within the restriction interval to form a new search space. The sparrow search algorithm is used to globally optimize the constructed search space, and the candidate printing control parameters that can maximize the printing quality are gradually determined through population optimization and iterative update. For each set of candidate printing control parameters, the quality of the printed product corresponding to the target printing machine is predicted and analyzed, and the parameter combination that meets the printing quality standard is screened out. Finally, based on these parameters that meet the standard, a printing parameter control plan is formulated to optimize the control parameter settings of the target visual printing machine in real time, thereby ensuring that the quality standards of the printing task are met and further optimizing the printing efficiency.

[0114] Figure 3 A pressure feedback analysis system 3 applied to a visual printer is provided in one embodiment of the present invention. The system comprises: a memory 31 and a processor 32. The memory 31 contains a pressure feedback analysis method program applied to a visual printer. When the pressure feedback analysis method program applied to a visual printer is executed by the processor 32, the following steps are implemented:

[0115] Obtain historical printing instances of the target model visual printing machine, extract features from each historical printing instance, evaluate the printing quality of each historical printing instance, and construct a historical printing instance dataset;

[0116] Analyzing the relationship between printing pressure, printing quality and printing control parameters based on the historical printing instance data set, and constructing a fusion topology structure diagram through the relationship between printing pressure, printing quality and printing control parameters;

[0117] Constructing a printing quality prediction model, training the model through the historical printing instance data set and the fusion topological structure diagram, obtaining real-time printing pressure monitoring information, predicting real-time printing quality and judging whether it meets the printing quality standard;

[0118] If the printing quality standard is not met, the current printing control parameters are optimized and a printing parameter control scheme is generated for control.

[0119] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0120] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0121] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0122] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0123] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0124] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A pressure feedback analysis method applied to a visual printing press, characterized in that: include: Obtain historical printing instances of the target model visual printing machine, extract features from each historical printing instance, evaluate the printing quality of each historical printing instance, and construct a historical printing instance dataset; Analyzing the relationship between printing pressure, printing quality and printing control parameters based on the historical printing instance data set, and constructing a fusion topology structure diagram through the relationship between printing pressure, printing quality and printing control parameters; Constructing a printing quality prediction model, training the model through the historical printing instance data set and the fusion topological structure diagram, obtaining real-time printing pressure monitoring information, predicting real-time printing quality and judging whether it meets the printing quality standard; If the printing quality standard is not met, the current printing control parameters are optimized and a printing parameter control scheme is generated for control; The step of constructing a fusion topology structure diagram specifically includes: Acquire a historical printing instance data set, extract printing pressure data corresponding to each historical printing instance according to the historical printing instance data set, and perform time series processing to obtain a printing pressure data sequence; Acquire printing quality evaluation information, and generate a printing image quality score sequence when each historical printing instance performs printing work at different printing time stages according to the printing quality evaluation information; Perform time-series matching on the printing pressure data sequence and the printing image quality score sequence, introduce a decision tree algorithm, take the printing pressure as the decision attribute and the printing image quality as the target attribute, and initialize the root node; According to the initialized root node, a plurality of branches are obtained by splitting based on the range of different printing pressure data, and a hierarchical structure of the decision tree is gradually constructed, and the best splitting node is selected according to the information gain, and a first topological structure diagram is generated according to the constructed decision tree structure; Subsequently, the printing control parameters corresponding to each historical printing instance are extracted through the historical printing instance data set, and the printing control parameters are used as decision attributes and the printing pressure is used as the target attribute. A decision tree is constructed to analyze the association between the printing control parameters and the printing pressure data, and a second topological structure diagram is generated; The first topology structure graph and the second topology structure graph are associated and fused to generate a fused topology structure graph.

2. The pressure feedback analysis method for a visual printer according to claim 1, characterized in that: The method of obtaining historical printing instances of a target model visual printing machine, extracting features of each historical printing instance, evaluating the printing quality of each historical printing instance, and constructing a historical printing instance data set specifically includes: Based on the big data network, historical printing instances of the target model visual printing machine are obtained, including the printed product images, printing control parameters and printing pressure data, to form a historical printing instance data set; Building a feature pyramid network, extracting the printed product image corresponding to each instance according to the acquired historical printing instances, inputting the image into the feature pyramid network, and performing a dilated convolution on the input printed product image according to a preset dilated rate; The feature maps outputted by each branch after the dilated convolution are obtained and horizontally connected to obtain the fused feature maps corresponding to each printed product image, and the image feature time series corresponding to each printed product image is generated according to the preset timestamp; The printing image quality evaluation index is set, and the weight of each printing image quality evaluation index is assigned using the intuitive fuzzy analysis method. The printing quality evaluation model is constructed by combining the weight assignment results and the printing image quality evaluation index. Inputting the image feature time series corresponding to each printed product into the printing quality evaluation model, analyzing the printing image quality when each historical printing instance performs printing work at different printing time stages, and obtaining printing quality evaluation information; The historical printing instance dataset is updated by using the printing quality evaluation information as an attached feature corresponding to the historical printing instance.

3. The pressure feedback analysis method for a visual printer according to claim 1, characterized in that: The construction of the printing quality prediction model, performing model training through the historical printing instance data set and the fusion topology structure diagram, obtaining real-time printing pressure monitoring information, predicting the real-time printing quality and judging whether it meets the printing quality standard, specifically includes: Acquire a fused topological structure diagram, and extract a relationship matrix between printing image quality, printing pressure data, and printing control parameters based on the fused topological structure diagram; A printing quality prediction model is constructed using a deep neural network, the relationship matrix is ​​input into the prediction model for coding learning, and model training and parameter adjustment are performed in combination with a historical printing instance data set to obtain a printing quality prediction model that meets expectations; Monitor the printing pressure of the target visual printer when performing a printing task, obtain real-time printing pressure monitoring information, perform data preprocessing on the real-time printing pressure monitoring information, and perform time series processing on the preprocessed real-time printing pressure data; Acquire a real-time printing pressure data sequence, input it into the printing quality prediction model, analyze the printing quality in a future time period under the current printing pressure, and obtain printing quality prediction information; Acquire the printing quality standard of the current printing task, judge the printing quality prediction information with the printing quality standard of the current printing task, and analyze whether it is necessary to adjust the printing control parameters; If the printing quality prediction information is greater than the printing quality standard, it means that the current printing control parameters do not need to be adjusted; if the printing quality prediction information is less than the printing quality standard, it means that the printing control parameters need to be adjusted.

4. The pressure feedback analysis method for a visual printer according to claim 1, characterized in that: If the printing quality standard is not met, the current printing control parameters are optimized and a printing parameter control scheme is generated for control, which specifically includes: Acquire printing quality prediction information, and search in a historical printing instance data set based on the printing quality prediction information to acquire historical printing control parameters corresponding to the historical printing instance under the current printing quality; Acquire the real-time printing control parameters of the target visual printer, and calculate the real-time printing control parameters with the retrieved historical printing control parameters to obtain the printing control parameter deviation; Add the printing control parameter deviation to the real-time printing control parameter to obtain the added printing control parameter, and construct a restriction interval with the real-time printing control parameter as the lower limit and the added printing control parameter as the upper limit; According to the restriction interval, historical printing instances whose historical printing control parameters are within the restriction interval are retrieved from the historical printing instance data set and a search space is formed, a sparrow search algorithm is introduced, and an initial sparrow population is randomly generated using the search space; Calculating the fitness value of each individual sparrow in the initial sparrow population, performing fitness sorting according to the fitness value to obtain a fitness sorting table, selecting a corresponding number of individuals as discoverers based on a preset discoverer ratio, and taking the remaining individuals as followers to update their positions; The position of the optimal sparrow individual is obtained by repeated iterative updating, the current optimal solution is output, and a number of candidate printing control parameters are generated according to the output current optimal solution.

5. The pressure feedback analysis method for a visual printer according to claim 1, characterized in that: If the printing quality standard is not met, the current printing control parameters are optimized, and a printing parameter control scheme is generated for control, which also includes: Acquire candidate printing control parameters, and analyze printing pressure data when the candidate printing control parameters are used by simulation software based on the candidate printing control parameters to obtain printing pressure simulation data; Inputting the printing pressure simulation data into a printing quality prediction model, predicting the quality of the printed product under the corresponding printing pressure, and judging it against the printing quality standard of the current printing task; If it is less than the printing quality standard of the current printing task, it means that the corresponding candidate printing control parameter is an unavailable parameter, and a restriction condition is generated based on the corresponding candidate printing control parameter, and the next candidate printing control parameter is obtained by re-iteration; If it is greater than the printing quality standard of the current printing task, it means that the corresponding candidate printing control parameter is an available parameter, and the corresponding printing quality prediction result is calculated with the printing quality prediction result of the real-time printing task; The calculated difference is used as an evaluation index to sort the candidate printing control parameters, and the candidate printing control parameter with the largest difference is selected as the final printing control parameter according to the sorting result; The final printing control parameters are subtracted from the real-time printing control parameters, and a printing parameter control scheme for the target visual printing machine is generated based on the calculation results.

6. A pressure feedback analysis system for a visual printing press, characterized in that: The system includes: a memory and a processor, wherein the memory contains a pressure feedback analysis method program applied to a visual printer, and when the pressure feedback analysis method program applied to a visual printer is executed by the processor, the steps of a pressure feedback analysis method applied to a visual printer as described in any one of claims 1-5 are implemented.

Images