Methods, devices, equipment and storage media for ink supply control of printing inks

Abstract

This invention relates to the field of artificial intelligence and discloses a method, apparatus, device, and storage medium for controlling the ink supply of printing ink, used to achieve intelligent ink supply control and improve the accuracy of ink supply control. The method includes: acquiring ink flow data of a target printing device and extracting ink attribute features to obtain multiple ink attribute features; acquiring printing standard data of a target item and creating a first ink supply control parameter combination; performing a printing operation on the target item and acquiring optical images of the target item through a preset optical sensor to obtain optical image data; inputting the optical image data into a printing image detection model for printing feature detection to obtain multiple ink optical features and multiple ink coating features; and optimizing the first ink supply control parameter combination based on the multiple ink optical features and multiple ink coating features to generate a corresponding second ink supply control parameter combination.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of artificial intelligence, and in particular to a printing ink supply control method, device, equipment and storage medium. BACKGROUND

[0002] With the rapid development of digital printing technology, the printing process has become more automated and precise. Digital printing equipment can monitor and adjust printing parameters in real time, but effective ink supply control is still needed to ensure printing quality. In the printing industry, ink supply control is one of the key quality control steps. In the printing process, the flow, color, viscosity and other properties of the ink must be precisely controlled to ensure that the output printed matter meets the standards and quality requirements.

[0003] Traditionally, the ink supply control process mainly relies on the experience and trial-and-error method of the operator, which can lead to inconsistent printing quality and waste of ink, resulting in low control accuracy of the existing solution. SUMMARY

[0004] The present application provides a printing ink supply control method, device, equipment and storage medium, which is used to realize intelligent printing ink supply control and improve the accuracy of ink supply control.

[0005] The first aspect of the present application provides a printing ink supply control method, which comprises:

[0006] Obtain ink flow data of printing ink in a target printing device through a pre-set flow sensor, and extract ink attribute features from the ink flow data to obtain a plurality of ink attribute features;

[0007] Obtain printing standard data of a target object, and create a first ink supply control parameter combination for the target printing device according to the plurality of ink attribute features and the printing standard data;

[0008] Perform a printing operation on the target object according to the first ink supply control parameter combination through the target printing device, and perform optical image acquisition on the target object through a pre-set optical sensor to obtain optical image data;

[0009] Input the optical image data into a pre-set printing image detection model to detect printing features, and obtain a plurality of ink optical features and a plurality of ink coating features;

[0010] Optimize the first ink supply control parameter combination based on the plurality of ink optical features and the plurality of ink coating features to generate a corresponding second ink supply control parameter combination.

[0011] In a first implementation manner of the first aspect, the ink flow data of the printing ink in the target printing device is acquired by a preset flow sensor, and ink attribute feature extraction is performed on the ink flow data to obtain a plurality of ink attribute features, including:

[0012] The ink flow data of the printing ink in the target printing device is acquired by a preset flow sensor, and data discretization is performed on the ink flow data to obtain discretized flow data;

[0013] The discretized flow data is subjected to matrix conversion to generate a target flow matrix, wherein each row in the target flow matrix represents data at a time point, and each column represents a different flow observation point;

[0014] A target covariance matrix corresponding to the target flow matrix is calculated, and eigenvalue decomposition is performed on the target covariance matrix to obtain eigenvalues and corresponding eigenvectors;

[0015] According to the sizes of the eigenvalues, N principal components are selected, and the discretized flow data is projected into a new low-dimensional space based on the N principal components to obtain a plurality of ink attribute features.

[0016] In a second implementation manner of the first aspect, the printing standard data of the target object is acquired, and a first ink supply control parameter combination of the target printing device is created based on the plurality of ink attribute features and the printing standard data, including:

[0017] The printing standard data of the target object is acquired, and standard feature extraction is performed on the printing standard data to obtain a plurality of printing standard features;

[0018] The feature correlation degrees between the plurality of printing standard features and the plurality of ink attribute features are respectively calculated;

[0019] According to the feature correlation degrees, a decision tree model between the plurality of printing standard features and the plurality of ink attribute features is constructed;

[0020] The first ink supply control parameter combination of the target printing device is created based on the decision tree model.

[0021] In a third implementation manner of the first aspect, the target object is subjected to a printing operation by the target printing device according to the first ink supply control parameter combination, and optical image data is obtained by performing optical image acquisition on the target object by a preset optical sensor, including:

[0022] According to the first ink supply control parameter combination, a parameter of the target printing device is set;

[0023] The target printing device is set by the parameter, and a printing operation is performed on the target object, and an optical positioning operation is performed on the target object by a preset optical sensor to obtain spatial position information of the target object;

[0024] Based on the spatial position information, optical image acquisition is performed on the target object by the optical sensor to obtain initial image data;

[0025] The initial image data is input into an optical image integrity analysis model, wherein the optical image integrity analysis model includes a residual layer, a convolution layer, a pooling layer, a full connection layer and a softmax function;

[0026] The optical image integrity analysis model is used to analyze the optical image integrity of the initial image data to obtain optical image integrity analysis data;

[0027] The optical image integrity analysis data and the spatial position information are checked to generate an image checking result;

[0028] According to the image checking result, the initial image data is preprocessed to obtain optical image data.

[0029] In combination with the first aspect, in a fourth implementation manner of the first aspect of the present application, the optical image data is input into a preset printing image detection model to detect printing features to obtain a plurality of ink optical features and a plurality of ink coating features, including:

[0030] The optical image data is input into a preset printing image detection model, wherein the printing image detection model includes a first convolution network, a first residual network, a first inverse residual network, a second residual network, a second inverse residual network and a second convolution network;

[0031] The first convolution network is used to perform convolution operation on the optical image data to obtain a first convolution feature map;

[0032] The first convolution feature map is input into the first residual network to extract optical features to obtain a first optical feature map, and the first inverse residual network is used to perform feature operation on the first optical feature map to obtain a second optical feature map;

[0033] The first convolution feature map is input into the second residual network to extract coating features to obtain a first coating feature map, and the second inverse residual network is used to perform feature operation on the first coating feature map to obtain a second coating feature map;

[0034] The second convolutional network outputs features from the second optical feature map and the second coating feature map to obtain multiple ink optical features and multiple ink coating features.

[0035] In conjunction with the first aspect, in a fifth implementation of the first aspect of the present invention, the step of optimizing the first ink supply control parameter combination based on the plurality of ink optical features and the plurality of ink coating features to generate a corresponding second ink supply control parameter combination includes:

[0036] Based on the multiple ink optical features and the multiple ink coating features, a corresponding target ink feature matrix is ​​generated;

[0037] Using the entropy weight evaluation model, the positive and negative ideal solutions for each element of the target ink feature matrix are calculated respectively.

[0038] Based on the positive ideal solution and the negative ideal solution, calculate the comprehensive score for each element of the feature matrix;

[0039] Based on the comprehensive score of each feature matrix element, the first ink supply control parameter combination is optimized using a preset genetic algorithm to generate the corresponding second ink supply control parameter combination.

[0040] In conjunction with the first aspect, in the sixth implementation of the first aspect of the present invention, the step of optimizing the first ink supply control parameter combination based on the comprehensive score of each feature matrix element using a preset genetic algorithm to generate a corresponding second ink supply control parameter combination includes:

[0041] Based on the comprehensive score of each feature matrix element, C candidate ink supply control parameter combinations corresponding to the first ink supply control parameter combination are created using a preset genetic algorithm.

[0042] Calculate the first fitness data for each candidate ink supply control parameter combination, and perform breeding, crossover, and mutation operations on the C candidate ink supply control parameter combinations based on the first fitness data to generate D candidate ink supply control parameter combinations, where C < D;

[0043] Calculate the second fitness data for each of the D candidate ink supply control parameter combinations, and select the corresponding second ink supply control parameter combination from the D candidate ink supply control parameter combinations based on the second fitness data.

[0044] A second aspect of the present invention provides an ink supply control device for printing ink, the ink supply control device for printing ink comprising:

[0045] The acquisition module is used to acquire ink flow data of printing ink in the target printing equipment through a preset flow sensor, and to extract ink attribute features from the ink flow data to obtain multiple ink attribute features.

[0046] A module is created to acquire printing standard data of the target item and create a first ink supply control parameter combination for the target printing equipment based on the multiple ink property characteristics and the printing standard data.

[0047] The acquisition module is used to perform printing operations on the target item through the target printing equipment according to the first ink supply control parameter combination, and to acquire optical images of the target item through a preset optical sensor to obtain optical image data;

[0048] The detection module is used to input the optical image data into a preset printing image detection model to detect printing features and obtain multiple ink optical features and multiple ink coating features.

[0049] The optimization module is used to optimize the first ink supply control parameter combination based on the multiple ink optical characteristics and the multiple ink coating characteristics, and generate a corresponding second ink supply control parameter combination.

[0050] A third aspect of the present invention provides an ink supply control device for printing ink, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor invokes the instructions in the memory to cause the ink supply control device for printing ink to execute the above-described ink supply control method for printing ink.

[0051] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described ink supply control method for printing ink.

[0052] The technical solution provided by this invention involves acquiring ink flow data of the target printing equipment and extracting ink attribute features to obtain multiple ink attribute features; acquiring printing standard data of the target item and creating a first ink supply control parameter combination; performing a printing operation on the target item and acquiring optical images of the target item through a preset optical sensor to obtain optical image data; inputting the optical image data into a printing image detection model for printing feature detection to obtain multiple ink optical features and multiple ink coating features; optimizing the first ink supply control parameter combination based on the multiple ink optical features and multiple ink coating features to generate a corresponding second ink supply control parameter combination. This invention, through the use of flow sensors and data processing, enables real-time monitoring and control of ink flow. This ensures precise ink supply, conforming to printing standards and reducing scrap rates. Multiple ink attribute features are extracted, including flow rate, ink color, and viscosity. This allows operators to have a more comprehensive understanding of the ink properties, facilitating more precise ink supply control. By analyzing the correlation between printing standard data and ink attribute features, the method can establish a decision tree model to link the two. This helps to automatically create the first ink supply control parameter combination based on the printing standard data. Image data, including information on color accuracy and sharpness, is captured in real-time during the printing process using optical sensors. This allows for timely detection of problems and the implementation of corrective measures. A printing image detection model automatically detects and extracts multiple ink optical and coating features. This eliminates the need for manual inspection and improves the efficiency of the printing process. Based on these ink optical and coating features, the method uses an entropy weight evaluation model and a genetic algorithm for parameter optimization. This ensures the optimization of ink supply control parameters for better print quality. This invention is not merely a one-time control measure but also includes a mechanism for continuous monitoring and adjustment. This allows the printing process to be continuously improved to adapt to different printing needs and conditions, thereby achieving intelligent ink supply control and improving its accuracy. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of one embodiment of the ink supply control method for printing ink in this invention.

[0054] Figure 2 This is a flowchart illustrating the creation of the first ink supply control parameter combination in an embodiment of the present invention;

[0055] Figure 3 This is a flowchart of optical image acquisition in an embodiment of the present invention;

[0056] Figure 4 This is a flowchart of printing feature detection in an embodiment of the present invention;

[0057] Figure 5This is a schematic diagram of one embodiment of the ink supply control device for printing ink in this invention.

[0058] Figure 6 This is a schematic diagram of one embodiment of the ink supply control device for printing ink in this invention. Detailed Implementation

[0059] This invention provides a method, apparatus, device, and storage medium for controlling the ink supply of printing ink, enabling intelligent ink supply control and improving the accuracy of ink supply control. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0060] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 One embodiment of the ink supply control method for printing ink in this invention includes:

[0061] S101. Obtain ink flow data of printing ink in the target printing equipment through a preset flow sensor, and extract ink attribute features from the ink flow data to obtain multiple ink attribute features.

[0062] It is understood that the executing entity of this invention can be an ink supply control device for printing ink, or it can be a terminal or a server; no specific limitation is made here. This embodiment of the invention will be described using a server as an example.

[0063] Specifically, the server uses pre-installed flow sensors to acquire ink flow data from the target printing equipment. This data includes information on ink flow variations over time. The acquired ink flow data is then discretized. Continuous flow data is divided into discrete time points or time intervals for easier processing and analysis. This step helps organize the data into a time series. Matrix transformation is performed on the discretized flow data to generate the target flow matrix. In the target flow matrix, each row represents data at a single time point, and each column represents a different flow observation point. This helps organize the data in matrix form for further analysis. The covariance matrix of the target flow matrix is ​​calculated. The covariance matrix measures the correlation and trend of change between different flow observation points. Eigenvalue decomposition is performed on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors. Eigenvalues ​​and eigenvectors provide important information about data variation. Based on the magnitude of the eigenvalues, N principal components are selected. Principal components are the eigenvectors of the covariance matrix; they represent the main directions of change in the data. Selecting N principal components can be used to reduce the dimensionality of the data, projecting the data into a new, lower-dimensional space. This helps simplify the data and reduce noise. Based on N selected principal components, the discretized flow data is projected into a new low-dimensional space, resulting in multiple ink attribute features. These features reflect the key attributes and trends of the ink flow data and can be used for further analysis and control of the ink supply process in printing. For example, suppose a digital printing press needs to ensure that the ink flow rate for each color is within the appropriate range to achieve the desired color quality. The server uses flow sensors to acquire flow data for various color inks and discretizes it. Through matrix transformation and eigenvalue decomposition, the server determines which flow observations are correlated and which trends are the most important. After selecting an appropriate number of principal components, the server projects the flow data into a low-dimensional space, obtaining multiple ink attribute features that can be used to dynamically adjust the ink supply of the printing press to maintain consistent print quality. This approach can help digital printing equipment to be more precise and automated in real-time monitoring and adjustment of printing parameters to meet the requirements of high-quality printing.

[0064] S102. Obtain the printing standard data of the target item, and create a first ink supply control parameter combination for the target printing equipment based on multiple ink property characteristics and the printing standard data.

[0065] Specifically, the server acquires the printing standard data for the target item. This printing standard data includes information such as the specifications, quality requirements, and color requirements of the printed item. This data is crucial to ensuring that the printing process meets specific standards. Standard feature extraction is performed on the printing standard data. Key features are extracted from the printing standard data to be combined with ink attribute features. For example, if the printing standard data includes color requirements, then color specifications can be a standard feature. Feature correlations are calculated between multiple printing standard features and multiple ink attribute features. This step is used to determine which standard features and ink attribute features are correlated. Correlation analysis helps the server understand the relationship between printing standards and ink attributes. Based on the feature correlations, a decision tree model is constructed. A decision tree model is a machine learning model used for classification and decision-making, making decisions based on input features. The server uses feature correlations as input features to construct a decision tree model that makes decisions based on the correlation between printing standard features and ink attribute features. Based on the decision tree model, a first combination of ink supply control parameters for the target printing equipment is created. The decision tree model can be used to determine appropriate ink supply control parameters based on the input of printing standard data and ink attribute features. These parameters can include ink flow rate, color adjustment, printing speed, etc., to ensure the printing process meets printing standard requirements. For example, suppose the printing standard data includes the book's size, color requirements, and quality requirements. The server also obtains ink property characteristics from the flow sensor, including color density and flow rate variations. The server extracts features from the printing standard data, such as the book's size and color specifications. The server calculates the correlation between these standard features and ink property characteristics (such as color density). If a high correlation is found between color density and color requirements, this can be reflected using a decision tree model. By constructing a decision tree model, the server creates an initial combination of ink supply control parameters for the printing equipment based on the input printing standard data and ink property characteristics. For example, if the book requires a specific color, the decision tree model can suggest appropriate ink flow rate and color adjustment to ensure the book's printing quality meets requirements.

[0066] S103. Using the target printing equipment, the target item is printed according to the first ink supply control parameter combination, and the target item is optically imaged by a preset optical sensor to obtain optical image data.

[0067] It should be noted that the target printing equipment is parameter-set according to the first ink supply control parameter combination. These parameters include ink flow rate, printing speed, printing pressure, etc., to ensure that the printing equipment operates in the required manner. The target printing equipment, with the parameter settings completed, performs the printing operation on the target item. During the printing process, a preset optical sensor simultaneously performs optical positioning to acquire the spatial position information of the target item. This helps track the item's position and position the printing. Based on the acquired spatial position information, optical images of the target item are acquired through the optical sensor. The image data obtained in this step includes information on the appearance and quality of the printed item. Optical image integrity analysis is performed on the acquired initial image data. This analysis typically includes several important steps, such as the application of residual layers, convolutional layers, pooling layers, fully connected layers, and the softmax function. These layers are used in deep learning neural networks to extract and analyze image features to evaluate image integrity. After analyzing the initial image data through the optical image integrity analysis model, optical image integrity analysis data is obtained. This data provides visual characteristics and quality information about the printed item and can be used to detect any defects or incompleteness in the image. The optical image integrity analysis data is then verified against the previously acquired spatial position information. This verification process aims to ensure that the analyzed data matches the location and orientation of the item, preventing misjudgments. Based on the image verification results, the initial image data undergoes image preprocessing. Image preprocessing may include operations such as noise reduction, color correction, and image enhancement to ultimately obtain optical image data that meets printing standards and has good visual quality.

[0068] S104. Input the optical image data into the preset printing image detection model to perform printing feature detection, and obtain multiple ink optical features and multiple ink coating features.

[0069] Specifically, the optical image data is input into a pre-set printing image detection model, which typically contains multiple layers, such as a first convolutional network, a first residual network, a first inverse residual network, a second residual network, a second inverse residual network, and a second convolutional network. These network layers together constitute a deep learning neural network for processing and analyzing the optical image data. The acquired optical image data is input into the first convolutional network for convolution operations, resulting in a first convolutional feature map. This convolution operation helps extract local features from the image. The first convolutional feature map is then passed to the first residual network for optical feature extraction. The residual network is a deep learning structure that can effectively learn and extract high-level features of the image, which often contain more abstract and complex information. This step yields the first optical feature map. Feature operations are then performed on the first optical feature map through the first inverse residual network to obtain a second optical feature map. The purpose of feature operations is to further enhance the representation and extraction of optical features. Simultaneously, the first convolutional feature map is passed to the second residual network for coating feature extraction. This step yields the first coating feature map. A second coating feature map is obtained by performing feature operations on the first coating feature map using a second inverse residual network. These feature operations help to further optimize the representation of the coating features. A second convolutional network is then used to output features from both the second optical feature map and the second coating feature map, resulting in multiple ink optical features and multiple ink coating features. These features describe printing characteristics, color consistency, sharpness, and coating uniformity in the image, aiding in the evaluation of printing and coating quality.

[0070] S105. Based on multiple ink optical characteristics and multiple ink coating characteristics, the first ink supply control parameter combination is optimized to generate the corresponding second ink supply control parameter combination.

[0071] Specifically, based on multiple ink optical and coating characteristics that reflect the quality of the printed matter and the ink distribution, these characteristics are integrated into a target ink feature matrix, where each row represents a printed matter and each column represents a feature. An entropy-weighted evaluation model is used to assess the importance of each feature, and positive and negative ideal solutions are calculated. These solutions are used for subsequent parameter optimization. For each feature, the positive ideal solution is the optimal value for that feature, used to maximize print quality. A positive ideal solution can be a specific target value or a range; conversely, a negative ideal solution is the worst value for that feature, representing minimizing the feature's adverse impact on print quality. A negative ideal solution can also be a specific target value or a range. Using the positive and negative ideal solutions, a comprehensive score can be calculated for each feature. The comprehensive score measures the performance of each feature in relation to the optimization objective. The score can be calculated using various methods, such as weighted summation or other custom functions. Based on the scores and the comprehensive score, a pre-defined genetic algorithm is used to optimize the first combination of ink supply control parameters. The genetic algorithm is a heuristic search algorithm used to find the optimal solution to a multi-objective optimization problem. In this scenario, the genetic algorithm attempts different parameter combinations to maximize the overall score of each feature and generates a second set of ink supply control parameters. For example, the server captures images of each printed item using optical sensors and extracts multiple ink optical and coating features. One optical feature might be color consistency, while a coating feature could be ink uniformity. Each feature has a positive ideal solution representing the target value under optimal conditions and a negative ideal solution representing the target value under worst conditions. Using an entropy-weighted evaluation model, the server calculates the importance of each feature and assigns a weight to each. The server calculates the overall score for each feature, reflecting the overall performance of print quality. The server applies the genetic algorithm to optimize the printing press's ink supply parameters to maximize print quality. The genetic algorithm searches through different parameter combinations and selects the optimal second set of ink supply control parameters to ensure high-quality and consistent print quality. This process automates print quality optimization, improving production efficiency and print quality.

[0072] Based on the comprehensive score of each feature matrix element, a pre-defined genetic algorithm is used to create C candidate ink supply control parameter combinations corresponding to the first ink supply control parameter combination. These candidate combinations serve as the starting population for the genetic algorithm. For each candidate ink supply control parameter combination, its first fitness data is calculated. The first fitness data is determined based on the comprehensive score of each candidate combination's feature matrix elements. This score represents the performance of each candidate combination in terms of various features. Using the first fitness data, the genetic algorithm operates on the C candidate ink supply control parameter combinations, including breeding, crossover, and mutation. These operations aim to generate D new candidate ink supply control parameter combinations, where D is greater than C. This process helps explore a wider parameter space to find better solutions. Breeding creates new combinations by selecting candidate combinations with higher first fitness data, which helps retain better-performing solutions; crossover involves exchanging some parameters of two candidate combinations to produce new combinations. This helps introduce diversity and accelerate the search process; mutation involves making small changes to the parameters of some candidate combinations to introduce randomness, which helps to escape local optima. For the generated D candidate ink supply control parameter combinations, their second fitness data is calculated. The second fitness data relates to print quality, ink efficiency, or other relevant metrics. This score is used to evaluate the performance of each candidate combination in actual production. After calculating the second fitness data, the combination with the highest second fitness data is selected from D candidate ink supply control parameter combinations. This combination becomes the second ink supply control parameter combination used in actual printing operations. For example, suppose the server creates an initial population containing C candidate ink supply control parameter combinations. By calculating the combined score of the feature matrix elements of each candidate combination, the server determines the first fitness data. The server runs a genetic algorithm, performing breeding, crossover, and mutation operations on these combinations to generate D new candidate combinations. The server calculates the second fitness data for each new combination, which relates to metrics such as print quality and production efficiency. The server selects the combination with the highest second fitness data as the second ink supply control parameter combination. This process allows digital printing plants to optimize their ink supply control parameters to achieve higher print quality and lower ink consumption in different printing jobs. The use of a genetic algorithm helps the server find the optimal parameter combination, improving production efficiency and print quality.

[0073] In this embodiment of the invention, ink flow data of the target printing equipment is acquired and ink attribute features are extracted to obtain multiple ink attribute features; printing standard data of the target item is acquired and a first ink supply control parameter combination is created; printing operation is performed on the target item, and optical image acquisition is performed on the target item through a preset optical sensor to obtain optical image data; the optical image data is input into a printing image detection model for printing feature detection to obtain multiple ink optical features and multiple ink coating features; based on the multiple ink optical features and multiple ink coating features, the first ink supply control parameter combination is optimized to generate a corresponding second ink supply control parameter combination. This invention, through the use of flow sensors and data processing, enables real-time monitoring and control of ink flow. This ensures precise ink supply, conforming to printing standards, thereby reducing scrap rates. Multiple ink attribute features are extracted, including flow rate, ink color, and viscosity. This allows operators to have a more comprehensive understanding of the ink properties, facilitating more precise ink supply control. By analyzing the correlation between printing standard data and ink attribute features, the method can establish a decision tree model to link the two. This helps to automatically create the first ink supply control parameter combination based on the printing standard data. Image data, including information on color accuracy and sharpness, is captured in real-time during the printing process using optical sensors. This allows for timely detection of problems and the implementation of corrective measures. A printing image detection model automatically detects and extracts multiple ink optical and coating features. This eliminates the need for manual inspection and improves the efficiency of the printing process. Based on these ink optical and coating features, the method uses an entropy weight evaluation model and a genetic algorithm for parameter optimization. This ensures the optimization of ink supply control parameters for better print quality. This invention is not merely a one-time control measure but also includes a mechanism for continuous monitoring and adjustment. This allows the printing process to be continuously improved to adapt to different printing needs and conditions, thereby achieving intelligent ink supply control and improving its accuracy.

[0074] In one specific embodiment, the process of performing step S101 may specifically include the following steps:

[0075] (1) The ink flow data of the printing ink in the target printing equipment is obtained by using a preset flow sensor, and the ink flow data is discretized to obtain discretized flow data.

[0076] (2) Perform matrix transformation on the discretized flow data to generate the target flow matrix, where each row of the target flow matrix represents data at a time point and each column represents a different flow observation point;

[0077] (3) Calculate the target covariance matrix corresponding to the target flow matrix, and perform eigenvalue decomposition on the target covariance matrix to obtain the eigenvalues ​​and corresponding eigenvectors;

[0078] (4) Select N principal components based on the magnitude of the eigenvalues, and project the discretized flow data into a new low-dimensional space based on the N principal components to obtain multiple ink attribute features.

[0079] Specifically, the flow rate of printing ink is monitored in real time within the target printing equipment using pre-installed flow sensors. These sensors measure the flow rate and volume of ink, typically recording it as a time series. The acquired flow data is usually continuous, and needs to be discretized for subsequent analysis. This can be achieved by sampling along a time axis or periodically recording the flow data. The continuous flow data is divided into different time points, and the flow value at each time point is recorded. This yields discretized flow data. The discretized flow data is then organized into a matrix, where each row represents data at a single time point, and each column represents a different flow observation point. This matrix is ​​called the target flow matrix. Each element represents the ink flow value at a specific time point and a specific observation point. For example, suppose a digital printing press contains three different ink supply points (observation points) and records ink flow data over a continuous time period. Based on this data, the server creates the target flow matrix. Using the generated target flow matrix, a covariance matrix is ​​calculated, reflecting the correlation between different flow observation points. Eigenvalue decomposition is performed on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors. Eigenvalue decomposition is a linear algebraic technique that decomposes the covariance matrix into combinations of eigenvalues ​​and eigenvectors. Eigenvectors describe the principal directions of the covariance matrix, while eigenvalues ​​represent the magnitude of the variance along those directions. Based on the magnitude of the eigenvalues, the top N principal components are selected. Generally, the larger the eigenvalue, the more important the corresponding principal component is in describing the data's variation. Therefore, the eigenvectors corresponding to the top N eigenvalues, i.e., the principal components, are selected to describe the main trends in data variation. These principal components can be used to represent a low-dimensional representation of the original data, thereby extracting multiple ink attribute features. Principal components are typically linear combinations, which can be used to reduce the dimensionality of the data while retaining most of its information.

[0080] In one specific embodiment, such as Figure 2 As shown, the process of executing step S102 can specifically include the following steps:

[0081] S201. Obtain the printing standard data of the target item, and extract standard features from the printing standard data to obtain multiple printing standard features.

[0082] S202. Calculate the feature correlation degree between multiple printing standard features and multiple ink attribute features respectively;

[0083] S203. Based on the feature correlation degree, construct a decision tree model between multiple printing standard features and multiple ink attribute features;

[0084] S204. Based on the decision tree model, create the first ink supply control parameter combination for the target printing equipment.

[0085] Specifically, the server collects printing standard data for the target item. This data includes printing speed, color requirements, viscosity standards, etc., used to define printing standards and quality requirements. Standard feature extraction is performed on the printing standard data to extract multiple printing standard features. These features can cover various requirements and parameters in the printing process, such as color accuracy, speed requirements, etc. Feature correlation degrees are calculated between multiple printing standard features and multiple ink attribute features. This can be done using statistical methods (such as correlation coefficients) to measure the correlation between different features. Feature correlation degree measures the relationship between printing standards and ink properties to determine their degree of influence on the printing process. Based on feature correlation degrees, a decision tree model is constructed between multiple printing standard features and multiple ink attribute features. A decision tree is a supervised learning algorithm that segments data into different decision paths and is used to make decisions based on input feature values. Using the decision tree model, a first combination of ink supply control parameters is created for the target printing equipment. This parameter combination is based on the correlation between printing standard features and ink attribute features to optimize the printing process to meet printing standards and quality requirements. For example, suppose the printing standard data includes color accuracy, resolution requirements, and speed requirements. The server also contains ink property data, including ink color, viscosity, and drying speed. Three standard features are extracted from the printing standard data: color accuracy, resolution requirements, and speed requirements. By analyzing the correlation between these printing standard features and ink property features, the server found a strong correlation between color accuracy and ink color, and a correlation between resolution requirements and ink viscosity and drying speed. Based on these correlations, the server constructs a decision tree model that selects appropriate ink color, viscosity, and printing speed according to the printing standard data and ink property data to meet the color accuracy, resolution, and speed requirements. This is the server's first ink supply control parameter combination, ensuring that the printing process meets standards and quality requirements.

[0086] In one specific embodiment, such as Figure 3 As shown, the process of executing step S103 can specifically include the following steps:

[0087] S301. Set the parameters of the target printing equipment according to the first ink supply control parameter combination;

[0088] S301. The target printing equipment, after parameter settings are completed, performs printing operations on the target item. At the same time, the target item is optically positioned using a preset optical sensor to obtain the spatial position information of the target item.

[0089] S302. Based on spatial location information, optical images of the target object are acquired using an optical sensor to obtain initial image data;

[0090] S303. Input the initial image data into the optical image integrity analysis model, wherein the optical image integrity analysis model includes: residual layer, convolutional layer, pooling layer, fully connected layer and softmax function;

[0091] S304. Perform optical image integrity analysis on the initial image data using the optical image integrity analysis model to obtain optical image integrity analysis data;

[0092] S305. Verify the integrity analysis data and spatial location information of the optical image and generate image verification results;

[0093] S306. Based on the image verification results, perform image preprocessing on the initial image data to obtain optical image data.

[0094] Specifically, parameters are set for the target printing equipment based on the first combination of ink supply control parameters. These parameters may include printing speed, ink flow rate, printhead position, etc. The purpose of parameter setting is to ensure that the printing process proceeds under the required conditions. The target printing equipment, with the parameters set, performs the printing operation on the target item. Simultaneously, optical positioning is performed on the target item using a preset optical sensor to obtain its spatial position information. Optical positioning helps ensure printing accuracy and positioning. Based on the spatial position information, optical images of the target item are acquired using the optical sensor. This step aims to obtain visual data of the printed target item. The acquired image will include the printed content as well as any defects or problems. The optical image integrity analysis model is a deep learning model used to detect any defects, errors, or problems in an image. It typically consists of multiple layers, including residual layers, convolutional layers, pooling layers, fully connected layers, and a softmax function. These layers work together to identify features in the image and detect image integrity. Initial image data is input into the optical image integrity analysis model, which analyzes the image to detect any problems or defects. This includes identifying printing errors, missing parts, or other issues in the image. Based on the results of the optical image integrity analysis, an image verification result is generated. If the image passes the verification, the result is "normal" or "no problem." If the image has problems, the verification result will include a detailed description of the problem, such as the location and type of the defect. Based on the image verification result, image preprocessing is performed on the initial image data. This may include operations such as repairing defects in the image, removing noise, and adjusting the image's contrast and brightness to obtain the final optical image data. For example, suppose that based on a first combination of ink supply control parameters, the server sets the printing speed, ink flow, and printhead position. The server's printing equipment begins printing and performs optical positioning operations on the printed material using preset optical sensors to ensure the accuracy of the printing position. The optical sensors acquire image data of the printed material. These images include text, images, and any defects on the printed material. The image data is input into the optical image integrity analysis model, which uses techniques such as convolutional neural networks to detect problems in the image. If the image passes the verification, the result will show that the image is complete and has no problems. If the image has problems, the verification result will indicate the nature and location of the problem. Based on the verification result, the image is preprocessed, such as repairing defects or removing noise, to ensure the quality and integrity of the printed material.

[0095] In one specific embodiment, such as Figure 4 As shown, the process of executing step S104 can specifically include the following steps:

[0096] S401. Input the optical image data into a preset printing image detection model, wherein the printing image detection model includes: a first convolutional network, a first residual network, a first inverse residual network, a second residual network, a second inverse residual network, and a second convolutional network.

[0097] S402. Perform a convolution operation on the optical image data through the first convolutional network to obtain the first convolutional feature map;

[0098] S403. Input the first convolutional feature map into the first residual network for optical feature extraction to obtain the first optical feature map, and perform feature operation on the first optical feature map through the first inverse residual network to obtain the second optical feature map;

[0099] S404. Input the first convolutional feature map into the second residual network for coating feature extraction to obtain the first coating feature map, and perform feature operation on the first coating feature map through the second inverse residual network to obtain the second coating feature map.

[0100] S405. The second convolutional network is used to output features from the second optical feature map and the second coating feature map respectively, to obtain multiple ink optical features and multiple ink coating features.

[0101] Specifically, server-side printed image detection models typically consist of multiple deep neural network layers, including convolutional networks, residual networks, and inverse residual networks. These networks play a crucial role in image processing and feature extraction. These networks are built and connected to construct a complete printed image detection model. The input optical image data is convolved using a first convolutional network. Convolution helps identify different features in the image, generating a first convolutional feature map. These features include edges, textures, and colors. The first convolutional feature map is then input into a first residual network for optical feature extraction. Residual networks are designed for deeper feature learning, helping to extract higher-level image features. This step yields a first optical feature map. Feature operations are then performed on the first optical feature map using a first inverse residual network. Feature operations include weighted summation and the application of activation functions to further refine feature information, resulting in a second optical feature map. The first convolutional feature map is then input into a second residual network for coating feature extraction. Similar to optical feature extraction, this step helps identify coating features in the image, such as ink distribution and color uniformity, resulting in a first coating feature map. The first coating feature map is processed by a second inverse residual network, similar to the process used for optical feature processing. This helps to further improve the representation of the coating features, resulting in a second coating feature map. The second optical feature map and the second coating feature map are then processed by a second convolutional network to output features. This step generates multiple ink optical features and ink coating features, which can be used for subsequent analysis and control.

[0102] In one specific embodiment, the process of executing step S105 may specifically include the following steps:

[0103] (1) Based on multiple ink optical features and multiple ink coating features, a corresponding target ink feature matrix is ​​generated;

[0104] (2) Using the entropy weight evaluation model, calculate the positive and negative ideal solutions for each element of the target ink feature matrix;

[0105] (3) Calculate the comprehensive score of each element of the feature matrix based on the positive ideal solution and the negative ideal solution;

[0106] (4) Based on the comprehensive score of each feature matrix element, the first ink supply control parameter combination is optimized by a preset genetic algorithm to generate the corresponding second ink supply control parameter combination.

[0107] Specifically, a target ink feature matrix is ​​created based on multiple ink optical and coating characteristics. Each row of this matrix represents a printing job, and each column represents a different feature. Each element represents the value of the corresponding feature in the corresponding printing job. The entropy weight evaluation model is a multi-criteria decision analysis method used to determine the weight of each feature. For each feature, its positive ideal solution and negative ideal solution are calculated. The positive ideal solution is the maximum value of each feature, and the negative ideal solution is the minimum value of each feature. These ideal solutions are used to determine the relative importance of each feature in different printing jobs. Based on the positive and negative ideal solutions, a comprehensive score is calculated for each feature. Generally, the comprehensive score can be calculated using the following formula: Comprehensive score = w * (positive ideal solution - feature value) / (positive ideal solution - negative ideal solution). Where w is the feature weight, the positive ideal solution is the maximum value of the feature, the negative ideal solution is the minimum value of the feature, and the feature value is the actual value in the target ink feature matrix. Based on the comprehensive score of each feature, an optimization objective for the first ink supply control parameter combination can be obtained. A pre-defined genetic algorithm can be used to search for the optimal second ink supply control parameter combination to maximize or minimize the optimization objective. A genetic algorithm is an optimization algorithm that simulates the mechanisms of natural selection and inheritance. It searches for the optimal solution by generating and evolving a set of candidate solutions. In this case, the candidate solutions are combinations of secondary ink supply control parameters. Through iterative generation, selection, crossover, and mutation operations, the genetic algorithm can find the optimal parameter combination to meet the requirements of print quality and ink coating.

[0108] In one specific embodiment, the process of executing step S106 may specifically include the following steps:

[0109] (1) Based on the comprehensive score of each feature matrix element, C candidate ink supply control parameter combinations corresponding to the first ink supply control parameter combination are created by a preset genetic algorithm;

[0110] (2) Calculate the first fitness data for each candidate ink supply control parameter combination, and perform breeding, crossover and mutation operations on the C candidate ink supply control parameter combinations based on the first fitness data to generate D candidate ink supply control parameter combinations, where C < D;

[0111] (3) Calculate the second fitness data of the D candidate ink supply control parameter combinations respectively, and select the corresponding second ink supply control parameter combination from the D candidate ink supply control parameter combinations based on the second fitness data.

[0112] Specifically, using a pre-built genetic algorithm, C candidate ink supply control parameter combinations are created based on the comprehensive score of each feature matrix element, corresponding to the first ink supply control parameter combination. These candidate combinations represent different ink supply strategies in the parameter space. First fitness data is calculated for each of these C candidate ink supply control parameter combinations. The first fitness data is used to evaluate the performance of each candidate combination in the current printing job. This evaluation can be based on factors such as actual print quality indicators, ink usage efficiency, and printing speed. The first fitness data determines which candidate combinations perform well and which perform poorly under the current conditions. Based on the first fitness data, the server performs evolutionary operations on the C candidate ink supply control parameter combinations, including reproduction, crossover, and mutation. The purpose of these operations is to generate new candidate ink supply control parameter combinations to introduce diversity and variation. Through these operations, D new candidate ink supply control parameter combinations are created, where D is typically greater than C. This process simulates the evolutionary mechanism in biological genetics, ensuring the diversity and fitness of the parameter combinations. Second fitness data is then used to evaluate these D candidate ink supply control parameter combinations. The second fitness data is calculated based on different performance metrics or objectives to ensure that each candidate combination is superior in various aspects. Based on the second fitness data, the server selects the optimal candidate ink supply control parameter combination, which becomes the second ink supply control parameter combination. This combination comprehensively considers the combined scores of multiple features and the actual needs of the printing job, thus ensuring optimal ink supply control during the printing process. For example, suppose C different ink supply control parameter combinations are generated, and their performance in the current job is calculated. Through the evolutionary operation of a genetic algorithm, D new parameter combinations are generated, each with a different ink supply strategy. Using the second fitness data, a second ink supply control parameter combination that ensures accurate color output while conserving ink is selected to meet the needs of the printing job.

[0113] The ink supply control method for printing ink in the embodiments of the present invention has been described above. The ink supply control device for printing ink in the embodiments of the present invention will be described below. Please refer to [link / reference]. Figure 5 One embodiment of the ink supply control device for printing ink in this invention includes:

[0114] The acquisition module 501 is used to acquire ink flow data of printing ink in the target printing equipment through a preset flow sensor, and to extract ink attribute features from the ink flow data to obtain multiple ink attribute features.

[0115] A creation module 502 is used to acquire printing standard data of the target item and create a first ink supply control parameter combination for the target printing equipment based on the multiple ink attribute characteristics and the printing standard data.

[0116] The acquisition module 503 is used to perform printing operations on the target item through the target printing equipment according to the first ink supply control parameter combination, and to acquire optical images of the target item through a preset optical sensor to obtain optical image data.

[0117] The detection module 504 is used to input the optical image data into a preset printing image detection model to perform printing feature detection, thereby obtaining multiple ink optical features and multiple ink coating features.

[0118] The optimization module 505 is used to optimize the first ink supply control parameter combination based on the plurality of ink optical characteristics and the plurality of ink coating characteristics, and generate a corresponding second ink supply control parameter combination.

[0119] Through the collaborative operation of the aforementioned components, the ink flow data of the target printing equipment is acquired and ink attribute features are extracted to obtain multiple ink attribute features; printing standard data of the target item is acquired and a first ink supply control parameter combination is created; printing operation is performed on the target item, and optical images of the target item are acquired through a preset optical sensor to obtain optical image data; the optical image data is input into a printing image detection model for printing feature detection to obtain multiple ink optical features and multiple ink coating features; based on the multiple ink optical features and multiple ink coating features, the first ink supply control parameter combination is optimized to generate a corresponding second ink supply control parameter combination. This invention, by using flow sensors and data processing, enables real-time monitoring and control of ink flow. This ensures precise ink supply, conforming to printing standards, thereby reducing scrap rates. Multiple ink attribute features are extracted, including flow rate, ink color, and viscosity. This allows operators to have a more comprehensive understanding of the ink properties, facilitating more precise ink supply control. By analyzing the correlation between printing standard data and ink attribute features, the method can establish a decision tree model to link the two. This helps to automatically create a first ink supply control parameter combination based on printing standard data. Image data, including information on color accuracy and sharpness, is captured in real-time during the printing process using optical sensors. This allows for timely detection of problems and the implementation of corrective measures. A printing image detection model automatically detects and extracts multiple ink optical and coating features. This eliminates the need for manual inspection and improves the efficiency of the printing process. Based on these ink optical and coating features, the method uses an entropy weight evaluation model and a genetic algorithm for parameter optimization. This ensures the optimization of ink supply control parameters for better print quality. This invention is not merely a one-time control measure but also includes a mechanism for continuous monitoring and adjustment. This allows the printing process to be continuously improved to adapt to different printing needs and conditions, thereby achieving intelligent ink supply control and improving its accuracy.

[0120] above Figure 5 The ink supply control device for printing ink in the embodiments of the present invention will be described in detail from the perspective of modular functional entities. The ink supply control device for printing ink in the embodiments of the present invention will be described in detail from the perspective of hardware processing.

[0121] Figure 6This is a schematic diagram of the structure of an ink supply control device for printing ink provided in an embodiment of the present invention. The ink supply control device 600 for printing ink can vary considerably due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 610 (e.g., one or more processors) and a memory 620, and one or more storage media 630 (e.g., one or more mass storage devices) storing application programs 633 or data 632. The memory 620 and storage media 630 can be temporary or persistent storage. The program stored in the storage media 630 may include one or more modules (not shown in the diagram), each module may include a series of instruction operations on the ink supply control device 600 for printing ink. Furthermore, the processor 610 may be configured to communicate with the storage media 630 and execute the series of instruction operations in the storage media 630 on the ink supply control device 600 for printing ink.

[0122] The ink supply control device 600 for printing ink may also include one or more power supplies 640, one or more wired or wireless network interfaces 650, one or more input / output interfaces 660, and / or one or more operating systems 631, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 6 The illustrated structure of the ink supply control device for printing ink does not constitute a limitation on the ink supply control device for printing ink. It may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0123] The present invention also provides an ink supply control device for printing ink, the ink supply control device for printing ink including a memory and a processor, the memory storing computer-readable instructions, when the computer-readable instructions are executed by the processor, causing the processor to perform the steps of the ink supply control method for printing ink in the above embodiments.

[0124] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when the instructions are executed on a computer, cause the computer to perform the steps of the ink supply control method for printing ink.

[0125] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0126] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0127] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An ink supply control method of a printing ink, characterized by, The ink supply control method of the printing ink comprises: acquiring ink flow data of the printing ink in a target printing device through a preset flow sensor, and performing ink attribute feature extraction on the ink flow data to obtain a plurality of ink attribute features; acquiring printing standard data of a target object, and creating a first ink supply control parameter combination of the target printing device according to the plurality of ink attribute features and the printing standard data; performing a printing operation on the target object according to the first ink supply control parameter combination through the target printing device, and performing optical image acquisition on the target object through a preset optical sensor to obtain optical image data; inputting the optical image data into a preset printing image detection model for printing feature detection to obtain a plurality of ink optical features and a plurality of ink coating features; specifically including: inputting the optical image data into a preset printing image detection model, wherein the printing image detection model comprises: a first convolutional network, a first residual network, a first anti-residual network, a second residual network, a second anti-residual network, and a second convolutional network; performing convolution operation on the optical image data through the first convolutional network to obtain a first convolutional feature map; inputting the first convolutional feature map into the first residual network for optical feature extraction to obtain a first optical feature map, and performing feature operation on the first optical feature map through the first anti-residual network to obtain a second optical feature map; inputting the first convolutional feature map into the second residual network for coating feature extraction to obtain a first coating feature map, and performing feature operation on the first coating feature map through the second anti-residual network to obtain a second coating feature map; performing feature output on the second optical feature map and the second coating feature map through the second convolutional network respectively to obtain a plurality of ink optical features and a plurality of ink coating features; performing parameter optimization on the first ink supply control parameter combination based on the plurality of ink optical features and the plurality of ink coating features to generate a corresponding second ink supply control parameter combination; specifically including: generating a corresponding target ink feature matrix based on the plurality of ink optical features and the plurality of ink coating features; calculating the positive ideal solution and the negative ideal solution of each feature matrix element according to the target ink feature matrix through an entropy weight evaluation model; calculating the comprehensive score of each feature matrix element according to the positive ideal solution and the negative ideal solution; creating C candidate ink supply control parameter combinations corresponding to the first ink supply control parameter combination through a preset genetic algorithm based on the comprehensive score of each feature matrix element; calculating the first fitness data of each candidate ink supply control parameter combination, and performing reproduction, crossover and mutation operations on the C candidate ink supply control parameter combinations according to the first fitness data to generate D candidate ink supply control parameter combinations, wherein C < D; calculating the second fitness data of the D candidate ink supply control parameter combinations, and selecting the corresponding second ink supply control parameter combination from the D candidate ink supply control parameter combinations according to the second fitness data.

2. The method of ink supply control of a printing ink according to claim 1, characterized in that, The ink flow data of the printing ink in the target printing equipment is acquired by a preset flow sensor, and ink attribute feature extraction is performed on the ink flow data to obtain a plurality of ink attribute features, including: The ink flow data of the printing ink in the target printing equipment is acquired by a preset flow sensor, and data discretization representation is performed on the ink flow data to obtain discretized flow data; The discretized flow data is subjected to matrix conversion to generate a target flow matrix, wherein each row in the target flow matrix represents data at a time point, and each column represents a different flow observation point; A target covariance matrix corresponding to the target flow matrix is calculated, and eigenvalue decomposition is performed on the target covariance matrix to obtain eigenvalues and corresponding eigenvectors; According to the size of the eigenvalues, N principal components are selected, and the discretized flow data is projected into a new low-dimensional space based on the N principal components to obtain a plurality of ink attribute features.

3. The method of ink supply control of a printing ink according to claim 1, characterized in that, The printing standard data of the target object is acquired, and a first ink supply control parameter combination of the target printing equipment is created based on the plurality of ink attribute features and the printing standard data, including: The printing standard data of the target object is acquired, and standard feature extraction is performed on the printing standard data to obtain a plurality of printing standard features; The feature correlation between the plurality of printing standard features and the plurality of ink attribute features is calculated respectively; According to the feature correlation, a decision tree model between the plurality of printing standard features and the plurality of ink attribute features is constructed; Based on the decision tree model, a first ink supply control parameter combination of the target printing equipment is created.

4. The method of ink supply control of a printing ink according to claim 1, characterized by, The target object is printed by the target printing equipment according to the first ink supply control parameter combination, and optical image data is obtained by optical image acquisition of the target object by a preset optical sensor, including: According to the first ink supply control parameter combination, the target printing equipment is parameterized; The target object is printed by the target printing equipment after parameterization, and the spatial position information of the target object is obtained by optical positioning of the target object by a preset optical sensor; Based on the spatial position information, the optical image of the target object is acquired by the optical sensor to obtain initial image data; The initial image data is input into an optical image integrity analysis model, wherein the optical image integrity analysis model includes a residual layer, a convolution layer, a pooling layer, a full connection layer, and a softmax function; The optical image integrity of the initial image data is analyzed by the optical image integrity analysis model to obtain optical image integrity analysis data; The optical image integrity analysis data and the spatial position information are verified to generate an image verification result; According to the image verification result, the initial image data is preprocessed to obtain optical image data.

5. An ink supply control device for a printing ink, characterized by The ink supply control device of the printing ink includes: An acquisition module is configured to acquire ink flow data of printing ink in a target printing device through a preset flow sensor, and perform ink attribute feature extraction on the ink flow data to obtain a plurality of ink attribute features. A creation module is configured to acquire printing standard data of a target object, and create a first ink supply control parameter combination of the target printing device according to the plurality of ink attribute features and the printing standard data. An acquisition module is configured to perform printing operation on the target object according to the first ink supply control parameter combination through the target printing device, and perform optical image acquisition on the target object through a preset optical sensor to obtain optical image data. A detection module is configured to input the optical image data into a preset printing image detection model to perform printing feature detection, and obtain a plurality of ink optical features and a plurality of ink coating features. Specifically, the optical image data is input into the preset printing image detection model, wherein the printing image detection model includes a first convolutional network, a first residual network, a first inverse residual network, a second residual network, a second inverse residual network, and a second convolutional network. The first convolutional network is used to perform convolution operation on the optical image data to obtain a first convolutional feature map. The first convolutional feature map is input into the first residual network to perform optical feature extraction, and a first optical feature map is obtained. The first inverse residual network is used to perform feature operation on the first optical feature map to obtain a second optical feature map. The first convolutional feature map is input into the second residual network to perform coating feature extraction, and a first coating feature map is obtained. The second inverse residual network is used to perform feature operation on the first coating feature map to obtain a second coating feature map. The second convolutional network is used to perform feature output on the second optical feature map and the second coating feature map respectively, and a plurality of ink optical features and a plurality of ink coating features are obtained. The optimization module is configured to perform parameter optimization on the first ink supply control parameter combination based on the plurality of ink optical characteristics and the plurality of ink coating characteristics, and generate a corresponding second ink supply control parameter combination. Specifically, the optimization module is configured to generate a corresponding target ink characteristic matrix based on the plurality of ink optical characteristics and the plurality of ink coating characteristics; calculate a positive ideal solution and a negative ideal solution of each characteristic matrix element according to the target ink characteristic matrix through an entropy weight evaluation model; calculate a comprehensive score of each characteristic matrix element according to the positive ideal solution and the negative ideal solution; create C candidate ink supply control parameter combinations corresponding to the first ink supply control parameter combination through a preset genetic algorithm based on the comprehensive score of each characteristic matrix element; calculate a first fitness data of each candidate ink supply control parameter combination, and perform reproduction, crossover and mutation operations on the C candidate ink supply control parameter combinations according to the first fitness data to generate D candidate ink supply control parameter combinations, wherein C < D; calculate a second fitness data of the D candidate ink supply control parameter combinations, and select a corresponding second ink supply control parameter combination from the D candidate ink supply control parameter combinations according to the second fitness data.

6. An ink supply control device for a printing ink, characterized by The ink supply control device of the printing ink comprises a memory and at least one processor, and the memory stores instructions. The at least one processor invokes the instructions in the memory to enable the ink supply control device of the printing ink to perform the ink supply control method of the printing ink according to any one of claims 1-4.

7. A computer-readable storage medium having stored thereon instructions, the computer-readable storage medium comprising: The instructions are executed by the processor to implement the ink supply control method of the printing ink according to any one of claims 1-4.

Images