A method and apparatus for layered printing of Chinese ink paintings based on artificial intelligence

Abstract

This invention discloses an artificial intelligence-based method and apparatus for layered printing of Chinese ink paintings. The printing apparatus includes a base, a support structure, a color management module, a micro-jet outlining printing module, a misting and blurring printing module, an oxidation light-liquid system module, and an electronic information module. Simultaneously, artificial intelligence algorithms and other technologies are used for image data acquisition, layer segmentation, and parameter adjustment preprocessing. Combined with the printing apparatus, the outlining layer and ink blurring layer are layered and timed on Xuan paper, and the entire image undergoes delayed oxidation treatment. Ultimately, a highly accurate reproduction of the Xuan paper ink blurring and blurring effect, along with a rustic historical feel, is achieved. Using this invention, the layered ink blurring effect on Xuan paper in Chinese ink paintings can be realistically reproduced, effectively improving problems such as rigid reflections, lack of depth, blurring defects, and color distortion currently existing in the field of Chinese painting reproduction printing. It has broad scientific research and cultural application value.

Description

Technical Field

[0001] This invention relates to the field of ink painting printing, and in particular to an artificial intelligence-based device and method for layered printing of Chinese ink paintings. Background Technology

[0002] Chinese ink painting is a quintessential representative of traditional Chinese art, possessing rich historical value and cultural significance. Throughout history, countless literati and artists have left us with precious and exquisite paintings. In the past, most original ink paintings were housed in museums or art galleries, and due to limitations in artifact preservation and time and space, they rarely had the opportunity to be displayed to the world. With the development of artifact digitization technology, an increasing number of traditional Chinese ink paintings are being scanned in high definition and compiled into comprehensive painting databases. The printing of high-definition ink painting reproductions aims to showcase the beauty of Chinese ink painting to audiences through multiple channels and up close, breaking down the limitations of time and space, bringing traditional art to life, and contributing to the inheritance and dissemination of Chinese culture.

[0003] Chinese patent document CN111524111A discloses a printing processing method and apparatus based on artworks, including: acquiring image data corresponding to an artwork captured by a preset shooting tool under a preset light source; determining measurement points of the artwork and acquiring spectral data of the measurement points using a point-measurement spectrophotometer; obtaining an image file to be printed based on the acquired characteristic curve and white field parameters of the preset shooting tool, the image data, and the spectral data of the measurement points; and printing the image file to be printed using a printer and media ICC preset file configured in the printer, so as to acquire the captured artwork on the media.

[0004] Chinese patent document CN103465664A discloses a method for reproducing ancient calligraphy and painting artworks. This method uses a high-resolution digital camera to photograph the artwork, generating a digital image of the original. A color inkjet printer outputs the digital image and prints it onto low-temperature Xuan paper, generating a printed draft. The printed draft is then irradiated with ultraviolet light in a dry, room-temperature environment to form a preliminary copy. A high-resolution digital camera is used to photograph the preliminary copy, generating a preliminary digital image. The preliminary digital image is then corrected based on the data from the original digital image to generate a digital copy image. The printer is then started, and the digital copy image is printed onto low-temperature Xuan paper to generate a copy. Finally, the copy is laminated to form a reproduction of the ancient calligraphy and painting artwork.

[0005] Existing technologies often employ high-definition inkjet printing to reproduce images on glossy cardstock. While this one-dimensional layer printing mode can present details in high definition, the overall visual effect is difficult to capture the multi-dimensional layering and blurring effects and brushstrokes of Chinese ink painting. It suffers from problems such as stiff texture, rigid reflection, and lack of depth.

[0006] Some technologies have attempted to use Xuan paper directly as a printing medium, but because Xuan paper's hydrophobicity and diffusion properties differ from glossy cardstock, printing high-definition color images directly onto it will result in severe dry-wet color deviations and diffusion-related deviations. Summary of the Invention

[0007] This invention provides a method and apparatus for layered printing of Chinese ink paintings based on artificial intelligence, which can realistically reproduce the ink and wash layering effect on Xuan paper in Chinese ink paintings.

[0008] First, this invention provides a Chinese ink painting layer printing device, comprising:

[0009] The base is used to support the Xuan paper being printed on;

[0010] The support structure, fixed on the base, is used to support and connect the various modules to ensure their normal operation.

[0011] The color management module is used to store and provide the dyes and bleed solvents required for different printing methods;

[0012] Micro-jet outlining printing module, used for line printing of the line outlining layer;

[0013] Atomized ink diffusion printing module for surface printing of water-based ink diffusion layers;

[0014] Oxidation photo-liquid system module for delayed oxidation treatment;

[0015] The electronic information module is used for data processing, transmission, and reception.

[0016] In addition, the present invention also provides an artificial intelligence-based method for layered printing of Chinese ink paintings, which, based on the aforementioned layered printing device for Chinese ink paintings, specifically includes the following steps:

[0017] (1) Based on the ink diffusion properties of Xuan paper, parameterized experiments were conducted and data were collected. The derived ratio and empirical formula for dry and wet color difference were derived, and an ink trace database was constructed.

[0018] (2) Based on the paper oxidation and discoloration law, parameterized experiments were conducted and data were collected. An empirical formula for photo-liquid oxidation time was derived and a photo-liquid oxidation database was constructed.

[0019] (3) The imported ink painting image is subjected to color layering, overlapping filling and attribute judgment by the ink painting brush color layering algorithm and ink painting brush overlapping segmentation algorithm, so as to realize the multi-attribute layering of outlining and shading.

[0020] (4) Compare the information of each layer after multi-attribute layering with the ink trace database, calculate the original and derived range of line color blocks through the ink stain brush boundary convergence algorithm, and perform printing boundary convergence.

[0021] The original color of the paper and the color parameters of the wet and dry lines are calculated by the ink dyeing brush color restoration algorithm to restore the printed color;

[0022] (5) For printing lines and color blocks, regional path planning and time sequence planning are carried out respectively. The micro-spraying outlining printing module and the atomizing blur printing module are used to layer and time-by-time superimpose the printing line outlining layer and the water ink blur layer on the rice paper.

[0023] (6) Compare the background layer with the light-liquid oxidation database, solve the optimal solution through the ink-dye brush oxidation algorithm, and use the oxidation light-liquid system module to perform delayed oxidation processing on the whole picture, finally obtaining a high degree of restoration of Xuan paper ink and wash overlay and ancient historical effect.

[0024] This invention uses artificial intelligence algorithms and other technologies to perform preprocessing on ink painting images, such as data acquisition, layer segmentation, and parameter adjustment, before printing. Combined with a supporting development device, it implements an artificial intelligence-based layered printing method for Chinese ink paintings, thereby realistically restoring the ink and wash layering effect on Xuan paper in Chinese ink paintings.

[0025] The specific process of step (1) is as follows:

[0026] (1-1) Set up a Lab color experimental environment, take isotropic Xuan paper with hydrophobic property differences of less than 5%, and measure the original RGB values;

[0027] (1-2) Divide 10g of water-based ink for calligraphy and painting printing into 100 portions. Mix and adjust the solvent according to the 5% increment of weight and print in an array. Use a high-definition full-color camera to collect the original area, the boundary of the smudging and the dry and wet color parameters of the experimental color blocks under different water-ink ratio conditions in real time.

[0028] (1-3) Based on the collected results, calculate and derive the smudge derivative ratio and the empirical formula for RGB dry and wet color difference;

[0029] (1-4) Repeated experiments were conducted on various commonly used water-based dyes for calligraphy and painting printing to form an ink trace database for multi-layer printing needs.

[0030] The specific process of step (2) is as follows:

[0031] (2-1) Set up a Lab color experiment environment, take a specific printed Xuan paper with isotropic properties and hydrophobic property differences of less than 5%, and measure the original RGB values;

[0032] (2-2) Use light with uniform irradiance, wavelength between 300-800 nm, and ultraviolet light density of 1.0 W / m². 2 A surface light source was used to irradiate the paper surface for a delayed oxidation test, and a 1% solvent was used to slightly moisten the paper every 60 seconds. Oxidation color data was collected using a high-definition full-color camera during the process.

[0033] (2-3) Based on the collected data, derive the empirical formula for photo-liquid oxidation time;

[0034] (2-4) Use ultraviolet light density of 1-10 W / m 2 The experiment was repeated using a surface light source and a solvent with a concentration of 1% to 10% to form a photo-liquid oxidation database.

[0035] The specific process of step (3) is as follows:

[0036] (3-1) Import the high-definition full-color ink painting image file that needs to be printed in layers;

[0037] (3-2) Using the ink-dye brush color layering algorithm, count the total number of colors in the picture, divide the color range into 10-20 intervals according to color classification and transparency value, perform color layering, and record the corresponding RGB intervals;

[0038] (3-3) Based on the closed-loop path of color distribution and the color gradient superposition rule, the color superposition area range and superposition order are determined by the ink brush overlap segmentation algorithm. The color ownership of the overlapping closed-loop area and the intersection area is determined, and the overlapping fill is performed to generate a new layer.

[0039] (3-4) Match the outlining and shading attributes according to the maximum aspect ratio of the lines and color blocks and the total proportion of the picture, and sort them in layers.

[0040] The specific process of step (4) is as follows:

[0041] (4-1) Import a layered layer and compare it with the ink trace database;

[0042] (4-2) Based on the smudging ratio parameters, the original range and derived boundary of the line color block are derived and calculated through the ink brush boundary convergence algorithm. Print boundary convergence is performed to generate new print boundaries and regions for print path planning.

[0043] (4-3) Based on the color attributes and RGB wet and dry color difference parameters, print color is restored by the ink brush color restoration algorithm. The original color and wet and dry line contrast and superposition parameters are derived for printing color adjustment.

[0044] (4-4) Repeat the above steps for the layered layer set to generate a layer set for layered printing.

[0045] The specific process of step (5) is as follows:

[0046] (5-1) Printing preparation: Place the printing paper of the same type as the experimental paper on the base of the printing device and flatten it;

[0047] (5-2) The electronic information module of the printing device receives a set of layers for layered printing. The image set contains a parametric printing path file and a color file.

[0048] (5-3) According to the layer set stacking order, the micro-jet outlining printing module of the printing device performs linear printing on the line outlining layer, and the atomization and blurring printing module performs surface printing on the water-ink blurring layer.

[0049] (5-4) Complete the layer printing and dry the whole thing with air.

[0050] The specific process of step (6) is as follows:

[0051] (6-1) Import the background layer with the largest color gamut range and compare it with the light-liquid oxidation database;

[0052] (6-2) The optimal delayed oxidation scheme is solved using the ink-dyeing brush oxidation algorithm;

[0053] (6-3) Oxidize the paper using the oxidation light-liquid system module in the printing device;

[0054] (6-4) Collect the color data of the oxidized paper and compare it with the original color data. Then, perform secondary oxidation or color correction to complete the artwork.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] 1. This invention uses an algorithm to perform multi-dimensional layering and overlapping area boundary filling on a one-dimensional high-definition image of Chinese ink painting. It can better handle the topological relationship of color block color junction. The micro-jet printing module realizes linear printing of the line outline layer, and the atomization and blurring printing module realizes planar printing of the ink blurring layer. During the layered printing process, the layered sense of the lines, color blocks and color overlay and blurring of the picture is restored, in order to realistically reproduce the layered effect of ink blurring on Xuan paper in Chinese ink painting.

[0057] 2. This invention obtains a large database of ink marks through parametric experiments, summarizes empirical formulas for the smudging and derivation ratio and RGB wet-dry color difference under specific Xuan paper conditions, performs data analysis, derivation boundary convergence, and wet-dry color restoration for each layer, and can accurately calculate the parametric adjustment scheme of color blocks for multi-layer printing on Xuan paper for each work, generating new printing areas and printing color parameters, which can effectively avoid the serious wet-dry color deviation and smudging deviation problems in Xuan paper printing under existing printing methods.

[0058] 3. This invention obtains a photo-liquid oxidation database through parametric experiments, summarizes an empirical formula for oxidation time under specific Xuan paper conditions, and uses a parameterized and controllable delayed oxidation system module to restore the aged texture of the paper, making the reproduction of Chinese ink paintings closer to history and removing the "new" feel of printing. Under consistent environmental conditions, no repetitive work is required, and this invention can be promoted and used as an industry standard, contributing to the scientific datafication and industrial scaling of this printing field. Attached Figure Description

[0059] Figure 1 This is a schematic diagram of the structure of a layered printing device for Chinese ink painting according to the present invention;

[0060] Figure 2 This is a flowchart of a layered printing method for Chinese ink painting based on artificial intelligence, according to the present invention.

[0061] Figure 3 This is a schematic diagram illustrating the multi-attribute layering of outlining and shading in an embodiment of the present invention;

[0062] Figure 4 This is a schematic diagram illustrating the preprocessing of ink painting images in an embodiment of the present invention. Detailed Implementation

[0063] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be noted that the embodiments described below are intended to facilitate the understanding of the present invention and do not constitute any limitation thereof.

[0064] like Figure 1 As shown, a layered printing device for Chinese ink painting includes a base 1, a support structure 2, a color management module 3, a micro-jet outlining printing module 4, a misting and blurring printing module 5, an oxidation light-liquid system module 6, and an electronic information module 7.

[0065] Among them, the base 1 is used to support the Xuan paper to be printed; the support structure 2 is fixed on the base and is used to support and connect the various modules to ensure that each module works normally; the color management module 3 is used to store and provide the dyes and diffusion solvents required for different printing methods; the micro-jet outlining printing module 4 is used for line printing of the line outlining layer; the atomized diffusion printing module 5 is used for surface printing of the water-ink diffusion layer; the oxidation light-liquid system module 6 is used for delayed oxidation treatment; and the electronic information module 7 is used for data processing and transmission and reception.

[0066] like Figure 2 As shown, an artificial intelligence-based method for layered printing of Chinese ink paintings includes the following steps:

[0067] Step (1): Parametric experiment to construct ink stain database.

[0068] By referencing the ink diffusion properties of specific Xuan paper, parameterized experiments were conducted and data was collected. Empirical formulas for derived ratios and wet-dry color differences were derived, and an ink stain database was constructed.

[0069] (1.1) Set up a Lab color experiment environment, take a specific printed Xuan paper with isotropic properties and hydrophobic property differences of less than 5%, and measure the original RGB values.

[0070] (1.2) 10g of a specific water-based ink for printing calligraphy and painting was evenly divided into 100 portions. The solvent was mixed and prepared according to a 5% increment in weight and then arrayed for printing. A high-definition full-color camera was used to collect the original area, the boundary of the smudging and the dry and wet color parameters of the experimental color blocks under different ink-water ratio conditions in real time.

[0071] (1.3) Based on the collection results, calculate and derive the smudge derivative ratio and the empirical formula for RGB dry and wet color difference.

[0072] (1.4) Repeated experiments were conducted on 21 commonly used water-based dyes for specific calligraphy and painting printing purposes to form an ink trace database for multi-layer printing needs.

[0073] Step (2): Parametric experiment, constructing a photo-liquid oxidation database.

[0074] Based on the oxidation and discoloration patterns of paper, parameterized experiments were conducted and data were collected. An empirical formula for photo-liquid oxidation time was derived, and a photo-liquid oxidation database was constructed.

[0075] (2.1) Set up a Lab color experiment environment, take a specific printed Xuan paper with isotropic properties and hydrophobic property differences of less than 5%, and measure the original RGB values.

[0076] (2.2) Use light with uniform irradiance, wavelength between 300-800 nm, and ultraviolet light density of 1.0 W / m². 2A surface light source was used to irradiate the paper surface for a delayed oxidation test, and a 1% concentration of a specific solvent (ferric chloride solution) was used for micro-moistening treatment every 60 seconds. Oxidation color data were collected using a high-definition full-color camera during the process.

[0077] (2.3) Based on the data collection results, derive the empirical formula for photo-liquid oxidation time.

[0078] (2.4) For ultraviolet light density of 1–10 W / m 2 The experiment was repeated using a surface light source and a specific solvent with a concentration of 1% to 10% to form a photo-liquid oxidation database.

[0079] Step (3): Import the image and layer it with multiple attributes.

[0080] By using the ink-dye brush color "layering" algorithm and the ink-dye brush overlap "segmentation" algorithm, the imported image is subjected to color layering, overlap filling and attribute judgment, realizing multi-attribute layering of "outlining and shading".

[0081] (3.1) Enter the high-definition full-color ink painting image file that needs to be printed in layers, in PSD or PNG format.

[0082] (3.2) Using the ink-dye brush color "layering" algorithm, the total number of colors in the picture is counted, and the color gamut is divided into 10-20 intervals according to the color classification and transparency value. Color layering is performed, and the corresponding RGB intervals are recorded.

[0083] In this embodiment, the specific steps of the ink-dye brush color "layering" algorithm are as follows:

[0084] The first step is to construct the corresponding color segmentation dataset.

[0085] The labelme (open source) annotation tool can be used to annotate different color layer regions, obtain the annotation information of the color layer regions, and perform one-hot encoding on the annotations. The value of non-color layer regions is 0, and other color layers are incremented sequentially according to the annotation number 1, 2, 3, ... n (n is an integer).

[0086] The second step is to construct the segmentation model.

[0087] The corresponding segmentation network model is constructed, mainly consisting of an encoder and a decoder. During the encoding stage, features from different scales are fused and added to the features from the decoder, enabling the model to cover multiple color layers and shapes.

[0088] The third step is to construct the loss function.

[0089] The loss function consists of cross-entropy loss and FocalLoss. Cross-entropy loss classifies each pixel in the image, while FocalLoss is used to address the problem of uneven color layer numbers.

[0090] The fourth step is model prediction.

[0091] The first three steps yield the corresponding color layering model. The image to be layered is then scaled and processed before being input into the network for prediction. The prediction results are converted into probability values ​​of 0, 1, 2, 3, ..., n. Color layering is then performed using these 1, 2, 3, ..., n values.

[0092] like Figure 3 As shown, this illustrates each layer of a traditional Chinese ink painting image after it has been layered using multiple attributes such as "outlining and shading".

[0093] (3.3) Based on the closed-loop path of color distribution and the color gradient superposition rule, the color superposition area range and superposition order are determined by the ink brush overlap "segmentation" algorithm, the color ownership of the overlapping closed-loop area and the intersection area is determined, and the overlapping fill is performed to generate a new layer.

[0094] In this embodiment, the specific steps of the ink-painting brush overlap "segmentation" algorithm are as follows:

[0095] Step 1: Construct the corresponding brush overlap segmentation dataset.

[0096] The labelme (open source) annotation tool can be used to annotate the overlapping areas, obtain the annotation information of the overlapping areas, and perform one-hot encoding on the annotations, marking the overlapping areas as 1 and other areas as 0.

[0097] Step 2: Constructing the segmentation model.

[0098] A U-Net-type neural network model is constructed, mainly consisting of an encoder and a decoder. Features are fused between encoders and decoders of the same scale. This allows the model to combine low-level visual features and high-dimensional semantic features to better detect overlapping regions.

[0099] Step 3: Construct the loss function for model convergence.

[0100] The loss function used is the binary cross-entropy loss function, which optimizes the model by minimizing the model's prediction results against the test data in the labeled dataset.

[0101] Step 4: Model prediction.

[0102] The first three steps yield the corresponding overlapping segmentation model. The image to be segmented is then scaled and processed before being input into the network for prediction. The prediction results are converted from probability values ​​to 0 and 1, and finally to 0 and 255. In this model, 255 represents the overlapping region.

[0103] (3.4) Match the “outline” and “blend” attributes according to the maximum aspect ratio of the line color blocks and the total proportion of the picture, and sort them in layers.

[0104] Step (4): Boundary convergence and color restoration.

[0105] By comparing the information of each layer with the ink trace database, the original and derived range of the line color blocks are inferred through the ink brush boundary "convergence" algorithm, and the printing boundary convergence is performed. The original color of the paper and the color parameters of the wet and dry lines are calculated and inferred through the ink brush color "restoration" algorithm, and the printing color is restored.

[0106] (4.1) Import a layered layer into the model and compare it with the ink trace database.

[0107] (4.2) Based on the smudging ratio parameters, the original range and derived boundary of the line color block are derived and calculated through the ink brush boundary "convergence" algorithm. Print boundary convergence is performed to generate new print boundaries and regions for print path planning.

[0108] In this embodiment, the specific steps of the ink brush boundary "convergence" algorithm are as follows:

[0109] The first step is to collect the corresponding boundary "convergence" data, which includes the paper material w1, the size of the ink-stained area s2 after stabilization, the ink concentration n1 when the pen is applied, and the size of the area where the pen is applied s1, and construct the corresponding dataset.

[0110] The second step is to construct a neural network model for boundary "convergence". Using the multilayer perceptron (MLP) basic model, the relationship between ink concentration n1 when the pen is applied, the size of the application area s1, the paper material w1, and the size of the ink-stained area s2 after stabilization is constructed.

[0111] The third step involves using model prediction and table lookup to predict the ink concentration n1 and the size of the pen application area s1 at the time of pen application.

[0112] (4.3) Based on the color attributes and RGB wet and dry color difference parameters, the printing color is restored by the ink brush color "restoration" algorithm, and the original color and wet and dry line contrast superposition parameters are derived for printing color adjustment.

[0113] In this embodiment, the specific steps of the ink-stained brush color "restoration" algorithm are as follows:

[0114] The first step is to collect the relevant color data, which includes the paper material W1, the ink concentration n2, and the final color c of the ink-stained area after stabilization, and to construct the corresponding dataset.

[0115] The second step involves constructing a neural network model for color "reproduction," using a multilayer perceptron (MLP) model to establish the relationship between ink concentration n2, material W1, and the final color c of the ink-stained area.

[0116] The third step is to predict the corresponding ink concentration using a model.

[0117] like Figure 4 As shown, this demonstrates the results of preprocessing an ink painting image, including color layering, overlap filling, boundary convergence, and color restoration.

[0118] Step (5): Path timing planning, layered printing.

[0119] Both printed lines and color blocks undergo regional path planning and time sequence planning. The micro-jet outlining printing module and the atomization blurring printing module are used to layer and time-stack printed line outlining layers and ink blurring layers on specific Xuan paper.

[0120] (5.1) Printing preparation: Place the specific printing paper of the same type as the experimental paper on the base and flatten it.

[0121] (5.2) Receive the layer set for layered printing using the electronic information module. The layer set contains a parametric print path file and a color file.

[0122] (5.3) According to the layer set stacking order, the micro-jet outlining printing module performs linear printing on the line outlining layer, and the atomization blurring printing module performs surface printing on the water-ink blurring layer.

[0123] (5.4) Complete the layer printing and dry the whole thing with air.

[0124] Step (6): Image background analysis and delayed oxidation.

[0125] By comparing the background layer with the light-liquid oxidation database and solving the optimal solution using the ink-dye brush "oxidation" algorithm, and then using the oxidation light-liquid system module to perform delayed oxidation processing on the entire image, a high degree of reproduction of the ink wash and the ancient and historical effect of Xuan paper can be obtained.

[0126] (6.1) Import the background layer with the largest color gamut range and compare it with the light-liquid oxidation database.

[0127] (6.2) The optimal delayed oxidation scheme is solved by using the ink-dyeing brush "oxidation" algorithm.

[0128] In this embodiment, the specific steps of the ink-stained brush "oxidation" algorithm are as follows:

[0129] The first step is to collect relevant brush "oxidation" data, including paper material W1, ultraviolet light intensity y1, light exposure time y2, and paper color c1 after stabilization, and construct the corresponding dataset.

[0130] The second step is to construct a neural network model for oxidation-reduction, using a multilayer perceptron (MLP) basic model to establish the relationship between ultraviolet light intensity y1, illumination time y2, paper material W1, and paper color c1.

[0131] The third step involves predicting ultraviolet light intensity y1 and illumination time y2 using model prediction and database lookup.

[0132] (6.3) The paper is oxidized using the oxidation light-liquid system module.

[0133] (6.4) Collect the color data of the oxidized paper and compare it with the original color data. Then, perform secondary oxidation or color correction to complete the artwork.

[0134] The embodiments described above provide a detailed explanation of the technical solutions and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, additions, and equivalent substitutions made within the scope of the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A layered printing method for Chinese ink paintings based on artificial intelligence, characterized in that, The device employs a layered printing system for Chinese ink painting, comprising: a base to support the Xuan paper to be printed; a support structure fixed to the base to support and connect the various modules to ensure their normal operation; a color management module to store and provide the dyes and diffusion solvents required for different printing methods; a micro-jet printing module for line printing of the line outline layer; a misting printing module for surface printing of the ink diffusion layer; an oxidation light-liquid system module for delayed oxidation treatment; and an electronic information module for data processing, transmission, and reception. The aforementioned method for layered printing of Chinese ink paintings specifically includes the following steps: (1) Based on the ink diffusion properties of Xuan paper, parameterized experiments were conducted and data were collected. The derived ratio and empirical formula for dry and wet color difference were derived, and an ink trace database was constructed. (2) Based on the paper oxidation and discoloration law, parameterized experiments were conducted and data were collected. An empirical formula for photo-liquid oxidation time was derived and a photo-liquid oxidation database was constructed. (3) The imported ink painting image is subjected to color layering, overlapping filling and attribute judgment by the ink painting brush color layering algorithm and ink painting brush overlapping segmentation algorithm, so as to realize the multi-attribute layering of outlining and shading. (4) Compare the information of each layer after multi-attribute layering with the ink trace database, calculate the original and derived range of line color blocks through the ink stain brush boundary convergence algorithm, and perform printing boundary convergence. The original color of the paper and the color parameters of the wet and dry lines are calculated by the ink dyeing brush color restoration algorithm to restore the printed color; (5) The area path planning and time sequence planning are carried out for the printed lines and color blocks. The micro-spraying outlining printing module and the atomizing blur printing module are used to overlay the printing line outlining layer and water ink blur layer on the rice paper in layers and time. (6) Compare the background layer with the light-liquid oxidation database, solve the optimal solution through the ink-dye brush oxidation algorithm, and use the oxidation light-liquid system module to perform delayed oxidation processing on the whole picture, finally obtaining a high degree of restoration of Xuan paper ink and wash overlay and ancient historical effect.

2. The method for layered printing of Chinese ink paintings based on artificial intelligence according to claim 1, characterized in that, The specific process of step (1) is as follows: (1-1) Set up a Lab color experimental environment, take isotropic Xuan paper with hydrophobic property differences of less than 5%, and measure the original RGB values; (1-2) Divide 10g of water-based ink for calligraphy and painting printing into 100 portions. Mix and adjust the solvent according to the 5% increment of weight and print in an array. Use a high-definition full-color camera to collect the original area, the boundary of the smudging and the dry and wet color parameters of the experimental color blocks under different water-ink ratio conditions in real time. (1-3) Based on the collected results, calculate and derive the smudge derivative ratio and the empirical formula for RGB dry and wet color difference; (1-4) Repeated experiments were conducted on various commonly used water-based dyes for calligraphy and painting printing to form an ink trace database for multi-layer printing needs.

3. The method for layered printing of Chinese ink paintings based on artificial intelligence according to claim 1, characterized in that, The specific process of step (2) is as follows: (2-1) Set up a Lab color experiment environment, take isotropic Xuan paper with hydrophobic property differences of less than 5%, and measure the original RGB values; (2-2) Use light with uniform irradiance, wavelength between 300-800 nm, and ultraviolet light density of 1.0 W / m². 2 A surface light source was used to irradiate the paper surface for a delayed oxidation test, and a 1% solvent was used to slightly moisten the paper every 60 seconds. Oxidation color data was collected using a high-definition full-color camera during the process. (2-3) Based on the collected data, derive the empirical formula for photo-liquid oxidation time; (2-4) Use ultraviolet light density of 1-10 W / m 2 The experiment was repeated using a surface light source and a solvent with a concentration of 1% to 10% to form a photo-liquid oxidation database.

4. The method for layered printing of Chinese ink paintings based on artificial intelligence according to claim 1, characterized in that, The specific process of step (3) is as follows: (3-1) Import the high-definition full-color ink painting image file that needs to be printed in layers; (3-2) Using the ink-dye brush color layering algorithm, count the total number of colors in the picture, divide the color range into 10-20 intervals according to color classification and transparency value, perform color layering, and record the corresponding RGB intervals; (3-3) Based on the closed-loop path of color distribution and the color gradient superposition rule, the color superposition area range and superposition order are determined by the ink brush overlap segmentation algorithm. The color ownership of the overlapping closed-loop area and the intersection area is determined, and the overlapping fill is performed to generate a new layer. (3-4) Match the outlining and shading attributes according to the maximum aspect ratio of the lines and color blocks and the total proportion of the picture, and sort them in layers.

5. The method for layered printing of Chinese ink paintings based on artificial intelligence according to claim 1, characterized in that, The specific process of step (4) is as follows: (4-1) Import a layered layer and compare it with the ink trace database; (4-2) Based on the smudging ratio parameters, the original range and derived boundary of the line color block are derived and calculated through the ink brush boundary convergence algorithm. Print boundary convergence is performed to generate new print boundaries and regions for print path planning. (4-3) Based on the color attributes and RGB wet and dry color difference parameters, print color is restored by the ink brush color restoration algorithm. The original color and wet and dry line contrast and superposition parameters are derived for printing color adjustment. (4-4) Repeat the above steps for the layered layer set to generate a layer set for layered printing.

6. The method for layered printing of Chinese ink paintings based on artificial intelligence according to claim 5, characterized in that, The specific process of step (5) is as follows: (5-1) Printing preparation: Place the printing paper of the same type as the experimental paper on the base of the printing device and flatten it; (5-2) The electronic information module of the printing device is used to receive a layer set for layered printing, which contains a parameterized print path file and a color file; (5-3) According to the layer set stacking order, the micro-jet outlining printing module of the printing device performs linear printing on the line outlining layer, and the atomization and blurring printing module performs surface printing on the water-ink blurring layer. (5-4) Complete the layer printing and dry the whole thing with air.

7. The method for layered printing of Chinese ink paintings based on artificial intelligence according to claim 1, characterized in that, The specific process of step (6) is as follows: (6-1) Import the background layer with the largest color gamut range and compare it with the light-liquid oxidation database; (6-2) The optimal delayed oxidation scheme is solved using the ink-dyeing brush oxidation algorithm; (6-3) Oxidize the paper using the oxidation light-liquid system module in the printing device; (6-4) Collect the color data of the oxidized paper and compare it with the original color data. Then, perform secondary oxidation or color correction to complete the artwork.

Images