An automatic ink dispensing control method and system for decorative paper

Abstract

The application discloses a kind of for decorative paper automatic ink control method and system, comprising: obtaining target picture, color space conversion and preprocessing are carried out to target picture, obtain standardization image;Color recognition processing is executed to standardization image, at least one main color contained in image is extracted, and the area proportion of each main color is determined;For each extracted main color, call preset base ink spectrum database, and the mixing quality proportion of each base ink is calculated using spectrum matching optimization algorithm;According to the area proportion of each main color, target printing area input by user and preset ink layer thickness, combined with mixing quality proportion, the cumulative total mass required for each base ink in entire production task is calculated;According to cumulative total mass, automatically generate ink dispensing instruction, and send ink dispensing instruction to automatic ink dispensing equipment.Can automatically identify multiple main colors and distribution, accurately calculate ink dispensing proportion, improve color reproduction accuracy and production efficiency.

Description

Technical Field

[0001] This invention relates to the field of decorative paper ink mixing technology, specifically to an automatic ink mixing control method and system for decorative paper. Background Technology

[0002] In printing, digital textile printing, ceramic inkjet printing, packaging printing, and art reproduction, multi-color ink mixing systems are key equipment for achieving high-fidelity color reproduction. Traditional ink mixing methods mainly rely on manual experience or simple color chart comparison: operators visually observe the target color, select several base inks based on experience, mix them according to an estimated ratio, and then iterate through trial printing, adjustments, and more trial printing to achieve an effect close to the target color. This method is heavily dependent on the operator's skill level, inefficient, and makes it difficult to guarantee batch-to-batch consistency.

[0003] With the development of computer-aided color matching technology, some ink mixing systems have begun to incorporate spectrophotometers to measure target color samples, and then calculate the base ink ratio through simple linear interpolation or tristimulus value matching. However, in the color recognition stage, existing systems typically require users to provide uniform color patch samples or standard color card numbers, and cannot directly process complex images (such as design drafts containing gradients, textures, or multiple color patches). Even if users provide images, existing systems often only take the average color of the entire image, or manually select the color of a point in the image, losing information about the multiple primary colors and their area distribution in the image, resulting in subsequent ink mixing failing to cover all the color requirements of the image. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art. This invention provides an automatic ink mixing control method and system for decorative paper. It acquires a target image, performs color conversion preprocessing, performs color recognition processing to extract the main color and area ratio, uses a spectral matching optimization algorithm to calculate the mixing quality ratio, calculates the cumulative total quality and generates ink mixing instructions. It accurately extracts the distribution characteristics of the main color through image analysis, optimizes the mixing ratio calculation by combining a spectral nonlinear superposition model, and realizes automatic accumulation of layered ink volume and instruction generation. It can automatically identify multiple main colors and their distribution ratios in the image, and accurately calculate the ink mixing ratio by combining a spectral matching optimization algorithm, thereby achieving high-precision automated ink mixing control and effectively improving color reproduction accuracy and production efficiency.

[0005] This invention provides an automatic ink dispensing control method for decorative paper, the control method comprising:

[0006] Acquire the target image, perform color space conversion and preprocessing on the target image to obtain a standardized image;

[0007] Perform color recognition processing on the standardized image to extract at least one primary color contained in the image and determine the area ratio occupied by each primary color;

[0008] For each extracted primary color, a preset base ink spectral database is called, and a spectral matching optimization algorithm is used to calculate the mixing mass ratio of each base ink.

[0009] Based on the area ratio of each primary color, the target printing area input by the user, and the preset ink layer thickness, combined with the mixing mass ratio, the total cumulative mass required for each base ink in the entire production task is calculated.

[0010] The ink mixing instruction is automatically generated based on the total accumulated mass and sent to the automatic ink mixing equipment to control the equipment to weigh each base ink in sequence and mix them.

[0011] Furthermore, the step of performing color recognition processing on the standardized image, extracting at least one primary color contained in the image, and determining the area proportion occupied by each primary color includes:

[0012] Convert the RGB value of each pixel in the standardized image to a CIE-Lab value, making each pixel a coordinate point in the CIE-Lab three-dimensional space;

[0013] Calculate the Euclidean distance between any two pixels in the CIE-Lab space, and use the Euclidean distance as a quantitative indicator of the color difference between the two pixels.

[0014] Randomly select a number of pixels as the initial cluster centers, assign each pixel to the cluster center with the closest Euclidean distance, and recalculate the center position of each cluster, that is, take the average of the CIE-Lab coordinates of all pixels in the cluster as the new cluster center;

[0015] Repeat the assignment and update steps until the location of the cluster centers no longer changes significantly. At this point, the center coordinates of each cluster represent a primary color.

[0016] Furthermore, the repeated assignment and update steps continue until the positions of the cluster centers no longer change significantly. At this point, the center coordinates of each cluster represent a primary color, including:

[0017] For each candidate color quantity within the interval, a complete clustering process is performed. The average distance between each pixel and other pixels in the same cluster is calculated as the cohesion, and the average distance between each pixel and the nearest other cluster center is calculated as the separation. The difference between the separation and cohesion is divided by the larger of the two to obtain the contour value of the pixel. The contour values ​​of all pixels are averaged to obtain the contour coefficient of the clustering result. The candidate color quantity with the largest contour coefficient is selected as the final primary color quantity, and the clustering result with this quantity is used as the final primary color extraction result.

[0018] Furthermore, before performing clustering, the normalized images are first subjected to superpixel segmentation:

[0019] The image is divided into several small regions, and the pixel color difference in each small region is less than a preset threshold, thereby forming superpixel blocks with uniform internal color.

[0020] Each superpixel block uses the CIE-Lab average of all pixels within it as the representative color of that block;

[0021] Subsequent clustering operations use superpixel blocks as the basic unit rather than individual pixels, treating each superpixel block as a sample point for K-Means clustering;

[0022] After clustering, each primary color corresponds to a group of superpixel blocks. By adding up the number of original pixels covered by all superpixel blocks in the group and dividing by the total number of pixels in the image, the area ratio occupied by that primary color can be obtained.

[0023] Furthermore, for each extracted primary color, the process of calling a preset base ink spectral database and using a spectral matching optimization algorithm to calculate the mixing mass ratio of each base ink includes:

[0024] A numerical sequence of reflectance at each wavelength point in the visible spectrum for each type of base ink is pre-established to form a base ink spectral database;

[0025] For the target primary color to be matched, an initial vector is set, which is composed of the mixing ratio of all base inks, where the ratio of each base ink is a positive number and the sum is a fixed value.

[0026] The reflectance values ​​of each base ink are weighted and multiplied according to their mixing ratio, and then summed to obtain the predicted reflectance sequence of the mixed ink.

[0027] The predicted CIE-Lab color values ​​are calculated by integrating the predicted reflectance sequence wavelength by wavelength using the standard colorimetric conversion method, combined with the preset spectral distribution of the illumination source and the standard observer color matching function.

[0028] The color difference between the predicted CIE-Lab value and the CIE-Lab value of the target primary color is calculated using the CIEDE2000 color difference formula.

[0029] With the goal of minimizing the color difference value, the mixing ratio of each base ink is repeatedly adjusted. After each adjustment, the predicted spectrum, predicted CIE-Lab value and color difference value are recalculated until the color difference value is less than the preset qualified threshold.

[0030] The final adjusted mixing ratio is taken as the optimal base ink ratio for this main color.

[0031] Furthermore, in the calculation of the predicted spectrum, if the base ink types include opaque or semi-transparent inks, the Kubelka-Munk model is used to process them, converting the reflectance value of each base ink into the corresponding absorption-scattering ratio value, and then weighting and superimposing the absorption-scattering ratios of each base ink according to the mixing ratio to obtain the absorption-scattering ratio of the mixed ink, and then converting it back to the reflectance value as the predicted spectrum of the mixed ink.

[0032] Furthermore, the step of repeatedly adjusting the mixing ratio of each base ink with the goal of minimizing color difference values ​​includes:

[0033] Apply a small increment to the proportion of each base ink, recalculate the color difference value, divide the change in color difference by the proportion increment to obtain the gradient approximation of that base ink, and combine the gradient values ​​of all base inks into a gradient vector.

[0034] Using the BFGS update rule, based on the gradient vector of the current iteration point and historical iteration information, an approximation of the inverse matrix of the Hessian matrix is ​​constructed, and then the quasi-Newton direction is calculated.

[0035] A one-dimensional line search is performed along the quasi-Newton direction to determine the optimal step size, thereby generating the next set of mixed scaling vectors;

[0036] Repeat the above process of gradient calculation, direction construction and step size search, and compare the current color difference value with the color difference value of the previous iteration after each iteration;

[0037] The iteration stops when the decrease in color difference value is less than the preset convergence threshold in multiple consecutive iterations, or when the color difference value itself is less than the qualified threshold.

[0038] After the iteration is completed, the proportion values ​​of each base ink are checked and interference data are filtered out. The proportions of the remaining base inks are then scaled up again to restore their sum to a fixed value, and the final mixing quality proportion vector is output.

[0039] Furthermore, the step of automatically generating an ink mixing instruction based on the accumulated total mass and sending the ink mixing instruction to the automatic ink mixing equipment, controlling the equipment to sequentially weigh and mix each base ink, includes:

[0040] The number of each base ink and its corresponding total accumulated mass are numerically encoded and organized into a data packet according to the preset instruction frame format;

[0041] The entire data packet is processed using a cyclic redundancy check algorithm to calculate a checksum, which is then appended to the end of the data packet.

[0042] The data packet is written to the designated register address of the automatic ink dispensing device using an industrial communication protocol.

[0043] Furthermore, before generating the ink mixing instruction, a layered ink volume accumulation calculation is performed:

[0044] Based on the preset inkjet process parameters, the entire inkjet process is decomposed into multiple functional layers, including a base layer, at least one color layer, a gradient layer, and a protective layer.

[0045] For each primary color, calculate its contribution to the amount of ink in the color layer;

[0046] For gradient layers, the area of ​​the gradient band on the boundary of adjacent main color areas is calculated, and the mixing ratio of each position in the gradient band is determined by linear interpolation according to distance weight. Then, the total amount of ink of each base ink in the gradient layer is calculated by integration.

[0047] For the base layer and the protective layer, calculate the amount of base ink and protective ink according to the fixed coverage parameters set by the user;

[0048] The total mass of each type of base ink is obtained by summing the ink volume values ​​of all layers containing the same base ink.

[0049] The present invention also provides an automatic ink dispensing control system for decorative paper, the control system being used to execute the control method, the control system comprising:

[0050] The image input module is used to acquire a target image, perform color space conversion and preprocessing on the target image, and output a standardized image.

[0051] The color recognition module, through any of the color recognition processes described above, extracts the main color and the area ratio of each main color;

[0052] The formula calculation module is used to perform any of the spectral matching optimization processes for each primary color and calculate the mixing mass ratio of each base ink.

[0053] The ink volume accumulation module is used to calculate the total accumulated mass of each base ink based on the area ratio of each main color, the target printing area, and the ink layer thickness, combined with the mixing mass ratio.

[0054] The instruction generation and communication module is used to generate the ink dispensing instruction and send the instruction to the automatic ink dispensing equipment through an industrial communication protocol.

[0055] This invention provides an automatic ink mixing control method and system for decorative paper. It involves acquiring a target image, performing color conversion preprocessing, executing color recognition processing to extract the main color and area ratio, using a spectral matching optimization algorithm to calculate the mixing quality ratio, calculating the accumulated total quality, and generating ink mixing instructions. Through image analysis, it accurately extracts the distribution characteristics of the main color, optimizes the mixing ratio calculation using a spectral nonlinear superposition model, and achieves automatic accumulation of layered ink amounts and instruction generation. It can automatically identify multiple main colors and their distribution ratios in an image, and accurately calculate the ink mixing ratio using a spectral matching optimization algorithm, achieving high-precision automated ink mixing control and effectively improving color reproduction accuracy and production efficiency. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 This is a flowchart of an automatic ink dispensing control method for decorative paper in an embodiment of the present invention;

[0058] Figure 2 This is a flowchart of a standardized image color recognition processing method in an embodiment of the present invention;

[0059] Figure 3 This is a schematic diagram of an automatic ink dispensing control system for decorative paper in an embodiment of the present invention. Detailed Implementation

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] Example 1:

[0062] Please refer to Figure 1 and Figure 2 This invention provides an automatic ink dispensing control method for decorative paper, the control method comprising:

[0063] S1: Obtain the target image, perform color space conversion and preprocessing on the target image to obtain a standardized image;

[0064] Target images can be acquired in various ways, such as by users manually uploading digital image files or by digitizing physical samples using a scanner. After acquiring the raw image, it needs to undergo color space conversion, such as converting the common RGB format to the more suitable Lab format for color analysis, or to the CMYK format commonly used in the printing industry. Simultaneously, preprocessing operations are performed, such as scaling the image to a uniform size or applying median filtering to remove potential image noise, ensuring the accuracy of subsequent color analysis.

[0065] S2: Perform color recognition processing on the standardized image to extract at least one primary color contained in the image and determine the area ratio of each primary color. This can be achieved by analyzing the color histogram of the image and identifying the color corresponding to the peak of the histogram as the primary color. After determining the primary color, the area ratio of each primary color in the image can be calculated by counting the number of pixels corresponding to each primary color and dividing it by the total number of pixels in the image.

[0066] Specifically, the step of performing color recognition processing on the standardized image, extracting at least one dominant color contained in the image, and determining the area proportion occupied by each dominant color includes:

[0067] S21: Convert the RGB values ​​of each pixel in the standardized image to CIE-Lab values, making each pixel a coordinate point in the CIE-Lab three-dimensional space.

[0068] Converting the RGB values ​​of each pixel in a standardized image to CIE-Lab values ​​aims to transform the device-dependent RGB color space into the perceptibly uniform CIE-Lab color space. The changes in values ​​in the RGB color space do not linearly correspond to the human eye's perception of color differences, while the CIE-Lab space aims to simulate the way the human eye perceives color. Here, L* represents brightness, a* represents the range from green to red, and b* represents the range from blue to yellow.

[0069] S22: Calculate the Euclidean distance between any two pixels in the CIE-Lab space, and use the Euclidean distance as a quantitative indicator of the color difference between the two pixels.

[0070] In the CIELab 3D space, the CIE-Lab coordinates of two pixels and Euclidean distance between It can be represented as: Because the CIE-Lab space has perceptual uniformity, this Euclidean distance can well reflect the degree of human eye's perception of color differences; the greater the distance, the more significant the color difference.

[0071] S23: Randomly select a number of pixels as the initial cluster centers, assign each pixel to the cluster center with the closest Euclidean distance, and recalculate the center position of each cluster, that is, take the average value of the CIE-Lab coordinates of all pixels in the cluster as the new cluster center.

[0072] The clustering algorithm iteratively performs the following operations: first, randomly select a number of pixels as initial cluster centers; then assign each pixel to the cluster center with the closest Euclidean distance; then recalculate the center position of each cluster, i.e., take the average of the CIE-Lab coordinates of all pixels in the cluster as the new cluster center; repeat the assignment and update steps until the position of the cluster center no longer changes significantly.

[0073] S24: Repeat the assignment and update steps until the location of the cluster centers no longer changes significantly. At this point, the center coordinates of each cluster represent a primary color.

[0074] Step S24 includes:

[0075] For each candidate color quantity within the interval, a complete clustering process is performed. The average distance between each pixel and other pixels in the same cluster is calculated as the cohesion, and the average distance between each pixel and the nearest other cluster center is calculated as the separation. The difference between the separation and cohesion is divided by the larger of the two to obtain the contour value of the pixel. The contour values ​​of all pixels are averaged to obtain the contour coefficient of the clustering result. The candidate color quantity with the largest contour coefficient is selected as the final primary color quantity, and the clustering result with this quantity is used as the final primary color extraction result.

[0076] Furthermore, in this embodiment, a target image is converted to a color space and preprocessed to obtain a standardized image. This image contains multiple colors, but its main color composition may be between 3 and 7, while the ink mixing device supports up to 8 base ink channels. In this case, the system can set the minimum number of possible colors to 3 and the maximum number of possible colors to 7. Then, the system will perform the following operations in sequence:

[0077] When the number of candidate colors is 3, the K-Means clustering algorithm is performed on the normalized image to cluster all pixels into 3 clusters. After clustering, the contour coefficient of these 3 clusters is calculated, for example, 0.65.

[0078] When the number of candidate colors is 4, the K-Means clustering algorithm is executed again to cluster all pixels into 4 clusters. The contour coefficient of these 4 clusters is calculated, for example, 0.72.

[0079] When there are 5 candidate colors, the K-Means clustering algorithm is executed, resulting in 5 clusters. The silhouette coefficient is calculated, for example, 0.78.

[0080] When there are 6 candidate colors, the K-Means clustering algorithm is executed, resulting in 6 clusters. The silhouette coefficient is calculated, for example, 0.70.

[0081] When the number of candidate colors is 7, the K-Means clustering algorithm is executed, resulting in 7 clusters. The silhouette coefficient is calculated, for example, 0.62.

[0082] Among all the clustering results for the number of candidate colors, the one with the highest silhouette coefficient is 0.78, corresponding to a number of primary colors of 5. Therefore, the system will automatically determine that the number of primary colors in the image is 5, and use the result of this clustering (i.e., clustered into 5 clusters) as the final primary color extraction result. That is, the normalized image is identified as containing 5 primary colors, each primary color is represented by its corresponding cluster center (CIELab value), and the area ratio occupied by each primary color can be calculated based on the number of pixels contained in each cluster.

[0083] By introducing the silhouette coefficient as a quantitative evaluation index for clustering quality, and combining it with preset color quantity ranges and device base ink channel limitations, the adaptive determination of the number of primary colors is achieved. This allows the system to automatically select the number of primary colors that best represent the visual effect of the image based on its color distribution characteristics, avoiding deviations caused by subjective judgment or lack of experience. This scheme, combined with color recognition processing that converts pixel RGB values ​​to CIE-Lab values ​​and performs clustering in the CIE-Lab space, significantly improves the accuracy and intelligence of color recognition. Due to the automatic optimization of the number of primary colors, the extracted primary colors can more accurately reflect the true color composition of the image, thus providing more reliable and accurate input data for subsequent base ink ratio calculations. This not only improves the accuracy of ink mixing and reduces trial-and-error costs and manual intervention, but also ensures color consistency between batches, thereby improving the efficiency and automation of the entire automatic ink mixing control method.

[0084] Specifically, after acquiring the target image and performing color space conversion and preprocessing to obtain a standardized image, the RGB values ​​of each pixel in the standardized image are converted. Since the RGB color space does not linearly correspond to human visual perception, these RGB values ​​are converted to CIE-Lab values ​​to more accurately quantify color differences. The CIE-Lab space has perceptual uniformity, meaning that the Euclidean distance between color points in this space can more accurately reflect the degree of human visual perception of color differences. Through this conversion, each pixel obtains a unique coordinate point in the CIE-Lab three-dimensional space, laying a foundation for accurate perceptual analysis in subsequent color analysis.

[0085] Specifically, before performing clustering, the standardized images are first subjected to superpixel segmentation:

[0086] The image is divided into several small regions, and the pixel color difference in each small region is less than a preset threshold, thereby forming superpixel blocks with uniform internal color. While preserving the local structural information of the image, this significantly reduces the amount of data for subsequent processing and lowers the computational complexity.

[0087] Each superpixel block uses the CIE-Lab average of all pixels within it as the representative color of the block; by calculating the average, color noise and minor variations within the superpixel block can be effectively smoothed, making the clustering process focus more on the dominant color of the region.

[0088] Subsequent clustering operations use superpixel blocks as the basic unit rather than individual pixels, treating each superpixel block as a sample point for K-Means clustering. First, representative colors of several superpixel blocks are randomly selected as initial cluster centers. Then, each superpixel block is assigned to the cluster center with the closest Euclidean distance to its representative color. Next, the average value of the representative colors of all superpixel blocks within each cluster is recalculated as the new cluster center. This assignment and update process is repeated until the positions of the cluster centers no longer change significantly or the preset maximum number of iterations is reached. Besides K-Means, the MeanShift clustering algorithm can also be used. This algorithm iteratively moves the representative color of each superpixel block to the average value of its neighborhood, thereby finding density peaks in the data space; these peaks represent different dominant color clusters.

[0089] After clustering, each primary color corresponds to a group of superpixel blocks. By adding up the number of original pixels covered by all superpixel blocks in the group and dividing by the total number of pixels in the image, the area ratio occupied by that primary color can be obtained.

[0090] By introducing superpixel segmentation technology, the original image is transformed into a set of uniformly colored superpixel blocks, thus significantly reducing the computational dimensionality of subsequent clustering algorithms while preserving image structural information. Using the CIELab average value of each superpixel block as the representative color effectively filters out local noise and subtle color fluctuations in the image, allowing the clustering process to focus on color regions with significant visual characteristics. Using superpixel blocks as the basic unit of clustering not only significantly improves the computational efficiency of dominant color extraction but also ensures the accuracy and robustness of dominant color area statistics by calculating the area ratio by summing the pixel counts of the superpixel blocks.

[0091] S3: For each extracted main color, call the preset base ink spectral database and use the spectral matching optimization algorithm to calculate the mixing mass ratio of each base ink;

[0092] The implementation of the spectral matching optimization algorithm may include: First, setting an initial base ink mixing ratio based on experience; then, using data from the base ink spectral database, predicting the reflection spectrum of the ink at this mixing ratio through a linear superposition model; next, converting the predicted spectrum into color values ​​and comparing them with the color values ​​of the target primary color to calculate the color difference between the two; finally, adjusting the base ink mixing ratio iteratively based on the magnitude of the color difference, repeating the above process until the color difference reaches an acceptable range.

[0093] Specifically, for each extracted primary color, the process of calling a preset base ink spectral database and using a spectral matching optimization algorithm to calculate the mixing quality ratio of each base ink includes:

[0094] Image analysis has identified a target primary color, such as a specific blue. A pre-established reflectance sequence for each base ink at various wavelengths within the visible spectrum is created, forming a base ink spectral database. The system retrieves reflectance sequences for all available base inks (e.g., cyan, magenta, yellow, black, white, blue, green, etc.) from this pre-established database. This database may store reflectance data for each base ink measured at 10nm intervals within the 380nm to 780nm wavelength range.

[0095] For the target primary color to be matched, an initial vector is set, which is composed of the mixing ratio of all base inks, where the ratio of each base ink is a positive number and the sum is a fixed value.

[0096] The system first calculates the predicted reflectance sequence of the mixed ink by weighting and multiplying the reflectance values ​​of each base ink according to their mixing ratio. Then, it sets an initial vector; for example, assuming there are six base inks, the initial mixing ratio of each base ink can be set to 1 / 6. The system then calculates the predicted spectrum based on this initial vector. For example, if a linear weighted superposition model is used, the reflectance value of each base ink at each wavelength is multiplied by its current mixing ratio, and then the weighted reflectance values ​​of all base inks at the same wavelength are summed to obtain the predicted reflectance sequence of the mixed ink.

[0097] The predicted CIE-Lab color values ​​are calculated by integrating the predicted reflectance sequence wavelength by wavelength using the standard colorimetric conversion method, combined with the preset spectral distribution of the illumination source and the standard observer color matching function.

[0098] The color difference between the predicted CIE-Lab value and the target primary color is calculated using the CIEDE2000 color difference formula. Subsequently, this predicted reflectance sequence is converted into CIE-Lab color values. This is typically done by integrating the spectral data with the CIED65 standard illuminator spectral distribution and the CIE1931 standard observer color matching function to first obtain the XYZ tristimulus values, and then converting the XYZ values ​​to CIE-Lab values. Next, the system calculates the color difference between this predicted CIE-Lab color value and the target primary color's CIE-Lab value.

[0099] Furthermore, this color difference value is calculated using the CIEDE2000 color difference formula, which takes into account differences in brightness, chroma, and hue, and applies weighted corrections to more accurately reflect human eye perception of color.

[0100] With the goal of minimizing the color difference value, the mixing ratio of each base ink is repeatedly adjusted. After each adjustment, the predicted spectrum, predicted CIE-Lab value, and color difference value are recalculated until the color difference value is less than the preset qualified threshold.

[0101] If the calculated color difference value exceeds a preset acceptable threshold (e.g., DeltaE2000 is less than 1.0), the system will enter the iterative optimization phase. The system will adjust the mixing ratio of each base ink based on the current color difference value and the proportion of the base ink using an optimization algorithm (e.g., gradient descent based on finite differences). For example, the algorithm might identify that increasing the proportion of a certain base ink can effectively reduce the color difference.

[0102] The final adjusted mixing ratio is taken as the optimal base ink ratio for the main color. When the optimization process converges, the system outputs the current mixing ratio as the optimal base ink ratio for the target main color.

[0103] By establishing a comprehensive base ink spectral database and employing an optimization algorithm that combines spectral physical properties with human visual perception, a more accurate base ink ratio calculation for the target primary color was achieved. Specifically, weighted summation of the reflectance sequences of the base inks more realistically simulates the optical performance after ink mixing; the spectral data is converted into CIELab values ​​using a standard colorimetric conversion method, ensuring that the assessment of color differences is highly consistent with human visual perception; and the introduction of the CIEDE2000 color difference formula, with its comprehensive weighted correction of brightness, chroma, and hue differences, significantly improves the accuracy of color difference quantification. This iterative optimization process, aimed at minimizing color difference, ensures that the final output base ink ratio meets the requirements for high-precision color reproduction, effectively solving the problem of large visual differences between the mixed ink color and the target color, and satisfying the need for high-fidelity color reproduction.

[0104] Specifically, in the calculation of the predicted spectrum, if the base ink types include opaque or semi-transparent inks, the Kubelka-Munk model is used to process them, converting the reflectance value of each base ink into the corresponding absorption-scattering ratio value, and then weighting and superimposing the absorption-scattering ratios of each base ink according to the mixing ratio to obtain the absorption-scattering ratio of the mixed ink, and then converting it back to the reflectance value as the predicted spectrum of the mixed ink.

[0105] The Kubelka-Munk model is a theoretical model describing the propagation behavior of light in scattering and absorbing media, and is particularly suitable for predicting the optical properties of opaque or translucent materials such as coatings and inks. It quantifies the material's ability to absorb and scatter light using the absorption coefficient K and the scattering coefficient S. This model can employ a two-constant Kubelka-Munk model, where K and S are the intrinsic optical constants of the material at a specific wavelength, obtained through experimental measurement or inversion calculation, and can accurately describe the multiple scattering and absorption processes of light in a homogeneous medium.

[0106] The directly measured reflectance data is converted into the internal optical parameters required for the Kubelka-Munk model, namely the ratio of the absorption coefficient K to the scattering coefficient S (K / S), because K and S are additive when inks are mixed, while reflectance is not. This conversion can be achieved by solving the inverse problem of the Kubelka-Munk equation. For a single-layer ink film of known thickness, its reflectance is measured, and then, using numerical iteration methods or lookup tables, the K and S values ​​of the ink film at various wavelengths are derived from the Kubelka-Munk equation, thus obtaining the K / S ratio.

[0107] For each base ink, at each wavelength, its absorption coefficient K and scattering coefficient S can be multiplied by its volume or mass proportion in the mixture. Then, the weighted K and weighted S values ​​of all base inks are summed to obtain the total absorption coefficient of the mixed ink at that wavelength. and total scattering coefficient .

[0108] The absorption-scattering ratio (or K and S values) of the mixed ink, obtained by weighted superposition, is converted back to reflectance values ​​for comparison with the reflectance of the target primary color or for subsequent colorimetric calculations. This conversion can be achieved using the forward solution of the Kubelka-Munk equation. The reflectance of the mixed ink at various wavelengths is known. and The values, along with the preset ink layer thickness, can be directly substituted into the Kubelka-Munk equation to calculate the reflectance sequence of the mixed ink at various wavelengths.

[0109] Specifically, the process of repeatedly adjusting the mixing ratio of each base ink with the goal of minimizing color difference includes:

[0110] A small increment is applied to the proportion of each base ink, the color difference value is recalculated, and the change in color difference is divided by the proportion increment to obtain the gradient approximation of that base ink. The gradient values ​​of all base inks are combined into a gradient vector. This gradient vector is a vector containing the gradient values ​​corresponding to the proportions of all base inks. It represents the steepest rising direction of the color difference function in the multidimensional base ink proportion space under the current mixing ratio. The optimization algorithm will search along the opposite direction of this vector to reduce the color difference.

[0111] Using the BFGS update rule, an approximation of the inverse of the Hessian matrix is ​​constructed based on the gradient vector of the current iteration point and historical iteration information, thereby calculating the quasi-Newton direction. BFGS (Broyden–Fletcher–Goldfarb–Shanno) is a commonly used quasi-Newton update rule. It does not directly calculate the Hessian matrix (second derivative matrix), but instead uses the gradient information and variable change information of each iteration to gradually construct an approximation of the inverse of the Hessian matrix. This approximate Hessian inverse matrix contains information about the curvature of the objective function, enabling the algorithm to more accurately estimate the "quasi-Newton direction" leading to the minimum.

[0112] A one-dimensional line search is performed along the quasi-Newton direction to determine the optimal step size, thereby generating the next set of mixed scaling vectors;

[0113] Repeat the above process of gradient calculation, direction construction and step size search, and compare the current color difference value with the color difference value of the previous iteration after each iteration;

[0114] The iteration stops when the decrease in color difference value is less than the preset convergence threshold in multiple consecutive iterations, or when the color difference value itself is less than the qualified threshold.

[0115] After the iteration is completed, the proportion values ​​of each base ink are checked and interference data are filtered out. The proportions of the remaining base inks are then scaled up again to restore their sum to a fixed value, and the final mixing quality proportion vector is output.

[0116] By introducing an optimization calculation method combining gradient descent and a quasi-Newton method, the efficiency and accuracy of the spectral matching optimization algorithm are significantly improved. This method begins with an initial base ink vector of equal-proportion mixing as the starting point for optimization. In each iteration, the system accurately calculates the rate of change of the color difference value relative to each base ink proportion at the current mixing ratio, i.e., the gradient. This gradient is approximated by applying a small increment to each base ink proportion and recalculating the color difference, thus avoiding the need for complex analytical differentiation and ensuring the accuracy of the gradient information. Subsequently, these gradient values ​​are combined into a gradient vector, indicating the steepest upward direction of the color difference function at the current point. To more efficiently approximate the minimum color difference, this scheme further utilizes the BFGS update rule to construct an approximation of the Hessian matrix inverse based on the gradient vector at the current iteration point and information accumulated during historical iterations. This approximate Hessian inverse matrix contains second-order information about the curvature of the objective function, enabling the algorithm to calculate a "quasi-Newton direction." This direction not only considers the information of the current gradient but also incorporates the characteristics of the function curvature, thus pointing to the global optimum more accurately and quickly. This effectively avoids the problem that traditional gradient descent methods may get stuck in local optima or converge slowly.

[0117] S4: Based on the area ratio of each main color, the target printing area input by the user, and the preset ink layer thickness, combined with the mixing mass ratio, calculate the total cumulative mass required for each base ink in the entire production task.

[0118] In this embodiment, when a primary color occupies 20% of the area in an image, the target printing area is 10 square meters, the preset ink layer thickness is 5 micrometers, and the primary color is composed of base ink A and base ink B mixed in a 70:30 mass ratio, then the total ink volume required for the primary color can be calculated first, then converted to mass based on ink density, and then allocated to base ink A and base ink B in a 70:30 ratio. After performing similar calculations for all primary colors, the total mass of the same base ink in all primary colors can be obtained by summing the amounts used in all primary colors.

[0119] S5: Automatically generate an ink mixing instruction based on the accumulated total mass, and send the ink mixing instruction to the automatic ink mixing equipment to control the equipment to weigh each base ink in sequence and mix them.

[0120] The ink mixing instructions can be generated by encapsulating the name of each base ink and its corresponding total mass in text or a predefined binary format. For example, the instruction could include information such as "Base Ink 1: 1000 grams, Base Ink 2: 500 grams". These instructions can be sent to the automatic ink mixing equipment via a standard industrial communication interface, such as an RS232 serial port. After receiving the instruction, the automatic ink mixing equipment will use its internal weighing system to weigh the base inks and inject them into a mixing container for stirring, thereby completing the automatic preparation of the required ink.

[0121] Specifically, S5 includes:

[0122] The number of each base ink and its corresponding total accumulated mass are numerically encoded and organized into a data packet according to the preset instruction frame format;

[0123] The entire data packet is processed using a cyclic redundancy check algorithm to calculate a checksum, which is then appended to the end of the data packet.

[0124] The data packet is written to the designated register address of the automatic ink dispensing device using an industrial communication protocol.

[0125] Based on the analysis of the target image, it was calculated that three base inks need to be mixed: red base ink (No. 01), with a total cumulative mass of 1500.5 grams; blue base ink (No. 02), with a total cumulative mass of 800.2 grams; and yellow base ink (No. 03), with a total cumulative mass of 2200.8 grams.

[0126] First, the system will encode this data numerically. For example, the base number can be encoded as a 1-byte hexadecimal number (0x01, 0x02, 0x03), and the total accumulated mass can be quantified as an integer in milligrams and represented by 4 bytes.

[0127] Therefore, 1500.5 grams will become 1500500 milligrams, 800.2 grams will become 800200 milligrams, and 2200.8 grams will become 2200800 milligrams. Next, the system will organize the data packets according to a preset instruction frame format. For example, an instruction frame can be defined as: a frame header (2 bytes, such as 0xAA55), an instruction type (1 byte, such as 0x01 indicating an ink mixing instruction), a base ink quantity (1 byte, such as 0x03), then the data block for each base ink (base ink number 1 byte + accumulated total mass 4 bytes), and finally a checksum (2 bytes).

[0128] When organizing the base ink data blocks, the system queries a pre-defined database of base ink physical properties. Assume the red base ink has a density of 1.05 g / cm³ and a viscosity of 50 cP; the blue base ink has a density of 1.10 g / cm³ and a viscosity of 60 cP; and the yellow base ink has a density of 1.02 g / cm³ and a viscosity of 45 cP. Following the rule of "lower density base inks first, higher density base inks last," the weighing order will be: yellow base ink (1.02 g / cm³), red base ink (1.05 g / cm³), and blue base ink (1.10 g / cm³). Therefore, the base ink data blocks in the data package will be arranged in this order. The target weighing mass value (milligram integer) for each base ink will be arranged according to a specified byte order (e.g., little-endian). For example, 1500500 (0x0016E3A4) would be represented as 0xA4,0xE3,0x16,0x00 in little-endian. At the end of the data packet, the system uses a cyclic redundancy check algorithm (e.g., CRC-16) to calculate the checksum of the entire data packet (excluding the checksum itself) and appends the result to the end of the data packet. Finally, this complete data packet, containing the frame header, instruction type, base ink quantity, sorted base ink data blocks (numbered and quantized quality values), and checksum, is sent via an industrial communication protocol (e.g., Modbus RTU). Specifically, the control system uses the RS-485 interface to write this data packet as the data part of a Modbus write multiple holding registers instruction to a preset starting register address inside the automatic ink dispensing device, such as consecutive registers starting at address 40001. After receiving the instruction, the automatic ink dispensing device first performs a CRC check. If the check passes, it parses the data packet and drives the weighing pump or valve to weigh and mix the base inks sequentially according to the specified weighing order and quantized quality values.

[0129] Specifically, before generating the ink mixing instruction, a layered ink volume accumulation calculation is performed:

[0130] Based on the preset inkjet process parameters, the entire inkjet process is decomposed into multiple functional layers, including a base layer, at least one color layer, a gradient layer, and a protective layer.

[0131] For each primary color, calculate its contribution to the amount of ink in the color layer;

[0132] For gradient layers, the area of ​​the gradient band on the boundary of adjacent main color areas is calculated, and the mixing ratio of each position in the gradient band is determined by linear interpolation according to distance weight. Then, the total amount of ink of each base ink in the gradient layer is calculated by integration.

[0133] For the base layer and the protective layer, calculate the amount of base ink and protective ink according to the fixed coverage parameters set by the user;

[0134] The total mass of each type of base ink is obtained by summing the ink volume values ​​of all layers containing the same base ink.

[0135] Based on preset inkjet process parameters, the entire inkjet process is decomposed into multiple functional layers, including a base layer, at least one color layer, a gradient layer, and a protective layer. For each primary color, its ink contribution in the color layer is calculated by multiplying the area ratio of the primary color by the target coverage area of ​​the color layer by the ink layer thickness of the layer, and then by the mixing ratio of a certain base ink in the primary color formula to obtain the ink amount of that base ink in that primary color layer. For the gradient layer, the gradient band area on the boundary of adjacent primary color areas is calculated, and the mixing ratio of each position within the gradient band is determined by linear interpolation with distance weights. Then, the total ink amount of each base ink in the gradient layer is calculated by integration. For the base layer and the protective layer, the amount of undercoat ink and protective ink is directly calculated according to the user-defined fixed coverage parameters. The ink amount values ​​of the same base ink in all layers are summed to obtain the total cumulative mass of each base ink. The generated ink mixing instruction also carries a layer sequence parameter, which specifies the printing order of each layer and the required drying time for each layer.

[0136] Example 2:

[0137] Please refer to Figure 3 The present invention also provides an automatic ink dispensing control system for decorative paper, the control system being used to execute the control method, the control system comprising:

[0138] Image input module 10 is used to acquire a target image, perform color space conversion and preprocessing on the target image, and output a standardized image;

[0139] Color recognition module 20, any of the color recognition processes described herein, extracts the main color and the area ratio of each main color;

[0140] The formula calculation module 30 is used to perform any one of the spectral matching optimization processes for each main color and calculate the mixing mass ratio of each base ink.

[0141] The ink volume accumulation module 40 is used to calculate the total accumulated mass of each base ink based on the area ratio of each main color, the target printing area and the ink layer thickness, combined with the mixing mass ratio.

[0142] The instruction generation and communication module 50 is used to generate the ink dispensing instruction and send the instruction to the automatic ink dispensing equipment through an industrial communication protocol.

[0143] Specifically, the image input module 10 converts and preprocesses the target image in color space, transforming the original image into a standardized data format. This provides a unified input benchmark for subsequent color analysis, avoiding recognition errors caused by differences in image formats. The color recognition module 20 extracts the main color and area ratio based on the standardized image, enabling it to directly process images with complex features such as gradients and textures. This overcomes the limitations of traditional methods that can only handle simple color blocks or rely on manual selection, ensuring that the ink mixing scheme can cover all color requirements of the image. The formula calculation module 30 performs spectral matching optimization on the extracted main color. By calculating the mixing mass ratio of each base ink, it solves the problem that traditional linear color mixing formulas cannot handle the nonlinear superposition of base ink spectra, improving the accuracy of color reproduction. The ink volume accumulation module 40 combines the main color area ratio, target printing area, ink layer thickness, and mixing mass ratio to accurately calculate the total accumulated mass of each base ink, ensuring the accuracy of ink supply in production tasks and avoiding waste or insufficiency caused by manual estimation. The instruction generation and communication module 50 converts the calculated cumulative total mass into instructions that can be recognized by the automatic ink dispensing equipment, and sends them directly through the industrial communication protocol. This achieves seamless connection from data calculation to equipment execution, eliminates errors that may be introduced by manual data transmission, and ensures the traceability of the production process.

[0144] This application's solution constructs a closed-loop automated control process from image input to device command output through modular integration. Specifically, firstly, the image input module 10 acquires the target image and preprocesses it, outputting a standardized image; secondly, the color recognition module 20 analyzes the standardized image, extracting the primary color and its area ratio; then, the formula calculation module 30 calls the base ink spectral database for each primary color and calculates the base ink mixing ratio through a spectral matching optimization algorithm; next, the ink accumulation module 40 calculates the total amount of base ink used based on the area ratio, printing area, and ink layer thickness; finally, the command generation and communication module 50 generates commands and sends them to the device. This end-to-end automated process not only solves the problem of difficult ink mixing for complex images but also significantly improves the efficiency and consistency of ink mixing through full-process automation.

[0145] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc.

[0146] Furthermore, the embodiments of the present invention have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for automatic ink dispensing control of decorative paper, characterized in that, The control method includes: Acquire the target image, perform color space conversion and preprocessing on the target image to obtain a standardized image; Perform color recognition processing on the standardized image to extract at least one primary color contained in the image and determine the area ratio occupied by each primary color; For each extracted primary color, a preset base ink spectral database is called, and a spectral matching optimization algorithm is used to calculate the mixing mass ratio of each base ink. Based on the area ratio of each primary color, the target printing area input by the user, and the preset ink layer thickness, combined with the mixing mass ratio, the total cumulative mass required for each base ink in the entire production task is calculated. The ink mixing instruction is automatically generated based on the total accumulated mass and sent to the automatic ink mixing equipment to control the equipment to weigh each base ink in sequence and mix them.

2. The control method according to claim 1, characterized in that, The step of performing color recognition processing on the standardized image, extracting at least one primary color contained in the image, and determining the area proportion occupied by each primary color includes: Convert the RGB value of each pixel in the standardized image to a CIE-Lab value, making each pixel a coordinate point in the CIE-Lab three-dimensional space; Calculate the Euclidean distance between any two pixels in the CIE-Lab space, and use the Euclidean distance as a quantitative indicator of the color difference between the two pixels. Randomly select a number of pixels as the initial cluster centers, assign each pixel to the cluster center with the closest Euclidean distance, and recalculate the center position of each cluster, that is, take the average of the CIE-Lab coordinates of all pixels in the cluster as the new cluster center; Repeat the assignment and update steps until the location of the cluster centers no longer changes significantly. At this point, the center coordinates of each cluster represent a primary color.

3. The control method according to claim 2, characterized in that, The repeated assignment and update steps continue until the positions of the cluster centers no longer change significantly. At this point, the center coordinates of each cluster represent a primary color, including: For each candidate color quantity within the interval, a complete clustering process is performed. The average distance between each pixel and other pixels in the same cluster is calculated as the cohesion, and the average distance between each pixel and the nearest other cluster center is calculated as the separation. The difference between the separation and cohesion is divided by the larger of the two to obtain the contour value of the pixel. The contour values ​​of all pixels are averaged to obtain the contour coefficient of the clustering result. The candidate color quantity with the largest contour coefficient is selected as the final primary color quantity, and the clustering result with this quantity is used as the final primary color extraction result.

4. The control method according to claim 3, characterized in that, Before performing clustering, superpixel segmentation is first performed on the normalized images: The image is divided into several small regions, and the pixel color difference in each small region is less than a preset threshold, thereby forming superpixel blocks with uniform internal color. Each superpixel block uses the CIE-Lab average of all pixels within it as the representative color of that block; Subsequent clustering operations use superpixel blocks as the basic unit rather than individual pixels, treating each superpixel block as a sample point for K-Means clustering; After clustering, each primary color corresponds to a group of superpixel blocks. By adding up the number of original pixels covered by all superpixel blocks in the group and dividing by the total number of pixels in the image, the area ratio occupied by that primary color can be obtained.

5. The control method according to claim 1, characterized in that, For each extracted primary color, a preset base ink spectral database is invoked, and a spectral matching optimization algorithm is used to calculate the mixing quality ratio of each base ink, including: A numerical sequence of reflectance at each wavelength point in the visible spectrum for each type of base ink is pre-established to form a base ink spectral database; For the target primary color to be matched, an initial vector is set, which is composed of the mixing ratio of all base inks, where the ratio of each base ink is a positive number and the sum is a fixed value. The reflectance values ​​of each base ink are weighted and multiplied according to their mixing ratio, and then summed to obtain the predicted reflectance sequence of the mixed ink. The predicted CIE-Lab color values ​​are calculated by integrating the predicted reflectance sequence wavelength by wavelength using the standard colorimetric conversion method, combined with the preset spectral distribution of the illumination source and the standard observer color matching function. The color difference between the predicted CIE-Lab value and the CIE-Lab value of the target primary color is calculated using the CIEDE2000 color difference formula. With the goal of minimizing the color difference value, the mixing ratio of each base ink is repeatedly adjusted. After each adjustment, the predicted spectrum, predicted CIE-Lab value and color difference value are recalculated until the color difference value is less than the preset qualified threshold. The final adjusted mixing ratio is taken as the optimal base ink ratio for this main color.

6. The control method according to claim 5, characterized in that, In the calculation of the predicted spectrum, if the base ink types include opaque or semi-transparent inks, the Kubelka-Munk model is used to process them, converting the reflectance value of each base ink into the corresponding absorption-scattering ratio value. Then, the absorption-scattering ratios of each base ink are weighted and superimposed according to the mixing ratio to obtain the absorption-scattering ratio of the mixed ink, and then converted back to the reflectance value as the predicted spectrum of the mixed ink.

7. The control method according to claim 5, characterized in that, The process of repeatedly adjusting the mixing ratio of each base ink with the goal of minimizing color difference values ​​includes: Apply a small increment to the proportion of each base ink, recalculate the color difference value, divide the change in color difference by the proportion increment to obtain the gradient approximation of that base ink, and combine the gradient values ​​of all base inks into a gradient vector. Using the BFGS update rule, based on the gradient vector of the current iteration point and historical iteration information, an approximation of the inverse matrix of the Hessian matrix is ​​constructed, and then the quasi-Newton direction is calculated. A one-dimensional line search is performed along the quasi-Newton direction to determine the optimal step size, thereby generating the next set of mixed scaling vectors; Repeat the above process of gradient calculation, direction construction and step size search, and compare the current color difference value with the color difference value of the previous iteration after each iteration; The iteration stops when the decrease in color difference value is less than the preset convergence threshold in multiple consecutive iterations, or when the color difference value itself is less than the qualified threshold. After the iteration is completed, the proportion values ​​of each base ink are checked and interference data are filtered out. The proportions of the remaining base inks are then scaled up again to restore their sum to a fixed value, and the final mixing quality proportion vector is output.

8. The control method according to claim 1, characterized in that, The step of automatically generating an ink mixing instruction based on the accumulated total mass, sending the ink mixing instruction to the automatic ink mixing equipment, and controlling the equipment to weigh and mix each base ink sequentially includes: The number of each base ink and its corresponding total accumulated mass are numerically encoded and organized into a data packet according to the preset instruction frame format; The entire data packet is processed using a cyclic redundancy check algorithm to calculate a checksum, which is then appended to the end of the data packet. The data packet is written to the designated register address of the automatic ink dispensing device using an industrial communication protocol.

9. The control method according to claim 8, characterized in that, Before generating the ink mixing instructions, perform the layered ink volume accumulation calculation: Based on the preset inkjet process parameters, the entire inkjet process is decomposed into multiple functional layers, including a base layer, at least one color layer, a gradient layer, and a protective layer. For each primary color, calculate its contribution to the amount of ink in the color layer; For gradient layers, the area of ​​the gradient band on the boundary of adjacent main color areas is calculated, and the mixing ratio of each position in the gradient band is determined by linear interpolation according to distance weight. Then, the total amount of ink of each base ink in the gradient layer is calculated by integration. For the base layer and the protective layer, calculate the amount of base ink and protective ink according to the fixed coverage parameters set by the user; The total mass of each type of base ink is obtained by summing the ink volume values ​​of all layers containing the same base ink.

10. An automatic ink dispensing control system for decorative paper, characterized in that, The control system is used to execute the control method as described in any one of claims 1 to 9, and the control system includes: The image input module is used to acquire a target image, perform color space conversion and preprocessing on the target image, and output a standardized image. The color recognition module, through any of the color recognition processes described above, extracts the main color and the area ratio of each main color; The formula calculation module is used to perform any of the spectral matching optimization processes for each primary color and calculate the mixing mass ratio of each base ink. The ink volume accumulation module is used to calculate the total accumulated mass of each base ink based on the area ratio of each main color, the target printing area, and the ink layer thickness, combined with the mixing mass ratio. The instruction generation and communication module is used to generate the ink dispensing instruction and send the instruction to the automatic ink dispensing equipment through an industrial communication protocol.

Images