A control and hardware imaging optimization system for edge sharpening of printed patterns

Abstract

The application discloses a printing pattern edge sharpening processing control and hardware imaging optimization system and relates to the technical field of printing pattern imaging processing.The system comprises a fragile feature pre-analysis module, an ambient light acquisition module, a structured light compensation imaging module and a layered sharpening module.The fragile feature pre-analysis module identifies the intersection of fine lines and high-saturation color blocks and generates a feature map.The ambient light acquisition module acquires illumination parameters and locates a strong light reflection area.The structured light compensation imaging module projects adaptive structured light and optimizes imaging.The layered sharpening module differentiates and sharpens each layer according to ambient light parameters and outputs a high-quality image.The application realizes the collaborative improvement of edge sharpening and imaging quality by accurately identifying fragile printing features and combining ambient light changes for imaging optimization.The application constructs an imaging strategy that adapts to the scene, differentiates and sharpens layers, guarantees accurate pattern restoration and meets the requirements of high-precision printing detection and display by relying on feature analysis and light parameter acquisition.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of printed pattern imaging processing, in particular to a control and hardware imaging optimization system for printed pattern edge sharpening processing. BACKGROUND

[0002] In the field of printed pattern imaging and processing, accurately restoring pattern details and ensuring edge clarity are one of the core requirements, especially in printed quality detection, product display and other scenarios, the presentation of pattern details directly affects the judgment of printed effect and subsequent application. With the continuous development of printing technology, pattern design is becoming more and more sophisticated, and features such as thin lines and high-saturation color block junctions are easily affected by external environment in the imaging process. The intensity, color temperature and distribution of ambient light as a key interference factor will cause problems such as strong light reflection and weak light detail loss during imaging, thereby affecting the accurate presentation of pattern edges and textures. At the same time, the parameter adaptation capability of the imaging device itself is limited, and it is difficult to dynamically adjust the imaging strategy according to different pattern features and light environment, so how to realize high-quality imaging and edge sharpening processing of printed patterns in complex light environment has become a technical direction that needs to be solved in the industry.

[0003] Traditional printed pattern imaging and sharpening processing technology has obvious shortcomings and cannot meet the high-precision requirements. In terms of feature recognition, traditional technology mainly uses single-dimensional analysis of pattern features, which cannot fully capture the key information of vulnerable areas such as thin lines and high-saturation color block junctions, resulting in lack of pertinence in subsequent processing. In terms of ambient light adaptation, traditional methods often use fixed imaging parameters and cannot adjust according to real-time light environment changes. In strong light scenes, reflection noise points may cover pattern details, and in weak light scenes, texture may be blurred due to insufficient gray scale range. In the sharpening processing link, traditional technology mainly uses unified sharpening for the whole image, which may produce noise points in the background layer due to excessive sharpening, and may not clearly show the edge and texture details due to insufficient sharpening, making it difficult to balance between highlighting details and overall texture, and ultimately affecting the accuracy and visual effect of printed pattern imaging. SUMMARY

[0004] The purpose of the present application is to make up for the shortcomings of the prior art, and provide a control and hardware imaging optimization system for printed pattern edge sharpening processing. The fragile feature pre-analysis module identifies features such as thin lines and high-saturation color block junctions, generates a feature map, the ambient light acquisition module obtains parameters such as illumination intensity and color temperature, locates the strong light reflection area, the structured light compensation imaging module projects adaptive structured light, synchronously adjusts the sensor parameters, the texture layering module divides the edge, texture and background layers, the layered sharpening module performs differential sharpening according to the ambient light and layer characteristics, and the imaging quality of the printed pattern is improved.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a control and hardware imaging optimization system for edge sharpening processing of printed patterns, the system comprising:

[0006] Vulnerable Feature Pre-analysis Module: Used to import the original design file of the printed pattern, identify fine lines with a line width of <0.5mm through pixel-level contour scanning and record the coordinates and direction, calculate the color difference value of adjacent color blocks through color value comparison analysis, define the area with a color difference value >50 as the boundary area of ​​high saturation color blocks, divide the feature priority according to two dimensions and count the area of ​​a single feature to generate a feature map;

[0007] The feature map is stored in XML format, and the stored content includes basic feature information, feature geometric attributes, and feature association information. The basic feature information includes feature type identifier, feature coordinate range, feature priority level, and area of ​​a single feature. The feature geometric attributes include the direction angle of thin lines, the radius of curvature of curved lines, and the edge smoothness of the boundary area of ​​high-saturation color blocks. The feature association information includes the layer affiliation of each vulnerable feature in the original design file and the relative positional relationship between each feature. After the feature map is generated, it is directly transmitted to the structured light compensation imaging module to provide complete data support for structured light projection path planning and compensation parameter calculation.

[0008] Ambient light acquisition module: This module integrates a multispectral sensor with an imaging device to acquire light intensity and color temperature parameters at a frequency of ≥30 frames / second. It uses image brightness analysis to locate the coordinates and area of ​​strong light reflection areas and organizes the parameters into JSON format for transmission to subsequent modules.

[0009] Structured light compensation imaging module: used to receive feature map and ambient light parameters, determine the light environment adaptation coefficient and structured light projection angle, call the adaptive structured light compensation amount calculation model to generate structured light pattern, control the micro digital micromirror device or tunable diffractive optical element to project structured light, synchronously adjust the white balance and exposure compensation of the imaging sensor, capture and transmit the optimized original image.

[0010] The pixel array of the micro digital micromirror device has a size of 1024×768, and the switching speed of a single micromirror is no less than 15 microseconds. It can generate structured light patterns with a minimum linewidth of 0.1 mm. The divergence angle of the adjustable diffraction optical element can be adjusted from 5 degrees to 30 degrees, and the coverage of the structured light can be dynamically adjusted according to the area size of the fragile feature.

[0011] Texture layering module: It receives the optimized original image, uses Gaussian filtering to filter reflection noise in strong light scenes, adjusts the grayscale range of the image through grayscale stretching in low light scenes, and divides the image into edge layer, texture layer and background layer through edge detection, texture complexity analysis and color consistency analysis, adds markers to the layers and transmits the layering results;

[0012] Layered sharpening module: It receives ambient light parameters and layering results, calculates the ambient light intensity ratio and color temperature deviation, calls the layered dynamic sharpening gain coefficient model to determine the sharpening gain of each layer, performs Sobel operator sharpening on the edge layer, performs bicubic B-spline interpolation sharpening on the texture layer, performs weakening sharpening on the background layer, integrates the layers through image alignment, and outputs the final image.

[0013] Furthermore, in the fragile feature pre-analysis module, feature priorities are divided according to two dimensions: feature type weight dimension and feature self-parameter level dimension. The feature priority value is obtained by multiplying the parameters of the two dimensions, and the priority value ranges from 1.0 to 5.0.

[0014] The feature type weight dimension is set as follows: the weight corresponding to the thin line feature is 2.0, and the weight corresponding to the high saturation color block boundary feature is 1.5.

[0015] The feature's own parameter level dimension is divided into three levels, and the corresponding level coefficients for each level are: 2.5 for level 1, 1.6 for level 2, and 0.8 for level 3.

[0016] The inherent parameter of fine line features is line width. Line width < 0.2mm is classified as Level 1, line width 0.2-0.3mm is classified as Level 2, and line width 0.3-0.5mm is classified as Level 3.

[0017] The intrinsic parameters of the boundary features of highly saturated color blocks are the color difference value and the area of ​​the continuous region. When the color difference value is >80 and the area of ​​the continuous region is >500 square pixels, it is classified as Level 1; when the color difference value is 65-80 and the area of ​​the continuous region is 300-500 square pixels, it is classified as Level 2; when the color difference value is 50-65 and the area of ​​the continuous region is 10-300 square pixels, it is classified as Level 3.

[0018] The feature priority is calculated as follows: Feature priority = Feature type weight × Feature's own parameter level coefficient;

[0019] The statistical method for the area of ​​a single feature is as follows: the area of ​​a thin line feature is the product of the line width and the line length; the area of ​​a high-saturation color block boundary feature is the product of the length and width of the bounding rectangle of the boundary region. The above feature priority values ​​and individual feature area information are all written into the feature map.

[0020] Furthermore, in the ambient light acquisition module, the multispectral sensor is integrated with the imaging device in the following way: the sensor and the lens of the imaging device are coaxially mounted, and the data acquired by the sensor includes the following:

[0021] Illumination intensity data: Includes illuminance in the visible light band and illuminance in the near-infrared light band. The detection wavelength range for the visible light band is 400nm-760nm, and the detection wavelength range for the near-infrared light band is 760nm-1000nm.

[0022] Color temperature data: includes the dominant color temperature value of the current environment and the uniformity of color temperature distribution. The uniformity of color temperature distribution is obtained by calculating the variance of the color temperature values ​​in different areas of the captured image.

[0023] Light intensity distribution data: contains the light intensity value of each pixel area in the captured image, and determines the area of ​​light intensity gradient change by the distribution of light intensity values;

[0024] Strong light reflection area data: includes the pixel coordinate range of the strong light reflection area, the average light intensity value within the area, and the standard deviation of the light intensity within the area.

[0025] Furthermore, in the ambient light acquisition module, when locating the coordinates and area of ​​the strong light reflection area through image brightness analysis, the acquired ambient image is first processed into grayscale, a brightness threshold of 240 is set, and areas with grayscale values ​​greater than or equal to the threshold are marked as candidate reflection areas. Then, the candidate reflection areas are filtered to remove isolated areas with an area of ​​less than 50 square pixels. Finally, the coordinate range of the strong light reflection area is determined and its area is calculated.

[0026] Furthermore, in the structured light compensation imaging module, the mathematical expression of the adaptive structured light compensation amount calculation model is: ;in, To compensate for the intensity of structured light, For light environment adaptation coefficient, Prioritize vulnerable features. Given the current ambient light intensity, The standard light intensity for printing imaging, For the area of ​​a single fragile feature, The total area of ​​the image. The angle between the direction of structured light projection and the direction of ambient light incidence.

[0027] Furthermore, in the structured light compensation imaging module, when adjusting the white balance parameters of the imaging sensor, a color temperature shift compensation method is adopted, shifting towards cooler tones by 200-500K in strong light scenes and towards warmer tones by 150-300K in weak light scenes; when adjusting exposure compensation, a local area independent adjustment strategy is adopted, performing exposure compensation of -0.5 to -1.0EV only on strong light reflection areas, while keeping the original exposure parameters stable in non-reflective areas; before projecting structured light, the shadow coverage range of each vulnerable feature is predicted by matching the feature map with the ambient light parameters, generating a structured light projection path planning map to ensure that the angle between the structured light projection direction and the shadow direction is controlled between 30° and 60°, and the structured light coverage range is 5% larger than the vulnerable feature area.

[0028] Furthermore, in the texture layering module, the specific details of using Gaussian filtering to filter reflection noise in strong light scenes and adjusting the image grayscale range through grayscale stretching in low light scenes are as follows:

[0029] In the high-light scene, a 3×3 kernel Gaussian filter is used to filter reflection noise. The weight matrix of the Gaussian filter kernel is as follows: This filtering process can specifically filter isolated reflection noise points in an image with a brightness value greater than or equal to 240, while preserving the contour information of fragile features.

[0030] In the low-light scenario, linear gray-scale stretching is used to adjust the gray-scale range of the image. First, the minimum and maximum gray-scale values ​​of the image are extracted, and then the original gray-scale range is mapped to the standard gray-scale range of 0 to 255 through linear mapping. During the mapping process, the linear relationship of gray-scale changes in each region of the image is maintained to ensure that the gray-scale difference of the image is increased to more than 80 in the low-light environment, thereby enhancing the recognizability of texture details.

[0031] Furthermore, the specific steps taken by the texture layering module to divide the edge layer, texture layer, and background layer through edge detection, texture complexity analysis, and color consistency analysis are as follows:

[0032] The edge detection is set with a fixed low threshold of 50 and a high threshold of 150 for gray-level gradient. Edge pixels with gray-level gradient values ​​exceeding the high threshold are extracted from the image. The region of the edge pixel and one pixel on each side is defined as the edge layer. The edge layer contains the boundary features of thin lines and highly saturated color blocks.

[0033] The texture complexity analysis uses a 100×100 pixel sliding window to traverse the image and calculates the number of times the pixel grayscale value changes within each window. When the number of changes is greater than 5, the window area is defined as a texture layer, which contains gradient patterns and detail patterns.

[0034] The color consistency analysis also uses a 100×100 pixel sliding window to calculate the standard deviation of the pixel grayscale value in each window. When the standard deviation is less than 10, the window area is defined as the background layer, which is the fabric substrate area.

[0035] Furthermore, in the layered sharpening module, the mathematical expression for the layered dynamic sharpening gain coefficient model is: ;in, This is the layer sharpening gain factor. For the base coefficients of the layer, Given the current ambient light intensity, The standard light intensity for printing imaging, This represents the ambient color temperature deviation value. This is the layer detail density coefficient.

[0036] Furthermore, in the layered sharpening module, the specific content of performing Sobel operator sharpening on the edge layer, bicubic B-spline interpolation sharpening on the texture layer, and weakening sharpening on the background layer is as follows:

[0037] When performing Sobel operator sharpening on the edge layer, the edge effect is enhanced by morphological dilation, with a dilation magnitude of 1 pixel to ensure that the edge contours are clearly distinguishable.

[0038] When performing bicubic B-spline interpolation sharpening on the texture layer, if the ambient color temperature is below 4500K, it is defined as a warm light environment, and the interpolation step size accuracy is set to 0.1 pixels. Interpolation calculation is performed using the gray values ​​of 16 adjacent pixels. If the ambient color temperature is above 6500K, it is defined as a cool light environment, and gray-level gradient amplification processing is added to amplify the area with a gray-level gradient difference of less than 5 to a gray-level gradient difference of greater than 10.

[0039] When performing weakening and sharpening on the background layer, the cutoff frequency of the Gaussian low-pass filter is adjusted. In strong light scenes, the cutoff frequency is reduced by 30% to 40%, and in weak light scenes, the cutoff frequency is reduced by 10% to 20%, preserving the original texture of the fabric substrate.

[0040] Compared with existing technologies, this control and hardware imaging optimization system for edge sharpening of printed patterns has the following advantages:

[0041] I. This invention accurately identifies key vulnerable features in printed patterns and performs targeted imaging optimization based on dynamic changes in ambient light, achieving a synergistic improvement in edge sharpening and image quality. Relying on multi-dimensional feature analysis and ambient light parameter acquisition, it constructs imaging strategies adapted to different scenarios, effectively avoiding interference from strong light reflection and dim light on pattern details, and maintaining clear outlines in easily distorted areas such as the intersection of fine lines and high-saturation color blocks. By dividing the image into different layers through a layered processing concept and adopting differentiated sharpening methods, it not only enhances the recognizability of key features but also avoids increased noise caused by over-sharpening, enabling printed patterns to still reproduce the original design intent in complex lighting environments and improving the accuracy and visual texture of the pattern presentation.

[0042] II. This invention achieves intelligent control of the entire process from feature recognition and light environment adaptation to sharpening through deep integration of hardware and algorithm models. It uses structured light compensation technology to specifically compensate for ambient light defects and optimizes imaging sensor parameters to ensure the original image quality, laying a good foundation for subsequent sharpening processing. The layered dynamic sharpening gain coefficient model combines ambient light intensity and color temperature changes to flexibly adjust the sharpening intensity of each layer. While highlighting edge and texture details, it retains the natural texture of the background layer, achieving a balance between detail presentation and overall harmony. The entire system can adapt to different lighting environments and pattern types without manual intervention, significantly improving the stability and adaptability of printed pattern imaging and meeting the actual needs of high-precision printed detection and display.

[0043] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0045] Figure 1 A flowchart of a control and hardware imaging optimization system for edge sharpening of printed patterns;

[0046] Figure 2 This is a schematic diagram of data transmission between various steps in a control and hardware imaging optimization system for edge sharpening of printed patterns.

[0047] Figure 3 This is a schematic diagram of data transmission for the structured light compensation imaging module. Detailed Implementation

[0048] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0049] Example 1:

[0050] Processing of patterns at the intersection of fine lines and highly saturated color blocks under strong light in a garment fabric printing workshop.

[0051] Import the original design file of the garment fabric containing the boundary pattern of fine lines and high-saturation color blocks. Through pixel-level contour scanning, accurately identify fine lines with line widths of 0.1mm, 0.3mm, and 0.45mm, record the coordinates and direction of these fine lines, and at the same time, use color value comparison analysis to calculate the color difference value of adjacent color blocks. Areas with color difference values ​​of 60, 75, and 90, and greater than 50, are defined as the boundary areas of high-saturation color blocks, providing a clear data foundation for subsequent targeted processing of vulnerable features. Feature priority is assigned based on a two-dimensional classification. Fine lines have a weight of 2.0, while high-saturation color block boundaries have a weight of 1.5. Fine lines are graded according to line width: 0.1mm line width is grade 1 with a coefficient of 2.5, 0.3mm line width is grade 2 with a coefficient of 1.6, and 0.45mm line width is grade 3 with a coefficient of 0.8. High-saturation color block boundaries are graded based on color difference and continuous area: a color difference value of 90 and a continuous area exceeding 500 square pixels is grade 1 with a coefficient of 2.5, and a color difference value of 75 and a continuous area between 300 and 500 square pixels is grade 2. A corresponding level coefficient of 1.6, a color difference value of 60, and a continuous area of ​​10-300 square pixels correspond to a level three level with a corresponding level coefficient of 0.8. The feature priority value is obtained by multiplying the feature type weight by its own parameter level coefficient. When calculating the area of ​​a single feature, the area of ​​thin lines is calculated as the product of the line width and line length, and the area of ​​the boundary feature of high-saturation color blocks is calculated as the product of the length and width of the bounding rectangle of the boundary area. Finally, an XML-formatted feature map containing basic feature information, geometric attributes, and correlation information is generated and transmitted to the structured light compensation imaging module, providing complete data support for the module's path planning and parameter calculation. Figure 1 As shown.

[0052] The multispectral sensor is coaxially integrated with the lens of the imaging device to collect parameters of the strong light environment in the workshop at a frequency of no less than 30 frames per second. The collected light intensity data covers the light intensity of the visible light band (400nm-760nm) and the near-infrared light band (760nm-1000nm), which can comprehensively reflect the spectral distribution of ambient light. The color temperature data includes the current dominant color temperature value of the environment and the uniformity of color temperature distribution obtained by calculating the variance of color temperature values ​​in different areas of the image, which can accurately grasp the ambient color temperature status. The light intensity distribution data records the light intensity value of each pixel area in the collected image, which helps to determine the areas of light intensity gradient change. The strong light reflection area data includes the pixel coordinate range, average light intensity value and light intensity standard deviation of the area, providing detailed basis for subsequent processing of strong light interference. The acquired environmental images are converted to grayscale, and a brightness threshold of 240 is set. Areas with grayscale values ​​greater than or equal to this threshold are marked as candidate reflection areas. Isolated areas with an area of ​​less than 50 square pixels are filtered out to accurately determine the coordinates and area of ​​strong light reflection areas. Finally, all parameters are organized into JSON format and transmitted to subsequent modules to ensure that subsequent modules can obtain accurate ambient light data.

[0053] The system receives feature maps and ambient light parameters, determines the light environment adaptation coefficient and the structured light projection angle, and uses an adaptive structured light compensation calculation model to generate a structured light pattern. This generated pattern adapts to the current ambient light and vulnerable features. The mathematical expression for the adaptive structured light compensation calculation model is as follows: ;in, To compensate for the intensity of structured light, For light environment adaptation coefficient, Prioritize vulnerable features. Given the current ambient light intensity, The standard light intensity for printing imaging, For the area of ​​a single fragile feature, The total area of ​​the image. The angle between the structured light projection direction and the ambient light incident direction. A miniature digital micromirror device with a pixel array size of 1024×768 and a single micromirror switching speed of no less than 15 microseconds is controlled to generate and project a structured light pattern with a minimum linewidth of 0.1 mm. Simultaneously, the white balance and exposure compensation of the imaging sensor are adjusted. The white balance adjustment adopts a color temperature shift compensation method, shifting towards cooler tones by 200-500K in strong light scenes to ensure accurate image color reproduction. The exposure compensation adopts a local area independent adjustment strategy, performing exposure compensation of -0.5 to -1.0 EV only on strong light reflection areas, while maintaining the original exposure parameters of non-reflective areas to avoid overall exposure imbalance. Before projecting structured light, the shadow coverage of each vulnerable feature is predicted through matching analysis of feature maps and ambient light parameters. A structured light projection path planning map is then generated to ensure that the angle between the structured light projection direction and the shadow direction is between 30° and 60°, and that the structured light coverage area is 5% larger than the vulnerable feature area. This effectively avoids the influence of shadows on the imaging of vulnerable features. Finally, the optimized original image is captured and transmitted, providing high-quality image material for subsequent texture layering. Figure 3 As shown.

[0054] After receiving the optimized original image, a 3×3 kernel Gaussian filter is used to filter reflection noise because it is in a strong light scene. This filter can specifically filter isolated reflection noise with a brightness value greater than or equal to 240 in the image, while preserving the contour information of fragile features. This eliminates noise interference caused by strong light without destroying key features. Subsequently, edge detection, texture complexity analysis, and color consistency analysis were used to divide the image into layers. For edge detection, a fixed low threshold of 50 and a high threshold of 150 were set for the grayscale gradient. Edge pixels with grayscale gradient values ​​exceeding the high threshold were extracted. The area of ​​the edge pixel and one pixel on each side was defined as the edge layer containing the boundary features of fine lines and highly saturated color blocks, accurately separating key edge features. For texture complexity analysis, a 100×100 pixel sliding window was used to traverse the image, calculating the number of times the pixel grayscale value changed within each window. Window areas with more than 5 changes were defined as texture layers containing gradient patterns and detailed designs, accurately identifying texture-rich areas. For color consistency analysis, a 100×100 pixel sliding window was also used to calculate the standard deviation of the pixel grayscale value within each window. Window areas with a standard deviation less than 10 were defined as the background layer of the fabric base area, clearly delineating the background portion. Marks were added to each layer, and the layering results were transmitted, laying the foundation for subsequent layered sharpening.

[0055] The system receives ambient light parameters and layering results, calculates the ambient light intensity ratio and color temperature deviation, and uses a layered dynamic sharpening gain coefficient model to determine the sharpening gain of each layer, ensuring that the sharpening process adapts to the current ambient light conditions. The mathematical expression for the layered dynamic sharpening gain coefficient model is as follows: ;in, This is the layer sharpening gain factor. For the base coefficients of the layer, Given the current ambient light intensity, The standard light intensity for printing imaging, This represents the ambient color temperature deviation value. This represents the layer detail density coefficient. After applying Sobel sharpening to the edge layer, morphological dilation is used to enhance the edge effect by 1 pixel, making the edge contours clearer and more discernible. Bicubic B-spline interpolation sharpening is applied to the texture layer. Given a strong light environment with a color temperature above 6500K (defined as a cool light environment), grayscale gradient amplification is added, increasing areas with a grayscale gradient difference less than 5 to a difference greater than 10, improving the clarity of texture details. A weakening sharpening is applied to the background layer by adjusting the cutoff frequency of a Gaussian low-pass filter. The weight matrix of this Gaussian filter kernel is: This filtering process can specifically filter isolated reflection noise points with a brightness value greater than or equal to 240 in the image, while preserving the contour information of fragile features. In strong light scenes, the cutoff frequency is reduced by 30% to 40%, preserving the original texture of the fabric substrate and avoiding over-sharpening that would damage the background effect. Finally, the layers are integrated through image alignment and the final image is output, resulting in a printed pattern image with clear edges, delicate texture, and a natural background.

[0056] In summary, in the strong light environment processing of garment fabric printing workshops, the following steps are taken: First, fragile feature pre-analysis accurately locates the boundary features between fine lines and highly saturated color blocks, generating a detailed feature map; then, comprehensive strong light environment parameters are obtained through ambient light acquisition, providing a basis for subsequent processing; structured light compensation imaging combines feature and environmental data to generate suitable structured light and adjust imaging parameters to obtain an optimized image; texture layering effectively filters noise and divides layers to separate key feature areas; layered sharpening adopts an adaptation method for different layers, ultimately outputting a printed image with clear edges, delicate texture, and a natural background. The entire process is closely integrated, fully adapts to the strong light environment, and ensures the processing effect of garment fabric printing patterns.

[0057] Example 2:

[0058] Gradient patterns and basic designs are processed in a low-light environment in the home textile fabric printing laboratory.

[0059] Import the original design file containing gradient patterns and basic designs from the home textile fabric. Identify fine lines with line widths of 0.2mm and 0.35mm through pixel-level contour scanning, record the coordinates and directions of these fine lines, and calculate the color difference value of adjacent color blocks by using color value comparison analysis. Areas with color difference values ​​of 55, 70 and greater than 50 are defined as high-saturation color block boundary areas, clearly locating the fragile features in the pattern and providing a clear target for subsequent processing. Feature priorities are assigned using a two-dimensional approach: fine lines have a weight of 2.0, and high-saturation color block boundaries have a weight of 1.5. Within fine lines, a 0.2mm line width is classified as a second-level feature with a corresponding coefficient of 1.6, and a 0.35mm line width is classified as a third-level feature with a corresponding coefficient of 0.8. For high-saturation color block boundaries, a color difference value of 55 with a continuous area of ​​200 square pixels is classified as a third-level feature with a corresponding coefficient of 0.8, and a color difference value of 70 with a continuous area of ​​400 square pixels is classified as a second-level feature with a corresponding coefficient of 1.6. Feature priority values ​​are obtained by multiplying the feature type weight by its own parameter coefficient. When calculating the area of ​​a single feature, the area of ​​fine lines is calculated as the product of line width and line length, and the area of ​​high-saturation color block boundaries is calculated as the product of the length and width of the bounding rectangle of the boundary area. This generates an XML-formatted feature map containing basic feature information, geometric attributes, and related information, which is then transmitted to the structured light compensation imaging module to provide accurate data support for its operation. Figure 2 As shown.

[0060] A multispectral sensor is coaxially integrated with the imaging device lens to acquire laboratory low-light environment parameters at a frequency of no less than 30 frames per second. Illumination intensity data covers the 400nm-760nm visible light band and the 760nm-1000nm near-infrared light band, comprehensively capturing light intensity information under low-light conditions. Color temperature data includes the primary color temperature value and the color temperature distribution uniformity obtained by calculating the variance of color temperature values ​​in different areas of the image, accurately reflecting the color temperature conditions of the low-light environment. Light intensity distribution data records the light intensity value of each pixel area, helping to determine areas of light intensity gradient change. Strong light reflection area data is acquired according to standards, providing complete ambient light data for subsequent processing. The acquired environmental images are grayscaled, a brightness threshold of 240 is set, candidate reflection areas are marked, and isolated areas with an area less than 50 square pixels are filtered out to accurately determine the coordinates and area of ​​strong light reflection areas. All parameters are compiled into JSON format and transmitted to subsequent modules to ensure that subsequent modules can operate based on accurate ambient light data.

[0061] The system receives feature maps and ambient light parameters, determines the light environment adaptation coefficient and the structured light projection angle, and uses an adaptive structured light compensation calculation model to generate a structured light pattern. The mathematical expression of the adaptive structured light compensation calculation model is as follows: ;in, To compensate for the intensity of structured light, For light environment adaptation coefficient, Prioritize vulnerable features. Given the current ambient light intensity, The standard light intensity for printing imaging, For the area of ​​a single fragile feature, The total area of ​​the image. The angle between the structured light projection direction and the ambient light incident direction is set to ensure the structured light pattern meets imaging requirements in low-light environments. An adjustable diffraction optical element with a divergence angle adjustable from 5 to 30 degrees is used, based on the size of the vulnerable feature area, to dynamically adjust the structured light coverage and projection, ensuring precise coverage of the vulnerable feature area. Simultaneously, the white balance and exposure compensation of the imaging sensor are adjusted. White balance adjustment uses a color temperature shift compensation method, shifting towards warmer tones by 150-300K in low-light scenes to improve the problem of cool colors in low-light images. Exposure compensation employs a local area independent adjustment strategy; if there are a few areas with strong light reflection, exposure compensation of -0.5 to -1.0 EV is applied, while non-reflective areas maintain their original exposure parameters, maintaining overall image exposure stability. Before projecting structured light, the coverage area of ​​vulnerable feature shadows is predicted by matching feature maps with ambient light parameters. A structured light projection path planning map is generated to ensure that the angle between the structured light projection direction and the shadow direction is between 30° and 60°, and that the coverage area of ​​the structured light is 5% larger than that of the vulnerable feature area. This reduces the impact of shadows on imaging, captures and transmits the optimized original image, and provides high-quality images for texture layering.

[0062] The system receives the optimized original image. Due to the low-light scene, the image grayscale range is adjusted by linear grayscale stretching. First, the minimum and maximum grayscale values ​​of the image are extracted. Then, the original grayscale range is mapped to the standard grayscale range of 0 to 255 through linear mapping. During the mapping process, the linear relationship of grayscale changes in each region of the image is maintained, ensuring that the grayscale difference of the image is increased to more than 80 in low-light environment. This significantly enhances the recognizability of texture details and solves the problem of blurred image details in low-light environment. Next, edge detection, texture complexity analysis, and color consistency analysis were used to divide the image into layers. Edge detection defined the edge layer containing the boundary features of fine lines and highly saturated color blocks based on the set grayscale gradient low threshold of 50 and high threshold of 150, accurately separating edge features. Texture complexity analysis used a 100×100 pixel sliding window to traverse the image, defining the window area where the pixel grayscale value changed more than 5 times as the texture layer containing gradient patterns, accurately identifying texture areas. Color consistency analysis used a 100×100 pixel sliding window to define the window area where the pixel grayscale value standard deviation was less than 10 as the background layer, clearly distinguishing the background part. Marks were added to each layer and the layering results were transmitted to prepare for layered sharpening.

[0063] The system receives ambient light parameters and layering results, calculates the ambient light intensity ratio and color temperature deviation, and uses a layered dynamic sharpening gain coefficient model to determine the sharpening gain of each layer, ensuring that the sharpening process conforms to the image characteristics in low-light environments. The mathematical expression for the layered dynamic sharpening gain coefficient model is as follows: ;in, This is the layer sharpening gain factor. For the base coefficients of the layer, Given the current ambient light intensity, The standard light intensity for printing imaging, This represents the ambient color temperature deviation value. This represents the layer detail density coefficient. After applying Sobel sharpening to the edge layer, a morphological dilation of 1 pixel is performed to further enhance the edge effect and make the edges clearer. For the texture layer, bicubic B-spline interpolation sharpening is applied. The current ambient color temperature is defined as a warm light environment below 4500K, and the interpolation step size is set to 0.1 pixels. Interpolation calculations are performed using the grayscale values ​​of 16 adjacent pixels to improve the fineness of texture details. For the background layer, a weakening sharpening is applied by adjusting the cutoff frequency of a Gaussian low-pass filter. The weight matrix of this Gaussian filter kernel is: This filtering process can specifically filter isolated reflection noise points with a brightness value greater than or equal to 240 in the image, while preserving the contour information of fragile features. In low-light scenes, the cutoff frequency is reduced by 10% to 20%, effectively preserving the original texture of the fabric substrate and avoiding excessive background sharpening that makes it appear rough. Finally, the layers are integrated through image alignment and the final image is output to obtain a high-quality image that meets the printing requirements of home textile fabrics.

[0064] In summary, in the low-light environment processing of home textile fabric printing laboratories, the pre-analysis of fragile features accurately identifies fragile features in the pattern, and the generated feature map provides data support for subsequent steps; ambient light acquisition comprehensively captures low-light environment parameters, assisting subsequent modules in precise operation; structured light compensation imaging generates appropriate structured light based on data and adjusts parameters to improve low-light imaging quality; texture layering enhances detail recognition through grayscale stretching and clearly delineates layers; layered sharpening, combined with the characteristics of the low-light environment, processes each layer specifically, ultimately outputting high-quality home textile printed images. The synergistic effect of each step effectively solves the problem of printing pattern processing in low-light environments and meets the printing needs of home textile fabrics.

[0065] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A control and hardware imaging optimization system for edge sharpening of printed patterns, characterized in that, The system includes: Vulnerable Feature Pre-analysis Module: Used to import the original design file of the printed pattern, identify fine lines with a line width of <0.5mm through pixel-level contour scanning and record the coordinates and direction, calculate the color difference value of adjacent color blocks through color value comparison analysis, define the area with a color difference value >50 as the boundary area of ​​high saturation color blocks, divide the feature priority according to two dimensions and count the area of ​​a single feature to generate a feature map; Ambient light acquisition module: This module integrates a multispectral sensor with an imaging device to acquire light intensity and color temperature parameters at a frequency of ≥30 frames / second. It uses image brightness analysis to locate the coordinates and area of ​​strong light reflection areas and organizes the parameters into JSON format for transmission to subsequent modules. Structured light compensation imaging module: used to receive feature map and ambient light parameters, determine the light environment adaptation coefficient and structured light projection angle, call the adaptive structured light compensation amount calculation model to generate structured light pattern, control the micro digital micromirror device or tunable diffractive optical element to project structured light, synchronously adjust the white balance and exposure compensation of the imaging sensor, capture and transmit the optimized original image. Texture layering module: It receives the optimized original image, uses Gaussian filtering to filter reflection noise in strong light scenes, adjusts the grayscale range of the image through grayscale stretching in low light scenes, and divides the image into edge layer, texture layer and background layer through edge detection, texture complexity analysis and color consistency analysis, adds markers to the layers and transmits the layering results; Layered sharpening module: It receives ambient light parameters and layering results, calculates the ambient light intensity ratio and color temperature deviation, calls the layered dynamic sharpening gain coefficient model to determine the sharpening gain of each layer, performs Sobel operator sharpening on the edge layer, performs bicubic B-spline interpolation sharpening on the texture layer, performs weakening sharpening on the background layer, integrates the layers through image alignment, and outputs the final image.

2. The control and hardware imaging optimization system for edge sharpening of printed patterns according to claim 1, characterized in that, In the fragile feature pre-analysis module, feature priorities are divided according to two dimensions: feature type weight dimension and feature self-parameter level dimension. The feature priority value is obtained by multiplying the parameters of the two dimensions, and the priority value ranges from 1.0 to 5.

0. The feature type weight dimension is set as follows: the weight corresponding to the thin line feature is 2.0, and the weight corresponding to the high saturation color block boundary feature is 1.

5. The feature's own parameter level dimension is divided into three levels, and the corresponding level coefficients for each level are: 2.5 for level 1, 1.6 for level 2, and 0.8 for level 3. The inherent parameter of fine line features is line width. Line width < 0.2mm is classified as Level 1, line width 0.2-0.3mm is classified as Level 2, and line width 0.3-0.5mm is classified as Level 3. The intrinsic parameters of the boundary features of highly saturated color blocks are the color difference value and the area of ​​the continuous region. When the color difference value is >80 and the area of ​​the continuous region is >500 square pixels, it is classified as Level 1; when the color difference value is 65-80 and the area of ​​the continuous region is 300-500 square pixels, it is classified as Level 2; when the color difference value is 50-65 and the area of ​​the continuous region is 10-300 square pixels, it is classified as Level 3. The feature priority is calculated as follows: Feature priority = Feature type weight × Feature's own parameter level coefficient; The statistical method for the area of ​​a single feature is as follows: the area of ​​a thin line feature is the product of the line width and the line length; the area of ​​a high-saturation color block boundary feature is the product of the length and width of the bounding rectangle of the boundary region. The above feature priority values ​​and individual feature area information are all written into the feature map.

3. The control and hardware imaging optimization system for edge sharpening of printed patterns according to claim 1, characterized in that, In the ambient light acquisition module, the multispectral sensor is integrated with the imaging device in the following manner: the sensor is coaxially mounted with the lens of the imaging device, and the data acquired by the sensor includes the following: Illumination intensity data: Includes illuminance in the visible light band and illuminance in the near-infrared light band. The detection wavelength range for the visible light band is 400nm-760nm, and the detection wavelength range for the near-infrared light band is 760nm-1000nm. Color temperature data: includes the dominant color temperature value of the current environment and the uniformity of color temperature distribution. The uniformity of color temperature distribution is obtained by calculating the variance of the color temperature values ​​in different areas of the captured image. Light intensity distribution data: contains the light intensity value of each pixel area in the captured image, and determines the area of ​​light intensity gradient change by the distribution of light intensity values; Strong light reflection area data: includes the pixel coordinate range of the strong light reflection area, the average light intensity value within the area, and the standard deviation of the light intensity within the area.

4. The control and hardware imaging optimization system for edge sharpening of printed patterns according to claim 1, characterized in that, In the ambient light acquisition module, when locating the coordinates and area of ​​the strong light reflection area through image brightness analysis, the acquired ambient image is first processed into grayscale, and a brightness threshold of 240 is set. Areas with grayscale values ​​greater than or equal to this threshold are marked as candidate reflection areas. Then, the candidate reflection areas are filtered to remove isolated areas with an area of ​​less than 50 square pixels. Finally, the coordinate range of the strong light reflection area is determined and its area is calculated.

5. The control and hardware imaging optimization system for edge sharpening of printed patterns according to claim 1, characterized in that, In the structured light compensation imaging module, the mathematical expression of the adaptive structured light compensation amount calculation model is: ;in, To compensate for the intensity of structured light, For light environment adaptation coefficient, Prioritize vulnerable features. Given the current ambient light intensity, The standard light intensity for printing imaging, For the area of ​​a single fragile feature, The total area of ​​the image. The angle between the direction of structured light projection and the direction of ambient light incidence.

6. The control and hardware imaging optimization system for edge sharpening of printed patterns according to claim 1, characterized in that, In the structured light compensation imaging module, when adjusting the white balance parameters of the imaging sensor, a color temperature shift compensation method is adopted. In strong light scenes, the color temperature shifts to cooler tones by 200-500K, and in weak light scenes, it shifts to warmer tones by 150-300K. When adjusting the exposure compensation, a local area independent adjustment strategy is adopted, and only the strong light reflection area is subjected to exposure compensation of -0.5 to -1.0EV, while the non-reflective area maintains the original exposure parameters. Before projecting structured light, the shadow coverage range of each vulnerable feature is predicted by matching the feature map with the ambient light parameters, and a structured light projection path planning map is generated to ensure that the angle between the structured light projection direction and the shadow direction is controlled between 30° and 60°, and the structured light coverage range is 5% larger than the vulnerable feature area.

7. The control and hardware imaging optimization system for edge sharpening of printed patterns according to claim 1, characterized in that, In the texture layering module, the specific implementation of Gaussian filtering to filter reflection noise in strong light scenes and grayscale stretching to adjust the image grayscale range in low light scenes is as follows: In the high-light scene, a 3×3 kernel Gaussian filter is used to filter reflection noise. The weight matrix of the Gaussian filter kernel is as follows: ; This filtering process can specifically filter isolated reflection noise points in an image with a brightness value greater than or equal to 240, while preserving the contour information of fragile features. In the low-light scenario, linear gray-scale stretching is used to adjust the gray-scale range of the image. First, the minimum and maximum gray-scale values ​​of the image are extracted, and then the original gray-scale range is mapped to the standard gray-scale range of 0 to 255 through linear mapping. During the mapping process, the linear relationship of gray-scale changes in each region of the image is maintained to ensure that the gray-scale difference of the image is increased to more than 80 in the low-light environment, thereby enhancing the recognizability of texture details.

8. The control and hardware imaging optimization system for edge sharpening of printed patterns according to claim 1, characterized in that, The texture layering module divides the edge layer, texture layer, and background layer through edge detection, texture complexity analysis, and color consistency analysis as follows: The edge detection is set with a fixed low threshold of 50 and a high threshold of 150 for gray-level gradient. Edge pixels with gray-level gradient values ​​exceeding the high threshold are extracted from the image. The region of the edge pixel and one pixel on each side is defined as the edge layer. The edge layer contains the boundary features of thin lines and highly saturated color blocks. The texture complexity analysis uses a 100×100 pixel sliding window to traverse the image and calculates the number of times the pixel grayscale value changes within each window. When the number of changes is greater than 5, the window area is defined as a texture layer, which contains gradient patterns and detail patterns. The color consistency analysis also uses a 100×100 pixel sliding window to calculate the standard deviation of the pixel grayscale value in each window. When the standard deviation is less than 10, the window area is defined as the background layer, which is the fabric substrate area.

9. The control and hardware imaging optimization system for edge sharpening of printed patterns according to claim 1, characterized in that, In the layered sharpening module, the mathematical expression for the layered dynamic sharpening gain coefficient model is: ;in, This is the layer sharpening gain factor. For the base coefficients of the layer, Given the current ambient light intensity, The standard light intensity for printing imaging, This represents the ambient color temperature deviation value. This is the layer detail density coefficient.

10. The control and hardware imaging optimization system for edge sharpening of printed patterns according to claim 1, characterized in that, The specific details of the layered sharpening module, which performs Sobel operator sharpening on the edge layer, bicubic B-spline interpolation sharpening on the texture layer, and weakening sharpening on the background layer, are as follows: When performing Sobel operator sharpening on the edge layer, the edge effect is enhanced by morphological dilation, with a dilation magnitude of 1 pixel to ensure that the edge contours are clearly distinguishable. When performing bicubic B-spline interpolation sharpening on the texture layer, if the ambient color temperature is below 4500K, it is defined as a warm light environment, and the interpolation step size accuracy is set to 0.1 pixels. Interpolation calculation is performed using the gray values ​​of 16 adjacent pixels. If the ambient color temperature is above 6500K, it is defined as a cool light environment, and gray-level gradient amplification processing is added to amplify the area with a gray-level gradient difference of less than 5 to a gray-level gradient difference of greater than 10. When performing weakening and sharpening on the background layer, the cutoff frequency of the Gaussian low-pass filter is adjusted. In strong light scenes, the cutoff frequency is reduced by 30% to 40%, and in weak light scenes, the cutoff frequency is reduced by 10% to 20%, preserving the original texture of the fabric substrate.

Images