Method for aligning electronic paper substrate with color filter inkjet layer and related assembly

Abstract

The application discloses a method for aligning an electronic paper substrate and a color filter inkjet layer and related components, and the method comprises the following steps: setting a detection window on the color filter inkjet layer, wherein the orthographic projection of the detection window on the substrate covers a Mark mark position of the substrate; the detection window comprises a plurality of complete irregular shape patterns; the center point coordinates of each complete irregular shape pattern are calculated, and the area weighted center coordinates of the detection window are calculated according to the center point coordinates of the plurality of irregular shape patterns; wherein at least three detection windows are arranged, and the Mark mark positions corresponding to the three detection windows do not lie on the same straight line. The application takes the actual form of the color filter inkjet layer itself as a benchmark, calculates the area weighted center coordinates according to the center point coordinates of the plurality of irregular shape patterns, and takes the area weighted center as an alignment benchmark, thereby fundamentally avoiding the recognition failure caused by the inkjet layer shielding the Mark point. The display color deviation and pixel misalignment problem are effectively solved.

Description

Technical Field

[0001] This invention relates to the field of electronic paper display technology, and in particular to a method and related components for aligning an electronic paper substrate with a color filter inkjet layer. Background Technology

[0002] In the manufacturing of color electronic paper display devices, after the color filter material is coated onto the surface of the ink screen by inkjet printing, it needs to be precisely aligned and bonded with the TFT substrate. This process directly determines the display accuracy and image consistency.

[0003] Currently, the industry generally adopts an alignment method based on preset optical mark positions on the TFT substrate, achieving position matching by identifying fixed mark points on the substrate. However, this method has significant drawbacks in actual industrial production: the inherent positional offset and uneven ink diffusion in the inkjet process often result in irregular edge contours of the inkjet layer, which can easily cover or obscure the substrate mark points, leading to recognition failure; at the same time, the TFT substrate is prone to slight deformation during processing, transportation, or temperature and humidity changes, causing the actual position of the preset mark points to shift, and the fixed mark points cannot adapt to this deformation, thus causing bonding deviations.

[0004] More importantly, the traditional method relies solely on a single fixed mark on the substrate side, completely disregarding the actual distribution of the inkjet layer itself. Even if the mark point is accurately identified, it is difficult to achieve precise matching between the inkjet layer and the TFT pixel, ultimately resulting in display defects such as color shift and pixel misalignment. Summary of the Invention

[0005] The purpose of this invention is to provide a method and related components for aligning an electronic paper substrate with a color filter inkjet layer, aiming to solve the problems of poor alignment accuracy between the inkjet layer and the TFT substrate in the prior art.

[0006] In a first aspect, embodiments of the present invention provide a method for aligning an electronic paper substrate with a color filter inkjet layer, comprising: A detection window is provided on the color filter inkjet layer, and the orthographic projection of the detection window on the substrate covers the Mark mark on the substrate; the detection window includes multiple complete irregular shape patterns; Calculate the center point coordinates of each complete irregular shape pattern, and calculate the area-weighted center coordinates of the detection window based on the center point coordinates of multiple irregular shape patterns; The detection window is provided with at least three, and the Mark flags covered by the three detection windows are not located on the same straight line.

[0007] Secondly, embodiments of the present invention provide an apparatus for aligning an electronic paper substrate with a color filter inkjet layer, comprising: The setting module is used to set a detection window on the color filter inkjet layer, wherein the orthographic projection of the detection window on the substrate covers the Mark mark position of the substrate; the detection window includes multiple complete irregular shape patterns. The calculation module is used to calculate the center point coordinates of each complete irregular shape pattern, and to calculate the area-weighted center coordinates of the detection window based on the center point coordinates of multiple irregular shape patterns. The detection window is provided with at least three, and the Mark flags covered by the three detection windows are not located on the same straight line.

[0008] Thirdly, embodiments of the present invention provide a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method for aligning an electronic paper substrate with a color filter inkjet layer as described in the first aspect.

[0009] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for aligning an electronic paper substrate with a color filter inkjet layer as described in the first aspect.

[0010] Fifthly, embodiments of the present invention also provide an electronic paper module, the electronic paper module comprising an electronic paper substrate and a color filter inkjet layer, wherein the electronic paper module and the color filter inkjet layer are bonded together using the method for aligning the electronic paper substrate and the color filter inkjet layer described in the first aspect above.

[0011] This invention discloses a method and related components for aligning an electronic paper substrate with a color filter inkjet layer. The method includes: setting a detection window on the color filter inkjet layer, the orthographic projection of the detection window onto the substrate covering a Mark marker on the substrate; the detection window comprising multiple complete irregular shape patterns; calculating the center point coordinates of each complete irregular shape pattern; and calculating the area-weighted center coordinates of the detection window based on the center point coordinates of the multiple irregular shape patterns; wherein at least three detection windows are provided, and the Mark markers covered by the three detection windows are not located on the same straight line. This invention abandons the traditional alignment method of fixing optical Mark points on TFT substrates, using the actual shape of the color filter inkjet layer as a reference. By calculating the area-weighted center coordinates based on the center point coordinates of multiple irregular shape patterns and using the area-weighted center as the alignment reference, it fundamentally avoids recognition failure caused by the inkjet layer obscuring the Mark points. This effectively solves the problems of color shift and pixel misalignment, significantly improving alignment accuracy and production yield, and can adapt to the manufacturing needs of color electronic paper with different resolutions. The present invention also provides an apparatus for aligning an electronic paper substrate with a color filter inkjet layer, a computer-readable storage medium, a computer device, and an electronic paper module, which have the above-mentioned beneficial effects, and will not be repeated here. Attached Figure Description

[0012] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a schematic diagram of a method for aligning an electronic paper substrate with a color filter inkjet layer. Figure 2 This is a schematic diagram of the TFT substrate structure; Figure 3 A schematic diagram of the structure of a color filter inkjet layer; Figure 4 This is a schematic diagram showing the alignment and bonding of the color filter inkjet layer with the TFT substrate. Figure 5 This is a schematic block diagram of a device for aligning an electronic paper substrate with a color filter inkjet layer. Detailed Implementation

[0014] 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, not all, of the embodiments of the present invention. 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.

[0015] It should be understood that, when used in this specification and the appended claims, the terms “comprising” and “including” indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more of its features, integrals, steps, operations, elements, components and / or collections thereof.

[0016] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0017] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the relevant listed items and all possible combinations, and includes such combinations.

[0018] Please see Figures 1-4 This embodiment provides a method for aligning an electronic paper substrate with a color filter inkjet layer, including: S101: A detection window is provided on the color filter inkjet layer, and the orthographic projection of the detection window on the substrate covers the Mark mark position of the substrate; the detection window includes multiple complete irregular shape patterns; at least three detection windows are provided, and the Mark mark positions covered by the three detection windows are not located on the same straight line.

[0019] Specifically, when setting detection windows on the color filter inkjet layer, the corresponding positions are first determined based on the original Mark markers at the four corners of the substrate, so that each detection window and a Mark marker are spatially associated, and the orthographic projection of the detection window onto the substrate covers the Mark marker. This correspondence does not mean that the window directly covers the Mark marker, but rather that the window is placed in the area on the color filter inkjet layer corresponding to the coordinates of the Mark marker, ensuring that the detection window can effectively capture the inkjet distribution pattern within that area, while avoiding recognition interference caused by the original Mark points on the substrate being blocked by the inkjet layer. Subsequently, at least three Mark markers are selected from the four corners or edge areas of the substrate. The positions of these Mark markers on the substrate plane must meet the condition that they are not on the same straight line, that is, any three Mark markers form a triangle rather than being collinear, to ensure that subsequent multi-point correction can simultaneously and effectively resolve deviations such as translation, rotation, and local deformation. The size of each detection window is set according to the electronic paper resolution and inkjet process characteristics. For example, for a high-resolution product with 300 PPI, the window can be set to a square area of ​​3 mm by 3 mm. However, the inside of the window is not uniformly covered by inkjet. Instead, it contains multiple inkjet patterns of different shapes formed by inkjet position offset, uneven ink diffusion, and the influence of the micro-morphology of the substrate surface. The edges of these patterns are irregular, the sizes are different, and there may be gaps or partial overlaps between them. These are the multiple irregular shaped patterns contained in the detection window.

[0020] During setup, a high-precision positioning platform is needed to precisely align the area-weighted center coordinates of the detection window with the coordinates of the substrate Mark marker. Simultaneously, a CCD camera pre-scan is used to confirm that the detection window position avoids functional areas such as substrate wiring and electrodes. If it does not avoid these areas, the color filter inkjet layer is moved to ensure that the detection window contains only effective distribution information of irregularly shaped patterns, avoiding interference from irrelevant structures in subsequent image analysis.

[0021] Due to the inherent randomness of inkjet printing, the number and specific shapes of complete irregular shapes within each detection window vary. These irregular shapes collectively constitute the feature information set of the detection window. In subsequent alignment operations, the system identifies each of these irregular shapes, extracts the outline of each irregular shape, and calculates the area-weighted center coordinates of all patterns within the entire window. These area-weighted center coordinates serve as the alignment reference point for the detection window. The reference point obtained through the detection window provides a reliable basis for subsequent deviation correction. Furthermore, the design of including multiple complete irregular shapes within the detection window, compared to a single regular pattern or fixed marker point, provides richer morphological statistical information, effectively reducing the impact of random noise and local defects on the calculation of the area-weighted center coordinates, and improving the stability and repeatability of the alignment reference point.

[0022] In some embodiments, setting a detection window on the color filter inkjet layer includes: Two detection windows are set symmetrically about a horizontal axis of symmetry on the color filter inkjet layer within at least three detection windows, and two detection windows are set symmetrically about a vertical axis of symmetry on the color filter inkjet layer within at least three detection windows.

[0023] Specifically, when setting detection windows on the color filter inkjet layer, it is necessary to determine the geometric center of the color filter inkjet layer itself and establish a planar coordinate system with this center as the origin, thereby clarifying two mutually perpendicular central axes of symmetry, namely the horizontal axis of symmetry and the vertical axis of symmetry. Then, at least three detection windows are selected in the edge region of the inkjet layer. The number of these detection windows is usually set according to the alignment accuracy requirements. For example, four windows can be selected and located in the four corner regions of the inkjet layer, but it is more flexible to set three windows in a triangular distribution. Regardless of the number, it is necessary to ensure that two adjacent detection windows are mirror-symmetric about one of the central axes of symmetry. For example, the upper left window and the upper right window are symmetric about the vertical axis of symmetry, the upper left window and the lower left window are symmetric about the horizontal axis of symmetry, and the upper right window and the lower right window also satisfy these two sets of symmetry relationships, and so on, so that all detection windows have strict symmetry in spatial arrangement.

[0024] To achieve this, before setting the detection windows, a high-resolution CCD camera is used to image the entire color filter inkjet layer. Image processing algorithms are then used to extract the overall outer contour of the color filter inkjet layer and calculate its geometric center. This center is the center of symmetry of the color filter inkjet layer, thus determining the horizontal and vertical axes of symmetry. Subsequently, using this center of symmetry as a reference point, candidate window positions are set at preset distances in each of the four quadrants of the color filter inkjet layer. These candidate positions must avoid existing wiring, electrodes, and functional areas on the substrate. After setting, a positioning platform and image feedback system are used to check the actual coordinates of each detection window against the axes of symmetry. The mirror image deviation of adjacent windows relative to their corresponding axes of symmetry is measured. If the deviation exceeds the allowable range, the window positions are fine-tuned until perfect symmetry is achieved. The window size can be uniformly set according to the electronic paper resolution and inkjet material characteristics. For example, a 3mm x 3mm window is used for 300PPI products, while a larger window can be used for lower resolution products. However, the size of all detection windows must remain consistent to ensure symmetry.

[0025] Once set up, each window contains multiple irregularly shaped patterns formed by inherent deviations in the inkjet process. Although these irregular patterns vary in shape, due to the symmetrical layout of the detection windows, the inkjet areas covered by each detection window have statistically similar characteristic distributions, which is beneficial to the stability of subsequent multi-point calibration. If only three detection windows are set, it is necessary to ensure that any two adjacent detection windows are mirror images about their corresponding axes of symmetry. For example, the three detection windows can be located at the upper left, lower left, and lower right of the color filter inkjet layer, respectively. The two lower windows are symmetrical about the vertical axis of symmetry, while the upper left and lower left windows are symmetrical about the horizontal axis of symmetry.

[0026] Whether using three or four detection windows, the symmetrical layout effectively offsets nonlinear errors caused by overall substrate tilt or global offset of the inkjet layer, making the alignment reference point more robust. The entire setup process is completed under automated control. The system automatically records the final coordinates of each detection window and repeatedly verifies the symmetry during each alignment, ensuring that the detection windows always meet the condition of symmetry about the central axis of symmetry. This lays the foundation for the precise alignment of the subsequent color filter inkjet layer with the TFT substrate pixels.

[0027] In some embodiments, the detection window can be square, circular, polygonal, or irregular in shape. Square windows are the preferred choice under normal circumstances because their regular boundaries make coordinate positioning and size adjustment easy. The shape of the detection window only needs to be set to ensure that it corresponds one-to-one with the Mark bits on the substrate.

[0028] S102: Calculate the center point coordinates of each complete irregular shape pattern, and calculate the area-weighted center coordinates of the detection window based on the center point coordinates of multiple irregular shape patterns; In this embodiment, the center point coordinates of each irregular shape pattern are calculated, and the area-weighted center coordinates of the detection window are calculated based on the center point coordinates of multiple irregular shapes patterns, including: Acquire the actual distribution image of all complete irregular shape patterns within the detection window; For each actual distribution image, irregular shape pattern contours are extracted to obtain the effective contour information of each detection window; Based on the effective contour information, the center point coordinates of each complete irregular shape pattern are calculated, and the area-weighted center coordinates of each detection window are calculated based on the center point coordinates of all complete irregular shape patterns.

[0029] Specifically, a color filter inkjet layer has been sprayed onto the surface of the ink screen, which is covered by a thin-film transistor substrate. A charge-coupled device (CCD) camera directly captures images of the surface carrying the color filter inkjet layer to obtain the actual distribution image of the irregular shape pattern within the detection window.

[0030] By fixing the thin-film transistor substrate to the support platform of the alignment platform, the spatial orientation of the platform is adjusted so that the 3 mm by 3 mm square detection windows are aligned with the imaging field of view of the charge-coupled device (CCD) camera. Each camera is equipped with a telecentric lens, with the lens optical axis strictly perpendicular to the substrate surface. The vertical working distance from the camera imaging surface to the substrate surface is set to a predetermined distance to eliminate image distortion caused by tilted shooting angles.

[0031] Turn on the coaxial reflective light source so that the incident light shines perpendicularly into the detection window area. The areas covered by irregularly shaped patterns have lower reflected light intensity due to the absorption and scattering of light by the ink, while the areas with exposed substrate have higher reflected light intensity, forming a clear optical contrast.

[0032] Under the control of a synchronous trigger signal, the camera simultaneously captures images of four detection windows, with the exposure time uniformly set to 5 milliseconds, obtaining four images showing the actual distribution of irregularly shaped patterns within each detection window. Each actual distribution image completely covers a 3mm x 3mm area of ​​its corresponding detection window. The grayscale value of each pixel in the image reflects whether there is an irregularly shaped pattern covering the corresponding location on the substrate surface and the difference in coverage thickness. Low grayscale areas correspond to locations covered by irregularly shaped patterns, high grayscale areas correspond to locations with exposed substrate, and grayscale gradient areas correspond to the edge diffusion transition zone of the irregularly shaped pattern. Each of the four actual distribution images is accompanied by a detection window location number and acquisition time sequence mark, and is transmitted in parallel to the industrial image processing system through the data acquisition channel as input data for subsequent inkjet contour extraction and area weighted center calculation.

[0033] Specifically, for 200PPI medium resolution electronic paper, the detection window is adjusted to 5mm×5mm, while other parameters remain unchanged, and the alignment deviation can be controlled within ±3μm; for 400PPI ultra-high resolution electronic paper, the detection window is adjusted to 2mm×2mm, and the CCD camera resolution is increased to 4096×4096, and the alignment deviation can be controlled within ±1μm, both of which can meet the manufacturing alignment requirements of color electronic paper with different resolutions.

[0034] In this embodiment, irregular shape pattern contours are extracted from the actual distribution image to obtain the effective contour information of each detection window, including: The actual distribution image is processed by grayscale and binarization to obtain each complete irregular shape pattern; Edge detection is performed on each complete irregular shape pattern to obtain the preliminary outline of each complete irregular shape pattern; The initial outlines of all complete irregular shapes are denoised to obtain effective outline information.

[0035] This embodiment transforms the optical contrast between irregularly shaped patterns and the exposed substrate background in the actual distribution image into a clear binary classification result through grayscale and binarization processing. The segmentation threshold is automatically determined by the valley of the grayscale histogram, avoiding the insufficient adaptability of fixed thresholds under different lighting conditions and accurately separating irregularly shaped patterns. When performing edge detection on irregularly shaped patterns, Gaussian filtering smoothing and gradient magnitude direction calculation are used, and the closed preliminary contour of the irregularly shaped patterns is extracted through double threshold division and edge connection operations, which completely preserves the detailed morphological features of the edges of the irregularly shaped patterns. Median filtering noise reduction processing is performed on the preliminary contour, replacing the original value with the median of the gradient magnitude in the neighborhood of the edge point, effectively filtering out isolated edge noise caused by ink splatter points and substrate surface impurities, while not damaging the continuity of the irregularly shaped pattern contour and the edge positioning accuracy. Finally, effective contour information reflecting the actual distribution morphology of the irregularly shaped patterns is obtained, providing high-quality input for the accurate calculation of the area weighting center.

[0036] In this embodiment, the actual distribution image is processed by grayscale and binarization to obtain various complete irregular shape patterns, including: A grayscale image is obtained by performing a weighted average of the three-channel values ​​of each pixel in the actual distribution image. Generate a grayscale histogram based on the grayscale value distribution of each pixel in the grayscale image; Extract the gray values ​​corresponding to the valley positions from the gray-level histogram as the segmentation threshold; The grayscale value of each pixel in the grayscale image is compared with the segmentation threshold. Pixels with grayscale values ​​greater than or equal to the segmentation threshold are assigned the first value, and pixels with grayscale values ​​less than the segmentation threshold are assigned the second value, thus obtaining a binary image. Identify connected regions in a binary image that are composed of the first numerical value, and treat the connected regions as complete irregular shape patterns.

[0037] This embodiment employs a weighted average method to convert the red, green, and blue channel values ​​to grayscale. The weighting coefficients for each channel are matched to the human eye's sensitivity to different wavelengths, resulting in a grayscale image that more realistically reflects the visual contrast difference between the irregularly shaped pattern and the substrate background. A grayscale histogram is generated based on the grayscale value distribution, which intuitively presents the bimodal distribution characteristics of the irregularly shaped pattern and the exposed substrate area within the detection window, providing a statistical basis for threshold selection. The segmentation threshold is extracted from the valley positions of the grayscale histogram, allowing the threshold to be dynamically determined based on the actual lighting conditions and contrast of each image. This avoids segmentation failure due to batch-to-batch process fluctuations caused by a fixed threshold, exhibiting strong adaptability. After comparing the pixel grayscale value with the segmentation threshold, a first or second value is assigned, converting the grayscale image into a binary image containing only two types of pixels. This completely eliminates the grayscale blurring caused by the inkjet edge diffusion transition zone, achieving rigid separation between the irregularly shaped pattern and the substrate background. In a binary image, the connected regions formed by the first value are identified as complete irregular shape patterns. This can automatically filter out isolated noise pixels caused by minor defects on the substrate surface, retaining only the coverage area of ​​complete irregular shape patterns with continuous spatial distribution, thus providing a complete and accurate foreground target area for subsequent edge detection.

[0038] Specifically, for each pixel in the actual distribution image, its brightness value is read in the red, green, and blue channels, with each channel's value ranging from 0 to 255. Based on the human eye's visual characteristics—highest sensitivity to green light, followed by red, and lowest sensitivity to blue—weighting coefficients are set at 0.299 for the red channel, 0.587 for the green channel, and 0.114 for the blue channel. Each pixel's red channel value is multiplied by 0.299, green channel value by 0.587, and blue channel value by 0.114, and these products are summed to obtain a weighted average, which is the grayscale value corresponding to that pixel. This weighted average calculation is performed on all pixels in the actual distribution image. The grayscale values ​​of all pixels together constitute a complete grayscale image, where the grayscale value of each pixel is between 0 and 255.

[0039] After generating the grayscale image, a grayscale histogram is generated based on the grayscale value distribution of each pixel in the image. The image processing system traverses all pixels in the grayscale image, counting the number of pixels appearing at each grayscale level from 0 to 255. A statistical graph of the grayscale value distribution is plotted with the grayscale level as the horizontal axis and the pixel frequency corresponding to each grayscale level as the vertical axis. Since the detection window contains only two main optical features—irregularly shaped patterns and background areas—the grayscale values ​​of the irregularly shaped patterns are concentrated in the lower grayscale range due to light absorption, while the grayscale values ​​of the background areas are concentrated in the higher grayscale range due to light reflection. These two types of areas each form a distribution peak on the grayscale histogram, resulting in a distinct bimodal shape. The transition area between the two peaks corresponds to the grayscale gradient pixels at the edges of the irregularly shaped patterns. This grayscale histogram visually displays the overall grayscale distribution characteristics of the image within the detection window, providing a statistical basis for the subsequent automatic extraction of the segmentation threshold.

[0040] Because the grayscale histogram exhibits a typical bimodal distribution, one peak corresponds to the clustered distribution of irregularly shaped pattern pixels in the low grayscale range, while the other peak corresponds to the clustered distribution of background pixels in the high grayscale range. The transition section between the two peaks corresponds to the grayscale gradient pixels in the edge diffusion zone of the irregularly shaped pattern. The image processing system first performs smoothing preprocessing on the original grayscale histogram, using a mean smoothing algorithm to perform a point-by-point moving average on the histogram curve with a window width of 5 grayscale levels. This eliminates local spikes caused by statistical fluctuations in pixels, making the valley shape between the two peaks more clearly discernible.

[0041] After smoothing, the image processing system scans the grayscale histogram from left to right, recording the smoothed pixel frequency value for each grayscale level. It then identifies the locations of the two maximum points on the curve: the left maximum point corresponds to the distribution peak of the irregular pattern, and the right maximum point corresponds to the distribution peak of the exposed substrate area. Within the grayscale interval between the two peaks, the image processing system performs a minimum value search on the smoothed curve, comparing the frequency values ​​of each grayscale level within this interval, and determining the grayscale level with the lowest frequency value as the trough location.

[0042] The grayscale value corresponding to this trough location is the segmentation threshold. It lies at the boundary between the grayscale distribution of the irregularly shaped pattern and the grayscale distribution of the background region, representing the grayscale boundary point that minimizes the probability of missegmentation of pixels in the two types of regions. The image processing system extracts this grayscale value and stores it as a dedicated segmentation threshold parameter for the current detection window image, which is used for pixel classification determination in subsequent binarization processing.

[0043] Next, each pixel in the grayscale image is traversed, and its grayscale value is read. This grayscale value is then compared with the segmentation threshold previously extracted from the valley positions of the grayscale histogram. Because irregular shapes absorb light, they exhibit lower grayscale values ​​in the grayscale image, while background areas reflect light and exhibit higher grayscale values. Therefore, using the segmentation threshold as a boundary, pixels with grayscale values ​​less than the threshold are identified as irregular shapes and assigned a first value of 0; pixels with grayscale values ​​greater than or equal to the threshold are identified as background areas and assigned a second value of 255. After all pixels in the full grayscale image have undergone the above comparison and assignment, a binary image is generated consisting only of the first value 0 and the second value 255. In this binary image, black pixels corresponding to the first value 0 represent irregular shapes, and white pixels corresponding to the second value 255 represent background areas. The boundaries between the two types of areas are distinct, with no grayscale transition.

[0044] After the binary image is generated, the system identifies connected regions formed by the first value in the binary image to obtain irregular shape patterns. The system uses an 8-connected component labeling algorithm to perform connected component analysis on the binary image. It scans each pixel in the binary image, and when a pixel with a value of 0 is detected, it checks the pixel values ​​of its eight neighboring pixels. Pixels with the same first value of 0 and spatially adjacent are grouped into the same connected set. Through row-by-row and pixel-by-pixel scanning and label merging operations, all connected regions formed by the first value of 0 are identified in the binary image. Since the coverage area of ​​the irregular shape pattern within the detection window is usually a continuous sheet distribution, the connected region with the largest area formed by the first value of 0 is the valid irregular shape pattern. The remaining fragmented areas with too small an area are sets of interfering pixels caused by substrate surface impurities or image noise. The image processing system retains the largest connected region as the irregular shape pattern of the detection window, and the boundary of this irregular shape pattern is the initial contour target to be extracted in subsequent edge detection operations.

[0045] In this embodiment, after acquiring the image of each detection window, the image often contains various inkjet-formed patterns. Due to the different random landing points and diffusion levels of the inkjet droplets, the outlines of these patterns may be partially within the window boundary or completely inside the window. Patterns located on the window boundary have their outlines truncated by the window edge, making it impossible to obtain a complete closed boundary. These patterns are incomplete irregular shapes. If they are included in the area and center calculations, the weighted center coordinates will be systematically biased due to the missing area. Therefore, it is necessary to distinguish between complete irregular shapes and incomplete patterns, retaining only complete patterns for subsequent calculations. In practice, the image processing system first performs connected component labeling on the binarized image of each window, identifying all independent inkjet regions, each region corresponding to an irregular shape pattern. The system then checks the pixel coordinate range of each connected component. If all pixels in the connected component are located inside the window boundary (i.e., none of the four sides of its smallest bounding rectangle touch the window boundary), it is considered a complete pattern. Conversely, if the coordinates of any pixel in the connected component coincide with or exceed the window boundary, it is considered an incomplete pattern and is removed from the valid dataset. This removal is not a physical deletion of image pixels, but rather the addition of invalid markers to the connected component in the image processing software, preventing it from participating in subsequent area and center coordinate calculations. To ensure accuracy, the window boundary must be set to a strict rectangle or a mathematical boundary of the corresponding shape, and manually marked lines on the window boundary are retained during edge detection to allow the system to accurately determine the relative position of the pattern to the boundary.

[0046] After screening, each detection window typically retains at least three complete irregularly shaped patterns. This is the minimum requirement to ensure that the area-weighted center coordinates are statistically representative. The more patterns there are, the more accurately the calculated weighted center reflects the actual distribution center of the inkjet layer within the window. This is because the area weights of multiple patterns are superimposed, effectively smoothing out the positioning noise caused by the randomness of inkjet printing in a single pattern.

[0047] In this embodiment, edge detection is performed on each complete irregular shape pattern to obtain the preliminary outline of each complete irregular shape pattern, including: Gradient calculation is performed on the complete irregular shape pattern to obtain the gradient magnitude image; Based on the gradient direction of each pixel in the gradient magnitude image, non-maximum suppression processing is performed to obtain a thinned edge image. For the refined edge image, a preset high threshold and a low threshold are used for dual threshold detection. Pixels with gradient magnitude greater than or equal to the high threshold are marked as strong edge points, pixels with gradient magnitude greater than the low threshold and less than the high threshold are marked as weak edge points, and pixels with gradient magnitude less than or equal to the low threshold are marked as non-edge points. The pixels adjacent to any strong edge point within a weak edge point are marked as strong edge points to obtain a preliminary outline.

[0048] This embodiment obtains a gradient amplitude image through gradient calculation, which can quantify the distribution of the degree of grayscale change at the boundary of a complete irregular shape pattern, providing an objective numerical basis for edge location determination. Non-maximum suppression processing is performed based on the gradient direction of each pixel, retaining the pixel with the largest amplitude in the gradient direction and suppressing its adjacent non-maximum amplitude pixels, refining the blurred gradient band region into candidate edges of single pixel width, significantly improving the spatial accuracy of edge localization. A dual-threshold detection is performed on the refined edge image using preset high and low thresholds. The high threshold filters out strong edge points with high confidence, while the low threshold retains weak edge points necessary for edge continuity, balancing the complete capture of the irregular shape pattern outline with the suppression of interference from fine textures on the substrate surface. Pixels spatially adjacent to any strong edge point among the weak edge points are marked as strong edge points, allowing broken strong edge segments to connect and close through weak edge paths, ultimately obtaining a preliminary outline of the irregular shape pattern without breaks or redundant branches, fully reflecting the actual distribution boundary morphology of the inkjet material within the detection window.

[0049] Specifically, after obtaining the complete irregular shape pattern, gradient calculation is performed on the complete irregular shape pattern to obtain a gradient magnitude image, which provides pixel-level grayscale change intensity information for edge detection.

[0050] Gradient calculation employs the Sobel operator to perform spatial convolution on the binary image containing the complete irregular shape pattern. Both horizontal and vertical Sobel operators are used to perform neighborhood weighted summation on each pixel in the binary image. The horizontal Sobel operator is a 3x3 matrix: the first row contains -1, 0, and +1; the second row contains -2, 0, and +2; and the third row contains -1, 0, and +1. This operator is convolved with the 3x3 neighborhood grayscale values ​​of each pixel to obtain the horizontal gradient component of that pixel. The vertical Sobel operator has the same structure: the first row contains -1, -2, and -1; the second row contains 0, 0, and +0; and the third row contains +1, +2, and +1. This operator is also convolved with the 3x3 neighborhood grayscale values ​​of each pixel to obtain the vertical gradient component of that pixel. After performing convolution operations in both directions on each pixel of the binary image, the square root of the sum of the squares of the horizontal and vertical gradient components is calculated for each pixel. This square root value is the gradient magnitude of that pixel. The gradient magnitudes of all pixels together constitute a gradient magnitude image. In the gradient magnitude image, the boundaries of complete irregular shapes and patterns have higher gradient magnitudes because the grayscale value changes abruptly from the second value 0 to the first value 255. The interiors of complete irregular shapes and patterns and the interiors of the substrate background have gradient magnitudes close to zero because the grayscale values ​​are consistent.

[0051] After obtaining the gradient magnitude image, non-maximum suppression is performed based on the gradient direction of each pixel in the gradient magnitude image to refine the blurred band-like region near the edge in the gradient magnitude image into candidate edges with a single pixel width.

[0052] Next, the gradient direction angle of each pixel in the gradient magnitude image is calculated. The gradient direction angle is determined by the arctangent of the horizontal and vertical gradient components, ranging from -90 degrees to +90 degrees. The gradient direction angle is discretized into four main direction categories: 0 degrees (horizontal), 45 degrees (diagonal), 90 degrees (vertical), and 135 degrees (anti-diagonal). The actual gradient direction angle of each pixel is assigned to the closest main direction category.

[0053] During non-maximum suppression, each pixel in the gradient magnitude image is traversed. In both the positive and negative directions of the gradient direction of that pixel, the gradient magnitudes of the two adjacent pixels are read. If the gradient magnitude of the current pixel is greater than or equal to the gradient magnitude of its adjacent pixel in the positive direction and simultaneously greater than or equal to the gradient magnitude of its adjacent pixel in the negative direction, the gradient magnitude of the current pixel is retained; otherwise, the gradient magnitude of the current pixel is set to zero. After traversing the entire image, only pixels at local maxima along the gradient direction of irregular shape patterns retain their gradient magnitudes. The gradient magnitudes of all other non-maximum pixels are cleared to zero, generating a thinned edge image. The candidate edge pixels retained in the thinned edge image form continuous or near-continuous boundary lines with a single pixel width.

[0054] After obtaining the thinned edge image, a dual threshold detection is performed on the thinned edge image using a preset high threshold and a low threshold, and candidate edge pixels are classified into two categories: strong edge points and weak edge points.

[0055] The preset values ​​for the high and low thresholds are determined in advance based on the overall statistical characteristics of the gradient magnitude image. In this embodiment, the image processing system calculates the cumulative distribution of gradient magnitudes for all non-zero pixels in the gradient magnitude image, takes the gradient magnitude corresponding to the position where the cumulative distribution reaches 70% as the high threshold, and takes the gradient magnitude corresponding to the position where the cumulative distribution reaches 30% as the low threshold. The high threshold is used to filter edge pixels with the strongest gradient response, and the low threshold is used to retain pixels with weaker gradient responses but which may form continuous edge paths.

[0056] During dual-threshold detection, each pixel in the refined edge image is traversed, and the gradient magnitude of that pixel is read. The gradient magnitude is compared to a high threshold. If the gradient magnitude is greater than or equal to the high threshold, the pixel is marked as a strong edge and assigned a strong edge identifier value in the edge labeling map. If the gradient magnitude is less than the high threshold but greater than the low threshold, the pixel is marked as a weak edge and assigned a weak edge identifier value in the edge labeling map. If the gradient magnitude is less than or equal to the low threshold, the pixel is considered a non-edge point and assigned a zero value in the edge labeling map, then discarded. After traversing the entire refined edge image, an edge labeling map containing both strong and weak edge point sets is generated.

[0057] After completing the dual-threshold classification and labeling, pixels adjacent to any strong edge point within a weak edge point are appended as strong edge points. An edge-connection operation is then used to obtain the preliminary outline of the complete irregular shape pattern. The image processing system traverses each weak edge point in the edge-labeled map, checking all neighboring pixels within its neighborhood. The neighborhood includes the four directly adjacent pixels above, below, left, and right, and the four diagonally adjacent pixels (top left, top right, bottom left, and bottom right). If at least one pixel already labeled as a strong edge point exists within the neighborhood, the weak edge point is determined to be located on the edge path of the complete irregular shape pattern outline, and is appended as a strong edge point, assigned a strong edge identifier value. If no strong edge points exist within the neighborhood, the weak edge point is determined to be a false edge response caused by fine surface texture or residual noise of the substrate, and is discarded with a value of zero. After a single traversal, some pixels that were originally weak edges are re-marked as strong edges. These newly enhanced edge points may then become adjacent strong edge points to other weak edge points. Therefore, the image processing system repeats the above traversal and appending marking operations on the edge marker map until no new weak edge points are appended in a certain traversal. Finally, all pixels marked as strong edge points are connected to each other, forming one or more closed continuous boundaries. These continuous boundaries constitute the preliminary outline of the complete irregular shape pattern within the detection window.

[0058] In this embodiment, the initial contour is denoised to obtain effective contour information, including: The median filtering algorithm is used to denoise the preliminary contour to obtain effective contour information.

[0059] Specifically, after obtaining the preliminary outline of the complete irregular shape pattern, a median filtering algorithm is used to denoise the preliminary outline to remove isolated edge noise caused by ink splatter points and impurities on the substrate surface, thereby obtaining effective outline information that reflects the true distribution of the irregular shape pattern.

[0060] In addition to the actual edge points of the irregularly shaped pattern boundary, the preliminary contour may also contain isolated small spot edges formed by trace amounts of ink splashing within the detection window during inkjet printing, as well as false edge fragments caused by residual dust particles on the substrate surface being misjudged as foreground after binarization. The gradient amplitude of these noise points is often close to that of the edge points of the irregularly shaped pattern contour, but their spatial distribution is discrete and isolated, and they have no spatial connection with the main continuous boundary of the irregularly shaped pattern contour.

[0061] During median filtering noise reduction, the image processing system traverses each edge point in the initial contour and obtains the coordinate position of that edge point in the image coordinate system. A square filtering window of size 5 pixels by 5 pixels is defined centered on that edge point. The gradient magnitudes corresponding to all edge points belonging to the initial contour within this filtering window are extracted and sorted in ascending order of value. The gradient magnitude located in the middle position after sorting, i.e., the 13th gradient magnitude in the sorted sequence, is selected as the replacement gradient magnitude for the center pixel of the filtering window.

[0062] After replacing the original gradient magnitude of the current edge point with the median replacement value, the image processing system compares the original gradient magnitude of the center pixel within the filtering window with the median replacement value. If the deviation exceeds a preset deviation tolerance, it indicates that the gradient magnitude of the edge point differs significantly from the gradient magnitudes of most edge points in the neighborhood, and it is highly likely to be an isolated noise point; therefore, the edge point is removed from the preliminary contour. If the deviation is within the preset tolerance range, it indicates that the gradient characteristics of the edge point are consistent with those of the neighboring edge points, and it is part of the boundary of an irregular shape pattern; therefore, it is retained.

[0063] The median filtering and deviation determination operations described above are performed on each edge point in the initial contour. The remaining edge points after all processing constitute the filtered edge point set. Based on the coordinate positions and connectivity of each edge point in the filtered edge point set, the image processing system redraws the boundary lines of the irregular shape pattern. The reconstructed boundary lines are the effective contour information of the detection window. This effective contour information eliminates isolated noise interference and completely preserves the true distribution boundary morphology of the irregular shape pattern within the detection window, providing accurate contour input data for subsequent area weighted center calculation.

[0064] In some embodiments, a contour image after median filtering and noise reduction is obtained. This image has removed most of the dot noise, but it still has defects such as minor edge breaks at the inkjet edges, tiny voids inside the contour, and fine burrs at the edges. Single filtering can easily weaken some of the fine edge structures and cannot completely guarantee the continuity of the contour. Small structural elements of suitable size are selected as morphological operation carriers to match the fine scale of irregular shape and pattern contours, avoiding the loss of detail caused by excessively large operation range.

[0065] Specifically, the morphological closing operation connects and fills tiny gaps in the contour that may have been created during noise reduction. It consists of a dilation operation and an erosion operation performed in series. During the dilation operation, a circular structuring element with a radius of 3 pixels is used. The center of this structuring element is aligned with each pixel in the median-filtered contour image. If at least one edge point belonging to the median-filtered contour exists within the structuring element's coverage area, that pixel is designated as an edge point. The dilation operation expands the contour lines outward and connects adjacent but separate contour segments. During the erosion operation, a circular structuring element of the same radius is used. The center of this structuring element is aligned with each pixel in the dilated contour image. If all pixels within the structuring element's coverage area are edge points, that pixel is retained as an edge point; otherwise, it is designated as a non-edge point. The erosion operation shrinks the contour lines inward, restoring them to near their original width, while preventing the gaps created during the dilation stage from breaking. After the closing operation, small gaps in the contour are bridged and closed, and tiny holes created by noise removal within the contour are filled.

[0066] Subsequently, the image processing system performs a morphological opening operation on the contour after the closing operation to remove scattered noise attached to the contour edges without breaking the contour backbone. The opening operation consists of a cascaded erosion operation and a dilation operation. The erosion operation uses a circular structuring element with a radius of 2 pixels, aligning the center of the structuring element with each pixel in the contour image. If all pixels within the structuring element's coverage area are edge points, they are retained; otherwise, they are treated as non-edge points. This step eliminates isolated scattered noise because it cannot completely cover the structuring element, while the contour backbone, due to its large continuous range, is only slightly thinned. The dilation operation then uses a circular structuring element of the same radius to expand and restore the eroded contour, restoring the contour backbone to a linewidth similar to that before the opening operation. After the opening operation, scattered isolated noise attached to the edges of the irregular shape pattern is effectively removed, while the continuity of the contour backbone and subtle edge features of the irregular shape pattern are fully preserved.

[0067] After a three-stage cascaded process involving median filtering, morphological closing, and morphological opening, the image processing system obtains optimized and denoised effective contour information. This effective contour information eliminates four common contour defects: isolated noise points, small broken edges, internal holes, and scattered edge noise. Simultaneously, it preserves to the maximum extent the subtle undulations caused by ink diffusion at the edges of irregularly shaped patterns. The contour integrity and boundary positioning accuracy are significantly improved compared to single median filtering, providing higher-quality contour input data for the accurate calculation of the subsequent area-weighted center.

[0068] Next, based on the effective contour information, the coordinates of the center point of each complete irregular shape pattern are calculated. Specifically, using Green's theorem (the paper "Fast and exact computation of Cartesian geometric moments using discrete Green's theorem" proposes a discrete Green's theorem for fast and accurate calculation of geometric moments in binary images. This method achieves the same accuracy as direct double summation while being faster), each closed contour is treated as a planar region of uniform density. The centroid (i.e., the physical center of gravity) is solved by measuring the overall distribution characteristics of its internal area, and this centroid is used as the center point of the irregular pattern, thus obtaining the center coordinates of the irregular shape pattern. When processing multiple targets, the edges of each connected region are independently topologically mapped and solved one by one without interference. This principle is universally applicable to complex shapes such as concave shapes. Center positioning focuses on the macroscopic spatial span enclosed by the contour, effectively weakening the interference caused by uneven local sampling density at the edges. Its effectiveness is predicated on the contour being strictly closed, without self-intersections, and without nested holes, to ensure the clarity of the region's topological relationships.

[0069] In this embodiment, based on the effective contour information, the center point coordinates of each complete irregular shape pattern are calculated, and the area-weighted center coordinates of each detection window are calculated based on the center point coordinates of all complete irregular shape patterns, including: The area-weighted center coordinates of each detection window are calculated using the following formula: Where X and Y represent the area-weighted center coordinates of each detection window; x i y i S represents the center coordinates of the i-th irregular shape pattern; i Let be the area of ​​the i-th irregular shape pattern; n is the total number of irregular shape patterns.

[0070] This embodiment uses the area of ​​each irregularly shaped pattern as a weighting coefficient, so that irregularly shaped patterns with larger areas contribute more to the final alignment reference point, while fragmented irregularly shaped patterns with smaller areas contribute less. The calculated area-weighted center coordinates can more accurately reflect the actual distribution center position of the detection window. Compared with directly taking the geometric center of the contour or a simple arithmetic mean, this formula can adaptively balance the influence weight of each sub-region even when the irregularly shaped patterns are irregularly distributed in multiple blocks due to uneven ink diffusion, avoiding excessive disturbance to the overall alignment reference point due to the offset of individual isolated small inkjet patches. At the same time, this calculation method has no preset limit on the total number n of irregularly shaped patterns, and can be applied to various distribution situations within the window, such as single continuous blocks, multiple discrete blocks, or complex connected forms, ensuring the calculation stability of the alignment reference point and the faithful representation of the actual shape of the irregularly shaped patterns.

[0071] In this embodiment, the coordinates of the center point of each irregular shape pattern are calculated, and the area-weighted center coordinates of the detection window are calculated based on the coordinates of the center points of multiple irregular shapes patterns, including: The coordinates of the Mark flag on the corresponding detection window on the acquisition substrate are obtained; Based on the coordinates of multiple area-weighted center points and the corresponding Mark flags, the positional deviation correction parameters between the substrate and the color filter inkjet layer are calculated using a multi-point averaging correction method. The color filter inkjet layer is aligned with the substrate and bonded using position deviation correction parameters.

[0072] This embodiment obtains the absolute position reference of the substrate in physical space by acquiring the coordinates of each Mark marker on the substrate. These coordinates, as inherent characteristics of the substrate itself, do not directly participate in the determination of the alignment reference point, but provide an indispensable fixed reference system for subsequent deviation quantification. Based on this, the weighted center coordinates of each detection window area obtained above are correlated and compared with the corresponding Mark marker coordinates. The multi-point averaging correction method, based on at least three sets of corresponding point pairs, can simultaneously quantify and compensate for the translational deviation between the substrate and the inkjet layer in the X and Y axes, and the rotational deviation around the vertical axis. This full-dimensional deviation calculation method completely eliminates the inherent defect of the single reference point correction method, which cannot distinguish between translational and rotational coupling errors. More importantly, this correction method uses the actual distribution state of the inkjet layer itself as an active reference reference, rather than passively matching the fixed markings on the substrate. This makes the correction parameters naturally adaptable to the random deviations of the inkjet process and the non-uniformity of ink diffusion, thereby fundamentally avoiding the problem of alignment failure caused by the inkjet layer covering the Mark points in the traditional method. Finally, the positional deviation correction parameter is used to drive a high-precision servo alignment platform to align and bond the inkjet layer to the substrate. At this point, each pixel area of ​​the inkjet layer achieves optimal spatial matching with the corresponding pixel electrode on the substrate. Furthermore, all calculations in this process are automated; correction parameters are generated in real time and directly output to the alignment platform control system without human intervention. This ensures high consistency in alignment accuracy during mass production and effectively reduces quality fluctuations caused by human error.

[0073] Specifically, after obtaining the area-weighted center coordinates of each detection window, the precise coordinates of the original Mark markers for the corresponding detection windows on the TFT substrate in the substrate's physical coordinate system need to be acquired simultaneously as a fixed reference system for subsequent deviation quantification calculations. The Mark markers on the substrate are typically cross-shaped or circular patterns. During actual acquisition, the same high-resolution CCD camera and telecentric lens as used for inkjet layer image acquisition are employed. Under the same lighting conditions and focal length parameters, the Mark markers at the four corners of the substrate are photographed separately to ensure clear images with sufficient contrast. The system identifies the cross-shaped outline of the Mark markers using a sub-pixel-level edge extraction algorithm, calculates the intersection point of the cross lines, and uses the coordinates of this intersection point as the reference coordinates for the corresponding Mark marker. The coordinate values ​​of the four Mark markers in a unified pixel coordinate system are recorded. To improve the reliability of the acquisition results, the system repeatedly acquires each Mark marker and takes the average value to eliminate measurement noise caused by random vibrations or ambient light fluctuations.

[0074] In this embodiment, based on the coordinates of multiple area-weighted center points and the corresponding Mark flags, the positional deviation correction parameters between the substrate and the color filter inkjet layer are calculated using a multi-point averaging correction method, including: Obtain the area-weighted center coordinates of the four detection windows, map the four area-weighted center coordinates to the same reference plane coordinate system, and obtain the normalized center coordinate set; Obtain the coordinates of the Mark flags corresponding to the coordinates of the four area-weighted center points, and map the coordinates of the four Mark flags to the reference plane coordinate system to obtain the theoretical reference point set; Calculate the offset vector between each coordinate point in the normalized center coordinate set and the corresponding reference point in the theoretical reference point set to obtain four deviation vectors; Using four deviation vectors as input, the least squares method is used to construct the affine transformation matrix between the TFT substrate and the color filter inkjet layer. The horizontal translation ΔX, the vertical translation ΔY, and the planar rotation angle Δθ are decomposed from the affine transformation matrix, and these three parameters are used as position deviation correction parameters.

[0075] This embodiment normalizes the area-weighted center coordinates of the four detection windows and their corresponding theoretical alignment reference point coordinates by mapping them to the same reference plane coordinate system. This eliminates coordinate reference differences introduced by the different spatial positions of each window, making the measured data and theoretical data of the four windows comparable under a unified reference framework. Using the offset vector between each measured coordinate point and its corresponding theoretical reference point as the input for deviation calculation, it can simultaneously reflect the alignment deviation status of each of the four corners, avoiding the limitation of single reference point correction failing to detect local substrate deformation and overall inkjet layer offset. The least squares method is used to construct the affine transformation matrix for the four deviation vectors, and the transformation parameters that minimize the overall residual are fitted through mathematical optimization. This effectively smooths out random fluctuations caused by local inkjet anomalies in individual windows, improving the overall representativeness and robustness of the correction parameters. By decomposing the horizontal translation, vertical translation, and planar rotation angle from the affine transformation matrix, the complex spatial deviation relationship is decoupled into three independent and controllable physical correction parameters. These parameters can directly drive the split-axis motion mechanism of the alignment platform to perform precise position compensation, thereby achieving synchronous correction of translational and rotational deviations between the TFT substrate and the color filter inkjet layer.

[0076] Specifically, the origin is the geometric center of the substrate pixel array region, the horizontal row direction of the pixel array is the positive X-axis, and the vertical column direction of the pixel array is the positive Y-axis, with the coordinate unit being micrometers. During coordinate mapping, the image processing system sequentially reads the area-weighted center coordinates of the four detection windows. For the area-weighted center coordinates of the upper left detection window, this coordinate value is currently in the image pixel coordinate system of the upper left camera, with the origin being the center of the upper left pixel of the camera imaging chip, and the coordinate unit being pixels. The system calls the pre-calibrated transformation parameters from the camera coordinate system to the reference plane coordinate system. These transformation parameters are obtained through the system calibration phase by shooting the standard calibration board and extracting corner points, and include the scaling factor from the camera pixel size to the physical size, the translation of the origin of the camera image coordinate system relative to the reference plane coordinate system, and the rotation angle of the coordinate axes.

[0077] Next, the x-coordinate pixel value of the area-weighted center coordinate of the top-left detection window is multiplied by a scaling factor to convert it into the physical size x-coordinate value; the y-coordinate pixel value is multiplied by a scaling factor to convert it into the physical size y-coordinate value; then, according to the translation of the camera image coordinate system relative to the origin of the reference plane coordinate system, translation correction values ​​are added to the converted physical size x-coordinate and y-coordinate respectively; finally, according to the coordinate axis rotation angle, a rotation transformation is applied to the translation-corrected x-coordinate and y-coordinate to obtain the final coordinate value of the area-weighted center coordinate in the reference plane coordinate system.

[0078] Then, the coordinate mapping operation described above is performed sequentially on the area-weighted center coordinates of the lower left, upper right, and lower right detection windows, with each window using the calibration transformation parameters of its corresponding camera. After all four area-weighted center coordinates are mapped, four coordinate points are obtained in the reference plane coordinate system. These four coordinate points constitute a normalized center coordinate set. Each coordinate point in the normalized center coordinate set is in micrometers on the substrate design layout plane, and the spatial relative positional relationship of each coordinate point corresponds one-to-one with the actual physical positional relationship on the substrate, providing accurate spatial position input for subsequent deviation vector calculation with the theoretical reference point set.

[0079] Furthermore, after extracting the area-weighted center coordinates of each detection window, the system simultaneously acquires the coordinates of the Mark markers corresponding to these four area-weighted center coordinates, i.e., the measured coordinate values ​​of the original optical Mark points on the substrate at the corners of each detection window. Since the four detection windows are located at the four corners of the substrate, and each window has a definite relative positional relationship with the Mark marker at its corresponding corner, a one-to-one correspondence can be established between the area-weighted center of each window and the Mark marker within that window area, thus forming four sets of corresponding coordinate pairs. Based on this, the alignment control system first uniformly transforms the measured coordinates of the four Mark markers into a reference plane coordinate system. The reference plane coordinate system is established based on the theoretical design dimensions of the substrate, with the geometric center of the substrate as the origin, and the horizontal and vertical edges of the substrate as the X and Y axes, respectively. The scale of the coordinate system corresponds to the actual physical dimensions. To achieve coordinate mapping, the system uses camera calibration parameters to convert the pixel coordinates of each Mark flag into physical coordinates, and then uses rotation and translation transformations to align the physical coordinates to the coordinate axis directions of the reference plane coordinate system. Finally, the system obtains the precise coordinate values ​​of the four Mark flags in the reference plane coordinate system. These four coordinate values ​​together constitute the theoretical reference point set.

[0080] After obtaining the normalized center coordinate set and the theoretical reference point set, the industrial image processing system calculates the offset vector between each coordinate point in the normalized center coordinate set and the corresponding reference point in the theoretical reference point set, thus obtaining four deviation vectors.

[0081] The normalized center coordinate set contains four area-weighted center coordinates calculated from measured inkjet profiles and mapped to the reference plane coordinate system: the measured coordinates of the upper left, lower left, upper right, and lower right detection windows. The theoretical reference point set contains four theoretical alignment reference point coordinates, which correspond one-to-one with the normalized center coordinate set according to window number.

[0082] Next, for the top-left detection window, the image processing system reads the measured x-coordinate and y-coordinate values ​​of the window in the normalized center coordinate set, and simultaneously reads the theoretical x-coordinate and y-coordinate values ​​of the window in the theoretical reference point set. Subtracting the theoretical x-coordinate value from the measured x-coordinate value yields the horizontal component of the window's offset vector; subtracting the theoretical y-coordinate value from the measured y-coordinate value yields the vertical component of the window's offset vector. These horizontal and vertical components together constitute the two-dimensional offset vector of the top-left detection window, with the positive and negative directions representing the offset direction of the measured inkjet center relative to the theoretical alignment position, respectively.

[0083] Subsequently, following the same method, the image processing system sequentially calculates the horizontal and vertical differences between the measured coordinates and the corresponding theoretical coordinates of the lower left detection window to obtain the offset vector of the lower left detection window; calculates the horizontal and vertical differences between the measured coordinates and the corresponding theoretical coordinates of the upper right detection window to obtain the offset vector of the upper right detection window; and calculates the horizontal and vertical differences between the measured coordinates and the corresponding theoretical coordinates of the lower right detection window to obtain the offset vector of the lower right detection window.

[0084] After calculating the offset vectors corresponding to each of the four detection windows, four deviation vectors are obtained. These four deviation vectors quantitatively describe the degree and direction of spatial offset of the actual distribution center of gravity of the color filter inkjet layer at the four corners of the substrate relative to the theoretical alignment reference point. Since the four detection windows are distributed at the four corners of the substrate and are relatively far apart, the numerical differences of the four deviation vectors can reflect the combined effect of two types of deviation factors: the overall offset of the inkjet layer and local deformation of the substrate. If the magnitudes and directions of the four deviation vectors are basically consistent, it indicates that the deviation is mainly manifested as the overall translation of the inkjet layer relative to the substrate; if there are significant differences in the magnitudes or directions of the four deviation vectors, it indicates that there is local deformation of the substrate or regional uneven offset of the inkjet layer. These four deviation vectors serve as input data for the subsequent construction of the affine transformation matrix, providing complete deviation distribution information for solving the horizontal translation, vertical translation, and planar rotation angle.

[0085] After obtaining the four deviation vectors, the industrial image processing system uses the four deviation vectors as inputs and employs the least squares method to construct the affine transformation matrix between the thin-film transistor substrate and the color filter inkjet layer.

[0086] The four deviation vectors correspond to the detection windows at the four corners of the substrate. Each deviation vector contains two values: a horizontal component and a vertical component. The four deviation vectors provide a total of eight scalar equations as observation data for solving the affine transformation matrix. The affine transformation matrix is ​​a 2x3 matrix containing six transformation parameters to be solved. Four parameters describe the rotation and scaling relationship between the substrate and the inkjet layer, and two parameters describe the translation relationship in the horizontal and vertical directions.

[0087] When constructing the affine transformation matrix, the image processing system first establishes a system of observation equations. For each detection window, the coordinates of its theoretical reference point in the reference plane coordinate system, multiplied by the affine transformation matrix, should equal the weighted center coordinates of the measured area corresponding to that window minus the deviation vector of that window. This establishes two equations for that window in the horizontal and vertical directions. A total of eight linear equations are established for the four detection windows. The unknowns in these equations are the six parameters of the affine transformation matrix to be solved. The number of equations is greater than the number of unknowns, forming an overdetermined system of linear equations.

[0088] Next, the least squares method is used to solve the overdetermined linear equation system. The eight equations are written in matrix form. The left side is the coefficient matrix, an 8x6 matrix composed of the coordinates of the theoretical reference points, with each row corresponding to the coefficients of the six parameters to be solved in one equation. The right side is the observation vector, an 8x1 column vector composed of the weighted center coordinates of the measured area. During the solution process, the product of the transpose of the coefficient matrix and the coefficient matrix is ​​calculated to obtain a 6x6 normal equation matrix. The product of the transpose of the coefficient matrix and the observation vector is calculated to obtain a 6x1 normal equation vector. The inverse of this 6x6 normal equation matrix is ​​then multiplied by the normal equation vector to obtain a 6x1 solution vector. The six elements of the solution vector are the six parameters of the affine transformation matrix. By optimizing the solution using the least squares method, six parameters minimize the sum of squared residuals between the weighted center coordinates of the measured areas of the four detection windows and the coordinates of the theoretical reference point after affine transformation, thereby determining the spatial transformation relationship between the substrate and the inkjet layer in the sense of overall optimization.

[0089] After obtaining the affine transformation matrix, the image processing system decomposes it into horizontal translation, vertical translation, and planar rotation angle. The element in the second row and third column of the affine transformation matrix is ​​the vertical translation ΔY.

[0090] The decomposition of the planar rotation angle is performed using a 2x2 submatrix describing rotation and scaling within the affine transformation matrix. The elements of the first row and first column of this submatrix are extracted, and the arctangent of the element in the first row and second column divided by the element in the first row and first column is calculated to obtain the rotation angle in radians. This radian value is then converted to an angle value to obtain the planar rotation angle Δθ. This rotation angle Δθ represents the overall rotation angle of the color filter inkjet layer relative to the thin-film transistor substrate within the bonding plane; a positive value indicates counter-clockwise rotation, and a negative value indicates clockwise rotation.

[0091] After decomposition, the horizontal translation ΔX, the vertical translation ΔY, and the planar rotation angle Δθ are used together as position deviation correction parameters. These position deviation correction parameters fully describe the spatial deviation between the thin-film transistor substrate and the color filter inkjet layer that needs to be compensated by the servo alignment platform. ΔX is the correction distance the platform needs to move in the horizontal direction, ΔY is the correction distance the platform needs to move in the vertical direction, and Δθ is the rotation correction angle the platform needs to perform. These three parameters are directly input into the motion controller of the alignment control system to drive the alignment platform to complete precise position compensation.

[0092] After obtaining the position deviation correction parameters, the alignment platform movement parameters are transmitted to the servo alignment control system. The position deviation correction parameters are used to align the color filter inkjet layer with the thin film transistor substrate and complete the bonding.

[0093] Before bonding, the thin-film transistor substrate is fixed on the bonding platform, and the platform's level is adjusted to ensure that the substrate surface is neither tilted nor displaced. Simultaneously, the bonding surface of the color filter inkjet layer is cleaned to remove surface dust and impurities, avoiding any impact on the bonding effect. Then, based on the position deviation correction parameters ΔX and ΔY, the bonding platform is driven to perform fine-tuning in the horizontal and vertical directions, initially aligning the actual position of the substrate with the preset position of the inkjet layer. Finally, based on the Δθ parameter, the platform's rotation angle is fine-tuned to correct misalignment caused by angular deviations, ensuring that the edge of the inkjet layer precisely corresponds to the pixel position of the substrate, without offset or misalignment.

[0094] After translation and rotation corrections are completed, the image processing system triggers four charge-coupled device (CCD) cameras to re-capture verification images of the four detection windows. The coordinates of the four area-weighted centers are recalculated and compared with the theoretical reference point set. Once the residual deviation is confirmed to be within the set threshold range, the bonding process begins.

[0095] During bonding, the upper platform of the alignment platform supports the e-ink substrate containing the color filter inkjet layer, while the lower platform supports the thin-film transistor substrate that has undergone positional correction. The upper and lower platforms remain parallel and opposite each other. The alignment control system drives the vertical lifting shaft to move the upper platform downwards at a low speed, gradually bringing the e-ink substrate closer to the thin-film transistor substrate. After the two substrates contact, the alignment platform applies a preset bonding pressure to tightly bond the color filter inkjet layer and the pixel array of the thin-film transistor substrate. The bonding pressure is maintained for a preset time and then released, completing the entire pixel alignment and bonding process between the color filter inkjet layer and the thin-film transistor substrate.

[0096] This embodiment abandons the traditional method of fixing optical markers on thin-film transistor substrates. Instead, it uses the area-weighted center of the inkjet layer itself as the alignment reference, fundamentally avoiding the problems of marker coverage or recognition failure caused by inkjet position offset and uneven ink diffusion, thus solving the core defects of existing alignment methods. Furthermore, the alignment reference point is calculated based on the actual distribution pattern of the inkjet layer, which can adapt to morphological changes caused by inkjet position offset and ink diffusion. Simultaneously, a four-point multi-point averaging correction method effectively compensates for slight substrate deformation, allowing the alignment and bonding deviation to be controlled within ±2 micrometers, far superior to the ±8 micrometer deviation level of traditional marker alignment, significantly improving alignment accuracy. This solution has strong process adaptability; the detection window size, charge-coupled device acquisition parameters, and image processing algorithms can all be flexibly adjusted according to the electronic paper resolution, inkjet material characteristics, and production process requirements. It is not limited by inkjet shape, diffusion degree, or whether the substrate is rigid or flexible, making it suitable for the manufacture of various types of color electronic paper. In terms of improving production yield and display quality, high-precision alignment and bonding effectively solves problems such as color shift, pixel misalignment, and uneven display in electronic paper, improving display uniformity by over 90%. It also significantly reduces product scrap due to alignment deviations, lowering the scrap rate from 15% to below 2%, thus reducing industrial production costs. Furthermore, the image acquisition, image processing, and multi-point correction processes for charge-coupled devices are all automated, allowing direct integration into existing color electronic paper production lines without large-scale modifications to existing equipment. This provides strong process compatibility and facilitates industrial-scale deployment.

[0097] Please see Figure 5 This embodiment provides an apparatus 500 for aligning an electronic paper substrate with a color filter inkjet layer, comprising: The setting module is used to set a detection window on the color filter inkjet layer, wherein the orthographic projection of the detection window on the substrate covers the Mark mark position of the substrate; the detection window includes multiple complete irregular shape patterns. The calculation module is used to calculate the center point coordinates of each complete irregular shape pattern, and to calculate the area-weighted center coordinates of the detection window based on the center point coordinates of multiple irregular shape patterns. The detection window is provided with at least three, and the Mark flags covered by the three detection windows are not located on the same straight line.

[0098] Furthermore, the setting module 501 includes: A window setting unit is configured to set two detection windows symmetrically arranged about a horizontal axis of symmetry on the color filter inkjet layer within at least three detection windows, and to set two detection windows symmetrically arranged about a vertical axis of symmetry on the color filter inkjet layer within at least three detection windows.

[0099] Furthermore, the computing module 503 includes: The image acquisition unit is used to acquire the actual distribution image of all complete irregular shape patterns within the detection window; The contour extraction unit is used to extract irregular shape pattern contours from each of the actual distribution images to obtain the effective contour information of each detection window. The coordinate calculation unit is used to calculate the center point coordinates of each complete irregular shape pattern based on the effective contour information, and to calculate the area-weighted center coordinates of each detection window based on the center point coordinates of all complete irregular shape patterns.

[0100] The contour extraction unit includes: The processing subunit is used to perform grayscale and binarization processing on the actual distribution image to obtain each complete irregular shape pattern; The edge detection subunit is used to perform edge detection on each complete irregular shape pattern to obtain the preliminary outline of each complete irregular shape pattern; The noise reduction processing subunit is used to perform noise reduction processing on the preliminary contours of all complete irregular shape patterns to obtain the effective contour information.

[0101] Furthermore, the processing subunit includes: The weighted average subunit is used to perform a weighted average calculation on the three-channel values ​​of each pixel in the actual distribution image to obtain a grayscale image. The histogram generation subunit is used to generate a grayscale histogram based on the grayscale value distribution of each pixel in the grayscale image. The threshold extraction subunit is used to extract the gray value corresponding to the valley position from the gray histogram as the segmentation threshold. The assignment subunit is used to compare the gray value of each pixel in the grayscale image with the segmentation threshold, assign a first value to the pixels whose gray value is greater than or equal to the segmentation threshold, and assign a second value to the pixels whose gray value is less than the segmentation threshold, so as to obtain a binary image. The region constitutes a sub-unit, used to identify connected regions formed by the first numerical value in the binary image, and to treat the connected regions as complete irregular shape patterns.

[0102] Furthermore, the edge detection subunit includes: The gradient calculation subunit is used to perform gradient calculation on the complete irregular shape pattern to obtain a gradient magnitude image. The suppression processing subunit is used to perform non-maximum suppression processing based on the gradient direction of each pixel in the gradient magnitude image to obtain a thinned edge image; The threshold detection subunit is used to perform dual threshold detection on the thinned edge image using a preset high threshold and a low threshold. Pixels with gradient magnitude greater than or equal to the high threshold are marked as strong edge points, pixels with gradient magnitude greater than the low threshold and less than the high threshold are marked as weak edge points, and pixels with gradient magnitude less than or equal to the low threshold are marked as non-edge points. An additional subunit is used to mark the pixels adjacent to any of the strong edge points within the weak edge points as strong edge points, thereby obtaining the preliminary contour.

[0103] Furthermore, the noise reduction processing subunit includes: The noise reduction subunit is used to perform noise reduction processing on the preliminary contour using a median filtering algorithm to obtain the effective contour information.

[0104] Furthermore, the computing module 504 includes: The coordinate calculation unit is used to calculate the area-weighted center coordinates of each detection window according to the following formula: Where X and Y represent the area-weighted center coordinates of each detection window; x i y i S represents the center coordinates of the i-th irregular shape pattern; i Let be the area of ​​the i-th irregular shape pattern; n is the total number of irregular shape patterns.

[0105] Furthermore, the computing module 504 includes: A coordinate acquisition unit is used to acquire the coordinates of the Mark flag bit on the substrate corresponding to the detection window; The parameter calculation unit is used to calculate the position deviation correction parameters between the substrate and the color filter inkjet layer by means of a multi-point averaging correction method, based on the coordinates of multiple area-weighted center coordinates and the corresponding Mark flags. The alignment unit is used to align the color filter inkjet layer with the substrate and complete the bonding using the position deviation correction parameters.

[0106] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed, can implement the methods provided in the above embodiments. The storage medium may include various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0107] The present invention also provides a computer device, which may include a memory and a processor. The memory stores a computer program, and when the processor calls the computer program in the memory, it can implement the methods provided in the above embodiments. Of course, the computer device may also include various network interfaces, power supplies, and other components.

[0108] The present invention also provides an electronic paper module, which includes an electronic paper substrate and a color filter inkjet layer. The electronic paper module and the color filter inkjet layer are bonded together using the method for aligning the electronic paper substrate and the color filter inkjet layer described in the above embodiments.

[0109] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section. It should be noted that those skilled in the art can make various improvements and modifications to this invention without departing from its principles, and these improvements and modifications also fall within the protection scope of the claims of this invention.

[0110] It should also be noted that, in this specification, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusivity.

[0111] The term "comprises" implies that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprises a..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A method for aligning an electronic paper substrate with a color filter inkjet layer, characterized in that, include: A detection window is provided on the color filter inkjet layer, and the orthographic projection of the detection window on the substrate covers the Mark mark on the substrate; The detection window includes multiple complete irregular shape patterns; Calculate the center point coordinates of each complete irregular shape pattern, and calculate the area-weighted center coordinates of the detection window based on the center point coordinates of multiple irregular shape patterns; The detection window is provided with at least three, and the Mark flags covered by the three detection windows are not located on the same straight line.

2. The method for aligning an electronic paper substrate with a color filter inkjet layer according to claim 1, characterized in that, The method of setting a detection window on the color filter inkjet layer includes: Two detection windows are symmetrically arranged about a horizontal axis of symmetry on the color filter inkjet layer within at least three detection windows, and two detection windows are symmetrically arranged about a vertical axis of symmetry on the color filter inkjet layer within at least three detection windows.

3. The method for aligning an electronic paper substrate with a color filter inkjet layer according to claim 1, characterized in that, The calculation of the center point coordinates of each of the irregular shapes and patterns, and the calculation of the area-weighted center coordinates of the detection window based on the center point coordinates of multiple irregular shapes and patterns, include: Acquire the actual distribution image of all complete irregular shape patterns within the detection window; For each of the actual distribution images, an irregular shape pattern contour is extracted to obtain the effective contour information of each detection window; Based on the effective contour information, the center point coordinates of each complete irregular shape pattern are calculated, and the area-weighted center coordinates of each detection window are calculated based on the center point coordinates of all complete irregular shape patterns.

4. The method for aligning an electronic paper substrate with a color filter inkjet layer according to claim 3, characterized in that, The step of extracting irregular shape and pattern contours from the actual distribution image to obtain effective contour information for each detection window includes: The actual distribution image is processed by grayscale and binarization to obtain each complete irregular shape pattern; Edge detection is performed on each complete irregular shape pattern to obtain the preliminary outline of each complete irregular shape pattern; The initial contours of all complete irregular shape patterns are subjected to noise reduction processing to obtain the effective contour information.

5. The method for aligning an electronic paper substrate with a color filter inkjet layer according to claim 4, characterized in that, The grayscale and binarization processing of the actual distribution image to obtain each complete irregular shape pattern includes: A grayscale image is obtained by performing a weighted average of the three-channel values ​​of each pixel in the actual distribution image. A grayscale histogram is generated based on the grayscale value distribution of each pixel in the grayscale image; Extract the gray values ​​corresponding to the valley positions from the gray-level histogram as the segmentation threshold; The grayscale value of each pixel in the grayscale image is compared with the segmentation threshold. Pixels with grayscale values ​​greater than or equal to the segmentation threshold are assigned a first value, and pixels with grayscale values ​​less than the segmentation threshold are assigned a second value to obtain a binary image. In the binary image, a connected region composed of the first value is identified, and the connected region is regarded as a complete irregular shape pattern.

6. The method for aligning an electronic paper substrate with a color filter inkjet layer according to claim 4, characterized in that, The step of performing edge detection on each complete irregular shape pattern to obtain the preliminary outline of each complete irregular shape pattern includes: Gradient calculation is performed on the complete irregular shape pattern to obtain a gradient magnitude image; Based on the gradient direction of each pixel in the gradient magnitude image, non-maximum suppression processing is performed to obtain a thinned edge image. The refined edge image is subjected to dual threshold detection using a preset high threshold and a low threshold. Pixels with gradient magnitude greater than or equal to the high threshold are marked as strong edge points, pixels with gradient magnitude greater than the low threshold and less than the high threshold are marked as weak edge points, and pixels with gradient magnitude less than or equal to the low threshold are marked as non-edge points. The pixels adjacent to any of the strong edge points within the weak edge points are marked as strong edge points to obtain the preliminary contour.

7. The method for aligning an electronic paper substrate with a color filter inkjet layer according to claim 4, characterized in that, The noise reduction process performed on the preliminary contour to obtain the effective contour information includes: The preliminary contour is denoised using a median filtering algorithm to obtain the effective contour information.

8. The method for aligning an electronic paper substrate with a color filter inkjet layer according to claim 3, characterized in that, The process of calculating the center point coordinates of each complete irregular shape pattern based on the effective contour information, and calculating the area-weighted center coordinates of each detection window based on the center point coordinates of all complete irregular shape patterns, includes: The area-weighted center coordinates of each detection window are calculated using the following formula: Where X and Y represent the area-weighted center coordinates of each detection window; x i y i S represents the center coordinates of the i-th irregular shape pattern; i Let be the area of ​​the i-th irregular shape pattern; n is the total number of irregular shape patterns.

9. The method for aligning an electronic paper substrate with a color filter inkjet layer according to claim 1, characterized in that, The process of calculating the center point coordinates of each of the irregular shapes and patterns, and then calculating the area-weighted center coordinates of the detection window based on the center point coordinates of multiple irregular shapes and patterns, includes: Collect the coordinates of the Mark flag bit on the substrate corresponding to the detection window; Based on the coordinates of multiple area-weighted center points and the corresponding Mark flags, the positional deviation correction parameters between the substrate and the color filter inkjet layer are calculated using a multi-point averaging correction method. The color filter inkjet layer is aligned with the substrate and bonded using the position deviation correction parameters.

10. An apparatus for aligning an electronic paper substrate with a color filter inkjet layer, characterized in that, include: The setting module is used to set a detection window on the color filter inkjet layer, wherein the orthographic projection of the detection window on the substrate covers the Mark flag bit of the substrate; The detection window includes multiple complete irregular shape patterns; The calculation module is used to calculate the center point coordinates of each complete irregular shape pattern, and to calculate the area-weighted center coordinates of the detection window based on the center point coordinates of multiple irregular shape patterns. The detection window is provided with at least three, and the Mark flags covered by the three detection windows are not located on the same straight line.

11. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for aligning an electronic paper substrate with a color filter inkjet layer as described in any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, causes the processor to perform the method for aligning an electronic paper substrate with a color filter inkjet layer as described in any one of claims 1 to 9.

13. An electronic paper module, characterized in that, The electronic paper module includes an electronic paper substrate and a color filter inkjet layer, and the electronic paper module and the color filter inkjet layer are bonded together using the method for aligning the electronic paper substrate and the color filter inkjet layer as described in any one of claims 1 to 9.

Images