Multi-screen seamless splicing glass substrate display optical consistency adjusting method

Abstract

The invention belongs to the technical field of optical display, and discloses a multi-screen seamless splicing glass substrate display optical consistency adjusting method. The method comprises the steps that optical parameters corresponding to each display unit are collected in real time and analyzed, and the display units are divided into reference units and adjusting units; dynamically correcting the optical parameter of each adjusting unit based on the optical parameter of the reference unit, and identifying a splicing seam area; obtaining the gradual change weight of each pixel point in the abutted seam area, and adjusting the optical parameter of each abutted seam area in real time; dividing all the display units into a plurality of sensing areas, configuring differentiated sensing weights for different sensing areas, and dynamically optimizing optical parameters of different sensing areas; according to the invention, highly reliable and intelligent multi-screen seamless splicing optical consistency adjustment is realized, and it is ensured that optimal visual experience can be provided in various application scenes, so that the overall viewing effect of a user is significantly enhanced.

Description

Technical Field

[0001] The present invention relates to the field of optical display technology, and more particularly to a method for adjusting the optical consistency of a glass-based display with seamless multi-screen splicing. Background Art

[0002] With the growing demand for information visualization, large-scale, high-resolution display systems have been widely used in a variety of fields, including security monitoring, command and dispatch, education and training, and commercial exhibitions. In particular, when constructing large-screen display walls, a method of splicing multiple small and medium-sized display units is usually adopted to meet the needs of ultra-large-scale displays. Among the current mainstream splicing display technologies, glass-based display panels (such as liquid crystal displays (LCDs), OLEDs, and MicroLEDs) are widely used due to their high brightness, high resolution, and high degree of industrialization. However, due to issues such as manufacturing batch differences, driver circuit differences, and changes in usage environments among individual display units, it is difficult for the spliced display system to maintain complete consistency in brightness, chromaticity, and contrast. This results in obvious "stitching effects" and visual discontinuities, which seriously affect the overall display quality and user experience. Therefore, an intelligent optical consistency adjustment method is urgently needed that can effectively adjust the optical consistency between each display unit without changing the display structure, thereby achieving a seamless, high-quality integrated display effect.

[0003] The patent with announcement number CN104656299B discloses a high-definition large-screen display with seamless splicing of arbitrary curved and flat surfaces; it comprises a plurality of brick-shaped fiber optic rear-projection display modules stacked together, the components of which include a variable-section fiber optic panel, a signal processing module, a liquid crystal display panel, an LED light source module, a fixing device and a shell, an anti-glare frosted coating, and a touch sensor panel; this invention can produce a seamlessly spliced, small-dot-pitch, high-definition large screen, which can display video images on any curved surface, any special shape, and any area; the principle of seamless splicing is native pixel-level splicing, the pixels are arranged evenly and tightly, and the display consistency of brightness and chromaticity is excellent; modular production is adopted, and the components are easy to repair and replace; a touch sensor panel can be embedded for precise positioning, and at the same time, it has no effect on the image quality and appearance of the display; it has good waterproof and impact resistance.

[0004] However, although the above-mentioned technology can achieve optical consistency in display when multiple screens are seamlessly connected, it mainly relies on physical splicing and material properties to achieve a seamless effect. Although it mentions the use of camera acquisition and computer algorithm processing to correct parameters, it does not explain in detail the specific algorithm principles and real-time adjustment mechanism. There is a lack of a systematic dynamic optimization solution for optical parameters, making it difficult to effectively deal with brightness and color inconsistencies that may occur during use, resulting in a decrease in the splicing effect over time in actual use. At the same time, the above-mentioned technology does not consider special processing algorithms for the splicing area, nor does it introduce optimization strategies based on human visual perception, making it difficult to solve the possible drift of the optical parameters of each display unit after long-term use.

[0005] In view of this, the present invention proposes a method for adjusting the optical consistency of a glass-based display with seamless multi-screen splicing to solve the above-mentioned problem. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned objectives, the present invention provides the following technical solution: a method for adjusting the optical consistency of a glass-based display with seamless multi-screen splicing, the method comprising: Step S1: collecting the optical parameters corresponding to each display unit in real time; Step S2: Analyze the optical parameters of each display unit and divide the display unit into a reference unit and an adjustment unit based on the minimum variance principle; Step S3: Based on the optical parameters of the reference unit, dynamically modify the optical parameters of each adjustment unit and identify the seam area in each display unit; Step S4: Using an edge fusion algorithm, perform pixel-level weight gradient processing on each seam area, obtain the gradient weight of each pixel in the seam area, and adjust the optical parameters of each seam area in real time based on the gradient weight; Step S5: Divide all display units into multiple perception areas according to the visual attention model, configure differentiated perception weights for different perception areas based on a pre-built human eye perception weight function, and dynamically optimize the optical parameters of different perception areas based on the perception weights.

[0007] Furthermore, the step of dividing the display unit into a reference unit and an adjustment unit includes: Step S201: randomly selecting a display unit that is not marked as a selected unit and marking it as the current unit, and taking all the remaining display units as a unit set; Step S202: Calculating a discrete set between the current unit and the unit set based on the optical parameters of each display unit, where the discrete set includes a discrete degree corresponding to each parameter in the optical parameters; Step S203: Preset a weight set, which includes a weight coefficient corresponding to each optical parameter; collect influencing parameters, and dynamically adjust the weight coefficients in the weight set based on the influencing parameters; Step S204: Based on the dynamically adjusted weight set, weighted summation is performed on the discrete degrees in the discrete set to obtain the unit discrete degree, and the current unit is marked as a selected unit; Step S205: looping steps S201 to S204 until all display units are marked as selected units, the loop ends, and the process proceeds to step S206; Step S206: comparing the discreteness of all units respectively, taking the display unit with the smallest discreteness as the reference unit, and taking all the display units that are not used as the reference unit as the adjustment units.

[0008] Furthermore, in step S203, the influencing parameters include application scenarios and content types; and the method for dynamically adjusting the weight coefficients in the weight set includes: Based on the influencing parameters, the corresponding scene labels and type labels are obtained; the weight set, scene labels, and type labels are used as an analysis set, and the analysis set is input into the trained weight adjustment model to obtain an adjustment set, which includes an adjustment amount corresponding to each weight coefficient; based on each adjustment amount in the adjustment set, the corresponding weight coefficient in the weight set is dynamically adjusted; the training process of the weight adjustment model includes: Pre-collection A different set of analyses Each group analysis set has a corresponding adjustment set. is an integer greater than 1, converting the analysis set and the corresponding adjustment set into a corresponding set of feature vectors; Each set of feature vectors is used as the input of the weight adjustment model, and the weight adjustment model takes a set of predicted adjustment sets corresponding to each set of analysis sets as output, and an actual adjustment set corresponding to each set of analysis sets as the prediction target, where the actual adjustment set is a pre-set adjustment set corresponding to the analysis set; minimizing the sum of the prediction errors of all analysis sets is used as the training goal; the weight adjustment model is trained until the sum of the prediction errors reaches convergence and the training is stopped.

[0009] Furthermore, the optical parameters of the reference unit are used as reference parameters, and the optical parameters of each adjustment unit are dynamically corrected to the reference parameters; The method for identifying the seam area within each display unit includes: Obtaining a display image corresponding to each display unit, performing grayscale processing on each display image to obtain a grayscale image; processing each grayscale image in turn using a gradient operator to calculate a horizontal gradient and a vertical gradient; taking the square root of the sum of the squares of the horizontal gradient and the vertical gradient corresponding to each grayscale image to obtain a gradient amplitude map corresponding to each grayscale image; In each gradient magnitude map, the gradient magnitudes of each row are summed vertically to obtain the total magnitude of the row gradients, and the gradient magnitudes of each column are summed horizontally to obtain the total magnitude of the column gradients; the maximum value of the total magnitudes of all row gradients is obtained, and the corresponding row is used as the segmentation row of the corresponding grayscale image; the maximum value of the total magnitudes of all column gradients is obtained, and the corresponding column is used as the segmentation column of the corresponding grayscale image; the grayscale image is divided into two regions according to each segmentation row or segmentation column; the number of pixels in each region is counted and marked as the number of pixels; the number of pixels in the two divided regions is compared each time, and the region with the smaller number of pixels is marked as the candidate stitching region, and the region with the larger number of pixels is marked as the non-stitching region; The optical parameters of each candidate seam area and non-seam area are collected, and the optical parameters of each candidate seam area and the optical parameters of the corresponding non-seam area are taken as a set of parameters; each set of parameters is input into the trained similarity analysis model to predict the corresponding similarity; a similarity threshold is preset, and each similarity is compared with the similarity threshold. The candidate seam areas with similarity less than the similarity threshold are marked as seam areas, and the candidate seam areas with similarity greater than or equal to the similarity threshold are not marked.

[0010] Furthermore, the method for obtaining the gradient weight of each pixel in the seam area includes: In each grayscale image, the overlapping area of two stitching regions is marked as the overlapping area, and the stitching region not marked as the overlapping area is marked as the separate area; connectivity analysis is performed on all overlapping areas, and adjacent overlapping areas are merged into the whole area; connectivity analysis is performed on all separate areas, and adjacent separate areas are merged into the joint area; The joint area and the overall area are collectively referred to as the analysis area. The index of each pixel point in each analysis area is obtained, and the index includes row index and column index. The row index of each pixel point is divided by the row index with the largest value in the corresponding analysis area to obtain the row weight. The column index of each pixel point is divided by the column index with the largest value in the corresponding analysis area to obtain the column weight. The row weight and column weight of each pixel point are used as the corresponding gradient weight.

[0011] Furthermore, the method for adjusting the optical parameters of each seam area in real time includes: Re-collecting the optical parameters of each display unit and marking them as re-collected parameters; marking the optical parameters of each seam area as seam parameters, and dividing each set of seam parameters by the corresponding re-collected parameters to obtain a correction weight set, wherein the correction weight set includes a correction weight corresponding to each parameter in the optical parameters; multiplying the optical parameters of each display unit by the correction weight set corresponding to the corresponding seam area to obtain the correction parameters of each display unit; Subtract the row weight of each pixel from 1 to obtain the reverse row weight of each pixel; subtract the column weight of each pixel from 1 to obtain the reverse column weight of each pixel; use the reverse row weight and reverse column weight of each pixel as the corresponding reverse gradient weight; For each overall area, the display units corresponding to the same overlapping areas of the overall area are merged to obtain an overall unit; based on the gradient weight and reverse gradient weight corresponding to each pixel and the correction parameter of the overall unit, a first adjustment parameter corresponding to each pixel is calculated, and the optical parameter of each pixel is adjusted in real time to the corresponding first adjustment parameter; For each joint area, the display units corresponding to the separate areas with the same corresponding joint area are obtained and merged to obtain a merged unit; the overall area adjacent to the joint area is obtained and merged with the merged unit to obtain an extended unit; according to the gradient weight, reverse gradient weight and correction parameter of the extended unit corresponding to each pixel point, the second adjustment parameter corresponding to each pixel point is calculated, and the optical parameter of each pixel point is adjusted to the corresponding second adjustment parameter in real time.

[0012] Furthermore, the method of dividing all display units into a plurality of sensing areas includes: A visual attention model is used to generate an attention distribution map, which includes the attention degree of each pixel in each display unit; wherein the visual attention model is a visual field center model; an eye movement heat map is obtained, which includes the attention degree of each pixel in each display unit; Based on the influencing parameters, the scaling factor is dynamically set; the scaling factor is subtracted from 1 to obtain the inverse factor; the product of the attention degree of each pixel and the scaling factor is added to the product of the corresponding attention degree and the inverse factor to obtain the fusion attention degree of each pixel; the perception threshold is preset, and the perception threshold includes the center threshold and the edge threshold, and the center threshold is greater than the edge threshold; The fusion attention of each pixel is compared with the perception threshold; if the fusion attention is greater than or equal to the center threshold, the corresponding pixel is marked as the center point; if the fusion attention is less than the center threshold and greater than the edge threshold, the corresponding pixel is marked as a transition point; if the fusion attention is less than or equal to the edge threshold, the corresponding pixel is marked as an edge point; Connected domain analysis is performed on the center points, transition points and edge points in all display units, and all display units are divided into a central visual area, an intermediate transition area and an edge perception area.

[0013] Furthermore, the method of dynamically setting the scale factor includes: Resetting different digital labels for different application scenarios and content types, and marking the reset digital labels of application scenarios as application labels, and marking the reset digital labels of content types as content labels; wherein both application labels and content labels are set according to the size of the corresponding scale factors of different application scenarios and content types; Using application tags and content tags as tag data, construct multiple fuzzy sets for each tag in the tag data; convert the tag data into the membership degree of each corresponding fuzzy set through fuzzification technology; All memberships are input into the trained membership analysis model to predict the corresponding proportion set; the membership analysis model is a deep neural network model, and the proportion set includes the membership of each scale factor level; set the proportion interval , divide the ratio interval evenly into Level ranges, is the number of scale factor levels; the average of the maximum and minimum values of each level interval is taken as the corresponding interval mean; according to the interval mean, the membership in the scale set is weighted and summed to obtain the total value of the scale; the membership in the scale set is added in sequence to obtain the total value of the membership; the quotient of the total value of the scale and the total value of the membership is taken as the scale factor.

[0014] Furthermore, the step of performing connected domain analysis on the center points, transition points, and edge points in all display units includes: Step S501: Pixels with the same perception type are grouped as a perception set, where the perception types include center points, transition points, and edge points. Step S502: randomly select a set of perception sets and mark them as the current set; Step S503: randomly select a pixel point that is not marked as a selected point from the current set and mark it as the current point; Step S504: Determine whether there are any pixels adjacent to the current point in all point sets; if there is only one pixel in the point set adjacent to the current point, add the current point to the corresponding point set; if there are pixels in multiple point sets adjacent to the current point, merge the multiple point sets into one point set and add the current point to the merged point set; if there are no pixels in the point set adjacent to the current point, create a new point set and add the current point to the new point set; Step S505: Mark the current point as a selected point; Step S506: looping steps S503 to S505 until all pixels in the current set are marked as selected points, then the loop ends and proceeds to step S507; Step S507: looping steps S502 to S506 until all the perception sets are marked as current sets, then the loop ends and proceeds to step S508; Step S508: Analyze all pixel points in each point set in turn; if all pixel points are center points, the corresponding point set is used as the central visual area; if all pixel points are transition points, the corresponding point set is used as the intermediate transition area; if all pixel points are edge points, the corresponding point set is used as the edge perception area.

[0015] Furthermore, the method for dynamically optimizing the optical parameters of different perception areas includes: According to the perception weight, an optimization algorithm is used to generate corresponding optimization strategies for different perception areas, and the optical parameters of different perception areas are dynamically optimized based on the optimization strategies.

[0016] The technical effects and advantages of the method for adjusting optical consistency of glass-based displays with seamless multi-screen splicing of the present invention are as follows: By collecting and analyzing the optical parameters of each display unit in real time, the display units are dynamically divided into reference units and adjustment units. Based on the optical parameters of the reference units, each adjustment unit is dynamically corrected, effectively solving the problem of inconsistent optical parameters caused by manufacturing differences, changes in the usage environment, etc.; the edge fusion algorithm is used to perform pixel-level weight gradient processing on the splicing area, effectively eliminating the visual faults and obvious gaps in the splicing area; the visual attention model and human eye perception weight function are introduced to divide the entire display area into the central visual area, the middle transition area and the edge perception area, and differentiated optical parameter optimization is performed according to the perceptual characteristics of different areas to achieve perception-driven intelligent adjustment, effectively improving the display quality and user experience of the spliced multi-screen display system; highly reliable and intelligent multi-screen seamless splicing optical consistency adjustment is achieved to ensure the best visual experience in various application scenarios, thereby significantly enhancing the user's overall viewing effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of the method for adjusting the optical consistency of a glass-based display with seamless multi-screen splicing according to Example 1 of the present invention. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] Example 1

[0020] See also Figure 1 As shown, the method for adjusting the optical consistency of a glass-based display with seamless multi-screen splicing in this embodiment includes: Step S1: collecting the optical parameters corresponding to each display unit in real time.

[0021] A display unit refers to each independent display module, which usually contains a glass-based display panel (such as a liquid crystal panel). Multiple display units can be combined through structural splicing to form a spliced multi-screen display system. Optical parameters are physical properties that describe the visual performance of display units, including brightness, chromaticity, color temperature, etc. Optical parameters are collected through optical measuring equipment, such as photometers, colorimeters, spectrometers, etc. The purpose of collecting optical parameters is to ensure the visual consistency and display quality of each display unit in the spliced multi-screen display system. By collecting optical parameters such as brightness, chromaticity, color temperature in real time, the visual performance of each display unit can be effectively adjusted and optimized, the visual differences of splicing gaps can be eliminated, and the uniform performance of the entire spliced multi-screen display system in dynamic or multi-scene applications can be ensured.

[0022] Step S2: Analyze the optical parameters of each display unit and divide the display unit into a reference unit and an adjustment unit based on the minimum variance principle.

[0023] The steps of dividing the display unit into the reference unit and the adjustment unit include: Step S201: randomly selecting a display unit that is not marked as a selected unit and marking it as the current unit, and taking all the remaining display units as a unit set; Step S202: Calculating a discrete set between the current unit and the unit set based on the optical parameters of each display unit, where the discrete set includes a discrete degree corresponding to each parameter in the optical parameters; Step S203: Preset a weight set, which includes a weight coefficient corresponding to each optical parameter. The weight set is pre-set by a person skilled in the art based on actual conditions; collect influencing parameters, and dynamically adjust the weight coefficients in the weight set based on the influencing parameters; Step S204: Based on the dynamically adjusted weight set, weighted summation is performed on the discrete degrees in the discrete set to obtain the unit discrete degree, and the current unit is marked as a selected unit; Step S205: looping steps S201 to S204 until all display units are marked as selected units, the loop ends, and the process proceeds to step S206; Step S206: comparing the discreteness of all units respectively, taking the display unit with the smallest discreteness as the reference unit, and taking all the display units that are not used as the reference unit as the adjustment units.

[0024] In the above step S202, the expression of the discrete degree is: Where, The current unit corresponds to The degree of dispersion of optical parameters, For the unit set The display unit corresponds to The values of the optical parameters, The current unit corresponds to The values of the optical parameters, is the number of displayed units in the unit set, ; In the above step S203, the influencing parameters include application scenarios and content types; application scenarios include commercial display screens, medical imaging diagnosis, security monitoring centers, theater background screens, etc.; content types include videos, graphics, medical images, surveillance images, etc. The influencing parameters are manually input by relevant staff; The method for dynamically adjusting the weight coefficients in the weight set includes: Based on the influencing parameters, corresponding scenario labels and type labels are obtained; wherein the scenario label is a numerical label corresponding to the application scenario, and the type label is a numerical label corresponding to the content type. Different application scenarios have different scenario labels, and different content types have different type labels. The weight set, scenario label, and type label are used as an analysis set, and the analysis set is input into a trained weight adjustment model to obtain an adjustment set, which includes an adjustment amount corresponding to each weight coefficient. Based on each adjustment amount in the adjustment set, the corresponding weight coefficient in the weight set is dynamically adjusted. The weight adjustment model is a deep neural network model, and the training process of the weight adjustment model includes: Pre-collection A different set of analyses Each group analysis set has a corresponding adjustment set. is an integer greater than 1, and the analysis set and the corresponding adjustment set are converted into a corresponding set of feature vectors; the adjustment set corresponding to the analysis set is collected by those skilled in the art in the process of historical dynamic adjustment of the weight set. Group analysis sets, analyze each analysis set in turn according to the actual situation, evaluate the adjustment set corresponding to each analysis set, and The group analysis sets are set up in turn with the corresponding adjustment sets; Each set of feature vectors is used as input to a weight adjustment model. The weight adjustment model outputs a set of predicted adjustment sets corresponding to each analysis set, and uses the actual adjustment set corresponding to each analysis set as the prediction target. The actual adjustment set is a pre-set adjustment set corresponding to the analysis set. The training objective is to minimize the sum of the prediction errors of all analysis sets. The prediction error is calculated as follows: ,in is the prediction error, is the group number of the corresponding eigenvector of the analysis set, For the The corresponding group analysis set A predicted adjustment amount, For the The corresponding group analysis set The actual adjustment amount, is the number of weight coefficients in the weight set; the weight adjustment model is trained until the sum of the prediction errors reaches convergence and the training is stopped.

[0025] It should be noted that the reason for dynamically adjusting the weight coefficient based on the influencing parameters is that different scenarios have different focuses on display quality. For example, medical imaging emphasizes the accuracy of brightness and color to ensure diagnostic reliability, while commercial displays pay more attention to color temperature and chromaticity to enhance visual appeal. Different content types also have different sensitive dimensions to image presentation. For example, dynamic images rely more on color continuity, while graphic content requires high brightness and clear contrast. Therefore, by dynamically adjusting the weights of optical parameters such as brightness, chromaticity, and color temperature, the spliced multi-screen display system can adapt to actual usage needs, ensuring the best visual consistency and display effects under various usage conditions, thereby improving the overall viewing experience and functional performance.

[0026] Step S3: Based on the optical parameters of the reference unit, dynamically modify the optical parameters of each adjustment unit and identify the seam area in each display unit.

[0027] Taking the optical parameters of the reference unit as the reference parameters, the optical parameters of each adjustment unit are dynamically corrected to the reference parameters; The method for identifying the seam area within each display unit includes: Obtain a display image corresponding to each display unit. The display image is obtained by frontally photographing the spliced multi-screen display system using a high-resolution camera array set by a person skilled in the art. The high-resolution camera corresponds one-to-one with the display unit to ensure the accuracy and completeness of the display image. Each display image is grayscaled to obtain a grayscale image. Grayscale processing is a prior art, and the specific processing process will not be described in detail here. A gradient operator (such as a Sobel operator, a Prewitt operator, a Laplacian operator, etc.) is used to process each grayscale image in turn to calculate the horizontal gradient and the vertical gradient. The square root of the sum of the squares of the horizontal gradient and the vertical gradient corresponding to each grayscale image is taken to obtain a gradient amplitude map corresponding to each grayscale image. The gradient amplitude map is a matrix of the same size as the grayscale image, that is, the elements in the gradient amplitude map correspond one-to-one to the pixels in the grayscale image, and the value of each element corresponds to the gradient amplitude of the corresponding pixel in the grayscale image. In each gradient amplitude map, the gradient amplitude of each row is summed in the vertical direction to obtain the total row gradient amplitude, and the gradient amplitude of each column is summed in the horizontal direction to obtain the total column gradient amplitude; the maximum value of the total gradient amplitude of all rows is obtained, and the corresponding row is used as the segmentation row of the corresponding grayscale image; the maximum value of the total gradient amplitude of all columns is obtained, and the corresponding column is used as the segmentation column of the corresponding grayscale image; the method for obtaining the maximum value is a prior art, and the specific process is not described in detail here; according to each segmentation row or segmentation column, the grayscale image is divided into two regions in turn; that is, the grayscale image is divided into two regions, upper and lower, with a segmentation row as the boundary, or the grayscale image is divided into two regions, left and right, with a segmentation column as the boundary; the number of pixels in each region is counted respectively and marked as the number of pixels; the number of pixels in the two divided regions each time is compared, and the region with a smaller number of pixels is marked as a candidate stitching region, and the region with a larger number of pixels is marked as a non-stitching region; The optical parameters of each candidate seam area and non-seam area are collected, and the optical parameters of each candidate seam area and the optical parameters of the corresponding non-seam area are used as a set of parameters; each set of parameters is input into a trained similarity analysis model to predict the corresponding similarity; a similarity threshold is preset, and the similarity threshold is pre-set by a person skilled in the art according to actual conditions; each similarity is compared with the similarity threshold, and the candidate seam areas with similarity less than the similarity threshold are marked as seam areas, and the candidate seam areas with similarity greater than or equal to the similarity threshold are not marked; the similarity analysis model is a deep neural network model, and the specific training process of the similarity analysis model is consistent with the training process of the weight adjustment model.

[0028] It should be understood that although all display units can be made consistent in various optical parameters by dynamically correcting the optical parameters of each adjustment unit, due to local inconsistency problems such as brightness attenuation, structural occlusion, frame reflection, viewing angle difference, and uneven pixel arrangement in the edge area of the display unit, there may still be splicing areas with uneven brightness, sudden color difference and other problems at the joints of adjacent display units, resulting in a visual sense of fault or gap.

[0029] Step S4: Using the edge fusion algorithm, perform pixel-level weight gradient processing on each seam area, obtain the gradient weight of each pixel in the seam area, and adjust the optical parameters of each seam area in real time based on the gradient weight.

[0030] The methods for obtaining the gradient weight of each pixel in the patchwork area include: In each grayscale image, the overlapping area of two stitching areas is marked as overlapping area, and the stitching areas not marked as overlapping areas are marked as separate areas; connectivity analysis is performed on all overlapping areas, and adjacent overlapping areas are merged into a whole area; connectivity analysis is performed on all separate areas, and adjacent separate areas are merged into a joint area; the joint area and the whole area are collectively referred to as analysis areas, and the index of each pixel point in each analysis area is obtained, and the index includes row index and column index; wherein, the minimum row index and the minimum column index of each analysis area are both 0, and the pixel point corresponding to the minimum row index and the minimum column index is the pixel point in the upper left corner of each analysis area; the row index of each pixel point is divided by the row index with the largest value in the corresponding analysis area to obtain the row weight; the column index of each pixel point is divided by the column index with the largest value in the corresponding analysis area to obtain the column weight; the row weight and column weight of each pixel point are used as the corresponding gradient weight.

[0031] The method for adjusting the optical parameters of each seam area in real time includes: Re-collecting the optical parameters of each display unit and marking them as re-collected parameters; marking the optical parameters of each seam area as seam parameters, and dividing each set of seam parameters by the corresponding re-collected parameters to obtain a correction weight set, wherein the correction weight set includes a correction weight corresponding to each parameter in the optical parameters; multiplying the optical parameters of each display unit by the correction weight set corresponding to the corresponding seam area to obtain the correction parameters of each display unit; Subtract the row weight of each pixel from 1 to obtain the reverse row weight of each pixel; subtract the column weight of each pixel from 1 to obtain the reverse column weight of each pixel; use the reverse row weight and reverse column weight of each pixel as the corresponding reverse gradient weight; For each overall area, the display units corresponding to the same overlapping areas of the overall area are merged to obtain an overall unit; based on the gradient weight and reverse gradient weight corresponding to each pixel and the correction parameter of the overall unit, a first adjustment parameter corresponding to each pixel is calculated, and the optical parameter of each pixel is adjusted in real time to the corresponding first adjustment parameter; The expression of the first adjustment parameter is: ; Where, is the first adjustment parameter, is the row weight, is the reverse row weight, is the column weight, is the reverse column weight, is the first parameter, is the second parameter, is the third parameter, is the fourth parameter; wherein the first parameter is the correction parameter of the display unit located in the upper left corner of the entire unit, the second parameter is the correction parameter of the display unit located in the upper right corner of the entire unit, the third parameter is the correction parameter of the display unit located in the lower left corner of the entire unit, and the fourth parameter is the correction parameter of the display unit located in the lower right corner of the entire unit; For each joint area, display units corresponding to the same individual areas as the joint area are obtained and merged to obtain a merged unit; an entire area adjacent to the joint area is obtained and merged with the merged unit to obtain an extended unit; a second adjustment parameter corresponding to each pixel is calculated based on the gradient weight and reverse gradient weight corresponding to each pixel and the correction parameter of the extended unit, and the optical parameter of each pixel is adjusted in real time to the corresponding second adjustment parameter; The expression of the second adjustment parameter is: ; Where, is the second adjustment parameter, is the fifth parameter, is the sixth parameter, is the seventh parameter, is the eighth parameter; wherein, the fifth parameter is the correction parameter of the display unit located on the left side of the expansion unit, the sixth parameter is the correction parameter of the display unit located on the right side of the expansion unit, the seventh parameter is the third adjustment parameter of the overall area located above the expansion unit, and the eighth parameter is the third adjustment parameter of the overall area located below the expansion unit; the third adjustment parameter of the overall area is the average value of the first adjustment parameters of all pixels in the overall area.

[0032] Step S5: Divide all display units into multiple perception areas according to the visual attention model, configure differentiated perception weights for different perception areas based on a pre-built human eye perception weight function, and dynamically optimize the optical parameters of different perception areas based on the perception weights.

[0033] Methods for dividing all display units into multiple sensing areas include: A visual attention model is used to generate an attention distribution map, which includes the attention level of each pixel in each display unit. The attention level is the degree to which different pixels should theoretically be paid attention to. The visual attention model is a visual field center model, which is constructed based on the perceptual characteristics of the human eye's gaze area attenuating with spatial distance. The visual field center model is an existing technology and its specific content will not be elaborated on here. An eye tracker is used to sample the gazes of multiple users in different application scenarios to obtain an eye movement heat map for characterizing actual visual attention behavior. The eye movement heat map includes the attention level of each pixel in each display unit. The attention level is used to reflect the frequency and duration of users' gazes at different pixels, representing the user's visual focus intensity during actual use. Based on the influencing parameters, a scaling factor is dynamically set; the scaling factor is subtracted from 1 to obtain the inverse factor; the product of the attention level of each pixel and the scaling factor is added to the product of the corresponding attention level and the inverse factor to obtain the fusion attention level of each pixel; a perception threshold is preset, which includes a center threshold and an edge threshold, and the center threshold is greater than the edge threshold. The perception threshold is pre-set by those skilled in the art based on actual conditions; The fusion attention of each pixel point is compared with the perception threshold; if the fusion attention degree is greater than or equal to the center threshold, the corresponding pixel point is marked as the center point; if the fusion attention degree is less than the center threshold and greater than the edge threshold, the corresponding pixel point is marked as the transition point; if the fusion attention degree is less than or equal to the edge threshold, the corresponding pixel point is marked as the edge point; the connected domain analysis is performed on the center points, transition points and edge points in all display units, and all display units are divided into the central visual area, the middle transition area and the edge perception area.

[0034] It should be noted that the reason for dynamically setting the scaling factor based on the influencing parameters is that the value of the scaling factor varies significantly under different application scenarios and content types. For example, for different application scenarios, control rooms or security monitoring scenarios rely more on the user's actual gaze behavior, so the scaling factor should be smaller. Exhibition displays or advertising retail interactive screens focus more on content layout and visual guidance, and the theoretical model is more dominant, so the scaling factor should be larger. For different content types, if the content type is a static image or document information with a clear structure, the human eye gaze path is highly regular, and the theoretical model has a strong predictive ability, so the scaling factor should be larger. On the contrary, for dynamic images or video content, the user's line of sight changes frequently, the uncertainty increases, and it should rely more on actual eye movements, so the scaling factor should be smaller.

[0035] Methods for dynamically setting the scale factor include: Reset different digital labels for different application scenarios and content types, and mark the reset digital labels of application scenarios as application labels, and mark the reset digital labels of content types as content labels; wherein both application labels and content labels are set according to the corresponding scale factors of different application scenarios and content types, that is, the larger the corresponding scale factors, the larger the application labels and content labels, and vice versa; Using application tags and content tags as tag data, construct multiple fuzzy sets for each tag in the tag data; for example, the fuzzy sets corresponding to application tags are high-scale factor scenarios, medium-scale factor scenarios, low-scale factor scenarios, etc.; the fuzzy sets corresponding to content tags are high-scale factor content, medium-scale factor content, low-scale factor content, etc.; the tag data are converted into the membership of each corresponding fuzzy set through fuzzification technology; fuzzification technology is the process of converting precise numerical values into the membership corresponding to fuzzy sets, such as triangular membership function and trapezoidal membership function; for example, if the numerical value of the application tag is high, it is inferred that the membership of the high-scale factor scenario is 0.9, the membership of the medium-scale factor scenario is 0.3, and the membership of the low-scale factor scenario is 0; All memberships are input into the trained membership analysis model to predict the corresponding proportion set; the membership analysis model is a deep neural network model, and the specific training process of the membership analysis model is consistent with the training process of the weight adjustment model; the proportion set includes the membership of each scale factor level, such as high, medium, and low; set the proportion interval , divide the ratio interval evenly into Level ranges, is the number of scale factor levels; the average of the maximum and minimum values of each level interval is taken as the corresponding interval mean; according to the interval mean, the membership in the scale set is weighted and summed to obtain the total value of the scale; the membership in the scale set is added in sequence to obtain the total value of the membership; the quotient of the total value of the scale and the total value of the membership is taken as the scale factor.

[0036] The steps of performing connected domain analysis on the center points, transition points, and edge points in all display units include: Step S501: Pixels with the same perception type are grouped as a perception set, where the perception types include center points, transition points, and edge points. Step S502: randomly select a set of perception sets and mark them as the current set; Step S503: randomly select a pixel point that is not marked as a selected point from the current set and mark it as the current point; Step S504: Determine whether there are any pixels adjacent to the current point in all point sets; if there is only one pixel in the point set adjacent to the current point, add the current point to the corresponding point set; if there are pixels in multiple point sets adjacent to the current point, merge the multiple point sets into one point set and add the current point to the merged point set; if there are no pixels in the point set adjacent to the current point, create a new point set and add the current point to the new point set; Step S505: Mark the current point as a selected point; Step S506: looping steps S503 to S505 until all pixels in the current set are marked as selected points, then the loop ends and proceeds to step S507; Step S507: looping steps S502 to S506 until all the perception sets are marked as current sets, then the loop ends and proceeds to step S508; Step S508: Analyze all pixel points in each point set in turn; if all pixel points are center points, the corresponding point set is used as the central visual area; if all pixel points are transition points, the corresponding point set is used as the intermediate transition area; if all pixel points are edge points, the corresponding point set is used as the edge perception area.

[0037] The human eye perception weight function is pre-set by technical personnel in this field in combination with the visual attention model and the eye movement heat map, and is used to differentiate the perception weights of different perception areas, thereby guiding the perception-driven regional image optimization strategy.

[0038] Methods for dynamically optimizing optical parameters of different perception areas include: According to the perception weight, an optimization algorithm (such as genetic algorithm, particle swarm optimization algorithm, simulated annealing algorithm, etc.) is used to generate corresponding optimization strategies for different perception areas. Based on the optimization strategy, the optical parameters of different perception areas are dynamically optimized to achieve perception-driven differentiated image adjustment. For example, the optimization strategy is as follows: the optical parameters of the central visual area remain unchanged, the optical parameters of the middle transition zone are reduced to 85% to 90% of the original value, and the optical parameters of the edge perception area are reduced to 70% to 80% of the original value.

[0039] It should be noted that the purpose of dynamic optimization of different perception areas based on perception weights is to make full use of the visual perception characteristics of the human eye and concentrate limited system resources on the user's most sensitive visual area, thereby improving the energy efficiency and performance of the spliced multi-screen display system while ensuring user experience, and realizing intelligent allocation of display resources and perception-driven differentiated optimization.

[0040] This embodiment collects and analyzes the optical parameters of each display unit in real time, dynamically divides the display units into reference units and adjustment units, and dynamically corrects each adjustment unit based on the optical parameters of the reference unit, effectively solving the problem of inconsistent optical parameters caused by manufacturing differences, changes in the usage environment, etc.; adopts an edge fusion algorithm to perform pixel-level weight gradient processing on the splicing area, effectively eliminating the visual faults and obvious gaps in the splicing area; introduces a visual attention model and a human eye perception weight function, divides the entire display area into a central visual area, an intermediate transition area, and an edge perception area, and performs differentiated optical parameter optimization based on the perception characteristics of different areas to achieve perception-driven intelligent adjustment, effectively improving the display quality and user experience of the spliced multi-screen display system; achieves highly reliable and intelligent multi-screen seamless splicing optical consistency adjustment, ensuring the best visual experience in various application scenarios, thereby significantly enhancing the user's overall viewing effect.

[0041] Example 2

[0042] This application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories may store computer-readable code that, when executed by the one or more processors, may implement the aforementioned method for adjusting optical consistency of a glass-based display for seamless multi-screen splicing.

[0043] The method or system according to the embodiment of the present application can also be implemented with the aid of the architecture of the electronic device shown in this application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, and the like. A storage device in the electronic device, such as a ROM or a hard disk, can store the optical consistency adjustment method for a multi-screen seamlessly spliced glass-based display provided in this application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in this application is only exemplary. When implementing different devices, one or more components in the electronic device shown in this application may be omitted according to actual needs.

[0044] Example 3

[0045] As shown, one embodiment of the present application discloses a computer-readable storage medium. Computer-readable instructions are stored on the computer-readable storage medium. When the computer-readable instructions are executed by a processor, the method for adjusting the optical consistency of a multi-screen seamlessly spliced glass-based display according to the embodiment of the present application described with reference to the above figures can be executed. The storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory. Non-volatile memory may include, for example, read-only memory (ROM), a hard disk, a flash memory, etc.

[0046] Furthermore, according to embodiments of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the present application provides a non-transitory machine-readable storage medium storing machine-readable instructions capable of being executed by a processor to perform the instructions corresponding to the steps of the method provided herein, such as a method for adjusting the optical consistency of a glass-based display for seamless multi-screen splicing. When executed by a central processing unit (CPU), this computer program performs the functions defined in the method of the present application.

[0047] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art will be able to modify the technical solutions described in the foregoing embodiments or to substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

[0048] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or apparatus comprising the element.

[0049] In the description of the present invention, it should be understood that the terms "first", "second", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0050] In the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0051] In the description of the present invention, “several” means one or more, and “a large number” means two or more.

[0052] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0053] The formulas in this manual are all dimensionless and calculated using numerical values. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters and thresholds in the formulas are set by technicians in this field based on actual conditions.

[0054] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the claims and their equivalents.

Claims

1. A method for adjusting optical consistency of glass-based displays with seamless multi-screen splicing, characterized in that: include: Step S1: collecting the optical parameters corresponding to each display unit in real time; Step S2: Analyze the optical parameters of each display unit and divide the display unit into a reference unit and an adjustment unit based on the minimum variance principle; Step S3: Based on the optical parameters of the reference unit, dynamically modify the optical parameters of each adjustment unit and identify the seam area in each display unit; Step S4: Using an edge fusion algorithm, perform pixel-level weight gradient processing on each seam area, obtain the gradient weight of each pixel in the seam area, and adjust the optical parameters of each seam area in real time based on the gradient weight; Step S5: Divide all display units into multiple perception areas according to the visual attention model, configure differentiated perception weights for different perception areas based on a pre-built human eye perception weight function, and dynamically optimize the optical parameters of different perception areas based on the perception weights.

2. The method for adjusting optical consistency of glass-based displays with seamless multi-screen splicing according to claim 1, characterized in that: The steps of dividing the display unit into the reference unit and the adjustment unit include: Step S201: randomly selecting a display unit that is not marked as a selected unit and marking it as the current unit, and taking all the remaining display units as a unit set; Step S202: Calculating a discrete set between the current unit and the unit set based on the optical parameters of each display unit, where the discrete set includes a discrete degree corresponding to each parameter in the optical parameters; Step S203: Preset a weight set, which includes a weight coefficient corresponding to each optical parameter; collect influencing parameters, and dynamically adjust the weight coefficients in the weight set based on the influencing parameters; Step S204: Based on the dynamically adjusted weight set, weighted summation is performed on the discrete degrees in the discrete set to obtain the unit discrete degree, and the current unit is marked as a selected unit; Step S205: looping steps S201 to S204 until all display units are marked as selected units, the loop ends, and the process proceeds to step S206; Step S206: comparing the discreteness of all units respectively, taking the display unit with the smallest discreteness as the reference unit, and taking all the display units that are not used as the reference unit as the adjustment units.

3. The method for adjusting optical consistency of glass-based displays with seamless multi-screen splicing according to claim 2, characterized in that: In step S203, the influencing parameters include application scenario and content type; The method for dynamically adjusting the weight coefficients in the weight set includes: Based on the influencing parameters, the corresponding scene labels and type labels are obtained; the weight set, scene labels, and type labels are used as an analysis set, and the analysis set is input into the trained weight adjustment model to obtain an adjustment set, which includes an adjustment amount corresponding to each weight coefficient; based on each adjustment amount in the adjustment set, the corresponding weight coefficient in the weight set is dynamically adjusted; the training process of the weight adjustment model includes: Pre-collection A different set of analyses Each group analysis set has a corresponding adjustment set. is an integer greater than 1, converting the analysis set and the corresponding adjustment set into a corresponding set of feature vectors; Each set of feature vectors is used as the input of the weight adjustment model, and the weight adjustment model takes a set of predicted adjustment sets corresponding to each set of analysis sets as output, and an actual adjustment set corresponding to each set of analysis sets as the prediction target, where the actual adjustment set is a pre-set adjustment set corresponding to the analysis set; minimizing the sum of the prediction errors of all analysis sets is used as the training goal; the weight adjustment model is trained until the sum of the prediction errors reaches convergence and the training is stopped.

4. The method for adjusting optical consistency of glass-based displays with seamless multi-screen splicing according to claim 3, characterized in that: Taking the optical parameters of the reference unit as the reference parameters, the optical parameters of each adjustment unit are dynamically corrected to the reference parameters; The method for identifying the seam area within each display unit includes: Obtaining a display image corresponding to each display unit, performing grayscale processing on each display image to obtain a grayscale image; processing each grayscale image in turn using a gradient operator to calculate a horizontal gradient and a vertical gradient; taking the square root of the sum of the squares of the horizontal gradient and the vertical gradient corresponding to each grayscale image to obtain a gradient amplitude map corresponding to each grayscale image; In each gradient magnitude map, the gradient magnitudes of each row are summed vertically to obtain the total magnitude of the row gradients, and the gradient magnitudes of each column are summed horizontally to obtain the total magnitude of the column gradients; the maximum value of the total magnitudes of all row gradients is obtained, and the corresponding row is used as the segmentation row of the corresponding grayscale image; the maximum value of the total magnitudes of all column gradients is obtained, and the corresponding column is used as the segmentation column of the corresponding grayscale image; the grayscale image is divided into two regions according to each segmentation row or segmentation column; the number of pixels in each region is counted and marked as the number of pixels; the number of pixels in the two divided regions is compared each time, and the region with the smaller number of pixels is marked as the candidate stitching region, and the region with the larger number of pixels is marked as the non-stitching region; The optical parameters of each candidate seam area and non-seam area are collected, and the optical parameters of each candidate seam area and the optical parameters of the corresponding non-seam area are taken as a set of parameters; each set of parameters is input into the trained similarity analysis model to predict the corresponding similarity; a similarity threshold is preset, and each similarity is compared with the similarity threshold. The candidate seam areas with similarity less than the similarity threshold are marked as seam areas, and the candidate seam areas with similarity greater than or equal to the similarity threshold are not marked.

5. The method for adjusting optical consistency of glass-based displays with seamless multi-screen splicing according to claim 4, characterized in that: The methods for obtaining the gradient weight of each pixel in the seam area include: In each grayscale image, the overlapping area of two stitching regions is marked as the overlapping area, and the stitching region not marked as the overlapping area is marked as the separate area; connectivity analysis is performed on all overlapping areas, and adjacent overlapping areas are merged into the whole area; connectivity analysis is performed on all separate areas, and adjacent separate areas are merged into the joint area; The joint area and the overall area are collectively referred to as the analysis area. The index of each pixel point in each analysis area is obtained, and the index includes row index and column index. The row index of each pixel point is divided by the row index with the largest value in the corresponding analysis area to obtain the row weight. The column index of each pixel point is divided by the column index with the largest value in the corresponding analysis area to obtain the column weight. The row weight and column weight of each pixel point are used as the corresponding gradient weight.

6. The method for adjusting optical consistency of glass-based displays with seamless multi-screen splicing according to claim 5, characterized in that: The method for adjusting the optical parameters of each seam area in real time includes: Re-collecting the optical parameters of each display unit and marking them as re-collected parameters; marking the optical parameters of each seam area as seam parameters, and dividing each set of seam parameters by the corresponding re-collected parameters to obtain a correction weight set, wherein the correction weight set includes a correction weight corresponding to each parameter in the optical parameters; multiplying the optical parameters of each display unit by the correction weight set corresponding to the corresponding seam area to obtain the correction parameters of each display unit; Subtract the row weight of each pixel from 1 to obtain the reverse row weight of each pixel; subtract the column weight of each pixel from 1 to obtain the reverse column weight of each pixel; use the reverse row weight and reverse column weight of each pixel as the corresponding reverse gradient weight; For each overall area, the display units corresponding to the same overlapping areas of the overall area are merged to obtain an overall unit; based on the gradient weight and reverse gradient weight corresponding to each pixel and the correction parameter of the overall unit, a first adjustment parameter corresponding to each pixel is calculated, and the optical parameter of each pixel is adjusted in real time to the corresponding first adjustment parameter; For each joint area, the display units corresponding to the separate areas with the same corresponding joint area are obtained and merged to obtain a merged unit; the overall area adjacent to the joint area is obtained and merged with the merged unit to obtain an extended unit; according to the gradient weight, reverse gradient weight and correction parameter of the extended unit corresponding to each pixel point, the second adjustment parameter corresponding to each pixel point is calculated, and the optical parameter of each pixel point is adjusted to the corresponding second adjustment parameter in real time.

7. The method for adjusting optical consistency of glass-based displays with seamless multi-screen splicing according to claim 6, wherein: Methods for dividing all display units into multiple sensing areas include: A visual attention model is used to generate an attention distribution map, which includes the attention degree of each pixel in each display unit; wherein the visual attention model is a visual field center model; an eye movement heat map is obtained, which includes the attention degree of each pixel in each display unit; Based on the influencing parameters, the scaling factor is dynamically set; the scaling factor is subtracted from 1 to obtain the inverse factor; the product of the attention degree of each pixel and the scaling factor is added to the product of the corresponding attention degree and the inverse factor to obtain the fusion attention degree of each pixel; the perception threshold is preset, and the perception threshold includes the center threshold and the edge threshold, and the center threshold is greater than the edge threshold; The fusion attention of each pixel is compared with the perception threshold; if the fusion attention is greater than or equal to the center threshold, the corresponding pixel is marked as the center point; if the fusion attention is less than the center threshold and greater than the edge threshold, the corresponding pixel is marked as a transition point; if the fusion attention is less than or equal to the edge threshold, the corresponding pixel is marked as an edge point; Connected domain analysis is performed on the center points, transition points and edge points in all display units, and all display units are divided into a central visual area, an intermediate transition area and an edge perception area.

8. The method for adjusting optical consistency of glass-based displays with seamless multi-screen splicing according to claim 7, characterized in that: Methods for dynamically setting the scale factor include: Resetting different digital labels for different application scenarios and content types, and marking the reset digital labels of application scenarios as application labels, and marking the reset digital labels of content types as content labels; wherein both application labels and content labels are set according to the size of the corresponding scale factors of different application scenarios and content types; Using application tags and content tags as tag data, construct multiple fuzzy sets for each tag in the tag data; convert the tag data into the membership degree of each corresponding fuzzy set through fuzzification technology; All memberships are input into the trained membership analysis model to predict the corresponding proportion set; the membership analysis model is a deep neural network model, and the proportion set includes the membership of each scale factor level; set the proportion interval , divide the ratio interval evenly into Level ranges, is the number of scale factor levels; the average of the maximum and minimum values of each level interval is taken as the corresponding interval mean; according to the interval mean, the membership in the scale set is weighted and summed to obtain the total value of the scale; the membership in the scale set is added in sequence to obtain the total value of the membership; the quotient of the total value of the scale and the total value of the membership is taken as the scale factor.

9. The method for adjusting optical consistency of glass-based displays with seamless multi-screen splicing according to claim 8, characterized in that: The steps of performing connected domain analysis on the center points, transition points, and edge points in all display units include: Step S501: Pixels with the same perception type are grouped as a perception set, where the perception types include center points, transition points, and edge points. Step S502: randomly select a set of perception sets and mark them as the current set; Step S503: randomly select a pixel point that is not marked as a selected point from the current set and mark it as the current point; Step S504: Determine whether there are any pixels adjacent to the current point in all point sets; if there is only one pixel in the point set adjacent to the current point, add the current point to the corresponding point set; if there are pixels in multiple point sets adjacent to the current point, merge the multiple point sets into one point set and add the current point to the merged point set; if there are no pixels in the point set adjacent to the current point, create a new point set and add the current point to the new point set; Step S505: Mark the current point as a selected point; Step S506: looping steps S503 to S505 until all pixels in the current set are marked as selected points, then the loop ends and proceeds to step S507; Step S507: looping steps S502 to S506 until all the perception sets are marked as current sets, then the loop ends and proceeds to step S508; Step S508: Analyze all pixel points in each point set in turn; if all pixel points are center points, the corresponding point set is used as the central visual area; if all pixel points are transition points, the corresponding point set is used as the intermediate transition area; if all pixel points are edge points, the corresponding point set is used as the edge perception area.

10. The method for adjusting optical consistency of glass-based displays with seamless multi-screen splicing according to claim 9, characterized in that: Methods for dynamically optimizing optical parameters of different perception areas include: According to the perception weight, an optimization algorithm is used to generate corresponding optimization strategies for different perception areas, and the optical parameters of different perception areas are dynamically optimized based on the optimization strategies.

Images