Pixel arrangement and driving method for special-shaped curved surface LED creative display screen

Abstract

This invention relates to the field of pixel driving technology, and more particularly to a pixel arrangement and driving method for irregularly curved LED creative displays. The method includes: determining the curvature deviation factor based on the normalized radial displacement deviation between each pixel of the target display and a standard reference screen to determine the required correction state of the target display; obtaining visual sensitivity weights and combining them with local pixel stretching factors to determine a weighted pixel stretching risk value to locate the left and right index boundaries and distinguish the region type of the target display; determining the final driving brightness of each pixel based on the smoothed brightness compensation coefficient and the final texture coordinates, and obtaining the local change characteristics of the final texture coordinates and combining them with the ideal reference gradient to determine the factor threshold for correcting the curvature deviation factor based on the boundary gradient anomaly. This invention achieves adaptive distortion-free driving for curved displays with different curvatures by diverting the curvature deviation factor and combining it with visual sensitivity-weighted regional progressive mapping and brightness compensation.

Description

Technical Field

[0001] This invention relates to the field of pixel driving technology, and in particular to a pixel arrangement and driving method for irregular curved LED creative displays. Background Technology

[0002] As LED display technology continues to expand into the creative display field, curved, wavy, and other irregularly shaped LED displays are increasingly used in stage performances, commercial spaces, and building facades. However, traditional LED driving methods are mainly designed for flat rectangular screens. Their core assumption is that physical pixels are arranged in a regular matrix and correspond one-to-one with the video source pixels, which makes it difficult to directly adapt to the geometric characteristics of curved screens.

[0003] For curved LED displays, the common approach is to approximate the curved screen as a flat surface, or to perform a one-time mapping calibration during production for a specific radius of curvature. The former approximation leads to significant stretching or compression distortion at the screen edges, severely impacting the geometric accuracy of content presentation. The latter lacks adaptability to curved screens with different curvatures; each change in curvature requires redeveloping and reprogramming the driver mapping parameters, resulting in high engineering adaptation costs and poor versatility. Furthermore, the uneven pixel density caused by geometric mapping on curved screens can lead to regional deviations in display brightness, and existing solutions rarely address the coordinated processing of geometric correction and brightness compensation. Methods that partially rely on manual visual calibration or 3D scanning to reconstruct pixel positions are complex and difficult to integrate into the automated operation chain of the driver system. Therefore, there is an urgent need for a curved LED display driver method that can automatically perform curvature deviation assessment, adaptive geometric correction, and brightness compensation based on the screen's own physical parameters, and possess parameter self-optimization capabilities.

[0004] Chinese Patent Publication No. CN118248078A discloses a display screen and a pixel driving method. The screen includes a pixel unit and a pixel driving unit, electrically connected to the pixel unit. The pixel unit includes at least three pixel groups. The pixel driving unit is used to: acquire an instruction to drive the pixel unit to display a full-color image within a display sensor; the display sensor includes at least three sub-sensors; upon receiving the instruction, within each sub-sensor, each pixel group is driven to emit light of a different color; and within different sub-sensors, the same pixel group is driven to emit light of different colors; the different colors of light include at least red, green, and blue light. Therefore, the pixel driving method has the following problems: The system cannot automatically assess curvature deviation based on the physical parameters of the curved LED display and adaptively distribute the correction to different precisions. Instead, it performs a region-by-region progressive adjustment of the stretching distortion in the screen edge area based on visual sensitivity, and simultaneously compensates for the uneven brightness caused by changes in pixel density. Summary of the Invention

[0005] To address this, the present invention provides a pixel arrangement and driving method for irregularly shaped curved LED creative displays, which overcomes the inability of existing technologies to assess the stretching distortion of the screen edge area caused by curvature deviation and to perform progressive regional adjustment based on visual sensitivity, thereby simultaneously compensating for the uneven brightness caused by changes in pixel density.

[0006] To achieve the above objectives, the present invention provides a pixel arrangement and driving method for irregularly shaped curved LED creative displays, comprising: The required correction state of the target display screen is determined based on the curvature deviation factor, wherein the curvature deviation factor is determined by normalizing the radial displacement deviation between each pixel of the target display screen and the standard reference screen. In response to the correction state, the linear mapping relationship between the display pixels and the video source pixels is adjusted, or the reference texture coordinates of each horizontal pixel index are determined based on the arc length coordinates and the ratio of the total arc length of the reference screen luminous area to the standard reference radius of curvature. The mapping deviation of each horizontal pixel is determined based on the reference texture coordinates of the target display pixel and the standard texture coordinates of the standard reference screen. The local pixel stretching factor is determined by combining the cross-texture distance to determine whether the image is stretched or compressed. The texture span distance is determined based on the reference texture coordinates of any pixel and its neighboring pixels, and the local pixel stretching factor is the ratio of the reference texture coordinates to the texture span distance. The visual sensitivity weight is obtained and combined with the local pixel stretching factor to determine the weighted pixel stretching risk value in order to locate the left index boundary and the right index boundary to distinguish the region type of the target display screen. The region type includes the core fidelity area and the edge transition area. In response to the region type being an edge transition region, the corresponding pixel fusion weight is determined by combining the reference texture coordinates and the standard texture coordinates to determine the final texture coordinates, adjusting the mapping relationship between pixels, and determining the smoothed brightness compensation coefficient of each pixel based on the local pixel stretching factor. The final driving brightness of each pixel is determined by combining the smoothed brightness compensation coefficient with the final texture coordinates, and the local change features of the final texture coordinates are obtained and the boundary gradient anomaly is determined by combining the ideal reference gradient, so as to correct the factor threshold of the curvature deviation factor.

[0007] Furthermore, the process of determining the required correction state for the target display screen includes, If the curvature deviation factor is less than or equal to the factor threshold, the target display screen is judged to be in a normal correction state, and the residual distortion of the video source image is within an acceptable range after global compensation. If the curvature deviation factor is greater than the factor threshold, the target display screen is judged to be in a depth correction state. The curvature deviation of the display screen is too large, and the residual pixel stretching or compression after global compensation in the edge area may affect the subjective visual quality.

[0008] Furthermore, a positive curvature deviation factor indicates that the target display screen is flatter than the standard reference screen, and the image will be stretched horizontally; a negative curvature deviation factor indicates that the target display screen is more curved than the standard reference screen, and the image will be compressed horizontally.

[0009] Furthermore, under normal correction conditions, the horizontal texture coordinates corresponding to the horizontal pixel index in the video source image are determined based on the arc length coordinates and the total arc length of the light-emitting area of ​​the reference screen, and the linear mapping relationship between the display screen pixels and the video source pixels is adjusted. Wherein, the cumulative arc length is the sum of several arc length intervals corresponding to the horizontal pixel index, and the horizontal pixel index is a single integer value that sequentially numbers the pixels arranged along the arc direction on the display screen.

[0010] Furthermore, in the depth correction state, if the local pixel stretching factor is greater than one, the image is determined to be stretched; if the local pixel stretching factor is less than one, the image is determined to be compressed.

[0011] Furthermore, the visual sensitivity weight is determined based on the horizontal center offset of the corresponding pixel; The index of the first weighted pixel on the left whose stretching risk value is greater than the stretching risk threshold is denoted as the left index boundary. The index of the first right-hand weighted pixel whose stretching risk value is greater than the stretching risk threshold is denoted as the right-hand index boundary.

[0012] Furthermore, the core fidelity region is the pixel area within the pixel index interval formed by the left index boundary and the right index boundary. The weighted pixel stretching risk value within the core fidelity region is less than or equal to the stretching risk threshold, indicating low stretching risk. The edge transition zone is a pixel region that exceeds the pixel index range. Within the edge transition zone, the weighted pixel stretching risk value is greater than the stretching risk threshold, indicating a high stretching risk. The mapping intensity is then gradually reduced.

[0013] Furthermore, in the core fidelity area, geometric mapping reference texture coordinates are used; In the edge transition region, the proportions of reference texture coordinates and standard texture coordinates in the final texture coordinate composition are determined by the fusion weights.

[0014] Furthermore, the brightness compensation coefficient is the local pixel stretching factor corresponding to each pixel; If the smoothed brightness compensation coefficient is higher than the upper and lower limits of brightness compensation, then the corresponding brightness compensation coefficient is the upper limit of brightness compensation. If the smoothed brightness compensation coefficient is lower than the upper and lower limits of brightness compensation, then the corresponding brightness compensation coefficient is the lower limit of brightness compensation.

[0015] The final driving brightness of each pixel is determined based on the smoothed brightness compensation coefficient and the original color brightness value, which is obtained by sampling from the video source according to the final texture coordinates.

[0016] Furthermore, the local variation features include the left boundary gradient peak and the right boundary gradient peak of the left and right texture coordinates; Obtain the average local gradient normalized to the left and right boundary gradient peaks within the core fidelity region to determine the boundary gradient anomaly. If the boundary gradient anomaly is less than or equal to the smoothness tolerance threshold, then the current transition zone boundary is determined to be smooth. If the boundary gradient anomaly is greater than the smoothness tolerance threshold, it is determined that there is a perceptible abrupt change in the mapping gradient at the boundary of the transition region, and the factor threshold is reduced according to the ratio of the smoothness tolerance threshold to the boundary gradient anomaly.

[0017] Compared with existing technologies, the advantages of this invention are as follows: This invention automatically corrects the curvature deviation factor to achieve adaptive adaptation to curved screens with different curvatures; in the normal correction state, a linear mapping is directly established, which is simple to calculate and has strong real-time performance; in the depth correction state, the reference texture coordinates and stretching factor are accurately derived by using curved surface geometric projection, and the boundaries of the core fidelity area and the edge transition area are located by combining visual sensitivity weighting, and then a progressive mapping transition is established between regions through continuous fusion weights to eliminate the visual abrupt changes in edge stretching distortion; the brightness compensation coefficient is generated in real time by the stretching factor and is smoothed and limited to simultaneously solve the brightness unevenness caused by changes in pixel density; finally, the factor threshold is corrected in a closed loop by the boundary gradient anomaly degree, and the flow accuracy is continuously optimized in long-term operation. The whole solution does not require manual calibration and is fully automated from parameter reading to drive output.

[0018] Furthermore, this invention achieves precise adaptation to curved LED displays with different curvatures through adaptive calculation of the curvature deviation factor δ and a factor threshold diversion mechanism. For screens with small curvature deviations, a linear mapping relationship is directly established, requiring only one division operation in the mapping table generation process, resulting in low computational complexity and high real-time performance. For screens with large curvature deviations, a depth correction state is automatically entered. Through region division based on pixel stretching risk values ​​and a gradual transition using continuous fusion weights, geometric accuracy is maintained in the core area while effectively suppressing the perception of stretching distortion in the edge area, eliminating abrupt boundary changes in the image. This method requires no manual calibration or configuration; the drive system can automatically complete all analysis and adjustment processes after reading the screen's factory parameters, significantly reducing the engineering complexity of adapting to irregularly shaped screens.

[0019] Furthermore, in the depth correction state, this invention achieves point-by-point quantitative evaluation of pixel stretching degree by constructing reference texture coordinates based on curved surface geometric projection and standard texture coordinates based on linear mapping; by introducing visual sensitivity weight and square-form stretching risk value calculation, it can accurately identify subjectively significant distortion locations in the screen edge region and automatically divide the core fidelity zone and the edge transition zone; through continuously changing fusion weight, a progressive transition mechanism is established between the two zones, effectively diluting the visual impact of edge stretching distortion while maintaining the geometric accuracy of the image center, and eliminating abrupt changes in the image at the regional boundary; the brightness compensation coefficient is calculated in real time based on the stretching factor and is smoothed and limited, simultaneously solving the problem of uneven brightness caused by changes in pixel density, ensuring smooth and consistent brightness of the entire screen display.

[0020] Furthermore, by establishing a direct mapping relationship between the pixel stretching factor and the brightness compensation coefficient, automatic correction of brightness unevenness caused by pixel density changes due to geometric mapping is achieved. The brightness compensation coefficient is taken from the real-time calculation result of the stretching factor, requiring no additional measurement or calibration, making the calculation process simple and efficient. By introducing a one-dimensional mean filter to smooth the compensation coefficient, brightness abruptness caused by small fluctuations in the stretching factor between adjacent pixels is effectively eliminated, ensuring a smooth and natural transition of screen brightness. At the same time, compensation upper and lower limits determined by hardware safety specifications are set, constraining the compensation amplitude within the range of [B_min, B_max], preventing the compensation current from exceeding the reliable operating range of the LED beads under extreme curvature, thus ensuring both display effect and hardware lifespan and operational safety. The dual processing mechanism of smoothing and limiting enables brightness compensation to have both response accuracy and output stability.

[0021] Furthermore, by monitoring the local gradient characteristics of the final texture mapping table at the boundary between the core fidelity area and the edge transition area, a closed-loop feedback mechanism for the accuracy of the factor threshold was established. Gradient peaks were extracted by five-point sampling of the boundary neighborhood and first-order numerical differentiation, and normalized based on the average gradient of the core area. The generated boundary gradient anomaly index can objectively quantify the smoothness of the transition area without relying on the statistical correctness judgment of the display effect. When the anomaly exceeds the smoothness tolerance threshold, it is inferred that the factor threshold is too large and it is quantitatively corrected, so that the factor threshold approaches the optimal dividing point for distinguishing between regular correction and deep correction in long-term operation. This avoids the problem of over-processing of screens with moderate curvature deviation and improves the adaptive ability and parameter self-optimization level of the driving method for screens with different curvatures. Attached Figure Description

[0022] Figure 1 This is a flowchart of a method for pixel arrangement and driving of an irregularly shaped curved LED creative display screen in an embodiment of the present invention; Figure 2This is a flowchart illustrating the process of determining the required correction state of the target display screen in an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the process of determining image stretching or compression in an embodiment of the present invention; Figure 4 This is a flowchart illustrating the process of distinguishing the region type of the target display screen in an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0024] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0025] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0026] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0027] Please see Figures 1-4 As shown, Figure 1 This is a flowchart of a method for pixel arrangement and driving of an irregularly shaped curved LED creative display screen in an embodiment of the present invention; Figure 2 This is a flowchart illustrating the process of determining the required correction state of the target display screen in an embodiment of the present invention; Figure 3 This is a schematic diagram illustrating the process of determining image stretching or compression in an embodiment of the present invention; Figure 4 This is a flowchart illustrating the process of distinguishing the region type of the target display screen in an embodiment of the present invention.

[0028] This invention provides a pixel arrangement and driving method for irregularly shaped curved LED creative displays, comprising: Step S1: Determine the required correction state of the target display screen based on the curvature deviation factor, wherein the curvature deviation factor is determined by normalizing the radial displacement deviation between each pixel of the target display screen and the standard reference screen. Step S2: In response to the correction state, adjust the linear mapping relationship between the display screen pixels and the video source pixels, or determine the reference texture coordinates of each horizontal pixel index based on the arc length coordinates and the ratio of the total arc length of the reference screen luminous area to the standard reference radius of curvature. Step S3: Determine the mapping deviation of each horizontal pixel based on the reference texture coordinates of the target display screen pixel and the standard texture coordinates of the standard reference screen, and combine the cross-texture distance to determine the local pixel stretching factor to determine whether the image is stretched or compressed. The texture span distance is determined based on the reference texture coordinates of any pixel and its neighboring pixels, and the local pixel stretching factor is the ratio of the reference texture coordinates to the texture span distance. Step S4: Obtain the visual sensitivity weight and combine it with the local pixel stretching factor to determine the weighted pixel stretching risk value in order to locate the left index boundary and the right index boundary to distinguish the region type of the target display screen. The region type includes the core fidelity area and the edge transition area. Step S5: In response to the region type being edge transition region, determine the pixel corresponding fusion weight, combine the reference texture coordinates and standard texture coordinates to determine the final texture coordinates, adjust the mapping relationship between pixels, and determine the smoothed brightness compensation coefficient of each pixel based on the local pixel stretching factor. Step S6: Based on the smoothed brightness compensation coefficient and the final texture coordinates, determine the final driving brightness of each pixel, and obtain the local change features of the final texture coordinates and combine them with the ideal reference gradient to determine the factor threshold of the boundary gradient anomaly correction curvature deviation factor.

[0029] Specifically, this invention achieves adaptive adaptation to curved screens with different curvatures by automatically correcting the curvature deviation factor. In the normal correction state, a linear mapping is directly established, which is simple to calculate and has strong real-time performance. In the depth correction state, the reference texture coordinates and stretching factor are accurately derived by using curved surface geometric projection. The core fidelity area and the edge transition area boundary are located by combining visual sensitivity weighting. Then, a progressive mapping transition is established between regions through continuous fusion weights to eliminate the visual abrupt changes in edge stretching distortion. The brightness compensation coefficient is generated in real time by the stretching factor and is smoothed and limited to solve the brightness unevenness caused by changes in pixel density. Finally, the factor threshold is corrected in a closed loop by the boundary gradient anomaly degree. The flow accuracy is continuously optimized in long-term operation. The whole solution does not require manual calibration and is fully automated from parameter reading to drive output.

[0030] In this embodiment, the target display screen is a curved LED display screen, which stores factory calibration data of its own inherent physical parameters.

[0031] Based on curvature deviation, global pre-compensation and state splitting are performed. According to the physical parameters of the target curved screen, the geometric deviation between it and the standard reference screen is calculated, a global pre-compensation is performed, and the state of conventional correction or deep correction is determined according to the degree of deviation.

[0032] Specifically, the distance from a pixel at arc length s on the target display screen to the center of the arc is always the target radius of curvature Rt, and the distance from the corresponding pixel on the standard reference screen to its center is always the standard reference radius of curvature Rr.

[0033] The radial displacement deviation ΔR(s) at this point is equal to the difference between the target radius of curvature Rt and the standard reference radius of curvature Rr. The global radial deviation is normalized to the curvature deviation factor δ. The curvature deviation factor δ is determined based on the ratio of the absolute value of the radial displacement deviation ΔR(s) to the standard reference curvature radius Rr.

[0034] It is understood that the curvature deviation factor δ is used to determine whether the curvature of the target curved screen will cause image distortion compared with the standard reference screen; a positive curvature deviation factor δ indicates that the target screen is flatter and the image will be stretched horizontally; a negative curvature deviation factor δ indicates that the target screen is more curved and the image will be compressed horizontally; the radial displacement deviation ΔR(s) directly reflects the degree of separation between the two curved surfaces of the target curved screen and the standard reference screen in the same arc length position space, and the radial deviation is divided by the reference radius to eliminate the dimension of physical scale.

[0035] If the curvature deviation factor δ is less than or equal to the factor threshold, it is determined that the video source image is in a normal correction state, and the residual distortion is within an acceptable range after global compensation. If the curvature deviation factor δ is greater than the factor threshold, it is determined that the system is in a depth correction state. If the deviation is too large, even if the edge region has undergone global compensation, the residual pixel stretching or compression will still seriously affect the subjective visual quality.

[0036] The preferred value range for the factor threshold is 0.05 to 0.15. The rationale is that when δ < 0.05, the curvature difference between the target screen and the reference screen is less than 5%, and the human eye can hardly perceive the image stretching or compression distortion caused by the curvature deviation. If it is still judged as a depth correction state at this time, unnecessary computational overhead will be introduced. When δ > 0.15, the curvature deviation exceeds 15%, and the pixel stretching factor at the screen edge significantly deviates from 1.0. If no regional fusion processing is performed, the subjective visual quality will be perceptibly degraded. Therefore, setting the factor threshold within this range achieves the optimal balance between computational efficiency and display quality.

[0037] Understandably, when the curvature deviation factor δ is less than or equal to the threshold, it means that the curvature of this curved screen is very close to that of the standard reference screen; the horizontal texture coordinate ui is the relative position of the pixel on the entire arc, a number between 0 and 1; for example, on a screen with a total arc length of 1 meter, the arc length of the 500th pixel is 0.5 meters, then its texture coordinate is 0.5 / 1=0.5, which means that the pixel should display the column of pixels in the very center of the video image.

[0038] In this embodiment, the inherent physical parameters of the target curved LED display screen accessed by the driving system are obtained. The inherent physical parameters include the actual radius of curvature Rt of the target screen, in millimeters; the total arc length La of the screen's light-emitting area, in millimeters; the physical height Hs of the screen, in millimeters; the number of physical pixels Npa along the arc direction and the number of physical pixels Nph in the vertical direction.

[0039] Calculate the one-dimensional arc coordinates of the physical pixels of the target display screen. In practice, the pixels are arranged on the arc surface with equal arc lengths. The arc length spacing Δs is determined based on the ratio of the total arc length of the light-emitting area to the value of the number of arc pixels minus one. The value of the number of arc pixels minus one is the difference between the number of physical pixels along the arc direction and one.

[0040] For each horizontal pixel index i, where index i ranges from 0 to the number of arc pixels, calculate its cumulative arc length si, which is calculated from one side of the arc surface. The cumulative arc length si is determined based on the product of the horizontal pixel index i and the arc length spacing Δs. The cumulative arc length si is the unique one-dimensional position identifier of the pixel on the arc surface.

[0041] Calculate the arc coordinates of the corresponding pixel on the standard reference screen, and preset a standard reference radius of curvature Rr.

[0042] The standard reference screen is designed to keep the total arc length La of the light-emitting area and the number of physical pixels along the arc direction unchanged. Therefore, the pixel spacing Δs is the same. Thus, pixels with the same index i on the reference screen also have the same cumulative arc length, that is, the cumulative arc length sir of the standard reference is equal to the cumulative arc length si.

[0043] Calculate the curvature deviation factor δ, using arc length s as the independent variable, and compare the radial displacement deviation ΔR(s) of the two curved surfaces of the target display screen and the standard reference screen at the same arc length s position.

[0044] Global compensation is performed based on the correction state of the target display screen. The curvature deviation is small in the normal correction state, and global compensation can be directly accomplished by establishing a linear mapping relationship from physical pixels to video source pixels.

[0045] Specifically, the total arc length La of the luminous area of ​​the target display screen and the standard reference screen are the same, as are the number of physical pixels along the arc direction. The arc length coordinates si of each physical pixel are determined.

[0046] A direct one-dimensional mapping relationship is established. For a horizontal pixel index i, its corresponding horizontal texture coordinate ui in the video source image is determined based on the ratio of the arc length coordinate si to the total arc length La of the luminous area of ​​the reference screen. It is understandable that the formula linearly maps physical pixels with equal arc length distribution to the normalized coordinates [0,1] interval of the video source image.

[0047] Output a mapping table containing the number of physical pixels along the arc direction. The mapping table includes the ratio of several horizontal pixel indices i to the number of arc pixels i.

[0048] This mapping table serves as the final driving parameter and is directly output to the pixel mapping module to complete pixel-by-pixel color sampling and driving.

[0049] The mapping table is equivalent to the horizontal texture coordinates UI, and is used to determine which coordinate of the video frame corresponds to which physical pixel on the screen.

[0050] Specifically, this invention achieves precise adaptation to curved LED displays with different curvatures through adaptive calculation of the curvature deviation factor δ and a factor threshold diversion mechanism. For screens with small curvature deviations, a linear mapping relationship is directly established, requiring only one division operation in the mapping table generation process, resulting in low computational complexity and high real-time performance. For screens with large curvature deviations, a deep correction state is entered. Through region division based on pixel stretching risk values ​​and a gradual transition using continuous fusion weights, geometric accuracy is maintained in the core area while effectively suppressing the perception of stretching distortion in the edge area, eliminating abrupt boundary changes in the image. This method requires no manual calibration or configuration; the drive system can automatically complete all analysis and adjustment processes after reading the screen's factory parameters, significantly reducing the engineering complexity of adapting to irregularly shaped screens.

[0051] In deep correction mode, the degree of pixel stretching in different areas of the screen is quantified to generate a continuously changing set of driving parameters, enabling fine-grained adjustment by region and pixel.

[0052] Specifically, the local pixel stretching factor is determined based on the reference texture distance di and the cross-texture distance dr(i), where the local pixel stretching factor is the ratio of the cross-texture distance dr(i) to the reference texture distance di.

[0053] If the local pixel stretching factor is greater than one, it is determined that the physical pixel at that location needs to display more video content than normal, and the image is stretched. If the local pixel stretching factor is less than one, the image is judged to be compressed, and the visual sensitivity weight Wi is calculated based on the horizontal center offset Di.

[0054] Understandably, the local pixel stretching factor is used to quantify the degree of stretching or compression that screen curvature brings to each specific pixel position. The reference texture distance di is the distance that adjacent pixels cross on the video source screen when the standard reference screen uses a simple linear mapping. It represents how much video screen span should correspond to one pixel interval when there is no distortion. The cross texture distance dr(i) is the actual distance that pixel i and pixel i-1 cross on the video source when the target screen is mapped according to precise geometric projection. Since the reference texture coordinates are projected using the arctangent function, the rate of change is different at different positions on the screen. Therefore, the cross texture distance dr(i) is no longer constant and there will be obvious differences at the edge and center of the screen.

[0055] In practice, each horizontal pixel index i is traversed, and the reference texture coordinate ur(i) of the target display pixel is calculated. The reference texture coordinate ur(i) is the horizontal coordinate of the video source that the physical pixel of the target screen should point to after geometric projection according to the standard reference curvature radius Rr of the reference screen.

[0056] Specifically, for a circular arc surface, the central angle θi of the pixel with arc length coordinate si on the screen is the ratio of arc length coordinate si to the standard reference radius of curvature Rr. The horizontal position of a pixel with arc length coordinate si projected onto the planar video source is proportional to tan(θi). In order to map to the [0,1] interval, normalization is required. Therefore, the reference texture coordinate ur(i) is the ratio of the tangent of the central angle θi to the tangent of the reference angle, where the reference angle is the ratio of the total arc length La of the light-emitting area of ​​the reference screen to the standard reference radius of curvature Rr. Calculate the standard texture coordinates ul(i) of the standard reference screen itself. The standard texture coordinates ul(i) are the coordinates of the reference screen itself when using linear mapping. The standard texture coordinates ul(i) are the ratio of the arc length coordinates si to the total arc length La of the light-emitting area of ​​the reference screen.

[0057] The mapping deviation Δu(i) is determined based on the reference texture coordinates ur(i) of the target display pixel and the standard texture coordinates ul(i) of the standard reference screen. The mapping deviation Δu(i) is the difference between the reference texture coordinates ur(i) and the standard texture coordinates ul(i) of the standard reference screen. If the target screen is mapped according to the reference texture coordinates ur(i), the texture distance dr(i) that adjacent pixels cross on the video source is the difference between the reference texture coordinates ur(i) of the current pixel and the reference texture coordinates ur(i-1) of the adjacent pixel. For i=0, dr(0) is defined as equal to dr(1). Under the linear mapping of the reference screen, the texture distance between adjacent pixels is a constant value, and the reference texture distance di is the ratio of 1 to the number of arc pixels minus one.

[0058] In this embodiment, when the local pixel stretching factor is less than one, and the image is determined to be compressed, the visual sensitivity weight Wi is calculated based on the horizontal center offset Di. The visual sensitivity weight Wi is determined based on one minus the square of the horizontal center offset Di. The horizontal center offset Di is determined based on the absolute value of the difference between twice and one of the center distance index, where the center distance index is the ratio of the horizontal pixel index i to the value of the number of arc pixels.

[0059] Specifically, the horizontal center offset Di is used to quantify the distance of the pixel index from the center. The center distance index transforms index i into a ratio between 0 and 1. The absolute value of the difference between twice and one maps the ratio to the interval [-1, 1], with -1 at the far left of the screen, 1 at the far right, and 0 in the center. The visual sensitivity weight Wi is used to convert the horizontal center offset Di into visual sensitivity. Since the decrease in human eye sensitivity is not linear, the sensitivity decays relatively slowly near the center and decays faster near the edge. The square relationship in the formula of visual sensitivity weight Wi approximates the nonlinear characteristics of vision.

[0060] In practice, the visual sensitivity weight Wi is defined as Wi=1-horizontal center offset Di^2, the visual sensitivity weight Wi=1 (highest weight) of the center pixel, and the visual sensitivity weight Wi=0 of the outermost pixel; the horizontal center offset Di of the pixel is |2*horizontal pixel index i / number of arc pixels minus one value -1|, and the horizontal center offset Di ranges from [0,1].

[0061] The weighted pixel stretching risk value Qi is determined based on the local pixel stretching factor and the visual sensitivity weight Wi: visual sensitivity weight Wi * (local pixel stretching factor - 1)^2; It is understandable that the square term (local pixel stretching factor - 1)^2 is used because both stretching and compression (local pixel stretching factor < 1) will reduce the visual quality, and the perceived distortion increases non-linearly with the deviation. The larger the weighted pixel stretching risk value Qi, the higher the risk that the stretching distortion at the corresponding pixel position will be perceived by the viewer.

[0062] Search from the horizontal pixel at the center of the screen outwards to both sides of the screen. The horizontal pixel index i corresponding to the horizontal pixel at the center of the screen is equal to the number of arc pixels minus one value / 2. The index of the first weighted pixel on the left whose stretching risk value Qi is greater than the stretching risk threshold Qt is denoted as the left index boundary. The pixel index of the first weighted pixel with a stretching risk value Qi greater than the stretching risk threshold Qt on the right is denoted as the right index boundary.

[0063] The target display screen's region type is distinguished based on the left and right index boundaries, and the region type includes a core fidelity region and an edge transition region. The pixel index range of the core fidelity zone is [left index boundary, right index boundary]. The weighted pixel stretching risk value Qi in the core fidelity zone is less than or equal to the stretching risk threshold Qt, indicating low stretching risk. The pixel index range of the edge transition zone is [0, left index boundary - 1] and [right index boundary + 1, number of physical pixels along the arc direction - 1]. The weighted pixel stretching risk value Qi in the edge transition zone is greater than the stretching risk threshold Qt, indicating a high stretching risk. Therefore, the intensity of the precise mapping needs to be gradually reduced.

[0064] The preferred value range for the stretching risk threshold Qt is 0.01 to 0.05. This value is determined by multiplying the weighted pixel stretching risk value by the visual sensitivity weight Wi and the square of the stretching factor deviation (Si-1)². When the local stretching factor Si deviates from 1.0 by approximately 10%, (Si-1)² is approximately 0.01. In the central area of ​​the screen, the visual sensitivity weight Wi is close to 1.0, and the weighted pixel stretching risk value reaches the level of 0.01. If the threshold is below 0.01, the core fidelity area will shrink excessively, and even minor edge fluctuations will trigger overprocessing, leading to unnecessary accuracy loss. If the threshold is above 0.05, it can tolerate approximately 22% stretching factor deviation, but significant edge distortion will still be classified as a core area, negating the purpose of regional adjustment. Therefore, setting the threshold within this range achieves a reasonable balance between stretching tolerance and fidelity area coverage.

[0065] In the core fidelity region, geometric mapping reference texture coordinates ur(i) are used; In the edge transition region, the transition is from precise mapping to linear mapping of standard texture coordinates ul(i) to reduce the visual impact of stretching distortion. This transition is controlled by a continuously varying fusion weight α(i).

[0066] Specifically, the core fidelity region fusion weight α(i) = 1.0, using a linear mapping; In the left edge transition area, the fusion weight α(i) decreases linearly from 1.0 at the boundary to αm at the outermost pixel i=0. The fusion weight α(i) = αm + (1.0-αm)*i / left index boundary. In the right edge transition area, α(i) decreases linearly from 1.0 at the boundary to αm at the outermost pixel, α(i) = αm + (1.0 - αm) * (number of arc pixels minus one value - i) / (number of arc pixels minus one value - right index boundary); Among them, αm is the fusion weight of the outermost pixel, which is a preset constant with a value range of [0,1]. In practice, αm=0.3, which determines the maximum allowable linearization degree of the edge transition area.

[0067] Specifically, the blending weight α(i) determines the proportion of exact mapping and linear mapping in the final texture coordinates of a pixel. Blending weight = 1 indicates a completely exact mapping, while blending weight = αm indicates a maximum bias towards linear mapping. In the left edge transition zone, when i equals zero, it is the outermost pixel, and the blending weight is the smallest. When i is at the boundary of the transition zone, the blending weight is 1.0, seamlessly connecting with the core fidelity zone. The blending weight of the middle pixels is linearly interpolated between αm and 1.0 based on their distance from the outermost edge. The principle of the right edge transition zone is symmetrical to that of the left edge transition zone. The closer to the screen edge, the smaller the blending weight, and the more inclined to use uniform linear mapping to dilute stretching distortion. At the same time, since αm determines the maximum intensity of linearization at the outermost edge of the screen, a balance needs to be found between two contradictory goals. αm = 0.3 means that edge pixels use 30% exact mapping and 70% linear mapping. The 70% linear mapping component is sufficient to effectively dilute stretching distortion, making the edge image look more uniform and natural. The 30% exact mapping component ensures that the basic structure of the image content is not lost, and there is no obvious sense of squashing or offset.

[0068] The final texture coordinates are determined based on the fusion weight α(i), the reference texture coordinate ur(i), and the standard texture coordinate ul(i); The final texture coordinate uf(i) for each pixel i is determined based on the product of the fusion weight α(i) and the reference texture coordinate ur(i), plus the weight deviation and the standard texture coordinate ul(i), wherein the weight deviation is a difference from the fusion weight α(i).

[0069] Understandably, the formula merges the two calculated mapping coordinates using continuously varying weights to output a final texture coordinate with a smooth transition.

[0070] The final mapping table for Npa pixels is output based on the final texture coordinates, which is equal to the final texture coordinates uf(i).

[0071] Understandably, the fusion weight α(i) changes continuously between 0 and 1, determining how much exact mapping and how much linear mapping is used for each pixel; the core fidelity area in the center of the screen has α=1.0, and uses 100% exact mapping ur(i); In the edge transition zone at the screen edge, α gradually decreases from 1.0 to αm=0.3, that is, from 100% accurate mapping, it gradually transitions to a mixture of 30% accurate mapping and 70% linear mapping; the final texture coordinates are determined based on the product of the fusion weight α(i) and the reference texture coordinate ur(i) plus the weighted deviation and the standard texture coordinate ul(i).

[0072] In practice, the final texture coordinate uf(i) = fusion weight α(i) * reference texture coordinate ur(i) + (1.0 - α(i)) * standard texture coordinate ul(i).

[0073] When α=1.0, the final texture coordinates are equal to the reference texture coordinates, achieving a completely accurate mapping; when α=0.3, the final texture coordinates = 0.3 * reference texture coordinates + 0.7 * standard texture coordinates, mostly using linear mapping, resulting in a more uniform image; intermediate values ​​provide a smooth transition without abrupt changes.

[0074] Specifically, under deep correction conditions, this invention achieves point-by-point quantitative evaluation of pixel stretching by constructing reference texture coordinates based on curved surface geometric projection and standard texture coordinates based on linear mapping. By introducing visual sensitivity weights and square-form stretching risk value calculations, it can accurately identify subjectively significant distortion locations in the screen edge region and automatically divide the core fidelity zone and the edge transition zone. Through continuously changing fusion weights, a gradual transition mechanism is established between the two zones, effectively diluting the visual impact of edge stretching distortion while maintaining the geometric accuracy of the image center and eliminating abrupt changes in the image at the regional boundaries. The brightness compensation coefficient is calculated in real time based on the stretching factor and is smoothed and limited, simultaneously solving the problem of uneven brightness caused by changes in pixel density and ensuring smooth and consistent brightness across the entire screen.

[0075] Brightness compensation adjustment based on pixel density is performed to further address the problem of uneven display brightness caused by pixel stretching / compression, building upon the adjustment of the geometric pixel position of the image.

[0076] The brightness compensation coefficient Br(i) of each pixel is determined based on the local pixel stretching factor, and the brightness compensation coefficient Br(i) is equal to the local pixel stretching factor corresponding to each pixel. The smoothed brightness compensation coefficient is determined based on the brightness compensation coefficient Br(i).

[0077] It is understandable that in order to compensate for the uneven brightness caused by changes in pixel density, each pixel needs to be multiplied by a brightness compensation coefficient that is proportional to the stretching factor. Directly using the brightness compensation coefficient may cause brightness jumps in adjacent pixels due to small fluctuations in the local pixel stretching factor. Therefore, a one-dimensional mean filter is applied to the brightness compensation coefficient Br(i) sequence.

[0078] The smoothed brightness compensation coefficient Bs(i) is compared with the preset brightness compensation upper and lower limits [Bmin, Bmax]. If the smoothed brightness compensation coefficient Bs(i) is higher than the upper and lower limits of brightness compensation, then the forced brightness compensation coefficient Bs(i) is the upper limit of brightness compensation Bmax. If the smoothed brightness compensation coefficient Bs(i) is lower than the upper and lower limits of brightness compensation, then the forced brightness compensation coefficient Bs(i) is the lower limit of brightness compensation Bmin.

[0079] Understandably, the purpose of limiting the current is to prevent excessive compensation current from damaging LED chips or causing local overheating under extreme curvature.

[0080] The smoothed brightness compensation coefficient is determined based on the brightness compensation coefficient Br(i). In practice, the Br(i) sequence is subjected to one-dimensional mean filtering and a moving average with a window size of W is used, where W=5. The smoothed brightness compensation coefficient Bs(i) is determined based on the product of the smoothing coefficient and the sum of the brightness compensation coefficients within the summation range. The smoothing coefficient is a ratio to the window size W. In practice, the smoothed brightness compensation coefficient Bs(i) = (1 / W)*ΣBr(i+k), and the summation range is from k=-(W-1) / 2 to (W-1) / 2. The boundary is filled with the nearest value.

[0081] In practice, the upper limit of brightness compensation (Bmax) and the lower limit of brightness compensation (Bmin) are determined by hardware safety specifications. The upper limit of brightness compensation (Bmax) is 1.5 to prevent overcurrent damage to LED beads. If the upper limit is higher than 1.5, it is easy to generate heat and accelerate light decay. The lower limit of brightness compensation (Bmin) is 0.5 to prevent excessive reduction of brightness, which would lead to loss of details in dark areas and color distortion. If the lower limit is lower than 0.5, such as 0.2, in areas of the image that need to be dark, the pixels may be pulled directly below the cutoff voltage.

[0082] For pixel (i,j), its final driving brightness Ld(i,j) = Lv(final texture coordinates uf(i),j) * smoothed brightness compensation coefficient Bs(i); Where Lv(u,j) is the original color brightness value sampled from the video source based on the final texture coordinates uf(i).

[0083] Specifically, by establishing a direct mapping relationship between the pixel stretching factor and the brightness compensation coefficient, automatic correction of brightness unevenness caused by pixel density changes due to geometric mapping is achieved. The brightness compensation coefficient is taken from the real-time calculation result of the stretching factor, requiring no additional measurement or calibration, making the calculation process simple and efficient. By introducing a one-dimensional mean filter to smooth the compensation coefficient, brightness abrupt changes caused by small fluctuations in the stretching factor between adjacent pixels are effectively eliminated, ensuring a smooth and natural transition of screen brightness. At the same time, compensation upper and lower limits determined by hardware safety specifications are set, constraining the compensation amplitude within the range of [Bmin, Bmax] to prevent the compensation current from exceeding the reliable operating range of the LED beads under extreme curvature, ensuring both display effect and hardware lifespan and operational safety. The dual processing mechanism of smoothing and limiting makes the brightness compensation both responsive and stable.

[0084] The boundary gradient anomaly E is determined based on the average local gradient normalization of the left and right boundary gradient peaks. Specifically, the boundary gradient anomaly is the rate of change of texture coordinates at the boundary of the transition zone, and the degree of deviation from the normal rate of change of the core fidelity zone. A boundary gradient anomaly close to 0 indicates that the boundary transition is very smooth. The larger the boundary gradient anomaly, the more obvious the abrupt change in the mapping gradient at the boundary, which may be visually manifested as a subtle bending sensation at a certain vertical line in the image.

[0085] If the boundary gradient anomaly is less than or equal to the smoothness tolerance threshold, then the current transition zone boundary is judged to be smooth, and the switch from the core fidelity zone to the edge transition zone is natural. If the boundary gradient anomaly is greater than the smoothness tolerance threshold, it is determined that there is a perceptible abrupt change in the mapping gradient at the boundary of the transition region, and the factor threshold is reduced according to the ratio of the smoothness tolerance threshold to the boundary gradient anomaly.

[0086] Understandably, the presence of perceptible abrupt changes in the mapping gradient at the transition zone boundary indicates that the boundary between the edge transition zone and the core fidelity zone of the screen currently in a deep correction state is not ideal. The transition start position may be too close to the center, causing a drastic mapping change in areas that should have transitioned smoothly. For the current screen with a curvature deviation of δ, the factor threshold is set too high, causing a moderately biased screen that should have been classified as undergoing regular correction to be incorrectly redirected to the deep correction branch. This leads to more aggressive region partitioning and fusion, resulting in unnecessary gradient abrupt changes at the fusion boundary.

[0087] In practice, the boundary gradient anomaly degree E is determined based on the ratio of the maximum gradient difference to the average local gradient, wherein the maximum gradient difference is the maximum absolute value of the difference between the left boundary gradient peak and the right boundary gradient peak and the average local gradient.

[0088] In practice, the boundary gradient anomaly E = max(|peak gradient of the left boundary - average local gradient|, |peak gradient of the right boundary - average local gradient|) / average local gradient; Specifically, a closed-loop feedback mechanism for the accuracy of the factor threshold was established by monitoring the local gradient characteristics of the final texture mapping table at the boundary between the core fidelity area and the edge transition area. The gradient peak was extracted by five-point sampling of the boundary neighborhood and first-order numerical differentiation, and normalized with the average gradient of the core area as the benchmark. The generated boundary gradient anomaly index can objectively quantify the smoothness of the transition area without relying on the statistical correctness judgment of the display effect. When the anomaly exceeds the smoothness tolerance threshold, it is inferred that the factor threshold is too large and it is quantitatively corrected. This makes the factor threshold approach the optimal dividing point for distinguishing between regular correction and deep correction in long-term operation, avoiding the problem of over-processing of screens with medium curvature deviation, and improving the adaptive ability and parameter self-optimization level of the driving method for screens with different curvatures.

[0089] In this embodiment, the local variation characteristics of the output final texture coordinate mapping table at the region boundary are determined; Obtain the left and right texture coordinates of the neighborhoods of the left and right index boundaries and their neighbors, respectively. In practice, the texture coordinates are the five adjacent pixels including the left and right index boundaries. Specifically, first-order numerical differentiation is performed on the left and right texture coordinates respectively, and the local gradient between each pair of adjacent points is calculated by combining the final texture coordinates; For example, the texture coordinates on the left include uf(left index boundary - 2), uf(left index boundary - 1), uf(left index boundary), uf(left index boundary + 1), and uf(left index boundary + 2); The local gradient between adjacent point pairs is gl(k) = uf(k) - uf(k-1), where k takes the values ​​of left index boundary -1, left index boundary, left index boundary +1, and left index boundary +2.

[0090] Take the maximum value of the local gradient gl(k) as the left boundary gradient peak, and similarly obtain the right boundary gradient peak.

[0091] Calculate the average local gradient within the core fidelity region as an ideal reference gradient when there is no transitional disturbance. In practice, one-third of the pixels in the core fidelity area are taken, and the difference between the final texture coordinate uf(i) and the adjacent final texture coordinate uf(i-1) is calculated for each pair of adjacent pixels in the interval. The average value is the average local gradient.

[0092] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A pixel arrangement and driving method for irregularly shaped curved LED creative displays, characterized in that, include: The required correction state for the target display screen is determined based on the curvature deviation factor. In response to the correction state being a normal correction state, the linear mapping relationship between the display screen pixels and the video source pixels is adjusted. In response to the correction state being a depth correction state, the reference texture coordinates of each horizontal pixel index are determined; The curvature deviation factor is determined by normalizing the radial displacement deviation between each pixel of the target display screen and the standard reference screen; the reference texture coordinates are determined based on the arc length coordinates and the ratio of the total arc length of the luminous area of ​​the reference screen to the standard reference curvature radius. Determine the mapping deviation of each horizontal pixel, combine the cross-texture distance to determine the local pixel stretching factor to determine whether the image is stretched or compressed, and determine the weighted pixel stretching risk value to locate the left index boundary and the right index boundary to distinguish the region type of the target display screen; The weighted pixel stretching risk value is determined based on visual sensitivity weight combined with local pixel stretching factor. The local pixel stretching factor is determined based on the ratio of the distance between adjacent pixels of the target display screen and the standard reference screen. The region type includes core fidelity region and edge transition region. In response to the region type being an edge transition area, the final texture coordinates are determined, the mapping relationship between pixels is adjusted, and the final driving brightness of each pixel is determined; The final texture coordinates are determined based on the pixel-corresponding fusion weight, combined with the reference texture coordinates and the standard texture coordinates. The final driving brightness is determined based on the smoothed brightness compensation coefficient of each pixel determined by the local pixel stretching factor, combined with the final texture coordinates. The boundary gradient anomaly degree is determined by combining the local variation features of the final texture coordinates with the ideal reference gradient to correct the factor threshold of the curvature deviation factor. Among them, the standard texture coordinates are the texture coordinates when the reference screen is linearly mapped, which are determined based on the ratio of the arc length coordinates to the total arc length of the light-emitting area of ​​the reference screen, and the reference texture coordinates are determined based on the ratio of the arc length coordinates to the total arc length of the light-emitting area of ​​the reference screen and the standard reference radius of curvature. The local variation features are the local gradients between each pair of adjacent points in the neighborhood of the left and right index boundaries based on the final texture coordinates, including the left boundary gradient peak and the right boundary gradient peak of the left and right texture coordinates. The ideal reference gradient is the average local gradient inside the core fidelity region. Obtain the average local gradient normalized to the left and right boundary gradient peaks within the core fidelity region to determine the boundary gradient anomaly.

2. The pixel arrangement and driving method for irregularly shaped curved LED creative displays according to claim 1, characterized in that, The process of determining the required correction state for the target display screen includes, If the curvature deviation factor is less than or equal to the factor threshold, the target display screen is judged to be in a normal correction state, and the residual distortion of the video source image is within an acceptable range after global compensation. If the curvature deviation factor is greater than the factor threshold, the target display screen is judged to be in a depth correction state. The curvature deviation of the display screen is too large, and the residual pixel stretching or compression after global compensation in the edge area may affect the subjective visual quality.

3. The pixel arrangement and driving method for irregularly shaped curved LED creative displays according to claim 2, characterized in that, A positive curvature deviation factor indicates that the target display screen is flatter than the standard reference screen, and the image will be stretched horizontally. A negative curvature deviation factor indicates that the target display screen is more curved than the standard reference screen, and the image will be compressed laterally.

4. The pixel arrangement and driving method for irregularly shaped curved LED creative displays according to claim 3, characterized in that, In the normal correction state, the horizontal texture coordinates of the horizontal pixel index in the video source image are determined based on the arc length coordinates and the total arc length of the light-emitting area of ​​the reference screen, and the linear mapping relationship between the display screen pixels and the video source pixels is adjusted. Wherein, the arc length coordinate is the sum of several arc length spacings corresponding to the horizontal pixel index, and the horizontal pixel index is a single integer value that sequentially numbers the pixels arranged along the arc direction on the display screen.

5. The pixel arrangement and driving method for irregularly shaped curved LED creative displays according to claim 4, characterized in that, The local pixel stretching factor is determined based on the ratio of the reference texture coordinates to the distance across the texture; The mapping deviation of each horizontal pixel is determined based on the reference texture coordinates of the target display pixel and the standard texture coordinates of the standard reference screen, and the cross-texture distance is determined based on the reference texture coordinates of any pixel and its adjacent pixels. In depth correction mode, if the local pixel stretching factor is greater than one, the image is determined to be stretched. If the local pixel stretching factor is less than one, the image is considered compressed.

6. The pixel arrangement and driving method for irregularly shaped curved LED creative displays according to claim 5, characterized in that, The visual sensitivity weight is determined based on the horizontal center offset of the corresponding pixel. The index of the first weighted pixel on the left whose stretching risk value is greater than the stretching risk threshold is denoted as the left index boundary. The index of the first right-hand weighted pixel whose stretching risk value is greater than the stretching risk threshold is denoted as the right-hand index boundary.

7. The pixel arrangement and driving method for irregularly shaped curved LED creative displays according to claim 6, characterized in that, The core fidelity zone is the pixel region within the pixel index interval formed by the left index boundary and the right index boundary. Within the core fidelity zone, the weighted pixel stretching risk value is less than or equal to the stretching risk threshold, indicating low stretching risk. The edge transition zone is a pixel region that exceeds the pixel index range. Within the edge transition zone, the weighted pixel stretching risk value is greater than the stretching risk threshold, indicating a high stretching risk. The mapping intensity is then gradually reduced.

8. The pixel arrangement and driving method for irregularly shaped curved LED creative displays according to claim 7, characterized in that, In the core fidelity area, geometric mapping references texture coordinates; In the edge transition region, the proportions of reference texture coordinates and standard texture coordinates in the final texture coordinate composition are determined by the fusion weights.

9. The pixel arrangement and driving method for irregularly shaped curved LED creative displays according to claim 8, characterized in that, The brightness compensation coefficient is the local pixel stretching factor corresponding to each pixel. If the smoothed brightness compensation coefficient is higher than the upper and lower limits of brightness compensation, then the corresponding brightness compensation coefficient is the upper limit of brightness compensation. If the smoothed brightness compensation coefficient is lower than the upper and lower limits of brightness compensation, then the corresponding brightness compensation coefficient is the lower limit of brightness compensation. The final driving brightness of each pixel is determined based on the smoothed brightness compensation coefficient and the original color brightness value, which is obtained by sampling from the video source according to the final texture coordinates.

10. The pixel arrangement and driving method for irregularly shaped curved LED creative display screens according to claim 9, characterized in that, If the boundary gradient anomaly is less than or equal to the smoothness tolerance threshold, then the current transition zone boundary is determined to be smooth. If the boundary gradient anomaly is greater than the smoothness tolerance threshold, it is determined that there is a perceptible abrupt change in the mapping gradient at the boundary of the transition region, and the factor threshold is reduced according to the ratio of the smoothness tolerance threshold to the boundary gradient anomaly.

Images