LED display screen local adjustment method and system and storage medium

Abstract

The invention relates to the technical field of display device manufacturing, in particular to an LED display screen local adjustment method and system and a storage medium, and the method comprises the steps: collecting displacement deviation and view field deviation characteristics of splicing boundary pixels, and carrying out the fusion to generate a geometric feature matrix; mapping the physical deviation to a driving circuit logic coordinate, establishing a driving energy level distribution field, integrating a generation power gain factor based on physical gap characteristics to remodel an equivalent light emitting envelope, and generating a light field reconstruction dimension; and finally, performing pulse width modulation stage processing in combination with the video signal, outputting error cooperative gain and performing dynamic pulse compensation. According to the invention, through digital mapping of physical tolerance and optical attenuation characteristics, hardware-level accurate compensation of energy loss of the splicing gap is realized, visual faults and uneven brightness caused by the splicing gap are inhibited from a physical light field level, local adjustment intellectualization is improved, and meanwhile, the energy consumption of the splicing gap is reduced. And the visual quality of the display screen in a complex application scene is ensured.

Description

Technical Field

[0001] This invention relates to the field of display device manufacturing technology, specifically to a method, system, and storage medium for local adjustment of an LED display screen. Background Technology

[0002] With the rapid development of LED display device manufacturing technology, ultra-large area, high resolution, and irregularly shaped display systems such as cylindrical and curved screens have been widely used in outdoor advertising, stage performances, and monitoring and command centers. LED displays are usually spliced ​​together by a large number of LED lights, modules, and receiver cards. Due to individual differences in light-emitting components, uneven heat dissipation, different aging levels, and mechanical errors in the splicing process, local areas of the display often exhibit problems such as inconsistent brightness and color, bright and dark lines at the splicing seams, or local display abnormalities, which seriously affect the overall visual effect.

[0003] Currently, the industry primarily employs graphical topology interaction or pixel-level fine-tuning technologies for localized adjustments of LED displays. Existing methods typically require obtaining the display's screen type and size, pre-drawing a screen distribution diagram, and then having the user manually select specific receiver cards, modules, or seam coordinates on a graphical interface before sending adjustment commands. Alternatively, some technologies focus on acquiring attribute information for all LEDs within a selected local area and providing single-point or batch modification modes to achieve pixel-level fine-tuning.

[0004] While the existing technologies mentioned above consider the hierarchical adjustment dimensions (from receiver card and module to lamp point level) and the coordinate positioning of the adjustment area, greatly improving the accuracy of local adjustment, they still have the following shortcomings: First, such methods rely heavily on manual intervention and preset topology maps. When facing display systems with extremely complex topologies or a large number of dynamic splicing elements, manual selection of areas is inefficient and has poor fault tolerance. Second, existing technologies are mostly based on static parameter compensation and do not fully consider the dynamic coupling relationship between display content characteristics and local hardware load, making it difficult to guarantee the consistency of adjustment effects under different display screens.

[0005] To address the bottlenecks in adjustment efficiency and insufficient dynamic adaptability in existing display device manufacturing technologies, a local adjustment method for LED displays is needed. This method should not only improve the intelligence of local adjustment but also ensure the visual quality of the display in complex application scenarios.

[0006] To address this, a method, system, and storage medium for local adjustment of an LED display screen are proposed. Summary of the Invention

[0007] The purpose of this invention is to provide a method, system, and storage medium for local adjustment of an LED display screen, which improves the intelligence of local adjustment while ensuring the visual quality of the display screen in complex application scenarios.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A method for local adjustment of an LED display screen, comprising: The theoretical pixel pitch and the actual pixel pitch of the center point of the splicing boundary pixel are obtained during the display manufacturing process. The pitch difference is calculated to generate a mechanical error operator. The luminous intensity data of the splicing boundary pixel under non-perpendicular viewing angle and standard viewing angle are compared to output the field deviation feature. The mechanical error operator and the field deviation feature are fused in the spatial coordinate system to generate a geometric feature matrix. Based on the geometric feature matrix, a driving energy level distribution field is established in the driving circuit control area; according to the physical gap features contained in the geometric feature matrix, the driving power compensation weight of the affected lamp points at the boundary is calculated to generate a power gain factor; based on the power gain factor, a nonlinear mapping operation is performed on the driving energy level distribution field to reshape the light emission envelope of the spliced ​​boundary pixels and generate the light field reconstruction dimension. Based on the light field reconstruction dimension, the input display signal is subjected to pulse width modulation hierarchical processing to generate a pulse mapping matrix; the pulse mapping matrix is ​​converted into a duty cycle adjustment sequence at the driver chip end to output error coordination gain; dynamic pulse compensation is performed on the pixel lamp points at the splicing boundary according to the error coordination gain to generate a visual effect smoothing dimension.

[0010] Preferably, the specific process of generating the mechanical error operator includes: extracting the theoretical pixel spacing determined by the display design file; obtaining the actual pixel spacing of the center point of the pixel at the splicing boundary of the display screen; calculating the spacing displacement vector that deviates from the theoretical pixel spacing of the actual pixel spacing; and mapping the spatial displacement vector into a mechanical error operator that quantifies the physical splicing deviation.

[0011] Preferably, the specific process of generating a geometric feature matrix by outputting the field of view deviation feature includes: acquiring the luminous intensity data of the splicing boundary pixels at the non-perpendicular viewing angle and the luminous intensity data at the standard viewing angle; comparing the distribution difference between the luminous intensity data at the non-perpendicular viewing angle and the luminous intensity data at the standard viewing angle, and outputting the field of view deviation feature; and fusing the mechanical error operator and the field of view deviation feature in the spatial coordinate system to generate a geometric feature matrix.

[0012] Preferably, the specific process of establishing the driving energy level distribution field and generating the power gain factor includes: extracting the mechanical error operator and field-of-view deviation features contained in the geometric feature matrix, obtaining the control pixel logic coordinates corresponding to the driving circuit control area, mapping the mechanical error operator and field-of-view deviation features to the control pixel logic coordinates, and generating driving energy level quantization parameters; aggregating the driving energy level quantization parameters to generate the driving energy level distribution field; analyzing the geometric feature matrix to obtain physical gap features characterizing the distribution state and energy loss trend of the splicing boundary gaps; determining the driving power compensation weights of the affected lamp points at the boundary based on the physical gap features, and integrating the driving power compensation weights to generate the power gain factor.

[0013] Preferably, the specific process of generating the light field reconstruction dimension includes: retrieving the power gain factor and the driving energy level distribution field; superimposing the power gain factor onto the driving energy level distribution field to perform energy level dynamic equalization processing and nonlinear mapping operation to generate an energy level correction matrix; and using the energy level correction matrix to reshape the emission envelope of the splicing boundary pixels to generate the light field reconstruction dimension.

[0014] Preferably, the specific process of generating the pulse mapping matrix includes: acquiring the input display signal from the video source; extracting the pixel grayscale data contained in the input display signal; performing pulse width modulation hierarchical processing on the pixel grayscale data based on the light field reconstruction dimension; calculating the pulse width adjustment parameters at the corresponding positions of the splicing boundary pixel regions; and generating the pulse mapping matrix.

[0015] Preferably, the specific process of generating a visual smoothing dimension based on the output error collaborative gain includes: converting the pulse mapping matrix into a duty cycle adjustment sequence composed of the proportion of light emission conduction time of the driving chip within the pulse modulation period; parsing the duty cycle adjustment sequence and outputting the error collaborative gain; and performing dynamic pulse compensation on the pixel lamp points at the splicing boundary through the error collaborative gain to offset the light emission energy loss caused by physical mechanical errors, suppress visual dark seams, and generate a visual smoothing dimension.

[0016] A local adjustment system for an LED display screen, comprising: The geometric feature extraction module is used to obtain the theoretical pixel spacing and the actual pixel spacing of the center point of the splicing boundary pixels during the display manufacturing process, calculate the spacing difference to generate a mechanical error operator; compare the luminous intensity data of the splicing boundary pixels under non-perpendicular viewing angle and standard viewing angle, and output the field deviation feature; couple the mechanical error operator and the field deviation feature to generate a geometric feature matrix. Driven energy field reconstruction module: Based on the geometric feature matrix, a drive energy level distribution field is established in the drive circuit control area; according to the physical gap features contained in the geometric feature matrix, the drive power compensation weight of the affected lamp points at the boundary is calculated to generate a power gain factor; based on the power gain factor, a nonlinear mapping operation is performed on the drive energy level distribution field to reshape the light emission envelope of the spliced ​​boundary pixels and generate the light field reconstruction dimension. Pulse sequence execution module: Based on the light field reconstruction dimension, it performs pulse width modulation hierarchical processing on the input display signal to generate a pulse mapping matrix; it converts the pulse mapping matrix into a duty cycle adjustment sequence at the driver chip end and outputs error coordination gain; it performs dynamic pulse compensation on the pixel lamp points at the splicing boundary according to the error coordination gain to generate a visual effect smoothing dimension.

[0017] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the LED display screen local adjustment method.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention generates a geometric feature matrix by coupling a mechanical error operator with field-of-view deviation characteristics, achieving a substantial correlation between the spatial displacement at the manufacturing end and the luminous characteristics at the observation end. Compared to simple manual static parameter settings, this multi-dimensional data fusion logic solves the problem of local brightness discontinuity caused by manufacturing tolerances, improving the accuracy of automated adaptation of local adjustments under complex observation angles.

[0019] 2. This invention utilizes the driving energy level distribution field in conjunction with the power gain factor to reshape the light-emitting envelope of the boundary pixels, achieving deep synergy between the energy topology distribution of the driving control area and the power compensation logic for physical gaps. This energy level reconstruction method directly performs light field alignment at the splicing position, suppressing the visual dark seam problem caused by the splicing boundary and ensuring the display continuity of the monitor in scenarios with physical mechanical errors.

[0020] 3. This invention combines a pulse mapping matrix generated by performing pulse width modulation (PWM) hierarchical processing on the input display signal, with a duty cycle adjustment sequence composed of the proportion of light-emitting conduction time of the driver chip within the pulse modulation period, to achieve precise real-time coupling between the original grayscale data at the video source and the pulse width executed by the terminal. This dynamic pulse compensation mechanism, while offsetting physical energy loss, ensures stable visual quality and grayscale performance of the display screen in complex dynamic application scenarios. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating the structure of a local adjustment method for an LED display screen according to the present invention. Figure 2This is a flowchart of the geometric feature extraction and coupling process according to an embodiment of the present invention; Figure 3 This is a schematic diagram of a local adjustment system for an LED display screen according to the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] Please see Figures 1 to 3 This invention provides a method, system, and storage medium for local adjustment of an LED display screen, the technical solution of which is as follows: Example 1 Reference Figure 1 This embodiment provides a specific application scenario for a local adjustment method for LED displays, specifically for aligning the physical boundaries of spliced ​​displays or irregularly shaped display terminals. In monitoring and command centers, stage performances, or commercial centers, the entire LED display area is formed by mechanically splicing modules. Due to manufacturing process errors, transportation displacement, and stress factors of mounting brackets, physical gap displacement occurs at the splicing boundaries, resulting in inconsistent brightness or visual gaps in the displayed image at the splicing positions. This embodiment introduces an adjustment method during the display manufacturing process to suppress visual breaks caused by physical splicing errors and ensure the integrity of the entire display area under complex viewing angles. The specific steps include: The theoretical pixel pitch and the actual pixel pitch of the center point of the splicing boundary pixel are obtained during the display manufacturing process. The pitch difference is calculated to generate a mechanical error operator. The luminous intensity data of the splicing boundary pixel under non-perpendicular viewing angle and standard viewing angle are compared to output the field deviation feature. The mechanical error operator and the field deviation feature are fused in the spatial coordinate system to generate a geometric feature matrix. Based on the geometric feature matrix, a driving energy level distribution field is established in the driving circuit control area; according to the physical gap features contained in the geometric feature matrix, the driving power compensation weight of the affected lamp points at the boundary is calculated to generate a power gain factor; based on the power gain factor, a nonlinear mapping operation is performed on the driving energy level distribution field to reshape the light emission envelope of the spliced ​​boundary pixels and generate the light field reconstruction dimension. Based on the light field reconstruction dimension, the input display signal is subjected to pulse width modulation hierarchical processing to generate a pulse mapping matrix; the pulse mapping matrix is ​​converted into a duty cycle adjustment sequence at the driver chip end to output error coordination gain; dynamic pulse compensation is performed on the pixel lamp points at the splicing boundary according to the error coordination gain to generate a visual effect smoothing dimension.

[0024] Furthermore, the specific process of generating the mechanical error operator includes: extracting the theoretical pixel spacing determined by the display design file; obtaining the actual pixel spacing of the center point of the pixel at the splicing boundary of the display screen; calculating the spacing displacement vector that deviates from the theoretical pixel spacing of the actual pixel spacing; and mapping the spatial displacement vector into a mechanical error operator that quantifies the physical splicing deviation.

[0025] Specifically, in a scenario where the monitoring center directs the display manufacturing process, the display manufacturing design file in the memory is retrieved beforehand. The theoretical pixel pitch value defined in the display manufacturing design file is read. In this embodiment, the theoretical pixel pitch value is set to 1.200mm. Simultaneously, a pixel acquisition device captures the image of the display splicing boundary, and a high-precision image acquisition device obtains the actual physical position of the center points of two adjacent pixel lamps at the display splicing boundary, thus obtaining the actual pixel pitch value. In this embodiment, the actual pixel pitch value is set to 1.245mm. Through numerical comparison calculation, the actual pixel pitch value is subtracted from the theoretical pixel pitch value to obtain a numerical difference of 0.045mm, and the direction of the numerical difference is determined to be the horizontal stretching direction, generating a pitch displacement vector. Subsequently, a preset mapping function is used... (in For quantization gain coefficients, The displacement vector (value) is converted into an energy compensation order recognizable by the driving logic. This energy compensation order is specifically recorded as the value obtained by discretizing the physical displacement according to a preset quantization step (e.g., 0.005 mm / order). This order serves as the core carrier of the mechanical error operator, representing the digital increment level that needs to be superimposed on the base brightness to compensate for gaps and dark lines. It is encapsulated in the form of a standardized quantized physical splicing deviation data format in the form of a three-dimensional feature vector to generate the mechanical error operator. This standardized quantized physical splicing deviation data format uses a 16-bit word length as the basic storage unit at the physical layer and is stored in the non-volatile memory of the display control system.

[0026] Specifically, the quantization gain coefficient k is a piecewise function that is positively correlated with the magnitude of the displacement vector. When the displacement vector... Within a preset small deviation range (e.g., 0.01mm-0.05mm), The value is set to a fixed constant of 128 to achieve linear reference compensation; when the displacement vector exceeds this interval... The value increases stepwise with the increase of displacement (ranging from 128 to 512) to compensate for nonlinear light intensity loss under large spacing. The specific values ​​are pre-calibrated based on the display pixel pitch specifications and the current resolution of the driver chip. The mechanical error operator uses 10-bit precision for three-dimensional feature vector encoding, recording the deviation weight and direction of the physical splicing position on the spatial coordinate axes. The highest bit (bit 9) is the direction flag, used to characterize the positive or negative attribute of the displacement vector; the middle 6 bits (bits 8-3) are the energy compensation order bits, recording the gain intensity calculated by the mapping function, corresponding to 64 levels of quantization precision; the last 3 bits (bits 2-0) are the spatial axis index bits, used to encode the X, Y, and Z three-dimensional coordinate axis attributes corresponding to the vector (e.g., 001 represents the horizontal axis, 010 represents the vertical axis, and 100 represents the normal flatness axis). Through this 10-bit standardized packaging, precise mapping from millimeter-level physical displacement to driver chip logic control instructions is achieved.

[0027] By quantitatively comparing the display manufacturing parameters with the actual physical state, the physical splicing error is transformed into a digital displacement vector that can be recognized by the driving circuit. This achieves a precise digital description of the manufacturing assembly deviation, solves the problem of visual discontinuity caused by black seams or bright lines at the splicing points, and provides a high-precision deviation data source for the subsequent local fine adjustment of the driving duty cycle.

[0028] Furthermore, the specific process of generating a geometric feature matrix by outputting the field of view deviation feature includes: acquiring the luminous intensity data of the stitching boundary pixels at the non-perpendicular viewing angle and the luminous intensity data at the standard viewing angle; comparing the distribution difference between the luminous intensity data at the non-perpendicular viewing angle and the luminous intensity data at the standard viewing angle, and outputting the field of view deviation feature; and fusing the mechanical error operator and the field of view deviation feature in the spatial coordinate system to generate a geometric feature matrix.

[0029] Specifically, an optical measuring device (such as a CCD luminance and colorimeter) is used to measure the luminous intensity of the pixels at the splicing boundary at a standard viewing angle perpendicular to the display plane, obtaining standard viewing angle luminous intensity data, which is then set as the luminance reference standard. Simultaneously, the luminous intensity of the pixels at the splicing boundary is measured at specified angles 30 degrees horizontally and 30 degrees vertically away from the central axis of the display plane, obtaining luminous intensity data from non-perpendicular viewing angles.

[0030] Using the standard viewing angle luminous intensity data as the denominator and the difference in luminous intensity between the standard viewing angle and the non-perpendicular viewing angle as the numerator, a division operation is performed to obtain the percentage decrease in brightness as the viewing angle increases, i.e., the brightness attenuation rate. This attenuation rate is used to quantify the degree of light energy loss at the current viewing angle; simultaneously, a cosine correction function is used... By performing spherical distribution fitting on the measured angle, the light intensity distribution curve across the entire viewing angle domain is derived, where... The measured value of luminous intensity is from a non-perpendicular viewing angle. The reference value for luminous intensity at a standard viewing angle perpendicular to the central axis of the display plane; The angle between the viewing angle and the central axis is denoted by , and 'n' is the directivity factor determined by the LED packaging structure. The luminous intensity data from the standard viewing angle is subtracted from the luminous intensity data from the non-perpendicular viewing angle to obtain the distribution difference in absolute energy values. This difference reflects the non-uniformity of energy distribution at that pixel in the spatial light field. Finally, the brightness attenuation rate is used as a gain weight and correlated with the distribution difference representing the absolute energy deviation to construct a field-of-view deviation feature that includes the viewing angle vector, attenuation weight, and energy level shift.

[0031] A display space coordinate system is established with the center of the first pixel at the top left corner of the display as the origin, and the horizontal and vertical directions of the pixel arrangement as the coordinate axes. Within this coordinate system, the displacement data recorded by the mechanical error operator and the brightness attenuation rate recorded by the field-of-view deviation feature are spatially aligned, ensuring that the physical position deviation and optical energy level characteristics are mapped at the same pixel index point. Subsequently, a weighted fusion algorithm is executed. The spatial position deviation recorded by the mechanical error operator is used as the basis for weight allocation: Specifically, a two-dimensional M×N geometric feature matrix is ​​established with the physical pixel array of the display screen as the dimension, where each matrix element... Each pixel logical address corresponds one-to-one with the splicing boundary, and each element uses {X of ,Y off ,R att D var The quaternion data structure encapsulates the normalized horizontal displacement X. off Normalized vertical displacement Y off Brightness attenuation rate R att and energy distribution difference D var In generating this matrix: as the displacement deviation increases, the correction weight of the brightness attenuation coefficient in the field-of-view deviation feature is simultaneously increased. Specifically, the displacement deviation is first divided by a preset maximum allowable deviation threshold (e.g., 0.5mm), and the physical offset scale is converted to the [0,1] interval to generate a normalized displacement deviation weight. Then, the normalized displacement deviation weight is used as a nonlinear operator, multiplied by the brightness attenuation rate in the field-of-view deviation feature, and the energy distribution difference component is added to calculate the comprehensive visual deviation value of the coordinate point at a specific viewing angle. That is, visual deviation value = (normalized horizontal displacement + normalized vertical displacement) × brightness attenuation rate + energy distribution difference. Finally, the comprehensive visual deviation values ​​of all coordinate points are arranged in an array to generate the geometric feature matrix. This matrix not only contains the offset information of the physical position but also characterizes the optical attenuation characteristics after integrating the influence of physical gaps through a weighted mechanism.

[0032] By integrating physical deviations at the splicing position with optical viewing angle characteristics across dimensions, a quantitative mapping relationship between physical spatial defects and visual perception at the observation end was established. Compared to traditional methods that rely solely on physical gap adjustments, this invention can compensate for viewing angle consistency deviations caused by mechanical deformation, ensuring smooth visual effects at the splicing boundary under different observation orientations and suppressing visual "jumps".

[0033] Furthermore, the specific process of establishing the driving energy level distribution field and generating the power gain factor includes: extracting the mechanical error operator and field-of-view deviation features contained in the geometric feature matrix, obtaining the control pixel logic coordinates corresponding to the driving circuit control area, mapping the mechanical error operator and field-of-view deviation features to the control pixel logic coordinates, and generating driving energy level quantization parameters; aggregating the driving energy level quantization parameters to generate the driving energy level distribution field; analyzing the geometric feature matrix to obtain physical gap features characterizing the distribution state and energy loss trend of the splicing boundary gaps; determining the driving power compensation weights of the affected lamp points at the boundary based on the physical gap features, and integrating the driving power compensation weights to generate the power gain factor.

[0034] Specifically, during the operation of the display screen driver control logic, the processing unit retrieves the geometric feature matrix. From the matrix, it extracts the displacement deviation value recorded by the mechanical error operator and the light intensity attenuation vector recorded by the field of view deviation feature. Subsequently, it obtains the logical coordinates of the control pixels corresponding to the driver circuit control area. The process of obtaining these logical coordinates involves a coordinate mapping transformation from a two-dimensional coordinate system in physical space to the linear address of the driver chip scan. Specifically, the processor maps the coordinate points in the physical array of the display screen to the chip number, port channel number, and time-division scan row index in the corresponding driver chipset according to a preset driver circuit topology (such as a "serpentine" or "Z-shaped" routing rule). In splicing screen scenarios, multiple driver chips are connected in a cascaded manner, and the logical coordinates follow a linear addressing rule based on differential signal chains, ensuring that each physical pixel at the splicing boundary can uniquely correspond to the register index address inside the driver chip.

[0035] After determining the logical coordinates, the energy adjustment order required for each affected lamp at a specific logical address is calculated using a multidimensional weighted mapping function, generating the driving energy level quantization parameters.

[0036] The specific process for generating the driving energy level quantization parameters includes: extracting physical feature weights based on the displacement deviation value and extracting visual sensitivity weights based on the light intensity attenuation vector; statically calibrating the energy compensation order using the physical feature weights and dynamically scaling the energy loss order using the visual sensitivity weights; and finally aligning and quantizing the calibrated energy compensation order and the scaled energy loss order using a nonlinear weighted summation algorithm to generate the driving energy level quantization parameters. The basic energy compensation order is the quantized physical stitching deviation data recorded in the mechanical error operator, obtained after discretizing the physical displacement according to a preset quantization step. The energy loss order is the viewing angle energy level correction value calculated based on the field-of-view deviation characteristics.

[0037] Specifically, the processor first analyzes the magnitude of the displacement deviation. The basic energy compensation order is not a fixed value, but is dynamically mapped from the spacing displacement vector to a preset photoelectric conversion coefficient, used to establish a preliminary brightness compensation benchmark. When the deviation exceeds a preset threshold (e.g., 0.05mm), the physical feature weight is increased to strengthen the priority of basic energy compensation. The increase in the weight shows a logarithmic growth trend positively correlated with the displacement amount; that is, the larger the displacement deviation, the more gradual but continuous the weight increase, with the maximum increase limited to 1.5 times the benchmark weight to ensure that the current drive is within a safe load range. This logarithmic growth relationship is implemented through a preset weight mapping table to ensure that the gain slope of the drive current remains smooth when the physical gap changes drastically, avoiding brightness overshoot. Simultaneously, a visual sensitivity weight is extracted based on the slope of the light intensity attenuation vector to characterize the drastic impact of viewing angle changes on human vision. When performing nonlinear weighted summation, the physical feature weight is applied to the basic energy compensation amount, and the visual sensitivity weight is applied to the loss caused by viewing angle attenuation; finally, the two are summed and locked.

[0038] The nonlinear weighted summation follows the formula: Comprehensive driving energy level quantization parameter = physical feature weight × energy compensation order + visual sensitivity weight × energy loss order. For example, in a low-brightness display environment, the visual sensitivity weight is dynamically scaled within the range of 0.2 to 0.5 to automatically suppress the viewing angle loss order, preventing pixel noise in low grayscale areas due to excessive dynamic compensation. In a high-brightness environment, full coupling is performed. Through the nonlinear weighted summation algorithm, the two types of energy orders are uniformly mapped to the 16-bit control space supported by the chip. The driving energy level quantization parameter is stored in 16-bit unsigned integer form, directly corresponding to the bit width precision of the driver chip's PWM register. By introducing dual weight adjustment of physical features and visual sensitivity, fine-grained on-demand allocation of compensation energy levels is achieved, effectively avoiding display distortion caused by the superposition of compensation amounts in boundary areas, and significantly enhancing the visual purity of splicing boundaries.

[0039] Based on the scanning timing and channel mapping relationship of the driving circuit, all driving energy level quantization parameters are spatially topologically integrated: the processor linearly maps the driving energy level quantization parameters of each pixel from two-dimensional physical space coordinates to memory data streams corresponding one-to-one with their driving chip ports, scan rows, and register addresses, according to the routing topology and cascading order of the driving circuit, constructing a driving energy level distribution field corresponding to the physical splicing boundary. The driving energy level distribution field is an M×N matrix, where M and N are the physical row and column numbers of the display screen, respectively. Each element in the matrix corresponds to the target value of the driving energy level of an affected LED. This distribution field, as a structured data matrix, is directly associated with the output buffer of the driving circuit through memory mapping, used to guide the driving chip to perform pulse modulation in real time. At the same time, the geometric feature matrix is ​​analyzed and physical gap features are extracted. These features quantitatively describe the geometric width distribution law (such as the width of the black gap) and the percentage trend of light emission energy loss of the splicing boundary gap.

[0040] The actual pixel spacing recorded in the physical gap features is divided by the theoretical pixel spacing to obtain the spacing deviation ratio. If the actual pixel spacing is greater than the theoretical pixel spacing, the spacing deviation ratio is used as the independent variable, and the corresponding energy compensation coefficient is calculated using a preset gain mapping function. Specifically, the preset gain mapping function adopts piecewise linear mapping logic, and the compensation coefficient is divided into three step intervals according to the size of the spacing deviation ratio: when the spacing deviation ratio is between 1.0 and 1.02, the mapping coefficient is set to 1.0, and a small gain compensation is performed proportionally; when the ratio is between 1.02 and 1.05, the mapping coefficient is switched to 1.8 to compensate for the rapid loss of light flux at the gap by increasing the slope; when the ratio exceeds 1.05, the mapping coefficient is further increased to 2.5 to forcibly cover the dark gap caused by the large spacing through high-rate current reference compensation.

[0041] The segmentation points (1.02, 1.05) and mapping coefficients (1.0, 1.8, 2.5) were pre-calibrated based on simulation data of luminous flux superposition of LED chips at different physical spacings. The specific calibration process was as follows: First, a Monte Carlo light trail simulation model based on Lambertian luminous characteristics was established. Physical parameters of specific LED chips (such as 1010 package size, phosphor layer refractive index, and chip center wavelength) were input to simulate the overlap of light energy fields between adjacent pixels at different spacings. Simulation results showed that when the deviation ratio exceeded 1.02, nonlinear collapse occurred in the edge light field overlap area, increasing the brightness attenuation slope; therefore, 1.8 was set as the first-level compensation slope. When the ratio exceeded 1.05, the total luminous flux per unit area dropped to below 85% of the normal level, triggering a 2.5x high-gain to maintain visual consistency.

[0042] To verify the accuracy of the simulation model, a distributed near-field spectral photometer and integrating sphere were used to perform actual light intensity measurements on the calibration samples. A calibration portability interface was provided for LED models with different pixel pitch specifications (e.g., P0.9, P1.2, P1.5) or from different brand suppliers: by inputting the rated brightness and half-value angle parameters of the model, the actual measured values ​​were fitted to the theoretical model using the least squares method, dynamically correcting the aforementioned segmentation points and mapping coefficients. This mechanism, based on a combination of physical simulation and actual measurement calibration, ensures the universality of compensation parameters under different display environments and LED batches.

[0043] Through the aforementioned segmented mapping, the spacing deviation ratio (e.g., 1.0375) is mapped to its corresponding mapping interval and combined with the positive linear correlation of luminous intensity, mapping the portion exceeding the theoretical spacing as a specific driving energy level increment. This increment is set as an energy compensation coefficient, serving as the driving power adjustment ratio. The energy compensation coefficient reflects the additional power intensity required due to the increased physical gap. Based on this driving power adjustment ratio, the driving power compensation weight of the affected lamp points at the splicing boundary is determined, and a power gain factor is generated.

[0044] The integrated power gain factor generation process involves the processor multiplying the energy compensation coefficient derived from piecewise linear mapping (as the first correction dimension) with the area sensitivity coefficient based on the coordinates of the affected lamp points (as the second correction dimension). To ensure the linearity of the hardware execution, this product result is processed by a preset power limiter, limiting its value to the adjustable range of the chip's constant current source reference (e.g., between 80% and 120% of the reference current). Finally, the integrated power gain factor is encapsulated as an 8-bit digital gain operator. This operator not only carries information about the energy increment due to the increased physical gap but also superimposes the address header of the driver chipset through bitwise operations. When this factor is applied to the constant current source reference terminal of the driver chip, it directly changes the equivalent potential across the external reference resistor or precisely fine-tunes the reference current value of the target output channel through the internal digital-to-analog converter, thereby achieving a complete closed loop from "physical gap calculation" to "physical-level current compensation".

[0045] Furthermore, the specific process of generating the light field reconstruction dimension includes: retrieving the power gain factor and the driving energy level distribution field; superimposing the power gain factor onto the driving energy level distribution field to perform energy level dynamic equalization processing and nonlinear mapping operation to generate an energy level correction matrix; and using the energy level correction matrix to reshape the emission envelope of the splicing boundary pixels to generate the light field reconstruction dimension.

[0046] The processor retrieves the power gain factor and the driving energy level distribution field. The power gain factor is superimposed onto the pixel coordinates recorded in the driving energy level distribution field, and dynamic energy level equalization is performed. Specifically, the process involves using the average driving energy level of pixels in the non-boundary region as the target benchmark, calculating the deviation between the energy level of each pixel at the boundary and the benchmark, and using the power gain factor to perform weighted compensation on this deviation, so that the driving energy level of the splicing boundary region reaches a preset equilibrium state consistent with the normal display area. This preset equilibrium state refers to adjusting the visual brightness of each pixel in the splicing boundary region to be consistent with the average brightness of the non-boundary display area (i.e., the center area of ​​the display screen).

[0047] Subsequently, a nonlinear mapping operation is performed. This operation aims to deeply couple the gray-level energy information contained in the driving energy level distribution field with the physical compensation weight contained in the power gain factor, generating the final energy level correction matrix. This process is achieved by: where the energy level correction matrix = the "correction parameter" raised to the power of (driving energy level distribution field × power gain factor) by the saturation truncation operator, constraining the calculation result within the peak range of a 16-bit integer to prevent pixel overflow caused by excessive compensation gain. Specifically, the calculation process involves: using a preset Gamma correction function, the balanced driving energy level is input into the correction model, which pre-stores hardware response curves obtained from actual measurements of different batches of LED chips. The correction parameter is dynamically optimized and calibrated within the range of 1.8 to 2.4 based on the actual electro-optical conversion efficiency of the LED chip; in this embodiment, the value of the correction parameter is 2.2. The calculation process includes: firstly, normalizing the driving energy level to obtain normalized energy level coefficients. Subsequently, the processor retrieves the hardware response mapping table of the current driver chip, uses the normalized energy level coefficient as the base, and performs a power function operation with a specific gamma value adapted to the response characteristics of the current LED chip as the exponent to calculate the theoretical brightness value that conforms to the linear perception of human vision. This dynamic adaptation method compensates for the nonlinear deviation of human vision in light intensity perception and the differences in physical response between different LED devices. After obtaining the theoretical brightness value, a quantization conversion from the floating-point domain to the integer domain is performed. To ensure that the mapped energy level correction matrix has a smooth grayscale transition, the processor uses error diffusion logic during the quantization process to distribute the remainder error generated by the quantization of the current pixel to the adjacent pixels to the right, below, and diagonally below in real time according to specific weights (e.g., 7 / 16, 3 / 16, 5 / 16, 1 / 16). Using the quantization principle of "rounding combined with error diffusion," the calculated high-precision theoretical brightness value is converted into a 16-bit integer grayscale value supported by the driver chip. For the tiny remainders generated during the quantization process, they are allocated to the pulse sequences of adjacent time steps for compensation according to a preset ratio, thereby ensuring that the quantized driving parameters can accurately reproduce the brightness details required by the nonlinear mapping while maintaining hardware compatibility.

[0048] Next, based on the response characteristics of the driver chip, the theoretical brightness value is converted into a correction value (i.e., the pulse width order between 0 and 65535 levels) to adapt to the grayscale grading requirements of the driver chip, generating an energy level correction matrix. The values ​​in this energy level correction matrix are quantized into integers to adapt to the 16-bit grayscale grading requirements of the driver chip, directly recording the pulse width parameters of each affected lamp point in the driver circuit.

[0049] Finally, the output current intensity of the driving chip at the splicing boundary is adjusted using an energy level correction matrix to reshape the equivalent emission envelope of the pixel lamps at the splicing boundary. This equivalent emission envelope is achieved by adjusting the light intensity distribution curves of adjacent lamps at the boundary, utilizing the principle of light energy superposition to form an energy compensation layer at the physical gap. This results in the physical connection of the light field energy at the splicing location, generating a reconstructed light field dimension with optical continuity characteristics.

[0050] By performing nonlinear coupling between the driving gain term and the energy level distribution field, a refined reshaping of the equivalent emission envelope at the splicing position is achieved. This method suppresses the visual tomography caused by the splicing gap at the optical distribution level, and ensures a smooth visual appearance of the spliced ​​area through the coordinated correction of the local light field morphology by the driving parameters.

[0051] Furthermore, the specific process of generating the pulse mapping matrix includes: acquiring the input display signal from the video source; extracting the pixel grayscale data contained in the input display signal; performing pulse width modulation hierarchical processing on the pixel grayscale data based on the light field reconstruction dimension; calculating the pulse width adjustment parameters at the corresponding positions of the splicing boundary pixel regions; and generating the pulse mapping matrix.

[0052] Specifically, during the video content display process, the processor receives the input display signal from the video source in real time and extracts the pixel grayscale data contained therein (usually 8-bit or 10-bit integers). To achieve fine-tuning, the original 8-bit pixel grayscale data is first linearly mapped to the 16-bit high-precision grayscale space supported by the driver chip.

[0053] Subsequently, the light intensity distribution compensation data (usually floating-point gain coefficients) recorded within the light field reconstruction dimension are called to perform pulse width modulation hierarchical processing.

[0054] The pulse width modulation hierarchical processing includes: dividing the display period of a single pixel into multiple non-uniformly weighted temporal sub-intervals based on the energy increment determined by the optical field reconstruction dimension, and inserting compensation pulses in the temporal sub-intervals according to the principle of high-frequency random distribution.

[0055] Specifically, after acquiring the 16-bit light intensity distribution compensation data output from the light field reconstruction dimension, the processor does not concentrate the compensation energy at the end of the cycle. Instead, it refines the 1 / 1920-second display refresh cycle (i.e., a complete frame cycle) into multiple time-domain slices with unequal weights. The number of time-domain slices is determined based on the balance between the depth of the shift register inside the driver chip and the data transmission bandwidth. In this embodiment, it is set to 16. Since mainstream LED driver chips have a limited number of logic flips that can be carried in a single refresh cycle at clock frequencies of 30MHz to 50MHz, if the number of slices is too small (e.g., 8), the energy cannot be effectively distributed, which can easily produce low-frequency visual flicker. If the number of slices is too large (e.g., 32), it will exceed the logic throughput limit of the chip, resulting in grayscale overflow. The 16 slices ensure energy discretization while perfectly matching the segmented control logic of 16-bit PWM modulation.

[0056] Because conventional PWM modulation of equal-width pulses generates strong odd-order harmonic electromagnetic interference, a nonlinear decreasing sequence with a golden ratio of 0.618 is used to ensure that the fundamental frequencies of each slice do not overlap, maximizing the broadening of the energy spectral density from an information theory perspective. This asymmetric architecture makes the spectral distribution at the pulse edges exponentially smooth, forcing the interference energy originally concentrated in the low-frequency band to the high-frequency background noise, thus reducing the EMI amplitude at its source.

[0057] The process of utilizing a decreasing sequence based on the golden ratio of 0.618 involves: using the duration of the first temporal slice as a baseline, the duration of subsequent slices decreases progressively by 0.618, thus forming an asymmetric distribution architecture that is wider at the beginning and narrower at the end in the temporal domain. This division method gives the compensation energy a natural nonlinear characteristic at the microscopic time scale, effectively breaking the energy superposition at fixed frequencies. When a 5% increase in brightness gain is detected due to a stitching gap, the algorithm does not directly extend the original pulse, but instead generates a bias sequence that conforms to a uniform distribution characteristic through a pseudo-random number generator. The seed of the random number generator is initialized based on the pixel coordinate hash value of the current frame, ensuring that the pulse distribution points of each frame and each pixel are different. The algorithm breaks down this 5% extra conduction time, using an extremely short pulse of 1 / 32 (meaning the pulse width is only 1 / 32 of the total length of a single time-domain slice) as the smallest compensation unit. This pulse is randomly distributed within the high-frequency band above 500kHz in the time-domain slice sequence. The 500kHz frequency limit is determined based on the transient light contrast sensitivity curve of the human eye; that is, brightness pulsations above this frequency completely exceed the sampling threshold of the human optic nerve, ensuring absolute visual smoothness at low grayscale. This randomized high-frequency distribution ensures that during low grayscale compensation, the physical switching action of the LED points is spread across a wide frequency band in the frequency domain, preventing the formation of continuous low-frequency peaks in the time domain, thus avoiding electromagnetic interference and visual flicker caused by pulse broadening. Through non-uniform time-domain slicing and high-frequency random pulse interpolation, a smooth superposition of minute energy increments is achieved, effectively suppressing display flicker in low-brightness compensation scenarios and improving the static image stability at splicing boundaries.

[0058] Next, the pulse width adjustment parameters at the corresponding positions of the splicing boundary pixel regions are calculated to generate a pulse mapping matrix.

[0059] The calculation process includes: based on the light intensity distribution compensation amount and the pixel grayscale data at the video source, performing the following steps: First, the processor upsamples the pixel grayscale data and maps it to the 16-bit (0-65535) high-precision dynamic range space corresponding to the internal logic of the driver chip. Then, it performs a product operation on the mapped grayscale data using the light intensity distribution compensation in the light field reconstruction dimension. If the original grayscale of a pixel at a stitching boundary is and the corresponding reconstruction compensation gain (e.g., 1.08) is obtained, the calculated target display level is the original grayscale multiplied by the reconstruction compensation gain.

[0060] Obtain the current refresh rate (e.g., 3840Hz) and clock frequency parameters of the driver chip. Convert the calculated target display energy level into a specific on-clock count value. Specific quantization process: Based on the driver chip's PWM control logic, divide the display period of a single pixel into 2... 16The pulse width adjustment parameter is equal to the total number of clock ticks corresponding to the target display energy level. For example, if the target display energy level is calculated to be 32768, then the conduction time of this pixel in one display cycle is strictly locked to 50%. Before finally generating the pulse width adjustment parameter, overflow verification and minimum pulse width limitation processing are first performed on the parameter to avoid the dead zone problem caused by the driver transistor not being able to conduct completely due to the pulse being too narrow. If the calculated value exceeds the bit width of the driver chip register, truncation or normalization processing is performed to ensure that each adjustment parameter is within the effective value range that can be executed by the hardware. Subsequently, the processor, based on the physical scanning architecture of the display screen and the topology relationship between row drive and column drive, fills the calculated pulse width adjustment parameter values ​​of all splicing boundary pixels into a two-dimensional data structure according to the coordinate index. By encapsulating the pulse width adjustment parameter of each row of affected pixels into a row of data packets and aggregating multiple data packets in scanning order, a pulse mapping matrix is ​​generated. This matrix serves as the underlying instruction stream and is directly sent to the register group of the driver chip through a high-speed differential bus (such as LVDS or Mini-LVDS).

[0061] By deeply coupling real-time video grayscale with the preset light field reconstruction dimension, synchronous response to the input signal content and physical splicing characteristics is achieved. A pulse width modulation conversion mechanism ensures precise allocation of display parameters in the time dimension, guaranteeing visual balance of the dynamically spliced ​​image at physical boundary positions.

[0062] Furthermore, the specific process of generating a visual smoothing dimension based on the output error collaborative gain includes: converting the pulse mapping matrix into a duty cycle adjustment sequence composed of the proportion of light emission conduction time of the driving chip within the pulse modulation period; parsing the duty cycle adjustment sequence and outputting the error collaborative gain; and performing dynamic pulse compensation on the pixel lamp points at the splicing boundary through the error collaborative gain to offset the light emission energy loss caused by physical mechanical errors, suppress visual dark seams, and generate a visual smoothing dimension.

[0063] Specifically, during the operation of the splicing screen display control process, the processor retrieves the aforementioned pulse mapping matrix and maps the recorded pulse width values ​​to the driver chip register. The duty cycle adjustment sequence is encapsulated using a standardized data frame structure. The control data for each pixel is defined as a 32-bit (4-byte) data packet: the first 16 bits record a high-precision count value of the duty cycle pulse width, and the last 16 bits contain the physical register address mapping of the driver chip and control flag bits. The data packets are arranged in little-endian byte order and use 16-bit cyclic redundancy check to ensure transmission accuracy in complex electromagnetic environments. At the communication protocol level, high-speed transmission is performed based on a differential bus, with the data transmission bit rate set between 600Mbps and 1.2Gbps to match the bandwidth requirements of high refresh rates. In terms of writing timing, the processor triggers a data synchronization pulse in the frame blanking area to ensure that the duty cycle adjustment sequences of all channels take effect synchronously at the start of the next frame. After receiving the sequence, the driver chip directly maps it to its internal constant current source driver register to control the duty cycle performance of the LEDs during the pulse modulation period. The duty cycle adjustment sequence reflects the proportion of the light-emitting conduction time of the driver chip in one refresh cycle to the total cycle.

[0064] During the output error collaborative gain process, the offset of the duty cycle adjustment sequence relative to the preset standard duty cycle is analyzed, and the incremental value used to correct the brightness loss of physical gaps is extracted, i.e., the output error collaborative gain. Subsequently, the error collaborative gain is sent to the execution end of the drive circuit corresponding to the splicing boundary. The execution end completes dynamic pulse compensation by extending the duration of the light-emitting conduction pulse. The additional brightness generated by the dynamic pulse compensation effectively offsets the luminous flux loss per unit area at the splicing position caused by physical mechanical errors. The visual dark seams generated at the physical gaps are suppressed by the real-time enhancement of light-emitting energy, ultimately enabling the entire splicing area to achieve a consistent brightness display state, generating visual effect smoothness dimension data characterizing the smooth distribution of the full-screen visual effect.

[0065] As a preferred embodiment, the generation of visual effect smoothing dimension further includes: calculating the brightness gradient difference between the splicing boundary pixel and the adjacent non-boundary pixel according to the error collaborative gain, and performing spatial convolution smoothing processing on the amplitude of the dynamic pulse compensation based on the brightness gradient difference to generate visual effect smoothing dimension data characterizing the continuity of the full-screen brightness distribution.

[0066] Specifically, after determining the error co-gain gain, the processor does not apply this gain constantly to all boundary pixels. Instead, it analyzes the brightness gradient between the boundary pixels and the surrounding normal pixels, using the center of the stitching seam as the axis. In this embodiment, based on the co-optimization result of the stitching boundary influence radius and computational overhead, a 3×3 spatial convolution kernel is used to weighted smooth the pulse compensation amplitude of each pixel. The weight allocation of the spatial convolution kernel follows a Gaussian distribution curve, and its standard deviation is set between 0.5 and 1.2 to control the smooth span of brightness transition. The selection of this range is based on the stimulated spatial frequency under the pixel physical spacing and typical viewing distance. When the pixel spacing is small (such as small-pitch LEDs below P0.9), a lower standard deviation (such as 0.5-0.8) is selected to shorten the brightness transition bandwidth and prevent excessive diffusion of compensation energy from causing image edge blurring. When the pixel spacing is large or the viewing distance is close, the standard deviation is increased (such as 0.9-1.2) to widen the transition band to match the low-frequency cutoff characteristics of the human eye's contrast sensitivity function, thereby visually eliminating "hard edges". For boundary pixels at the seam edges, a mirror-fill method is used to process the convolution boundaries. This involves copying the compensation features of neighboring pixels to participate in the computation, ensuring that the brightness at the convolutional edges does not experience abrupt energy drops. The gradient decrease of the compensation energy level follows an exponential decay law with the center of the seam as the peak value. For example, when the center pixel receives an 8% duty cycle enhancement, the algorithm uses a Gaussian kernel-based weighted mapping to give first-order adjacent pixels approximately 5% enhancement (approximately 0.6 times the peak value), and second-order adjacent pixels approximately 2% enhancement (approximately 0.25 times the peak value), until the energy increment converges to zero and is perfectly aligned with the average brightness of the entire screen. This processing method generates visual smoothing dimension data that records the final luminous energy level of each coordinate point within the entire screen after a smooth transition. Through this non-linearly decreasing energy level mapping from the seam core to the periphery, the brightness jump between the compensation area and the regular display area is effectively eliminated, ensuring that the compensation action does not introduce new visual hard edges at the physical boundaries, thus guaranteeing the optical continuity of the entire display area. By introducing a spatial gradient convolution algorithm to construct a visual smoothing dimension, the problem of brightness connection between the compensation area and the regular area was successfully solved, achieving seamless visual integration of the splicing boundary in the spatial dimension and improving the overall sense of the displayed image.

[0067] By mapping logic pulses to the duty cycle control logic of the underlying driver chip, precise hardware-level compensation for energy loss at the splicing position is achieved. Dynamic pulse adjustment is used to suppress visual dark seams caused by physical gaps, ensuring a high degree of visual uniformity between the splicing area and the regular display area.

[0068] This invention solves the problem of visual discontinuity at the boundaries caused by traditional splicing screens relying solely on physical calibration by establishing a full-link digital mapping model of "physical gap - observation angle - driving energy level". It achieves synchronous compensation for displayed content and physical defects, effectively suppressing visual flicker and "dark seam" residue in dynamic images. It also enables sub-pixel-level precise adjustment of the hardware output current. This method significantly enhances the visual coherence of the splicing area, making the large splicing screen appear as a seamless, smooth logical display plane, greatly improving the image quality of high-end displays.

[0069] Example 2 Reference Figure 3 This embodiment provides a specific application scenario for a local adjustment system for an LED display screen, specifically in the automated assembly manufacturing and factory calibration environment of an ultra-high-definition display panel.

[0070] At the end of the ultra-high-definition LED display panel production line, the LED display local adjustment system operates within an industrial control computer, communicating with automated mechanical assembly stations and optical measurement stations via a high-speed data interface. This system intelligently compensates for splicing gaps caused by mechanical assembly tolerances through the coordinated operation of a geometric feature extraction module, a drive energy field reconstruction module, and a pulse sequence execution module.

[0071] The geometric feature extraction module extracts the theoretical pixel pitch of the display manufacturing design from the automated mechanical assembly station using a high-precision charge-coupled device (CCD) measurement sensor. Simultaneously, it detects the actual physical coordinates of the center point of the pixel at the splicing boundary to obtain the actual pixel pitch value. The module then uses subtraction logic to obtain the displacement of the actual pixel pitch from the theoretical pixel pitch, generating a mechanical error operator reflecting the physical geometric deviation. At the optical measurement station, the module acquires the luminous intensity data of the pixel at the splicing boundary at a non-perpendicular viewing angle and the luminous intensity data at a standard viewing angle in the vertical plane. By comparing the luminous intensity values ​​at the two locations, the module outputs a field-of-view deviation feature describing the attenuation law of the viewing angle light intensity. The mechanical error operator and the field-of-view deviation feature undergo nonlinear fusion processing in the spatial coordinate system to generate a geometric feature matrix encompassing both physical position information and viewing angle optical information.

[0072] The driving energy field reconstruction module retrieves the geometric feature matrix from the geometric feature extraction module. Based on the physical arrangement of the pixel lamps at the splicing boundary in the display driving circuit, the driving energy field reconstruction module obtains the logical coordinates of the control pixels corresponding to the control area of ​​the driving circuit. The driving energy field reconstruction module maps the displacement deviation recorded by the mechanical error operator and the light intensity attenuation vector recorded by the field of view deviation feature to the logical coordinates of the control pixels, determines the energy adjustment amount of the target pixel lamps, and establishes a driving energy level distribution field in the control area of ​​the driving circuit. The driving energy field reconstruction module analyzes the geometric feature matrix to extract physical gap features reflecting the physical width of the gap and the trend of energy loss. Based on the physical gap features, the driving energy field reconstruction module calculates the driving power ratio of the lamps affected by the splicing boundary, determines the driving power compensation weight, and integrates to generate a power gain factor for current reference adjustment. The driving energy field reconstruction module superimposes the power gain factor onto the driving energy level distribution field to perform nonlinear mapping, reshapes the emission envelope of the boundary pixel lamps, and generates a light field reconstruction dimension with continuous spatial light field characteristics.

[0073] The pulse sequence execution module processes the input display signal acquired from the video signal generator based on the light field reconstruction dimension. It extracts the pixel grayscale data from the input display signal and, combined with the light intensity distribution compensation requirements recorded in the light field reconstruction dimension, performs pulse width modulation (PWM) hierarchical processing. The module calculates the pulse conduction count at the corresponding coordinate positions of the pixel region at the splicing boundary, generating a pulse mapping matrix. It then converts the pulse logic recorded in the pulse mapping matrix into a duty cycle adjustment sequence composed of the proportion of light emission conduction time of the driver chip within the pulse modulation period, outputting a quantized error coordination gain. This error coordination gain is sent to the underlying display driver circuit via a synchronous serial port. The driver chip applies the error coordination gain to perform dynamic pulse compensation on the pixel lamps at the splicing boundary, extending the light emission conduction time to offset the energy loss caused by physical mechanical errors. Visual dark seams generated at the physical splicing position are suppressed under the light field reconstruction logic. The LED display local adjustment system ultimately achieves visual alignment between the splicing boundary region and the regular display area, generating a smooth visual effect dimension.

[0074] By coupling mechanical error operators with field-of-view deviation characteristics to generate a geometric feature matrix, a substantial match between manufacturing physical tolerances and viewing angle attenuation characteristics is achieved. This solves the brightness banding problem caused by mechanical splicing spacing deviation. By utilizing the driving energy level distribution field and power gain factor to perform light field reconstruction and reshape the luminous envelope, a deep synergy between driving energy distribution and physical loss trends is achieved. Light energy compensation is performed on splicing seams, suppressing visual dark seams at the splicing location and ensuring the optical continuity of the display panel under the influence of mechanical manufacturing defects. Dynamic pulse compensation is performed by combining a duty cycle adjustment sequence generated by pulse hierarchical processing, ensuring real-time coupling between the video signal and the physical compensation logic. Mapping the logic grayscale to the precise on-time of the driving chip within the pulse modulation cycle ensures the consistency of the overall display visual effect.

[0075] Example 3 This embodiment provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the LED display screen local adjustment method.

[0076] Specifically, the computer-readable storage medium provides a storage carrier for the computer program implementing the local adjustment method for LED displays. Using this medium, the computer program can be transferred and installed between different display driver control terminals. When the computer program is executed by a processor, the processor can accurately execute the entire action flow, including extracting manufacturing deviation description parameters, generating geometric feature matrices, reshaping the light-emitting envelope, and adjusting the output duty cycle sequence. This storage method ensures that, in complex splicing screen application scenarios, different display driver hardware can achieve the function of suppressing physical seams by executing the same computer program.

[0077] The computer program stored on the computer-readable storage medium, during execution, can intelligently align parameters to address manufacturing defects in the display screen, ensuring that different display devices achieve the same level of visual smoothness. This approach not only facilitates the promotion of local adjustment technology but also guarantees the stability of the overall visual effect of the display screen, resolving the impact of physical and mechanical errors on display quality.

[0078] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for local adjustment of an LED display screen, characterized in that, include: The theoretical pixel pitch and the actual pixel pitch at the center point of the splicing boundary pixels are obtained during the display manufacturing process, and the pitch difference is calculated to generate a mechanical error operator; By comparing the luminance data of the stitched boundary pixels under non-perpendicular viewing angle and standard viewing angle, the field of view deviation feature is output. The mechanical error operator and the field of view deviation feature are fused in the spatial coordinate system to generate a geometric feature matrix; Based on the geometric feature matrix, a driving energy level distribution field is established in the driving circuit control area; according to the physical gap features contained in the geometric feature matrix, the driving power compensation weight of the affected lamp points at the boundary is calculated to generate a power gain factor; based on the power gain factor, a nonlinear mapping operation is performed on the driving energy level distribution field to reshape the light emission envelope of the spliced ​​boundary pixels and generate the light field reconstruction dimension. Based on the optical field reconstruction dimension, pulse width modulation hierarchical processing is performed on the input display signal to generate a pulse mapping matrix; The pulse mapping matrix is ​​converted into a duty cycle adjustment sequence at the driver chip level, and the output error cooperative gain is generated. Dynamic pulse compensation is performed on the pixel lights at the stitching boundary based on the error collaborative gain to generate a visually smooth dimension.

2. The method for local adjustment of an LED display screen according to claim 1, characterized in that, The specific process of generating the mechanical error operator includes: extracting the theoretical pixel spacing determined by the display design file; obtaining the actual pixel spacing of the center point of the pixel at the splicing boundary of the display screen; calculating the spacing displacement vector that deviates from the theoretical pixel spacing of the actual pixel spacing; and mapping the spatial displacement vector into a mechanical error operator that quantifies the physical splicing deviation.

3. The method for local adjustment of an LED display screen according to claim 1, characterized in that, The specific process of generating a geometric feature matrix from the output field of view deviation features includes: acquiring the luminous intensity data of the stitching boundary pixels at the non-perpendicular viewing angle and the luminous intensity data at the standard viewing angle; comparing the distribution difference between the luminous intensity data at the non-perpendicular viewing angle and the luminous intensity data at the standard viewing angle, and outputting the field of view deviation features; and fusing the mechanical error operator and the field of view deviation features in the spatial coordinate system to generate a geometric feature matrix.

4. The method for local adjustment of an LED display screen according to claim 1, characterized in that, The specific process of establishing the driving energy level distribution field and generating the power gain factor includes: extracting the mechanical error operator and field deviation features contained in the geometric feature matrix, obtaining the control pixel logic coordinates corresponding to the driving circuit control area, mapping the mechanical error operator and field deviation features to the control pixel logic coordinates, and generating driving energy level quantization parameters; aggregating the driving energy level quantization parameters to generate the driving energy level distribution field; analyzing the geometric feature matrix to obtain the physical gap features characterizing the distribution state and energy loss trend of the splicing boundary gaps; determining the driving power compensation weights of the affected lamp points at the boundary based on the physical gap features, and integrating the driving power compensation weights to generate the power gain factor.

5. The method for local adjustment of an LED display screen according to claim 1, characterized in that, The specific process of generating the light field reconstruction dimension includes: retrieving the power gain factor and the driving energy level distribution field; superimposing the power gain factor onto the driving energy level distribution field to perform energy level dynamic equalization processing and nonlinear mapping operation to generate the energy level correction matrix; and using the energy level correction matrix to reshape the emission envelope of the splicing boundary pixels to generate the light field reconstruction dimension.

6. The method for local adjustment of an LED display screen according to claim 1, characterized in that, The specific process of generating the pulse mapping matrix includes: acquiring the input display signal from the video source; extracting the pixel grayscale data contained in the input display signal; performing pulse width modulation hierarchical processing on the pixel grayscale data based on the light field reconstruction dimension; calculating the pulse width adjustment parameters at the corresponding positions of the splicing boundary pixel regions; and generating the pulse mapping matrix.

7. The method for local adjustment of an LED display screen according to claim 1, characterized in that, The specific process of generating a visual smoothing dimension based on the output error collaborative gain includes: converting the pulse mapping matrix into a duty cycle adjustment sequence composed of the proportion of light emission conduction time of the driving chip within the pulse modulation period; parsing the duty cycle adjustment sequence and outputting the error collaborative gain; and performing dynamic pulse compensation on the pixel lamp points at the splicing boundary through the error collaborative gain to offset the light emission energy loss caused by physical mechanical errors, suppress visual dark seams, and generate a visual smoothing dimension.

8. A local adjustment system for an LED display screen, characterized in that, include: Geometric feature extraction module: used to obtain the theoretical pixel spacing and the actual pixel spacing of the center point of the splicing boundary pixels during the display manufacturing process, and calculate the spacing difference to generate a mechanical error operator; By comparing the luminance data of the stitched boundary pixels under non-perpendicular viewing angle and standard viewing angle, the field of view deviation feature is output. Couple the mechanical error operator with the field of view deviation feature to generate a geometric feature matrix; Driven energy field reconstruction module: Based on the geometric feature matrix, a drive energy level distribution field is established in the drive circuit control area; according to the physical gap features contained in the geometric feature matrix, the drive power compensation weight of the affected lamp points at the boundary is calculated to generate a power gain factor; based on the power gain factor, a nonlinear mapping operation is performed on the drive energy level distribution field to reshape the light emission envelope of the spliced ​​boundary pixels and generate the light field reconstruction dimension. Pulse sequence execution module: Based on the optical field reconstruction dimension, it performs pulse width modulation hierarchical processing on the input display signal to generate a pulse mapping matrix; The pulse mapping matrix is ​​converted into a duty cycle adjustment sequence at the driver chip level, and the output error cooperative gain is generated. Dynamic pulse compensation is performed on the pixel lights at the stitching boundary based on the error collaborative gain to generate a visually smooth dimension.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the local adjustment method according to any one of claims 1-7.

Images