Screen color uniformity on-line detection system based on dynamic rotating polarizing spectrum

Abstract

This invention relates to the field of optoelectronic detection technology and discloses an online detection system for screen color uniformity based on dynamic rotating polarization spectrum. The system includes an online transmission module for carrying and driving the illuminated screen under test to move continuously along a set direction at a set speed; and a combined optical encoding module, positioned above the screen under test, comprising a static spatial-spectral phase delay mask and a continuously rotating polarization modulator arranged sequentially along the beam propagation direction. By driving the screen to continuously translate using the online transmission module, and cooperating with the continuously rotating polarization modulator and hyperspectral acquisition module to acquire dynamic spatiotemporal aliasing data, the accuracy of the spectral dimension can be guaranteed. Furthermore, by extracting high-precision chromaticity parameter matrices and multidimensional polarization distortion feature matrices from the decoupled and reconstructed optimal polarization spectral tensor solution, the system's ability to detect complex internal quality problems can be significantly enhanced.

Description

Technical Field

[0001] This invention relates to the field of photoelectric detection technology, specifically to an online detection system for screen color uniformity based on dynamic rotating polarization spectrum. Background Technology

[0002] With the rapid iteration of high-end display technologies (such as OLED and Micro-LED), smart devices have placed extremely high demands on the color uniformity and optical quality of display panels. In the panel manufacturing process, variations in the coating thickness of multilayer micro-nano optical films (such as polarizers and phase retardation films) or the internal residual stress generated during the encapsulation process often lead to hidden color unevenness on the screen.

[0003] Traditional two-dimensional colorimeters or photometers can only collect macroscopic light intensity and colorimetric information, making it difficult to accurately trace and detect optical distortions caused by microscopic thin-film structures or stress. Therefore, the industry has gradually introduced optical detection technologies that combine hyperspectral and polarization state analysis, aiming to achieve accurate assessment of deep optical uniformity defects in screens by acquiring the complete three-dimensional colorimetric space and four-dimensional Stokes polarization parameters of the light field.

[0004] However, the current conventional approach to obtaining the polarization state of a target is to install a mechanically stepped polarizing device at the front end of the spectral detector. This requires the target object to remain absolutely still during measurement to ensure that the same physical coordinate point can undergo multiple polarization shots from different angles. However, on a high-speed conveyor line with continuous transmission, the screen is in a state of continuous translation. If accurate data is to be obtained, the conveyor line must frequently perform the tedious action of "moving-hovering-rotating polarizer-shooting-moving again". This extremely slow detection cycle caused by the forced alignment of physical space and time makes the technology limited to offline sampling inspection and completely loses the engineering feasibility of online full inspection.

[0005] To avoid the hovering waiting time caused by mechanical movement, some existing technologies attempt to use electro-optic modulators based on electro-birefringence (such as liquid crystal retarder retarders, LCVRs) to replace mechanical rotation mechanisms, attempting to achieve pseudo-dynamic polarization acquisition through millisecond-level voltage switching. However, such purely electro-controlled modulation materials inevitably suffer from severe spectral dispersion effects, meaning that the polarization phase delays they produce for beams of different wavelengths are difficult to keep consistent. This can easily lead to huge nonlinear errors in the calculated polarization characteristics when performing broadband spectral analysis on the entire visible light spectrum emitted by the screen, thus affecting the accuracy of the spectral dimensions. Therefore, this invention designs an online screen color uniformity detection system based on dynamic rotating polarization spectrum to address the aforementioned problems. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides an online screen color uniformity detection system based on dynamic rotating polarization spectrum. This system solves the problem that existing optical detection technologies suffer from severe physical signal aliasing caused by spatial displacement, time passage, and polarization angle switching on a high-speed production line where the screen under test is continuously translated, thus making it impossible to extract complete broadband polarization spectral features with high precision for online color uniformity assessment.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an online screen color uniformity detection system based on dynamic rotating polarization spectrum, comprising:

[0008] The online transmission module is used to carry and drive the screen under test, which is lit up, to move continuously along a set direction at a set speed.

[0009] The joint optical coding module, located above the screen under test, includes a static spatial-spectral phase delay mask and a continuously rotating polarizing modulator arranged sequentially along the beam propagation direction, for jointly performing spatial-wavelength phase coding and time-polarization continuous amplitude modulation on the optical signal emitted by the screen under test.

[0010] The hyperspectral acquisition module is set on the outgoing optical path of the joint optical coding module and is used to continuously expose and acquire the modulated optical signal to obtain the observation tensor containing spatiotemporal and polarization aliasing features.

[0011] The data processing decoupling module is communicatively connected to the hyperspectral acquisition module. It is used to construct a topological dictionary prior based on the physical sub-pixel arrangement of the screen under test, and to construct a spatiotemporal decoupling objective function in combination with the modulation parameters of the joint optical coding module. Through optimization, the observation tensor is asynchronously matrix decoupled to reconstruct a four-dimensional polarization spectral tensor that eliminates spatiotemporal motion ambiguity.

[0012] The color uniformity evaluation module is communicatively connected to the data processing decoupling module. It is used to extract the chromaticity parameter matrix and polarization distortion feature matrix based on the four-dimensional polarization spectral tensor, and to comprehensively determine and output the color uniformity defect distribution of the screen under test.

[0013] Preferably, the static spatial-spectral phase delay mask has a non-uniformly distributed phase delay and fast axis azimuth angle in the first spatial dimension and spectral wavelength dimension, which is used to add independent optical phase encoding for incident beams with different microscopic spatial positions and different wavelengths, and generate a mask transformation matrix accordingly.

[0014] Preferably, the continuously rotating polarizing modulator is controlled by a drive motor to rotate continuously at a constant angular velocity, which is used to convert its transmission axis azimuth angle into a polarization transformation matrix that changes continuously with time, thereby dynamically polarizing the light signal passing through the static spatial-spectral phase delay mask.

[0015] Preferably, the hyperspectral acquisition module is a linear array hyperspectral imager that performs line scan exposure at a constant sampling frequency; the data processing decoupling module has a preset spatiotemporal binding mapping relationship, which linearly binds the global continuous time of the system to the longitudinal physical coordinates of the screen under test driven by the online transmission module according to the set speed, thereby establishing the spatiotemporal coordinate system of the observation tensor.

[0016] Preferably, the data processing decoupling module performs the following configuration when constructing the topology dictionary prior:

[0017] Read the design size, center distance, and black matrix distribution rules of the luminescent subpixels of the screen under test, and construct a binary physical mapping dictionary matrix;

[0018] A spatial structure weight tensor is generated based on the binary physical mapping dictionary matrix, such that the value of the spatial structure weight tensor inside the effective luminous region of the sub-pixel approaches zero, and the value of the tensor at the physical boundary of the luminous region and the non-luminous black matrix region approaches infinity.

[0019] Preferably, the data processing decoupling module encapsulates a global observation operator before constructing the spatiotemporal decoupling objective function;

[0020] The global observation operator is a forward physical projection model constructed by performing matrix joint operations on the projection measurement parameters of the hyperspectral acquisition module, the polarization transformation matrix of the continuously rotating polarizer, the mask transformation matrix of the static spatial-spectral phase delay mask, and the spatiotemporal binding mapping relationship.

[0021] Preferably, the spatiotemporal decoupling objective function constructed by the data processing decoupling module is an optimization evaluation function that combines data fidelity terms and prior constraint penalty terms;

[0022] The data fidelity term is used to measure the degree of error approximation between the polarization spectral tensor to be reconstructed and the observed tensor through a global observation operator;

[0023] The prior constraint penalty term is used to apply a weighted constraint penalty to the spatial gradient of the polarization spectrum tensor to be reconstructed using the spatial structure weight tensor, so as to force the reconstructed solution to strictly conform to the physical emission boundary of the screen under test in terms of spatial physical structure.

[0024] Preferably, the data processing decoupling module uses the alternating direction multiplier method to optimize the spatiotemporal decoupling objective function. By introducing auxiliary variables and Lagrange multipliers, it performs iterative convergence calculations in computer memory and outputs the optimal polarization spectral tensor solution containing four Stokes parameter components.

[0025] Preferably, the color uniformity evaluation module includes the following features when extracting:

[0026] The zeroth-order Stokes component is extracted from the optimal polarization spectral tensor solution, and integral operations are performed within a set band using the standard observer color matching function to obtain the chromaticity parameter matrix.

[0027] The three higher-order Stokes components in the optimal polarization spectral tensor solution are extracted, and the polarization distortion feature matrix, including the degree of linear polarization, the degree of circular polarization, and the azimuth angle of linear polarization, is obtained by calculation in the polarization vector space.

[0028] Preferably, the color uniformity evaluation module includes the following components in the comprehensive judgment:

[0029] The first spatial derivative of the extracted chromaticity parameter matrix is ​​calculated, and a joint threshold determination is performed in conjunction with the polarization distortion feature matrix.

[0030] The identified chromaticity anomaly two-dimensional coordinates and polarization mapping anomaly two-dimensional coordinates are logically combined to mark the physical coordinates of non-uniform defects caused by internal stress or film thickness variation in the screen under test, and the quality inspection results are generated.

[0031] This invention provides an online screen color uniformity detection system based on dynamic rotating polarization spectrum. It has the following advantages:

[0032] 1. This invention drives the screen to continuously translate through an online transmission module, and works with a continuous rotating polarizer and a hyperspectral acquisition module to acquire dynamic spatiotemporal aliasing data, thereby achieving true online detection without interruption of the production line and with continuous screen transmission, thus ensuring the accuracy of the spectral dimension.

[0033] 2. This invention extracts high-precision chromaticity parameter matrix and multidimensional polarization distortion feature matrix from the decoupled and reconstructed optimal polarization spectral tensor solution, and performs joint thresholding and logical union judgment on these two types of feature matrices with different physical meanings. This enables the system to detect ordinary backlight or emission brightness unevenness, significantly enhancing the system's ability to detect complex internal quality problems.

[0034] 3. This invention constructs a sub-pixel topology dictionary prior by reading the specific process specifications of the screen under test, and generates a spatial structure weight tensor for forcibly constraining the boundary between the light-emitting area and the black matrix based on this prior. It abandons the solution method in traditional computational photography that simply relies on the general smoothing assumption, so that the evolution direction of matrix decoupling can accurately fit the real microscopic semiconductor light-emitting array law of the display panel, fundamentally suppressing cross-pixel spatial crosstalk caused by screen movement, and ensuring the authenticity of the reconstructed polarization spectral signal in the spatial microstructure.

[0035] 4. This invention integrates a global observation operator, which includes hyperspectral acquisition parameters, mask transformation matrix, and polarization transformation matrix, with prior penalties to construct a spatiotemporal decoupling objective function. It then uses optimization algorithms such as the alternating direction multiplier method for iterative solution. This successfully transforms the physical bottleneck of traditional detection, which is limited by the hardware switching speed, into the advantage of asynchronous tensor inverse operation in the backend software. Without the need for complex and expensive high-speed tracking mechanical structures, the complete Stokes polarization parameter, which eliminates motion blur, is stably calculated through hardware and software collaboration.

[0036] 5. This invention innovatively introduces a static spatial-spectral phase delay mask into the joint optical coding module. For incident beams with different microscopic spatial positions and different wavelengths, independent and non-uniform phase delay and fast axis azimuth are pre-applied. This enables high-dimensional physical joint coding of spatial and wavelength information at the physical optics front end. This effectively avoids the physical degradation and loss of light intensity and polarization characteristics during continuous dynamic scanning, and provides a solid physical observation foundation for the back end to reverse decouple multidimensional polarization parameters in an underdetermined state. Attached Figure Description

[0037] Figure 1 This is a schematic diagram of the system architecture of the present invention;

[0038] Figure 2 This is one of the system flow diagrams of the present invention;

[0039] Figure 3 This is the second schematic diagram of the system flow of the present invention;

[0040] Figure 4 This is the third system flow diagram of the present invention;

[0041] Figure 5 This is the fourth system flow diagram of the present invention;

[0042] Figure 6 This is the fifth system flowchart of the present invention. Detailed Implementation

[0043] The technical solutions in 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.

[0044] Please see the appendix Figure 1 -Appendix Figure 6 This invention provides an online screen color uniformity detection system based on dynamic rotating polarization spectrum, comprising:

[0045] The online transmission module is used to carry and drive the screen under test, which is lit up, to move continuously along a set direction at a set speed.

[0046] The joint optical coding module, located above the screen under test, includes a static spatial-spectral phase delay mask and a continuously rotating polarizing modulator arranged sequentially along the beam propagation direction. It is used to jointly perform spatial-wavelength phase coding and time-polarization continuous amplitude modulation on the optical signal emitted from the screen under test.

[0047] The hyperspectral acquisition module is located on the outgoing optical path of the joint optical coding module and is used to continuously expose and acquire the modulated optical signal to obtain the observation tensor containing spatiotemporal and polarization aliasing features.

[0048] The data processing decoupling module communicates with the hyperspectral acquisition module to construct a topological dictionary prior based on the physical sub-pixel arrangement of the screen under test, and constructs a spatiotemporal decoupling objective function by combining the modulation parameters of the joint optical coding module. Through optimization, the observation tensor is asynchronously matrix decoupled to reconstruct a four-dimensional polarization spectral tensor that eliminates spatiotemporal motion ambiguity.

[0049] The color uniformity evaluation module is connected to the data processing decoupled module. It is used to extract the chromaticity parameter matrix and polarization distortion feature matrix based on the four-dimensional polarization spectral tensor, and to comprehensively determine and output the color uniformity defect distribution of the screen under test.

[0050] Specifically, in actual operation, the screen under test (such as OLED, Micro-LED and other display panels) in the lit state is transported into the detection area by the online transmission module. At this time, the light field emitted by the screen surface not only contains traditional spectral energy information, but also contains complex polarization state characteristics caused by the screen's multi-layer optical thin films (such as polarizers, phase retardation films and other materials) and internal stress. Before being captured by the camera, these composite light signals first pass through the joint optical encoding module and are subjected to optical aliasing and modulation at the physical level. Then, they are recorded as discrete observation data by the hyperspectral acquisition module. Finally, the data processing decoupling module and the color uniformity evaluation module reconstruct the real polarization characteristics in the computer through reverse mathematical derivation and parameter extraction, and complete the defect judgment.

[0051] The static spatial-spectral phase delay mask has a non-uniformly distributed phase delay and fast axis azimuth angle in the first spatial dimension and spectral wavelength dimension. It is used to add independent optical phase encoding for incident beams with different microscopic spatial positions and wavelengths, and generate a mask transformation matrix accordingly. The continuously rotating polarizing modulator is controlled by a drive motor to rotate continuously at a constant angular velocity. It is used to convert its transmission axis azimuth angle into a polarization transformation matrix that changes continuously with time, thereby dynamically polarizing the light signal passing through the static spatial-spectral phase delay mask. The hyperspectral acquisition module is a linear array hyperspectral imager, which performs line scan exposure at a constant sampling frequency. The data processing decoupling module has a preset spatiotemporal binding mapping relationship. According to the set speed, the global continuous time of the system is linearly bound to the longitudinal physical coordinates of the screen under test driven by the online transmission module, thereby establishing the spatiotemporal coordinate system of the observation tensor.

[0052] Specifically, the online transmission module has a precise servo drive unit, configured to carry the edge of the screen under test. Axial direction at a constant speed The hyperspectral acquisition module employs a linear array hyperspectral imager, which performs continuous translational motion. Its photosensitive pixels are distributed in a one-dimensional array along the horizontal direction (X-axis) and sampled at a constant linear scanning frequency. During continuous exposure, a global continuous clock is established during system initialization. Because the screen is constantly shifting, at any given scanning moment... Physical row coordinates captured by the linear hyperspectral imager Satisfy linear binding mapping relationship This binding relationship strictly maps a one-dimensional time series to the physical space coordinates of the vertical direction of the screen being tested;

[0053] Modulation principle of static spatial-spectral phase delay mask (SSPR mask):

[0054] Along the beam propagation path, the first stage of the joint optical coding module is a static spatial-spectral phase delay mask. This mask surface is etched with micro / nano optical structures, which are located in the first spatial dimension (X-axis) and the spectral wavelength dimension (…). The phase delay is discrete and non-uniformly distributed. and fast axis azimuth In continuous detection, without preprocessing the incident light, polarized light of different wavelengths and spatial positions will completely aliased. This system uses a mask to pre-add independent optical phase codes to incident beams of different microscopic spatial positions and wavelengths. Let the Stokes vector of the incident light be... Then the mask plate is equivalent to a space-wavelength dependent mask transformation matrix (Muller matrix). ;

[0055] After passing through the mask, the light beam enters the second stage of the joint optical coding module—a continuously rotating polarizing modulator, which is controlled by a brushless drive motor at a constant angular velocity. Continuous rotation around the system's optical axis at any time Its transmission axis azimuth angle This transforms the polarization transformation matrix (Mueller matrix) of the device into a time-varying matrix. Continuously changing periodic function Through the combined action of a mask and a rotating polarizer, the original three-dimensional composite optical signal is modulated and projected onto a linear hyperspectral imager. The projection measurement parameter of the imager (the sensor only responds to light intensity) is denoted as... Thus far, the imager has captured the aliased observation tensor. It contains strongly coupled spatiotemporal and polarization characteristics.

[0056] When constructing the topology dictionary prior, the data processing decoupling module executes the following configuration:

[0057] The design dimensions, center distances, and black matrix distribution rules of the luminous subpixels of the screen under test are read to construct a binary physical mapping dictionary matrix. Based on this matrix, a spatial structure weight tensor is generated, such that its value approaches zero within the effective luminous region of the subpixel and approaches infinity at the physical boundaries of the luminous region and in the non-luminous black matrix region. Before constructing the spatiotemporal decoupling objective function, the data processing decoupling module encapsulates a global observation operator. This global observation operator is a forward physical projection model constructed by performing matrix joint operations on the projection measurement parameters of the hyperspectral acquisition module, the polarization transformation matrix of the continuously rotating polarizer, the mask transformation matrix of the static spatial-spectral phase delay mask, and the spatiotemporal binding mapping relationship. The spatiotemporal decoupling objective function constructed by the decoupling module is an optimization evaluation function that combines a data fidelity term and a prior constraint penalty term. The data fidelity term is used to measure the degree of error approximation between the polarization spectral tensor to be reconstructed and the observed tensor through a global observation operator. The prior constraint penalty term is used to apply a weighted constraint penalty to the spatial gradient of the polarization spectral tensor to be reconstructed using a spatial structure weight tensor, so as to force the reconstructed solution to strictly conform to the physical luminescence boundary of the screen under test in terms of spatial physical structure. The data processing decoupling module uses the alternating direction multiplier method to optimize the spatiotemporal decoupling objective function. By introducing auxiliary variables and Lagrange multipliers, it performs iterative convergence calculations in computer memory and outputs the optimal polarization spectral tensor solution containing four Stokes parameter components.

[0058] Specifically, the data processing decoupling module is the part of this system that enables the inverse recovery of four-dimensional polarization spectral characteristics (horizontal coordinates) from a one-dimensional strongly aliased signal. Vertical coordinates ,wavelength The core calculation unit for polarization state parameters;

[0059] Topological dictionary priors and structure weight tensors: due to Coordinates and Time In this case, the traditional point-by-point multi-angle static polarization solution equation fails due to severe underdeterminacy. To overcome the limitation of non-full rank matrices, this invention utilizes the physical characteristics of the screen under test as a "discrete light-emitting array" to construct a priori model. The system pre-reads the process specifications of the screen under test and establishes a binary physical mapping dictionary matrix. This dictionary precisely describes the physical size, pitch, and non-divergent matrix distribution of sub-pixels (such as R, G, and B sub-pixels), and generates a spatial structure weight tensor based on this. The assignment principle of this tensor is: its value approaches a minimum (zero) within the effective light-emitting area of ​​the sub-pixel, while its value approaches infinity at the physical boundary of the light-emitting pixel and in the black matrix region.

[0060] Construction of the global observation operator and objective function: Define the global four-dimensional polarization spectrum tensor to be reconstructed as follows: Project the camera vector Polarization transformation matrix Mask Real Matrix and spatiotemporal binding mapping relationship Unified encapsulation as a global observation operator Therefore, the spatiotemporal decoupling objective function is constructed as follows:

[0061] ;

[0062] In this function, the first term is the data fidelity term, which uses the Frobenius norm to measure the reconstructed tensor's resemblance to the actual observed tensor after passing through the forward physics model. The residual; the latter term is the prior constraint penalty term, using the Hadamard product ( Combining spatial structure weight tensors and two-dimensional spatial difference operator The reconstructed spectral signal strictly conforms to the physical topological emission boundary of the screen in terms of spatial distribution.

[0063] ADMM Algorithm Optimization Solution: For the aforementioned large-scale and non-smooth convex optimization objective function, the system employs the Alternating Direction Multiplier Method (ADMM) in the data processing and understanding module, introducing auxiliary variables. And Lagrange multipliers (dual variables) An augmented Lagrangian function is constructed. In computer memory, three sub-steps—S-update, Z-update (soft threshold shrinkage), and U-update—are executed alternately in a loop until the calculated residuals satisfy a preset convergence threshold. The final output is the optimal solution. This means that the motion ambiguity is eliminated and the physical boundaries are clear in the complete Stokes polarization spectral tensor.

[0064] The color uniformity assessment module extracts features by: extracting the zero-order Stokes component from the optimal polarization spectral tensor solution, performing integration within a set band using the standard observer color matching function to obtain the chromaticity parameter matrix; extracting the three higher-order Stokes components from the optimal polarization spectral tensor solution, and obtaining polarization distortion feature matrices including linear polarization degree, circular polarization degree, and linear polarization azimuth angle through polarization vector space calculations. The color uniformity assessment module performs comprehensive judgment by: calculating the first-order spatial derivative of the extracted chromaticity parameter matrix and performing joint threshold judgment in conjunction with the polarization distortion feature matrix; and performing logical union processing on the identified chromaticity anomaly two-dimensional coordinates and polarization mapping anomaly two-dimensional coordinates to mark the physical coordinate positions of non-uniformity defects caused by internal stress or film thickness variations in the screen under test, and generating quality inspection result output.

[0065] Specifically, obtaining the reconstructed solution Subsequently, the color uniformity evaluation module performs specific industrial quality inspection judgments and extracts multi-dimensional feature parameters:

[0066] Chromaticity parameter matrix extraction: Extracting the zeroth-order Stokes component representing the absolute spectral power distribution. Combined with the International Commission on Illumination (CIE) standard observer color matching function By performing integral operations in the visible light band, a high-precision chromaticity parameter matrix (such as the CIE XYZ spatial coordinate matrix) of the entire screen area can be calculated.

[0067] Polarization distortion feature matrix extraction: Extracting higher-order Stokes components corresponding to the polarization state. , and The degree of linear polarization is obtained by calculating using the polarization optics vector space formula. Circular polarization degree and linear polarization azimuth angle To form a polarization distortion feature matrix.

[0068] Comprehensive Judgment and Result Output: Since conventional brightness or chromaticity uniformity defects (such as backlight unevenness) are manifested as gradual numerical changes in the chromaticity parameter matrix, while latent color non-uniformity defects (color mura) caused by uneven stress release or film thickness variation in the internal structure of the screen often manifest as abrupt changes in the polarization parameter, the system performs first-order spatial derivative calculation on the chromaticity parameter matrix to find gradient anomaly regions; at the same time, it sets a benchmark deviation threshold for the polarization distortion feature matrix for judgment, and performs logical union processing of the two recognition results in two-dimensional coordinates. Finally, the system accurately locates and marks the two-dimensional physical coordinates of the non-uniformity defects caused by film thickness variation or stress, and generates quality inspection results containing defect category and location mapping, which are output to the upstream production line control end.

[0069] In summary, this invention provides an online screen color uniformity detection system based on dynamic rotating polarized light spectrum. By driving continuous translational motion of the screen through an online transmission module, and cooperating with a continuous rotating polarized light modulator and a hyperspectral acquisition module for dynamic spatiotemporal aliasing data acquisition, it completely breaks the inertial hardware logic of existing high-precision polarization spectral detection, which relies on a stepper motor's "move-stop-shoot-move again" process. This achieves true online detection without interruption of the production line and with continuous screen transmission, thus ensuring the accuracy of the spectral dimension. Furthermore, by extracting high-precision chromaticity parameter matrices and multidimensional polarization distortion feature matrices (linear polarization degree, circular polarization degree, and azimuth angle) from the decoupled and reconstructed optimal polarization spectral tensor solution, and performing joint thresholding and logical union judgment on these two types of feature matrices with different physical meanings, the system can not only detect conventional backlight or luminous brightness unevenness, but also accurately capture hidden color non-uniformity defects that are difficult to detect by traditional pure chromaticity analysis, such as microscopic stress unevenness and film thickness variation in the multi-layer thin films inside the panel. This significantly enhances the system's ability to detect complex internal quality problems.

[0070] 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 screen color uniformity online detection system based on dynamic rotating polarized light spectrum, characterized in that, include: The online transmission module is used to carry and drive the screen under test, which is lit up, to move continuously along a set direction at a set speed. The joint optical coding module, located above the screen under test, includes a static spatial-spectral phase delay mask and a continuously rotating polarizing modulator arranged sequentially along the beam propagation direction, for jointly performing spatial-wavelength phase coding and time-polarization continuous amplitude modulation on the optical signal emitted by the screen under test. The hyperspectral acquisition module is set on the outgoing optical path of the joint optical coding module and is used to continuously expose and acquire the modulated optical signal to obtain the observation tensor containing spatiotemporal and polarization aliasing features. The data processing decoupling module is communicatively connected to the hyperspectral acquisition module. It is used to construct a topological dictionary prior based on the physical sub-pixel arrangement of the screen under test, and to construct a spatiotemporal decoupling objective function in combination with the modulation parameters of the joint optical coding module. Through optimization, the observation tensor is asynchronously matrix decoupled to reconstruct a four-dimensional polarization spectral tensor that eliminates spatiotemporal motion ambiguity. The color uniformity evaluation module is communicatively connected to the data processing decoupling module. It is used to extract the chromaticity parameter matrix and polarization distortion feature matrix based on the four-dimensional polarization spectral tensor, and to comprehensively determine and output the color uniformity defect distribution of the screen under test.

2. The screen color uniformity online detection system based on dynamic rotating polarization spectrum according to claim 1, characterized in that, The static spatial-spectral phase delay mask has a non-uniformly distributed phase delay and fast axis azimuth angle in the first spatial dimension and spectral wavelength dimension. It is used to add independent optical phase encoding for incident beams with different microscopic spatial positions and different wavelengths, and generate a mask transformation matrix accordingly.

3. The screen color uniformity online detection system based on dynamic rotating polarized light spectrum according to claim 1, characterized in that, The continuously rotating polarizing modulator is controlled by a drive motor to rotate continuously at a constant angular velocity. It is used to convert the azimuth angle of its transmission axis into a polarization transformation matrix that changes continuously with time, thereby dynamically polarizing the light signal passing through the static spatial-spectral phase delay mask.

4. The screen color uniformity online detection system based on dynamic rotating polarization spectrum according to claim 1, characterized in that, The hyperspectral acquisition module is a linear array hyperspectral imager that performs line scanning exposure at a constant sampling frequency. The data processing decoupling module has a preset spatiotemporal binding mapping relationship. This mapping relationship linearly binds the global continuous time of the system with the longitudinal physical coordinates of the screen under test driven by the online transmission module according to the set speed, thereby establishing the spatiotemporal coordinate system of the observation tensor.

5. The screen color uniformity online detection system based on dynamic rotating polarization spectrum according to claim 1, characterized in that, When constructing the topology dictionary prior, the data processing decoupling module performs the following configuration: Read the design size, center distance, and black matrix distribution rules of the luminescent subpixels of the screen under test, and construct a binary physical mapping dictionary matrix; A spatial structure weight tensor is generated based on the binary physical mapping dictionary matrix, such that the value of the spatial structure weight tensor inside the effective luminous region of the sub-pixel approaches zero, and the value of the tensor at the physical boundary of the luminous region and the non-luminous black matrix region approaches infinity.

6. The screen color uniformity online detection system based on dynamic rotating polarization spectrum according to claim 1, characterized in that, Before constructing the spatiotemporal decoupling objective function, the data processing decoupling module encapsulates a global observation operator. The global observation operator is a forward physical projection model constructed by performing matrix joint operations on the projection measurement parameters of the hyperspectral acquisition module, the polarization transformation matrix of the continuously rotating polarizer, the mask transformation matrix of the static spatial-spectral phase delay mask, and the spatiotemporal binding mapping relationship.

7. The screen color uniformity online detection system based on dynamic rotating polarization spectrum according to claim 1, characterized in that, The spatiotemporal decoupling objective function constructed by the data processing decoupling module is an optimization evaluation function that combines data fidelity terms and prior constraint penalty terms. The data fidelity term is used to measure the degree of error approximation between the polarization spectral tensor to be reconstructed and the observed tensor through a global observation operator; The prior constraint penalty term is used to apply a weighted constraint penalty to the spatial gradient of the polarization spectrum tensor to be reconstructed using the spatial structure weight tensor, so as to force the reconstructed solution to strictly conform to the physical emission boundary of the screen under test in terms of spatial physical structure.

8. The screen color uniformity online detection system based on dynamic rotating polarization spectrum according to claim 7, characterized in that, The data processing decoupling module uses the alternating direction multiplier method to optimize the spatiotemporal decoupling objective function. By introducing auxiliary variables and Lagrange multipliers, it performs iterative convergence calculations in computer memory and outputs the optimal polarization spectral tensor solution containing four Stokes parameter components.

9. The screen color uniformity online detection system based on dynamic rotating polarization spectrum according to claim 1, characterized in that, The color uniformity evaluation module includes the following features during feature extraction: The zeroth-order Stokes component is extracted from the optimal polarization spectral tensor solution, and integral operations are performed within a set band using the standard observer color matching function to obtain the chromaticity parameter matrix. The three higher-order Stokes components in the optimal polarization spectral tensor solution are extracted, and the polarization distortion feature matrix, including the degree of linear polarization, the degree of circular polarization, and the azimuth angle of linear polarization, is obtained by calculation in the polarization vector space.

10. The screen color uniformity online detection system based on dynamic rotating polarization spectrum according to claim 1, characterized in that, The color uniformity evaluation module includes the following aspects in the comprehensive judgment: The first spatial derivative of the extracted chromaticity parameter matrix is ​​calculated, and a joint threshold determination is performed in conjunction with the polarization distortion feature matrix. The identified chromaticity anomaly two-dimensional coordinates and polarization mapping anomaly two-dimensional coordinates are logically combined to mark the physical coordinates of non-uniform defects caused by internal stress or film thickness variation in the screen under test, and the quality inspection results are generated.

Images