Detection method of peep-proof light guide plate and related equipment

Abstract

The invention discloses a detection method of a peep-proof light guide plate and related equipment. The detection method comprises the following steps: writing a self-disappearing phase grid on the surface of a base material through roller coating of a photosensitive phase change coating; on the basis of the phase reference, carrying out same-grid-point superposition processing on a guided-mode resonance spectrum in a laser etching process to obtain dynamic data of the depth and the inclination angle of the microprism; the iron-gallium alloy nano-wire network is used for capturing stress distribution and executing time-space synchronous mapping with a thermal radiation gray field; performing topology persistent coherence dimension reduction modeling on the multi-physics field coupling data to generate a dynamic privacy view field distribution diagram; and performing regional fine adjustment control on the process parameters and triggering a phase grid elimination reaction. According to the scheme, real-time monitoring and accurate control of the whole manufacturing process of the peep-proof light guide plate can be achieved, direct association between technological parameter fluctuation and final optical performance in the microstructure forming process is established, the hysteresis problem of traditional off-line detection is solved, the product yield and performance stability are remarkably improved, and meanwhile material and energy waste is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of real-time detection and quality control of optical elements, and in particular to a detection method for an anti-peep light guide plate and related equipment. Background Art

[0002] The anti-peep light guide plate is a special optical component used in electronic device displays. Its core function is to limit the viewing angle of the screen content through a precisely designed microstructure, so that only viewers from the front can clearly see the displayed content, while viewers from the side cannot obtain the screen information, thereby effectively protecting the privacy and security of users. The anti-peep light guide plate is mainly composed of three parts: a precision microstructure layer, which is composed of a micron or submicron prism array. These prisms have a specific depth, inclination angle and period, and are the core component for achieving directional light control; an optical substrate layer, usually made of high-transmittance polycarbonate or acrylic material, which provides support for the microstructure and ensures light transmission efficiency; functional coatings, including anti-glare, anti-fingerprint and hardened coatings, enhance product durability and user experience. The performance quality of the anti-peep light guide plate directly depends on the processing accuracy and assembly quality of these components, so strict quality inspections must be carried out to ensure that it meets the design specifications and application requirements.

[0003] Currently, the industry's testing of anti-peep light guide plates is mainly concentrated in the offline darkroom testing stage after production is completed, that is, using dedicated optical instruments to perform angle scanning, brightness measurement, and light leakage analysis on the finished product. This back-end testing method has a fundamental flaw: the detection is completely separated from the production process, and it is impossible to establish a direct correlation between the fluctuations in process parameters during the microstructure molding and bonding process and the deviation in the final optical performance. In actual production, instantaneous fluctuations in process parameters such as laser etching energy density, imprint cavity pressure, UV curing shrinkage, and bonding tension will have a complex impact on the depth, inclination, period, and distribution of the microprism, which in turn determines the anti-peep effect and display quality of the finished product. However, the existing offline detection mode can only detect problems but cannot prevent them, resulting in a large number of unqualified products being screened out after completing the entire process, causing serious waste of materials and energy. Summary of the Invention

[0004] The main purpose of the present invention is to solve the technical problems of the existing offline detection method of anti-peep light guide plates, in which the detection and production processes are completely separated and a direct correlation between process parameter fluctuations and final optical performance deviations cannot be established.

[0005] A first aspect of the present invention provides a method for detecting an anti-peep light guide plate, the method comprising: By rolling a photosensitive phase change coating and writing a planar moiré fringe pattern under a low-energy mask light field, an optical phase grating writing process is performed on the surface of the optical substrate before entering the micro-engraving station to obtain a phase reference data matrix with dual-dimensional calibration of spatial coordinates and energy dose; Based on the phase reference data matrix, a grid point superposition process is performed on the guided mode resonance spectrum sequence obtained during the laser etching process to obtain a resonance spectrum feature data set that characterizes the dynamic evolution of the microprism depth and tilt angle; Based on the resonance spectrum characteristic data set, a time-space synchronous mapping process is performed on the Barkhausen transition signal sequence and the thermal radiation grayscale field generated by the iron-gallium alloy nanowire grid formed at the root of the microprism groove to obtain a multi-physics field coupling data set representing the stress-geometry-optical correlation; Based on the multi-physics field coupling data set, a topological persistent coherence dimensionality reduction modeling process is performed on the privacy viewing angle half parallax angle, brightness half-decay width, and lateral stray light threshold to obtain a dynamic privacy viewing angle field distribution map representing the continuous distribution of the panel surface; According to the dynamic privacy viewing angle field distribution diagram, regional fine-tuning control is implemented on the laser pulse energy, imprinting pressure and curing dosage, and the optical phase grating elimination reaction is triggered to obtain the quality judgment result of the anti-peep light guide plate.

[0006] Preferably, the process of roll-coating a photosensitive phase change coating and writing a planar moiré fringe pattern under a low-energy mask light field, performing an optical phase grating writing process on the surface of the optical substrate before entering the microlithography station, and obtaining a phase reference data matrix calibrated in two dimensions of spatial coordinates and energy dose, comprises: Performing temperature gradient measurement processing on the surface of the optical substrate to obtain a temperature compensation parameter array distributed along the conveying direction; Performing roller coating thickness equalization processing on the photosensitive phase change coating according to the temperature compensation parameter array to obtain coating thickness distribution data; According to the coating thickness distribution data, the photosensitive phase change coating is exposed to a low-energy mask light field and a phase modulation writing process is performed to obtain a planar moiré fringe phase image containing biaxial periodic information; According to the plane moiré fringe phase image, performing phase difference calibration processing on the position coordinates of the substrate surface to obtain a spatial coordinate mapping matrix; According to the spatial coordinate mapping matrix and the energy absorption dose-phase response curve of the photosensitive phase change coating, the coating is subjected to energy dose calibration processing to obtain a phase reference data matrix.

[0007] Preferably, the guided mode resonance spectrum sequence obtained during the laser etching process is subjected to same-grid point superposition processing based on the phase reference data matrix to obtain a resonance spectrum feature data set characterizing the dynamic evolution of the microprism depth and tilt angle, including: Performing time window segmentation processing on the guided mode resonance spectrum sequence obtained during the laser etching process to obtain a spectrum frame set containing etching displacement indexes; According to the phase reference data matrix, performing grid point coordinate mapping and intensity superposition processing on the spectrum frame set to obtain an aligned spectrum grid frame set; According to the aligned spectral grid frame set, frequency shift vector extraction processing is performed on the cross-time spectrum line of the same grid point to obtain a depth-inclination change parameter matrix; A multi-order sequence fitting process is performed according to the depth-tilt variation parameter matrix and the corresponding time index to obtain a resonance spectrum feature data set.

[0008] Preferably, the method performs spatiotemporal synchronous mapping processing on the Barkhausen transition signal sequence and the thermal radiation grayscale field generated by the iron-gallium alloy nanowire grid formed at the root of the microprism groove based on the resonance spectrum characteristic data set to obtain a multi-physics field coupling data set characterizing the stress-geometry-optical correlation, including: The Barkhausen transition signal sequence generated by the iron-gallium alloy nanowire grid formed at the root of the microprism groove is subjected to time segmentation processing to obtain a magnetoelastic signal frame set containing etching displacement index; performing grid point coordinate alignment and amplitude superposition processing on the magnetoelastic signal frame set according to the resonance spectrum feature data set to obtain an aligned magnetoelastic grid frame set; Perform position-time synchronous sampling processing on the thermal radiation grayscale field to obtain a temperature grayscale frame set; According to the aligned magnetoelastic grid frame set, the temperature grayscale frame set is subjected to same-grid point registration processing to obtain a thermal-magnetic combined grid frame set; According to the thermal-magnetic joint grid frame set, a co-correlation quantization process is performed on the magnetoelastic amplitude, temperature grayscale and depth-dip parameters to obtain a stress-geometry-optical coupling matrix; A multi-dimensional joint fusion process is performed based on the stress-geometry-optical coupling matrix and the spatial temperature gradient matrix calculated from the temperature grayscale frame set to obtain a multi-physics field coupling data set.

[0009] Preferably, the step of performing co-correlation quantization processing on the magnetoelastic amplitude, temperature grayscale, and depth-dip parameters according to the thermal-magnetic joint grid frame set to obtain a stress-geometry-optical coupling matrix includes: Normalization is performed on the magnetoelastic amplitude, temperature grayscale and depth-inclination parameters to obtain a set of standardized feature tensors; According to the standardized characteristic tensor set, performing directional wave packet coherence calculation processing on the magnetoelastic amplitude and depth-dip parameters to obtain an anisotropic coherence spectrum matrix; Performing multi-core information entropy mapping processing according to the anisotropic coherence spectrum matrix and the temperature grayscale parameter to obtain a cross-domain weight tensor; A Tucker decomposition and fusion process is performed on the cross-domain weight tensor and the standardized feature tensor set to obtain the stress-geometry-optical coupling matrix.

[0010] Preferably, the method performs topological persistent coherence dimensionality reduction modeling on the privacy viewing angle half parallax angle, brightness half-width, and lateral stray light threshold based on the multi-physics field coupling data set to obtain a dynamic privacy viewing angle field distribution map representing the continuous distribution of the panel surface, including: Performing privacy performance parameter extraction processing on the multi-physics field coupling data set to obtain a privacy parameter tensor set including a half-parallax angle, a brightness half-width, and a lateral stray light threshold; Performing multi-scale filter complex construction processing on the panel grid adjacency relationship according to the privacy parameter tensor set to obtain a filter complex sequence; According to the filter complex sequence, performing bar coding calculation processing on the persistence of the homology group to obtain a persistence bar matrix; According to the persistent strip matrix, performing a center-of-gravity embedding mapping process on the topological skeleton nodes to obtain a privacy perspective vector field; According to the privacy viewing angle vector field, continuous splicing processing is performed on the panel position coordinates to obtain a dynamic privacy viewing angle field distribution map.

[0011] Preferably, performing a gravity center embedding mapping process on the topological skeleton nodes according to the persistent strip matrix to obtain a privacy perspective vector field includes: Performing scale gating on the persistent bar matrix to obtain a threshold filtered bar set; Filter the bar set according to the threshold value, perform Laplace eigenmapping processing on the topological skeleton nodes, and obtain a skeleton spectral coordinate matrix; According to the skeleton spectral coordinate matrix, vector interpolation processing is performed on the panel grid coordinates to obtain a privacy perspective vector field.

[0012] Preferably, the method of implementing regional fine-tuning control on the laser pulse energy, imprinting pressure, and curing dosage according to the dynamic privacy viewing angle field distribution diagram, and triggering the optical phase grating elimination reaction to obtain the quality judgment result of the anti-peep light guide plate includes: Perform error vector field extraction processing on the dynamic privacy viewing angle field distribution map to obtain the fine-tuning instruction matrix of the coverage panel grid; According to the fine-tuning instruction matrix, a space-time solution process is performed on the laser pulse energy sequence, the imprint pressure sequence and the curing dose sequence to obtain a synchronous parameter modulation profile; According to the synchronization parameter modulation profile, regional modulation processing is performed on the roll-to-roll process time axis, and an optical phase grid elimination reaction is synchronously triggered at the corresponding plate grid to obtain a process execution record; The privacy viewing angle field snapshot is recalculated according to the process execution record, and a tolerance comparison process is performed with the original dynamic privacy viewing angle field distribution map to obtain a quality determination result.

[0013] A second aspect of the present invention provides a detection device for an anti-peep light guide plate, the detection device for an anti-peep light guide plate comprising: The phase grating writing module is used to write a planar moiré fringe pattern by roll-coating a photosensitive phase change coating and writing it under a low-energy mask light field. This module performs an optical phase grating writing process on the surface of the optical substrate before entering the micro-engraving station, thereby obtaining a phase reference data matrix calibrated in two dimensions: spatial coordinates and energy dose. A resonance spectrum analysis module is used to perform a same-grid point superposition process on the guided mode resonance spectrum sequence obtained during the laser etching process based on the phase reference data matrix to obtain a resonance spectrum feature data set that characterizes the dynamic evolution of the microprism depth and tilt angle; a multi-physics field coupling module for performing spatiotemporal synchronous mapping processing on the Barkhausen transition signal sequence and the thermal radiation grayscale field generated by the embedded Fe-Ga alloy nanowire according to the resonance spectrum characteristic data set, thereby obtaining a multi-physics field coupling data set representing the stress-geometry-optical correlation; A topological dimensionality reduction modeling module is used to perform topological persistent coherence dimensionality reduction modeling on the privacy viewing angle half parallax angle, brightness half-width, and lateral stray light threshold based on the multi-physics field coupling data set to obtain a dynamic privacy viewing angle field distribution map representing the continuous distribution of the panel surface; The adaptive control module is used to implement regional fine-tuning control of laser pulse energy, imprinting pressure and curing dosage according to the dynamic privacy viewing angle field distribution map, and trigger the optical phase grating elimination reaction to obtain the quality judgment result of the anti-peep light guide plate.

[0014] The third aspect of the present invention provides a detection device for an anti-peep light guide plate, comprising: a memory and at least one processor, wherein instructions are stored in the memory, and the memory and the at least one processor are interconnected through a line; the at least one processor calls the instructions in the memory so that the detection device for the anti-peep light guide plate performs the steps of the above-mentioned anti-peep light guide plate detection method.

[0015] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the steps of the above-mentioned method for detecting an anti-peep light guide plate.

[0016] During the manufacturing process of anti-peep light guide plates, the existing detection methods have the problem of complete separation of detection and production processes, resulting in the inability to establish a direct correlation between process parameter fluctuations during microstructure molding and bonding and final optical performance deviations. Problems can only be discovered after production is completed but cannot be prevented, resulting in serious waste of materials and energy.

[0017] The anti-peep light guide plate inspection method proposed in this invention aims to address the aforementioned issues. Before the optical substrate enters the microlithography station, the method implements an optical phase grating writing process by roll-coating a photosensitive phase change coating and writing a planar moiré pattern under a low-energy mask light field. This process generates a phase reference data matrix calibrated in both spatial coordinates and energy dose dimensions. This step creates an in-situ reference coordinate system that disappears with the production process, enabling precise comparison and correlation of inspection data from subsequent stages within a unified spatial and energy dose reference system.

[0018] During the laser etching process, the obtained guided mode resonance spectrum sequence is subjected to a grating point superposition process based on the aforementioned phase reference data matrix to generate a resonance spectral feature dataset that characterizes the dynamic evolution of the microprism depth and tilt angle. The periodic structure formed by the microprism immediately generates a waveguide grating coupled cavity. When illuminated by appropriate incident light, a resonant transmission peak determined by the microprism's geometric parameters appears. Because the resonance peak has subnanometer sensitivity to changes in microstructure depth and tilt angle, this step enables real-time, high-precision monitoring of the microstructure geometry formation process, far exceeding the accuracy of traditional scattering measurement methods without affecting the quality of the final product.

[0019] The method then performs spatiotemporal synchronous mapping of the Barkhausen transition signal sequence and thermal radiation grayscale field generated by the iron-gallium alloy nanowire grid embedded at the root of the microprism groove to construct a multi-physics field coupling dataset that characterizes the stress-geometry-optical correlation. Iron-gallium alloys have excellent magnetoelastic properties. When the local stress in the microstructure changes, detectable Barkhausen transitions are generated, accompanied by thermoelastic coupled radiation. By correlating these signals with the aforementioned spectral features, the method achieves precise monitoring of the evolution of stress distribution within the microstructure, revealing the complex correlation mechanism between stress field, geometric parameters, and optical properties, a microscopic physical process that is inaccessible to traditional detection methods.

[0020] Based on a multi-physics coupled dataset, this method performs topologically persistent homology dimensionality reduction modeling on key performance parameters such as the privacy viewing angle half-parallax angle, brightness half-width, and lateral stray light threshold, generating a dynamic privacy viewing angle field distribution map representing the continuous distribution of the panel surface. Topologically persistent homology theory can extract essential geometric features from high-dimensional complex data, significantly reducing the data dimension while preserving key topological structure information. This process establishes a direct mapping between microscopic physical parameters and macroscopic privacy protection performance, enabling the system to predict final product performance based on real-time monitoring data, thereby achieving preventive quality control.

[0021] Finally, based on the dynamic privacy viewing angle field distribution map, the method implements regional fine-tuning of laser pulse energy, imprint pressure, and curing dose, while simultaneously triggering the optical phase grating elimination reaction to complete the quality assessment process for the anti-peep light guide plate. This step implements prediction-based multi-parameter coordinated control, converting privacy performance prediction results into precise process parameter adjustment instructions, allowing timely intervention in the microstructure formation process to ensure that the final product performance meets design requirements. Furthermore, the self-vanishing nature of the optical phase grating ensures that the inspection process does not affect the final product quality.

[0022] This approach fundamentally addresses the hysteresis problem of traditional offline inspection by establishing a unified spatiotemporal reference system, enabling real-time multi-physics monitoring of the microstructure formation process, mapping microscopic parameters to macroscopic performance, and implementing prediction-based closed-loop control. This approach closely integrates the inspection process with the production process, enabling the system to promptly detect and correct defects before they escalate, significantly improving product yield and reducing material and energy waste. The accumulation of multi-physics data also provides a scientific basis for process optimization and product development. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying any creative work.

[0024] Figure 1 Schematic diagram of an embodiment of a method for detecting an anti-peep light guide plate according to an embodiment of the present invention; Figure 2 Schematic diagram of an embodiment of a detection device for an anti-peep light guide plate according to an embodiment of the present invention; Figure 3 Schematic diagram of an embodiment of a detection device for an anti-peep light guide plate in an embodiment of the present invention.

[0025] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0027] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative position relationship, movement status, etc. between the various components under a certain specific posture. If the specific posture changes, the directional indications will also change accordingly.

[0028] In addition, the descriptions of "first", "second", etc. in the present invention are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, "and / or" in the full text includes three solutions. Taking A and / or B as an example, it includes technical solution A, technical solution B, and technical solution that satisfies both A and B. In addition, the technical solutions between the various embodiments can be combined with each other, and must be based on the ability of ordinary technicians in this field to implement. When the combination of technical solutions is mutually contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0029] An embodiment of the present application provides a method for detecting an anti-peep light guide plate. Figure 1 A flow chart of a method for detecting a privacy-preventing light guide plate according to an embodiment of the present application. In this embodiment, the method includes: See also Figure 1 By rolling a photosensitive phase change coating and writing a planar moiré fringe pattern under a low-energy mask light field, an optical phase grating writing process is performed on the surface of the optical substrate before entering the micro-engraving station to obtain a phase reference data matrix with dual-dimensional calibration of spatial coordinates and energy dose; In one embodiment of the present invention, the process of roll-coating a photosensitive phase change coating and writing a planar moiré fringe pattern under a low-energy mask light field, performing an optical phase grating writing process on the surface of the optical substrate before entering the microlithography station, and obtaining a phase reference data matrix calibrated in two dimensions of spatial coordinates and energy dose, comprises: Performing temperature gradient measurement processing on the surface of the optical substrate to obtain a temperature compensation parameter array distributed along the conveying direction; Performing roller coating thickness equalization processing on the photosensitive phase change coating according to the temperature compensation parameter array to obtain coating thickness distribution data; According to the coating thickness distribution data, the photosensitive phase change coating is exposed to a low-energy mask light field and a phase modulation writing process is performed to obtain a planar moiré fringe phase image containing biaxial periodic information; According to the plane moiré fringe phase image, performing phase difference calibration processing on the position coordinates of the substrate surface to obtain a spatial coordinate mapping matrix; According to the spatial coordinate mapping matrix and the energy absorption dose-phase response curve of the photosensitive phase change coating, the coating is subjected to energy dose calibration processing to obtain a phase reference data matrix.

[0030] The following is a detailed description of the steps involved in the above embodiment: The temperature gradient on the optical substrate surface can be measured using an array of infrared thermal imagers. Specifically, along the roll-to-roll production line, an infrared thermal imager is installed every 20 cm along the optical substrate conveyor direction, forming a temperature monitoring network covering the entire production area. Each thermal imager has a temperature resolution of 0.1°C and a sampling frequency of 100Hz, enabling real-time capture of substrate surface temperature changes. The system collects temperature data at multiple points along the substrate conveyor direction and uses a spline interpolation algorithm to generate a continuous temperature distribution curve, calculating the temperature gradient between each point. These temperature gradients are then converted into temperature compensation parameters, forming an array of temperature compensation parameters distributed along the conveyor direction. This temperature gradient measurement enables subsequent processes to accurately compensate for temperature non-uniformity on the substrate surface, avoiding variations in coating thickness and photosensitivity reaction rates caused by temperature fluctuations, thereby ensuring consistent and accurate phase grid writing. For example, if the substrate temperature is detected to be 28°C in one section and 25°C in another, the system calculates the corresponding compensation parameters, assigning a smaller coating thickness to the higher-temperature area and a larger coating thickness to the lower-temperature area.

[0031] Roll-to-roll coating thickness balancing is performed using a precision coating system equipped with a closed-loop controlled micro-dispensing head and a high-precision roller. First, the photosensitive phase-change coating material (a mixture consisting of 5-15wt% azobenzene derivative, 1-3wt% photoinitiator, and solvent) is introduced into a reservoir. Based on the aforementioned temperature compensation parameter array, the system dynamically adjusts the dispensing head aperture and the roller speed, reducing the coating amount in higher temperature areas and increasing it in lower temperature areas, thereby forming a coating with a controlled thickness distribution on the substrate surface. The coating thickness is monitored in real time by an integrated optical interferometer thickness gauge, and the measured data is processed through a moving average filter to generate coating thickness distribution data. This temperature gradient-based roll-to-roll coating thickness control mechanism ensures that the photosensitive phase-change coating maintains uniform photothermal response properties even under non-uniform temperature conditions. For example, if a high temperature is detected in a certain area, the system precisely controls the coating thickness in that area within a range of 200±5nm, while the lower temperature area is controlled within a range of 240±5nm, accurately compensating for temperature variations.

[0032] Phase modulation writing is achieved through a digital micromirror device (DMD) projection system. The system uses a DMD chip with a 1024×768 pixel resolution and a low energy density (5-10mW / cm 2 ) uses a 405nm wavelength laser light source. Based on the aforementioned coating thickness distribution data, the system dynamically adjusts the switching state and dwell time of each micromirror on the DMD chip to generate a spatially modulated mask light field. When the photosensitive phase change coating is exposed to this light field, the azobenzene molecules in the coating undergo cis-trans isomerization, causing a localized refractive index change and forming a phase difference pattern. The system uses an algorithm to control the DMD projection pattern, resulting in a light field with two sinusoidal intensity distributions in different directions, with periods of 50μm and 75μm, respectively. The superposition of these two light fields creates a moiré interference pattern in the coating, ultimately yielding a planar moiré phase map with biaxial periodicity (i.e., different period lengths in the x and y directions). This biaxial periodicity ensures spatial uniqueness of the phase map, providing a rigorous reference for subsequent spatial coordinate establishment. For example, when the x-direction period is set to 50μm and the y-direction period to 75μm, a non-repeating phase pattern can be generated within a 100mm×100mm area, ensuring a unique phase signature at every location.

[0033] Phase difference calibration is performed using a high-resolution phase microscope equipped with a 50x objective and a polarization analyzer, capable of detecting phase differences caused by molecular alignment changes in photosensitive phase change coatings. The system scans and measures the moiré fringe phase pattern along a pre-set grid of points on the substrate surface, acquiring the phase difference value at each measurement point. This phase difference data is then compared with a theoretically designed moiré fringe pattern, and a least-squares fit is used to establish a mapping between the actual phase difference and the expected position. After distortion correction and coordinate transformation, this mapping relationship forms a spatial coordinate mapping matrix that accurately represents the substrate surface position, with an absolute position accuracy of ±5μm. This phase difference calibration method avoids the interference of physical markings on the substrate while providing a highly accurate spatial reference system. For example, a 10×10 calibration grid is set up on a 100mm×100mm substrate area. By measuring the phase difference value at each grid point, the system can accurately identify the spatial distribution characteristics of the moiré fringe and establish a precise micron-level coordinate system.

[0034] The energy dose calibration process is completed through the spectral analysis system. First, the dose response characteristics of the photosensitive phase change coating are pre-calibrated using a standard light source to obtain the energy absorption dose-phase response curve of the photosensitive phase change coating, which describes the phase delay relationship produced by the coating under different energy doses. Subsequently, according to the position information determined by the spatial coordinate mapping matrix, the system performs spectral analysis on each grid point in the Moire fringe phase diagram and measures the movement of the characteristic peak in the reflection spectrum. These spectral features are compared with the pre-calibrated dose-phase response curve, and the energy dose actually received at each position is inverted and calculated. In this way, the system integrates the spatial position information and the energy dose information into a two-dimensional matrix, namely the phase reference data matrix, which provides a unified spatiotemporal energy reference system for subsequent microstructure processing. For example, when the phase delay of a grid point is detected to be 0.25π, according to the pre-calibrated response curve, it can be known that the energy dose received by the point is approximately 8mJ / cm 2 , thus realizing the dual-dimensional calibration of position and energy.

[0035] Please continue reading Figure 1 , based on the phase reference data matrix, the guided mode resonance spectrum sequence obtained during the laser etching process is subjected to the same grid point superposition processing to obtain a resonance spectrum feature data set that characterizes the dynamic evolution of the microprism depth and tilt angle; In one embodiment of the present invention, the guided mode resonance spectrum sequence obtained during the laser etching process is subjected to same-grid point superposition processing based on the phase reference data matrix to obtain a resonance spectrum feature data set characterizing the dynamic evolution of the microprism depth and tilt angle, including: Performing time window segmentation processing on the guided mode resonance spectrum sequence obtained during the laser etching process to obtain a spectrum frame set containing etching displacement indexes; According to the phase reference data matrix, performing grid point coordinate mapping and intensity superposition processing on the spectrum frame set to obtain an aligned spectrum grid frame set; According to the aligned spectral grid frame set, frequency shift vector extraction processing is performed on the cross-time spectrum line of the same grid point to obtain a depth-inclination change parameter matrix; A multi-order sequence fitting process is performed according to the depth-tilt variation parameter matrix and the corresponding time index to obtain a resonance spectrum feature data set.

[0036] The following is a detailed description of the steps involved in the above embodiment: Time window segmentation is a data preprocessing step for the guided mode resonance (GMR) spectrum sequence collected during laser etching. This GMR spectrum sequence is generated by injecting broadband, tunable, tilted incident light into the backside of the substrate while the microprism structure is being formed by laser etching. The resulting periodic structure immediately forms a one-dimensional waveguide grating coupled cavity, generating a characteristic spectral data stream. In specific implementation, a high-speed spectrometer continuously records the output spectrum at a sampling rate of 1000 frames / second, while a photoelectric encoder simultaneously records the precise displacement of the laser etching head. Time window segmentation first divides the continuous spectral data into discrete frames at fixed time intervals (5 milliseconds). Each frame is then assigned a corresponding etching displacement index value, representing the precise spatial coordinates of the laser head at that moment. For example, when the laser head moves at a speed of 10 mm / s, every 5 milliseconds corresponds to a displacement of 0.05 mm. The system then labels each spectral frame with the corresponding position coordinates. This time window segmentation process converts the continuous spectral data stream into discrete data sets with clear temporal and spatial correspondences, eliminating data inconsistencies caused by fluctuations in laser etching speed and ensuring that subsequent analysis can accurately correlate spectral changes with the location of microstructure formation.

[0037] The same-grid point coordinate mapping and intensity superposition processing is the process of spatially aligning the segmented spectral frame set with the pre-established phase reference data matrix. The phase reference data matrix contains the precise grid point coordinate information, and the etching displacement index in the spectral frame set represents the acquisition position. During implementation, the etching displacement index is first converted into a position in the phase reference grid coordinate system through a coordinate transformation algorithm. Taking into account the possible deviation between the laser etching path and the spatial orientation of the phase grid, the conversion process is corrected using a six-parameter affine transformation model. Then, for each grid point, multiple spectral frames with the closest spatial distance to the point are extracted from the spectral frame set, and the intensity is weighted averaged according to the distance weight to obtain a synthetic spectrum representing the grid point. For example, when the grid point spacing is 100μm, the system will select all spectral frames within 150μm from the target grid point and use 1 / r 2(r is the distance) is the weight coefficient for superposition averaging. This same-grid point processing ensures that spectral data collected at different times can be compared and analyzed in a unified spatial reference system, eliminating errors caused by sampling unevenness and position drift, and improving the accuracy of subsequent microstructure geometric parameter extraction.

[0038] Frequency shift vector extraction is the core data processing step for identifying and quantifying changes in resonant peak positions from a collection of aligned spectral grid frames. During the microprism formation process, characteristic peaks in the resonant spectrum shift in frequency as the structure depth and inclination angle change. In implementation, a Gaussian fitting algorithm is used to precisely locate the center wavelength of the resonant peak within the time series spectral frame at each grid point, generating a trajectory of the resonant peak wavelength over time. The wavelength shift amount and direction between adjacent time points are then calculated to construct frequency shift vectors. These frequency shift vectors are directly correlated to changes in microstructure geometric parameters: a wavelength shift toward shorter wavelengths typically corresponds to an increase in structural depth, while a wavelength shift toward longer wavelengths corresponds to a change in inclination angle. Through multi-peak correlation analysis, the system separates the depth and inclination change components from the frequency shift vectors, forming a depth-inclination change parameter matrix. For example, if the resonant peak at a particular grid point shifts from 532.5 nm to 531.8 nm, the system calculates a depth increase of 15 nm and a inclination change of 0.2 degrees. This frequency shift vector extraction method achieves quantitative mapping from spectral changes to geometric parameter changes, enabling the system to accurately monitor the geometric properties of microstructures in the early stages of formation. The peak position measurement accuracy reaches ±0.05nm, and the corresponding depth change detection sensitivity is better than 1nm.

[0039] Multi-order sequence fitting is a data processing step that performs time-domain analysis of the depth-tilt variation parameter matrix. Because microstructure formation during laser etching exhibits nonlinear dynamic characteristics, time series analysis is necessary to capture its evolution. In this implementation, a third-order B-spline function is fitted to the depth and tilt variation parameter sequences at each grid point, generating a smooth and continuous parameter variation curve. The first-order derivative (rate of change) and second-order derivative (acceleration) of the curve are then calculated to analyze the dynamic characteristics of the parameter variation. The system further extracts characteristic points (such as inflection points and extreme points) and their time indices to construct a set of feature vectors characterizing the dynamics of microstructure formation. Ultimately, the feature vectors from all grid points are integrated into a resonant spectral feature dataset, encompassing multidimensional information such as spatial distribution, temporal evolution, and depth-tilt relationship. For example, when analyzing the microprism formation process in a specific area, the system can identify that the depth growth rate reaches a maximum of 3 nm / ms 12 milliseconds after etching begins, while the tilt angle stabilizes after 20 milliseconds. This multi-order sequence fitting analysis reveals the dynamic process and spatial distribution characteristics of microstructure formation. It can not only detect abnormal change points in a timely manner, but also predict the final structural parameters, providing richer and more in-depth information than traditional static detection.

[0040] Please continue reading Figure 1 Based on the resonance spectrum characteristic data set, a spatiotemporal synchronous mapping process is performed on the Barkhausen transition signal sequence and the thermal radiation grayscale field generated by the iron-gallium alloy nanowire grid formed at the root of the microprism groove to obtain a multi-physics field coupling data set representing the stress-geometry-optical correlation; In one embodiment of the present invention, the Barkhausen transition signal sequence generated by the iron-gallium alloy nanowire grid formed at the root of the microprism groove is subjected to spatiotemporal synchronous mapping processing based on the resonance spectrum characteristic data set, and the thermal radiation grayscale field is obtained to obtain a multi-physics field coupling data set characterizing the stress-geometry-optical correlation, including: The Barkhausen transition signal sequence generated by the iron-gallium alloy nanowire grid formed at the root of the microprism groove is subjected to time segmentation processing to obtain a magnetoelastic signal frame set containing etching displacement index; performing grid point coordinate alignment and amplitude superposition processing on the magnetoelastic signal frame set according to the resonance spectrum feature data set to obtain an aligned magnetoelastic grid frame set; Perform position-time synchronous sampling processing on the thermal radiation grayscale field to obtain a temperature grayscale frame set; According to the aligned magnetoelastic grid frame set, the temperature grayscale frame set is subjected to same-grid point registration processing to obtain a thermal-magnetic combined grid frame set; According to the thermal-magnetic joint grid frame set, a co-correlation quantization process is performed on the magnetoelastic amplitude, temperature grayscale and depth-dip parameters to obtain a stress-geometry-optical coupling matrix; A multi-dimensional joint fusion process is performed based on the stress-geometry-optical coupling matrix and the spatial temperature gradient matrix calculated from the temperature grayscale frame set to obtain a multi-physics field coupling data set.

[0041] The following is a detailed description of the steps involved in the above embodiment: Time segment segmentation is a data preprocessing step for the Barkhausen transition signal sequence. The root of the microprism groove refers to the area where the bottom of the microprism structure formed by laser etching or embossing is connected to the side wall. There is a large stress concentration at this area, which is a key position for monitoring structural deformation. The Fe-Ga alloy nanowire grid refers to the Fe-Ga alloy nanowire grid formed by gas-solid co-deposition at the root of the microprism groove. 80 Ga 20The alloy nanowire network has a wire diameter of 20-50 nanometers, a length of 5-10 microns, and a surface density of 200-300 wires per square millimeter. This type of nanowire has an extremely low content and is discontinuously distributed. After UV curing is completed, it is completely embedded in the substrate and will not affect the optical performance and structural stability of the anti-peep light guide plate. The Barkhausen transition signal sequence refers to the pulse signal stream generated by the discontinuous movement of the internal magnetic domains of the iron-gallium alloy nanowire when it is subjected to local stress changes. This signal is highly sensitive to small stress changes. During implementation, a small magnetic induction coil array is used to continuously collect Barkhausen transition signals at a sampling rate of 50kHz, and the displacement data of the laser etching head is recorded at the same time. The time segment segmentation process divides the continuous signal into discrete segments according to a fixed time window of 10 milliseconds, and assigns a corresponding etching displacement index to each segment. For example, when the etching speed is 5mm / s, each 10-millisecond signal segment corresponds to a physical distance of 0.05mm on the substrate. This processing converts the continuous Barkhausen transition signal into discrete data frames with a clear spatial correspondence, which facilitates subsequent precise alignment and analysis with other sensor data. The change in the transition signal amplitude directly reflects the stress evolution during the microstructure formation process.

[0042] The same-grid coordinate alignment and amplitude superposition processing is the process of accurately spatially aligning the magnetoelastic signal frame set with the resonance spectral feature data set. During implementation, the physical position corresponding to each signal frame in the magnetoelastic signal frame set is first determined, and then these positions are mapped to the grid coordinate system used by the resonance spectral feature data set. Due to the different installation positions and sampling frequencies of the acquisition equipment of the two sets of data, there are spatial and temporal offsets, which need to be corrected through coordinate transformation and time interpolation. The specific method is to use the least squares method to fit the spatial transformation matrix and convert the position coordinates of the magnetoelastic signal into spectral grid coordinates. Subsequently, for each grid point, all magnetoelastic signal frames within 100μm of the point are extracted, and the weighted average is calculated with the inverse square of the distance as the weight coefficient to obtain the synthetic magnetoelastic signal representing the grid point. For example, when three magnetoelastic signal acquisition points are located around a grid point (10.5 mm, 15.2 mm) at distances of 50 μm, 70 μm, and 90 μm, respectively, with amplitudes of 0.8 mV, 1.2 mV, and 0.6 mV, the composite amplitude calculated through weighted average is 0.92 mV. This alignment and superposition process ensures the spatial consistency of the magnetoelastic signal and the microstructure's geometric parameters, providing an accurate data foundation for subsequent analysis of stress-geometry correlations. The spatial alignment accuracy is better than ±5 μm.

[0043] Position-time synchronous sampling is a data acquisition and preprocessing step for the thermal radiation grayscale field. The thermal radiation grayscale field refers to the heat distribution field generated by laser heating and material cooling during the microprism formation process, captured as a temperature distribution image by an infrared thermal imager. In implementation, a high-speed infrared thermal imager with a resolution of 640×480 pixels and a temperature resolution of 0.05°C was used to continuously acquire thermal images at a rate of 100 frames per second. Considering the large volume and redundancy of thermal image data, the system employs a sparse sampling strategy in both position and time dimensions. In the spatial dimension, key monitoring points are selected based on a grid distribution. In the temporal dimension, a dynamic sampling interval is set based on the thermal diffusion time constant, with intensive sampling (e.g., every 10 ms) during periods of rapid thermal change and sparse sampling (e.g., every 50 ms) during periods of slow change. Each sampling point records the real-time temperature value, its spatial coordinates, and a timestamp, forming a collection of temperature grayscale frames. For example, during the laser etching process, the system focuses on monitoring temperature changes within a 2mm radius around the etching head, recording the entire process from initial room temperature to peak temperature (approximately 80°C) and then cooling to stabilization. This synchronous sampling process ensures that sufficiently dense temperature data is obtained at key locations and time points, while avoiding data redundancy and improving processing efficiency. The time resolution of the temperature data is better than the characteristic time scale of the thermal diffusion process (typically tens of milliseconds).

[0044] Co-registration is a data processing step that integrates the aligned magnetoelastic grid frame set and the temperature grayscale frame set into a unified reference frame. Because the two data sets are acquired using different sensors and differ in spatial distribution and temporal sequence, precise co-registration is required for joint analysis. First, a common spatial reference grid network with a 50μm grid spacing is established for both data sets. Bilinear interpolation is then performed on each data point in the temperature grayscale frame set, mapping the discrete temperature data onto the unified grid. For each grid point, the system extracts the closest magnetoelastic and temperature data based on the timestamp to form a time-synchronized data pair. In the event of a time mismatch, cubic spline interpolation is used to estimate the values at intermediate moments. The resulting thermal-magnetic joint grid frame set contains time-synchronized magnetoelastic signal amplitude and temperature grayscale values at each grid point. For example, at the grid point located at (12.5mm, 18.7mm), at time t = 105ms, the associated magnetoelastic signal amplitude is 1.25mV and the temperature is 65.8°C. This registration process enables the comparison and correlation analysis of measurement data from different physical fields within the same spatiotemporal framework, providing a data basis for revealing the interaction mechanism between stress, temperature, and geometry. The spatial accuracy after registration is better than 1 / 10 of the grid spacing, and the temporal accuracy is better than 1 / 5 of the sampling interval.

[0045] Co-correlation quantification is a data analysis step that analyzes the relationship between various physical parameters in a set of thermal-magnetic joint grid frames. During implementation, the magnetoelastic amplitude, temperature grayscale and depth-inclination parameters (extracted from the resonance spectrum feature data set) are first normalized so that the value range of each parameter is unified to the [0,1] interval. Then the time delay correlation function between the parameters is calculated to determine the temporal relationship and causal relationship between different physical processes. In particular, the correlation between the change in magnetoelastic amplitude and the change in depth-inclination is analyzed through a sliding time window to quantify the immediate effect of the stress field on the geometric formation of the microstructure. At the same time, the modulation effect of the thermal field on the stress distribution is revealed through the correlation analysis of temperature grayscale and magnetoelastic amplitude. These correlation analysis results are organized into a stress-geometry-optical coupling matrix, where the matrix element a ij The quantification process represents the degree and direction of the influence of the i-th physical quantity on the j-th physical quantity. For example, when the temperature of a certain area increases by 10°C, the corresponding magnetoelastic amplitude increases by 0.5mV, the microprism depth growth rate slows by 15%, and the inclination angle increases by 0.3 degrees. This co-correlation quantification reveals the complex interactions between multiple physical fields, enabling the system to understand the coordinated response mechanism of materials during microstructure formation. The time resolution of the correlation analysis is better than 5ms, capable of capturing rapidly changing transient processes.

[0046] Multidimensional joint fusion processing is a data fusion step that integrates the stress-geometry-optical coupling matrix and the spatial temperature gradient matrix. The spatial temperature gradient matrix is calculated from a collection of temperature grayscale frames and represents the rate of change of temperature over a spatial distribution. In implementation, the temperature difference between adjacent grid points is first calculated and then divided by the grid spacing to obtain the temperature gradient components in the x and y directions. The fusion process uses tensor decomposition techniques to combine the stress-geometry-optical coupling matrix (local interaction relationship) and the temperature gradient matrix (spatial distribution characteristics) into a unified multiphysics coupled dataset. This dataset is a high-dimensional tensor containing multiple dimensions, including spatial coordinates (x, y), time t, physical parameter type p, and their mutual influence coefficient α. For example, when analyzing the relationship between stress fields and geometric parameters during microprism formation in a specific area, cross-sectional data along specific directions and time periods can be extracted from this dataset to demonstrate how the temperature gradient modulates the stress distribution and influences the geometric formation of the microstructure. This multi-dimensional joint fusion processing realizes the transition from discrete physical fields to unified multi-physical field description. It not only retains the spatial distribution and time evolution characteristics of each physical quantity, but also captures the complex coupling relationship between them, providing a comprehensive data basis for subsequent topological analysis and performance prediction. The information density of the data set is 3-5 times higher than that of a single physical field, which greatly enhances the system's understanding of the microstructure formation mechanism.

[0047] In one embodiment of the present invention, performing co-correlation quantization processing on the magnetoelastic amplitude, temperature grayscale, and depth-dip parameters according to the thermal-magnetic joint grid frame set to obtain a stress-geometry-optical coupling matrix includes: Normalization is performed on the magnetoelastic amplitude, temperature grayscale and depth-inclination parameters to obtain a set of standardized feature tensors; According to the standardized characteristic tensor set, performing directional wave packet coherence calculation processing on the magnetoelastic amplitude and depth-dip parameters to obtain an anisotropic coherence spectrum matrix; Performing multi-core information entropy mapping processing according to the anisotropic coherence spectrum matrix and the temperature grayscale parameter to obtain a cross-domain weight tensor; A Tucker decomposition and fusion process is performed on the cross-domain weight tensor and the standardized feature tensor set to obtain the stress-geometry-optical coupling matrix.

[0048] The following is a detailed description of the steps involved in the above embodiment: Normalization is a data preprocessing step that standardizes physical quantities of different dimensions and numerical ranges, such as magnetoelastic amplitude, temperature grayscale, and depth-tilt parameters. Magnetoelastic amplitude refers to the intensity value of the Barkhausen transition signal generated when the iron-gallium alloy nanowire is subjected to micro-stress changes. It is usually measured in millivolts (mV). In the production process of anti-peep light guide plates, the typical range is 0.1-5mV. Temperature grayscale refers to the temperature value of each point in the temperature image collected by the infrared thermal imager. It is measured in degrees Celsius (°C) and usually varies in the range of 25-85°C in the laser-etched area. The depth-tilt parameter is a two-dimensional vector that describes the geometric characteristics of the microprism, where the depth represents the vertical distance from the base to the vertex of the prism, usually in the range of 10-50 microns; the tilt angle represents the angle between the side wall of the prism and the base, usually in the range of 20-45 degrees. During implementation, a minimum-maximum normalization transformation is applied to each physical quantity. This involves subtracting the observed minimum from the original value and dividing it by the difference between the maximum and minimum values, so that the processed values are mapped to the interval [0,1]. These normalized parameters are organized into a four-dimensional tensor structure based on spatial coordinates (x,y), time t, and parameter type p, known as a normalized feature tensor set. For example, if the original magnetoelastic amplitude at a location (15.2mm, 22.3mm) at t = 150ms is 2.3mV, and the amplitude range in this region is 0.5-4.0mV, the normalized value is (2.3-0.5) / (4.0-0.5) = 0.51. This normalization process not only eliminates dimensional differences between different physical quantities, making the data comparable, but also suppresses the influence of outliers, improving the stability and accuracy of subsequent analysis, while preserving the relative variation characteristics and spatial distribution patterns of each parameter.

[0049] Directional wave packet coherence calculation is a signal processing step that analyzes the spatiotemporal correlation between magnetoelastic amplitude and depth-inclination parameters in a set of standardized characteristic tensors. Directional wave packets are localized signal patterns with specific directional and scale characteristics in the time-frequency domain. During microprism formation, they manifest as physical field fluctuations propagating along a specific direction. Coherence is a statistical measure of the similarity between two signals in the time or frequency domain, with perfect correlation equal to 1, no correlation equal to 0, and negative correlation equal to -1. In implementation, a continuous wavelet transform is first applied to the time series of magnetoelastic amplitude and depth-inclination parameters at each spatial grid point, decomposing the time series into wave packet coefficients at different scales and temporal locations. The cross-correlation function between the wave packet coefficients at different orientations (e.g., 0°, 45°, 90°, and 135°) is then calculated to obtain a coherence spectrum that depends on direction, scale, and time. These direction-dependent coherence spectra are organized into a four-dimensional matrix structure, where the matrix element c(d, s, t, θ) represents the coherence between the magnetoelastic amplitude and the depth-inclination parameter along the direction d at direction θ, scale s, and time t. This structure is called the anisotropic coherence spectrum matrix. For example, in the early stages of microprism formation, the magnetoelastic wave packet along the 45° direction exhibits a high coherence of 0.85 with the depth-varying wave packet along the same direction on small scales (1-5 ms), while it exhibits a low coherence of 0.3 with the inclination variation along the 90° direction on large scales (10-20 ms). This directional wave packet analysis reveals the direction-dependent coupling relationship between different physical fields, captures the anisotropic characteristics of the microstructure formation process, and provides quantitative evidence for understanding the immediate impact of stress fields along specific directions on geometric formation. The temporal resolution of the coherence measurement is better than 2 ms, and the directional resolution is better than 15°.

[0050] Multi-kernel information entropy mapping is a data analysis step that evaluates the information dependency between the anisotropic coherence spectrum matrix and the temperature grayscale parameters. Information entropy is a statistical physics quantity that measures signal uncertainty or information content; larger values indicate richer information or higher uncertainty. Multi-kerneling refers to the use of multiple kernel functions to simultaneously process data associations of different types and scales. In implementation, the anisotropic coherence spectrum matrix is first sliced along the time and direction dimensions at each spatial point to obtain characteristic spectra representing local coherence properties. The mutual information between these characteristic spectra and the temperature grayscale time series at the corresponding locations is then calculated using different kernel functions, such as Gaussian, Laplacian, and polynomial kernels, to capture linear and nonlinear correlations. This mutual information is organized into a three-dimensional tensor structure, where the element w(x,y,k) represents the mutual information calculated using the kth kernel function at position (x,y), i.e., the modulation strength of the stress-geometry coherence relationship due to temperature variation. This structure is called a cross-domain weight tensor. For example, when the temperature in a certain area suddenly increased by 10°C, the coherence between the magnetoelastic amplitude and depth variation dropped from 0.7 to 0.4. The system calculated a Gaussian kernel mutual information value of 1.8 bits in this area, indicating that the temperature change strongly affected the stress-geometry coupling relationship. This multi-kernel information entropy mapping process quantifies the modulation of the temperature field on the stress-geometry coherence, revealing high-order nonlinear correlations between multiple physical fields. The kernel bandwidth range for the mutual information calculation was selected to be 0.1-1.0 to ensure that information dependencies at different scales can be captured.

[0051] Tucker decomposition and fusion processing is a data dimensionality reduction and fusion step that integrates cross-domain weight tensors and a set of standardized feature tensors into a unified description. Tucker decomposition is a tensor decomposition technique that represents high-dimensional tensors as the product of a core tensor and a factor matrix along each dimension. It can be regarded as a generalization of principal component analysis (PCA) for multidimensional data. In implementation, a joint high-dimensional tensor consisting of a set of standardized feature tensors and a cross-domain weight tensor is first constructed, and then Tucker decomposition is applied to decompose it into a low-rank representation. Specifically, the main eigenvectors are extracted for the four dimensions of space, time, parameter type, and weight type, forming four sets of factor matrices. These factor matrices are combined with the reduced core tensor to form the stress-geometry-optical coupling matrix. This matrix is a spatially distributed second-order tensor field, where the matrix element M(i,j)(x,y) represents the strength and mode of coupling of the i-th type of physical quantity to the j-th type of physical quantity at position (x,y). For example, at a distance of 5 mm from the etching starting point, the positive coupling coefficient of stress to depth is 0.72, indicating that increased stress at this location significantly promotes depth growth; whereas the coupling coefficient of stress to inclination angle is -0.35, indicating that increased stress slightly suppresses changes in inclination angle. The Tucker decomposition rank is chosen to be r1=15 for spatial dimension, r2=8 for temporal dimension, r3=5 for parameter dimension, and r4=3 for weight dimension. This reduces the data size by approximately 86% while retaining 95% of the information. This Tucker decomposition fusion process transforms high-dimensional complex data into a low-dimensional intrinsic representation, significantly reducing data redundancy and improving computational efficiency. It also extracts the essential characteristics of multi-physics field interactions, providing a streamlined and information-rich data foundation for subsequent topological analysis and performance prediction. The precise quantification of coupling strength enables the system to distinguish between weakly and strongly coupled regions and adjust process parameters accordingly. The condition number of the coupling matrix is kept below 20, ensuring numerical stability and the rationality of the physical interpretation.

[0052] Please continue reading Figure 1 Based on the multi-physics field coupling data set, a topological persistent coherence dimensionality reduction modeling process is performed on the privacy viewing angle half parallax angle, brightness half-decay width and lateral stray light threshold to obtain a dynamic privacy viewing angle field distribution map representing the continuous distribution of the panel surface; In one embodiment of the present invention, the topological persistent coherence dimensionality reduction modeling is performed on the privacy viewing angle half parallax angle, brightness half-width, and lateral stray light threshold based on the multi-physics field coupling dataset to obtain a dynamic privacy viewing angle field distribution map representing the continuous distribution of the panel, including: Performing privacy performance parameter extraction processing on the multi-physics field coupling data set to obtain a privacy parameter tensor set including a half-parallax angle, a brightness half-width, and a lateral stray light threshold; Performing multi-scale filter complex construction processing on the panel grid adjacency relationship according to the privacy parameter tensor set to obtain a filter complex sequence; According to the filter complex sequence, performing bar coding calculation processing on the persistence of the homology group to obtain a persistence bar matrix; According to the persistent strip matrix, performing a center-of-gravity embedding mapping process on the topological skeleton nodes to obtain a privacy perspective vector field; According to the privacy viewing angle vector field, continuous splicing processing is performed on the panel position coordinates to obtain a dynamic privacy viewing angle field distribution map.

[0053] The following is a detailed description of the steps involved in the above embodiment: The privacy performance parameter extraction process involves converting the multi-physics coupled dataset into the actual optical performance metrics of the privacy light guide. The half-parallax angle (HPA) refers to the viewing angle within which the observer can clearly see the screen content, typically within ±30° of the frontal viewing angle. The luminance half-fall width (HWFW) is the viewing angle width at which the brightness drops to half its maximum value, which determines the sharpness of the privacy protection effect. The lateral stray light threshold (LSL) is the residual brightness level within the intended obscured viewing angle, which determines the thoroughness of the privacy protection. During implementation, an optical physics transfer function algorithm is used to convert the microstructure geometric parameters (depth, tilt angle) and material optical properties from the multi-physics coupled dataset into these three privacy performance parameters. The process involves scanning the virtual viewing angle space from -80° to +80° in 5° steps. For each viewing angle, the intensity distribution of light passing through the microprism structure is calculated, resulting in a viewing angle-luminance curve. From this curve, the half-parallax angle (the angle at which the curve drops to 50%), the luminance half-fall width (full width at half maximum (FWHM), and the lateral stray light threshold (relative brightness value at ±60°) are extracted. For example, when the depth of the microprisms in a certain area is 25μm and the tilt angle is 38°, the calculated half-parallax angle is ±28°, the brightness half-width is 12°, and the lateral stray light threshold is 2.5%. These parameters are organized into a three-dimensional tensor structure based on spatial grid points, called the privacy parameter tensor set. Through this parameter extraction process, the system achieves a precise mapping from microscopic physical parameters to macroscopic optical performance, making the anti-peeping effect predictable and quantifiable, with the performance parameter calculation accuracy better than ±1°.

[0054] The multi-scale filter complex construction process is a computational topology method that analyzes the topological structure of the privacy protection performance of a panel based on a set of privacy parameter tensors. The panel grid adjacency refers to the connections between adjacent sampling points on the surface of the privacy light guide panel, forming a grid graph structure. The filter complex is a set of adjacent points selected based on a threshold parameter, forming a geometric structure including points, edges, faces, and high-dimensional simplexes. In implementation, the privacy parameter tensor set is first discretized into a uniform 100×100 grid of points in the spatial dimension, with each point containing three privacy performance parameter values. A series of thresholds ε are then defined, increasing from 0 to the maximum parameter difference. For each threshold ε, when the Euclidean distance between two adjacent grid points is less than ε, a connection is established between the two points, forming an edge. When three points are connected, a triangular face is formed. This process continues to construct a high-dimensional simplex. As the threshold ε increases, the number of connections gradually increases, and the complex structure continuously changes, forming a series of nested geometric structures, namely, a filter complex sequence. For example, when ε = 0.1, only points with very similar parameters are connected, forming multiple isolated islands. When ε = 0.3, these islands begin to connect, forming larger connected regions and ring structures. When ε = 0.5, almost all points are connected into a whole, but some holes still remain. This multi-scale filter complex construction reveals the continuous variation pattern and topological characteristics of anti-peek performance parameters on the board surface, capturing structural properties at different scales and laying the geometric foundation for subsequent topological analysis. The threshold increment during the construction process is set to 0.05 to ensure that subtle topological changes in the parameter space can be captured.

[0055] The computational processing of homology group persistence barcodes is the core step in computational topology for analyzing the topological stability characteristics of filter complex sequences. Homology groups are algebraic tools for describing the structure of holes in topological space. 0-dimensional homology represents the number of connected components, 1-dimensional homology represents the number of loop holes, and 2-dimensional homology represents the number of cavities. Persistence refers to the duration of these topological features during the change of filtering parameters. During implementation, the birth and death history of homology groups of each dimension is calculated for the filter complex sequence generated above. Specifically, when a topological feature (such as a connected component or loop hole) is affected by the parameter ε, birth When the parameter ε death disappears when , and its persistence is defined as ε death Subtract ε birth. These birth and death events and persistence values are organized into barcode representations, where the starting point of each bar is the threshold for the appearance of the feature, the end point is the threshold for the disappearance of the feature, and the length of the bar is the persistence. The barcodes of all dimensions are integrated into a matrix form, with each row corresponding to a topological feature, containing the dimension, generation threshold, extinction threshold and persistence value, to form a persistent bar matrix. For example, connected component No. 1 is generated at ε=0.05, merges with other connected components and disappears at ε=0.35, with a persistence of 0.3; while ring hole No. 5 is formed at ε=0.25, is filled and disappears at ε=0.6, with a persistence of 0.35. Through this barcode calculation, the system extracts topological features with significant persistence in the anti-peeping performance distribution, distinguishes stable features from noise features, and sets the length threshold of the persistent bar to 0.2 to ensure that only topological features that are truly structurally meaningful are retained.

[0056] The centroid embedding mapping process is a dimensionality reduction visualization technique that transforms the abstract topological information in a persistent bar matrix into a concrete spatial vector field. A topological skeleton node is a key point that represents a topological feature (such as a connected component or loop hole), typically selected as the centroid position within the feature's lifetime. The privacy perspective vector field refers to the direction and magnitude of change in the privacy performance parameter at each point on the anti-peep light guide plane. In implementation, bars with a persistence greater than a threshold (typically 0.2) are first selected and the corresponding topological features are extracted from the persistent bar matrix. For each feature, its "centroid" in parameter space is calculated—the weighted average position of the grid points involved in the feature's lifetime, with the weight proportional to the point's importance within the feature. A Laplacian matrix is then constructed to describe the connectivity between the skeleton nodes, and its eigenvectors are calculated as the low-dimensional coordinates representing the nodes. Finally, these coordinates are mapped back to the original space, forming a privacy perspective vector field covering the entire panel surface. The direction of the vector indicates the direction of the fastest parameter change, and the length of the vector represents the rate of change. For example, in the center of the panel, the vector points to the upper right and has a length of 0.8, indicating that the privacy performance in this area changes rapidly in the upper right direction. In contrast, the vectors in the edge areas are shorter and have different directions, indicating that the changes are slower and more irregular. This centroid embedding mapping process transforms high-dimensional topological information into an intuitive vector field representation, revealing the changing trend and gradient distribution of anti-peeping performance. The eigenvector dimension is selected as the vector corresponding to the first three non-zero eigenvalues to ensure that the main topological change direction is captured.

[0057] Continuous splicing processing is a data interpolation and visualization step that integrates discrete privacy view vector fields into a global continuous distribution. The dynamic privacy view field distribution map refers to the continuous privacy performance parameter distribution covering the entire anti-peep light guide surface, which intuitively displays the spatial distribution characteristics of the anti-peep effect. During implementation, the privacy view vector field obtained in the previous step is first subjected to a radial basis function (RBF) interpolation algorithm to extend the vector values on the discrete grid points to a continuous plane. Specifically, a multi-quadratic function kernel is used, and its smoothing parameter is set to 2 times the grid spacing to ensure that the interpolation result is smooth and retains local features. Then, a high-density (500×500) regular sampling grid is generated, and the interpolated vector value is calculated at each grid point. Finally, based on the characteristics of the vector field, features such as contour lines, singular points, and streamlines are extracted and superimposed on the board coordinate map to form a color-coded dynamic privacy view field distribution map. For example, red areas in the figure indicate areas with large semi-parallax angles (>30°), yellow indicates moderate semi-parallax angles (25°-30°), and green indicates small semi-parallax angles (<25°). Black streamlines show the main directions of change in the field of view; blue dots mark singular points in the field of view. This continuous stitching process transforms discrete data points into a continuously distributed field, providing an intuitive and clear global view, facilitating the identification of spatial distribution characteristics and abnormal areas of anti-peeping performance. The interpolation error is kept within 5% of the original vector amplitude, ensuring the accuracy and consistency of the visual representation.

[0058] In one embodiment of the present invention, performing a gravity center embedding mapping process on topological skeleton nodes according to the persistent strip matrix to obtain a privacy perspective vector field includes: Performing scale gating on the persistent bar matrix to obtain a threshold filtered bar set; Filter the bar set according to the threshold value, perform Laplace eigenmapping processing on the topological skeleton nodes, and obtain a skeleton spectral coordinate matrix; According to the skeleton spectral coordinate matrix, vector interpolation processing is performed on the panel grid coordinates to obtain a privacy perspective vector field.

[0059] The following is a detailed description of the steps involved in the above embodiment: Scale gating is a data filtering step that selects the persistence bar matrix. The persistence bar matrix is a data structure representing the range of topological features. Each row contains the generation value, extinction value, and persistence of a topological feature (such as a connected component or loop hole). Scale gating is a processing method that selects relevant data based on a specific scale range, similar to bandpass filtering in signal processing. In implementation, a persistence threshold (typically 0.2-0.3) is set to retain only bars with a persistence greater than this threshold, thereby filtering out short-lived topological features. Next, a lifetime range is set to retain bars with generation values between 0.1-0.4 and extinction values between 0.5-0.8. These ranges correspond to physically meaningful microstructural variations in the privacy light guide. Finally, a categorized selection is performed based on topological dimension, processing 0-dimensional (connected components), 1-dimensional (loop holes), and 2-dimensional (cavities) topological features separately, as different dimensional features reflect different aspects of privacy protection performance. This triple screening process results in a threshold-filtered bar set containing truly physically meaningful and statistically significant topological features. For example, when analyzing the privacy parameter distribution of a region, holes with a persistence of 0.15 are filtered out (considered noise), while holes with a persistence of 0.35 (indicating a stable annular distribution of privacy performance in that region) are retained. This scale gating process effectively removes random noise and unstable structures from the data, retaining data that truly represents the essential topological features of the microstructure formation process. The number of bars after screening is typically reduced by 60-70%, while retaining approximately 95% of the key topological information.

[0060] Laplacian eigenmapping is a dimensionality reduction step that computes a topological skeleton geometric representation based on a thresholded, filtered set of strips. Topological skeleton nodes are key representative points representing topological features, such as the centroid of a connected component or the representative point of a loop hole. Laplacian eigenmapping is a nonlinear dimensionality reduction technique that maps high-dimensional data to a low-dimensional space by computing the Laplacian matrix and its eigenvectors between data points. In implementation, the point sets corresponding to each topological feature are first extracted from the thresholded, filtered set of strips. Each topological feature (such as a loop hole) is represented by a set of grid points contributing to its formation. A weighted adjacency graph is then constructed, where the connection weights between points are based on their Euclidean distance in the original parameter space, converted to similarity using a Gaussian kernel function. The Laplacian matrix of the graph is then calculated, and its eigenvectors and eigenvalues are determined. The two to three eigenvectors corresponding to the smallest non-zero eigenvalue are selected to form the skeleton spectral coordinate matrix, with each row representing the coordinates of a topological skeleton node in the low-dimensional spectral space. For example, when analyzing a prominent ring structure in the privacy view distribution, the system extracts 16 key points on the ring and calculates its Laplacian eigenvectors, forming a 16×3-dimensional spectral coordinate matrix that preserves the spatial distribution characteristics of the ring structure on the board. This Laplacian eigenmapping process transforms complex topological structures into a concise geometric representation, preserving the basic shape of the topological features while providing a low-dimensional representation that facilitates subsequent calculations. The reduced spectral coordinates retain approximately 92% of the original topological information.

[0061] Vector interpolation is a spatial interpolation step that extends the skeleton spectral coordinate matrix across the entire panel surface. The privacy view vector field refers to the direction and intensity of the change in privacy protection performance at each location on the privacy light guide panel. During implementation, each node in the skeleton spectral coordinate matrix is first treated as a sampling point on the panel surface, with its spectral coordinate value serving as the vector field value (direction and magnitude) at that point. The entire panel surface is then interpolated using the moving least squares method. This method assigns distance-based weights to surrounding sampling points when calculating the vector value of the target point. Specifically, quadratic basis functions and exponentially decaying weights are used, with the weight decay radius set to three times the grid spacing to ensure a balance between local smoothness and global continuity. For each grid point on the panel surface, the distance to each skeleton node is calculated, a weight is determined based on the distance, and the vector value at that point is calculated using a weighted average. Finally, a continuous vector field distribution is formed across the entire panel surface, known as the privacy view vector field. For example, for a grid point located at (25.6mm, 42.3mm) that is not a skeleton node, the vector value obtained by calculating the weighted average of this point and the eight surrounding skeleton nodes is (0.65, -0.42, 0.18), indicating the local variation in anti-peeping performance at that location. This vector interpolation process extends the discrete topological skeleton to a continuous vector field, providing a fine-grained directional description of the anti-peeping performance of the entire panel. The interpolation error is controlled within 7% of the original value, meeting the accuracy requirements of process control.

[0062] Please continue reading Figure 1 According to the dynamic privacy viewing angle field distribution map, the laser pulse energy, imprinting pressure and curing dosage are regionally fine-tuned and controlled, and the optical phase grating elimination reaction is triggered to obtain the quality judgment result of the anti-peep light guide plate.

[0063] In one embodiment of the present invention, the laser pulse energy, imprinting pressure, and curing dosage are fine-tuned regionally based on the dynamic privacy viewing angle field distribution map, and an optical phase grating elimination reaction is triggered to obtain a quality determination result of the anti-peep light guide plate, including: Perform error vector field extraction processing on the dynamic privacy viewing angle field distribution map to obtain the fine-tuning instruction matrix of the coverage panel grid; According to the fine-tuning instruction matrix, a space-time solution process is performed on the laser pulse energy sequence, the imprint pressure sequence and the curing dose sequence to obtain a synchronous parameter modulation profile; According to the synchronization parameter modulation profile, regional modulation processing is performed on the roll-to-roll process time axis, and an optical phase grid elimination reaction is synchronously triggered at the corresponding plate grid to obtain a process execution record; The privacy viewing angle field snapshot is recalculated according to the process execution record, and a tolerance comparison process is performed with the original dynamic privacy viewing angle field distribution map to obtain a quality determination result.

[0064] The following is a detailed description of the steps involved in the above embodiment: Error vector field extraction is the data analysis step that converts the dynamic privacy viewing angle field distribution map into specific process control instructions. The dynamic privacy viewing angle field distribution map is a continuous two-dimensional field that represents the distribution of privacy protection performance on the privacy light guide surface. It contains information such as the half-parallax angle, half-brightness half-fall width, and lateral stray light threshold. During implementation, this distribution map is first compared with the preset target performance specifications, and the parameter deviation at each grid point is calculated. The target specifications are typically a half-parallax angle of ±30°±1.5°, a half-brightness half-fall width of 12°±1°, and a lateral stray light threshold of less than 3%. Then, based on the direction and magnitude of the deviation, an error vector field is generated. The direction of the vector indicates the direction of correction, and the magnitude of the vector indicates the correction intensity. Next, the error vector is converted into specific fine-tuning instructions, including correction values for three dimensions: laser energy adjustment, imprint pressure adjustment, and curing dose adjustment. Finally, these adjustments are organized into a fine-tuning instruction matrix based on a grid (typically 200×200 points) on the panel surface. For example, when the half-parallax angle at a certain position (35.2mm, 58.4mm) is 26.5°, 30° below the target value, the system generates fine-tuning instructions: increase laser energy by 8%, decrease imprint pressure by 5%, and increase curing dose by 3%. This error vector field extraction process achieves a precise mapping from performance parameters to process parameters, providing quantitative instructions for real-time closed-loop control. The relative accuracy of the error-process mapping is better than ±2%, ensuring the accuracy and effectiveness of the control instructions.

[0065] The space-time solution process is a data conversion step that converts the fine-tuning instruction matrix into timing parameters that can be executed by the production equipment. The laser pulse energy sequence is the time sequence that controls the power of each laser pulse, which is usually in the range of 5-15W; the imprint pressure sequence is the numerical sequence that controls the pressure applied by the mold head in each area, which is usually in the range of 0.5-2.5MPa; the curing dose sequence is the numerical sequence that controls the intensity of the UV lamp, which is usually in the range of 50-200mJ / cm 2During implementation, the mapping relationship between the grid points on the board surface on the time axis is first calculated based on the operating speed of the roll-to-roll production line (usually 5-10m / min) and the width of the board. Then, the spatially distributed adjustment values in the fine-tuning instruction matrix are rearranged into a time series according to the board feed direction. For each process parameter (laser energy, imprint pressure, curing dose), an independent time series control curve is generated to form a parameter-time two-dimensional matrix called a synchronous parameter modulation profile. For example, when the production line runs at a speed of 8m / min and processes a 100mm×100mm board, the adjustment value in row 50 and column 75 of the fine-tuning instruction matrix will be mapped to the control value at 3.75 seconds after the start of production. This space-time solution process converts static spatial distribution adjustment requirements into dynamic time series control signals, enabling process equipment to precisely control specific areas of the board at the appropriate time. The time resolution accuracy is better than 5ms, meeting the synchronous control requirements of high-speed production lines.

[0066] Regional modulation processing is the execution step of controlling the production equipment in real time according to the synchronous parameter modulation profile. The roll-to-roll process timeline refers to the time dimension along the material conveying direction during the continuous roll production process; the optical phase grating elimination reaction refers to the process of triggering the azobenzene molecules in the photosensitive phase change coating to return to the ground state through ultraviolet light irradiation of a specific wavelength, causing the phase grating pattern to disappear. During implementation, the synchronous parameter modulation profile is first loaded into the control system, and the trigger time point and execution cycle are set. Then, as the roll is continuously conveyed, the control system sends modulation signals to the laser, imprinting module and UV curing lamp at the preset time point to dynamically adjust their working parameters. In particular, the laser energy is achieved by modulating the driving current; the imprinting pressure is achieved by an array of adjustable stiffness air-floating pressure heads; and the curing dose is achieved by pulse width modulation of the UV LED lamp. At the same time, before the sheet passes through the final curing area, the trigger wavelength is 365nm and the power density is 80mW / cm 2 An ultraviolet light source is used with an irradiation time of 1-2 seconds to trigger an optical phase grating elimination reaction. The system records the execution time, parameter value and corresponding position of each modulation action to form a process execution record. For example, when the coil runs to the 125th second, the system adjusts the laser energy in the 72nd section of the plate area from 10W to 10.8W, and records the adjustment behavior and its corresponding spatial position. This regional modulation processing realizes refined real-time control of the continuous production process, so that different areas of each anti-peep light guide plate can receive personalized process adjustments, while ensuring that the phase grating used for detection is completely eliminated in the final product. The spatial accuracy of regional regulation is better than ±0.5mm, ensuring the accuracy of process adjustment and the complete elimination of the phase grating.

[0067] Tolerance comparison is a decision-making step for evaluating the effectiveness of process adjustments and determining product quality. A privacy field snapshot refers to a recalculated privacy protection performance parameter distribution map after process adjustments. Tolerance comparison compares the difference between actual performance and target specifications with the allowable tolerances. During implementation, the adjusted privacy field parameter distribution is recalculated based on the actual process parameter adjustment data in the process execution record using the previously established physical field mapping relationship, creating a privacy field snapshot. This snapshot is then compared pixel-by-pixel with the original dynamic privacy field distribution map to calculate a difference map, representing the performance change resulting from the process adjustment. This difference map is then compared with a pre-defined tolerance band (half-parallax angle ±1.5°, luminance half-width ±1°, and lateral stray light threshold +1%) to count the number and distribution of out-of-tolerance points. Finally, based on the proportion and concentration of out-of-tolerance points, a quality assessment is generated, including "accepted," "partially out-of-tolerance," and "unacceptable," with the location and parameter type of the problem area marked. For example, after process adjustments, a batch of privacy-protected light guide plates saw the half-parallax angle out-of-tolerance ratio drop to 1.2% (below the 2% acceptance threshold), with dispersed distribution. The system then determined the product to be "acceptable." Meanwhile, another batch saw only 1.8% out-of-tolerance points, but they were concentrated at the edges, forming a continuous area. This tolerance comparison process provides an objective and quantitative quality evaluation standard, avoiding the crude judgment of the traditional "pass / fail" dichotomy while pinpointing specific problem areas and types. This provides a precise basis for quality traceability and process optimization, with an accuracy rate exceeding 95%, significantly improving the reliability of quality control.

[0068] The above describes the detection method of the anti-peep light guide plate in the embodiment of the present invention. The following describes the detection device of the anti-peep light guide plate in the embodiment of the present invention. Figure 2 An embodiment of the detection device of the anti-peep light guide plate of the present invention includes: Phase grating writing module 101 is used to perform optical phase grating writing on the surface of the optical substrate before entering the micro-engraving station by roll-coating a photosensitive phase change coating and writing a planar moiré fringe pattern under a low-energy mask light field, thereby obtaining a phase reference data matrix calibrated in two dimensions: spatial coordinates and energy dose; The resonance spectrum analysis module 102 is configured to perform a same-grid point superposition process on the guided mode resonance spectrum sequence obtained during the laser etching process based on the phase reference data matrix to obtain a resonance spectrum feature data set that characterizes the dynamic evolution of the microprism depth and tilt angle; A multi-physics field coupling module 103 is configured to perform spatiotemporal synchronous mapping processing on the Barkhausen transition signal sequence and the thermal radiation grayscale field generated by the embedded Fe-Ga alloy nanowire according to the resonance spectrum characteristic data set, thereby obtaining a multi-physics field coupling data set representing stress-geometry-optical correlation; A topological dimensionality reduction modeling module 104 is configured to perform topological persistent coherence dimensionality reduction modeling on the privacy viewing angle half parallax angle, brightness half-width, and lateral stray light threshold based on the multi-physics field coupling dataset to obtain a dynamic privacy viewing angle field distribution map representing the continuous distribution of the panel surface; The adaptive control module 105 is used to implement regional fine-tuning control of the laser pulse energy, imprinting pressure and curing dosage according to the dynamic privacy viewing angle field distribution map, and trigger the optical phase grating elimination reaction to obtain the quality judgment result of the anti-peep light guide plate.

[0069] above Figure 2 The detection device of the anti-peep light guide plate in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The detection device of the anti-peep light guide plate in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0070] Figure 3 Figure 2 is a schematic diagram of the structure of a device for detecting a privacy-preventive light guide plate provided by an embodiment of the present invention. The device 200 for detecting a privacy-preventive light guide plate may vary significantly depending on its configuration or performance. It may include one or more processors 210 (e.g., one or more processors), a memory 220, and one or more storage media 230 (e.g., one or more mass storage devices) storing applications 233 or data 232. The memory 220 and storage medium 230 may be either transient or persistent storage. The program stored in the storage medium 230 may include one or more modules (not shown), each of which may include a series of instructions for operating the device 200. Furthermore, the processor 210 may be configured to communicate with the storage medium 230, executing the series of instructions stored in the storage medium 230 on the device 200 to implement the steps of the aforementioned method for detecting a privacy-preventive light guide plate.

[0071] The anti-peep light guide plate detection device 200 may further include one or more power supplies 240, one or more wired or wireless network interfaces 250, one or more input and output interfaces 260, and / or one or more operating systems 231, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 3 The structure of the detection device for the anti-peep light guide plate shown does not constitute a limitation on the detection device for the anti-peep light guide plate provided by the present invention, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0072] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions, and when the instructions are run on a computer, the computer executes the steps of the anti-peep light guide plate detection method.

[0073] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0074] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0075] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. All equivalent structural transformations made by using the contents of the present invention description and drawings under the inventive concept of the present invention, or direct / indirect application in other related technical fields are included in the patent protection scope of the present invention.

Claims

1. A method for detecting an anti-peep light guide plate, characterized in that: include: By rolling a photosensitive phase change coating and writing a planar moiré fringe pattern under a low-energy mask light field, an optical phase grating writing process is performed on the surface of the optical substrate before entering the micro-engraving station to obtain a phase reference data matrix with dual-dimensional calibration of spatial coordinates and energy dose; Based on the phase reference data matrix, a grid point superposition process is performed on the guided mode resonance spectrum sequence obtained during the laser etching process to obtain a resonance spectrum feature data set that characterizes the dynamic evolution of the microprism depth and tilt angle; Based on the resonance spectrum characteristic data set, a time-space synchronous mapping process is performed on the Barkhausen transition signal sequence and the thermal radiation grayscale field generated by the iron-gallium alloy nanowire grid formed at the root of the microprism groove to obtain a multi-physics field coupling data set representing the stress-geometry-optical correlation; Based on the multi-physics field coupling data set, a topological persistent coherence dimensionality reduction modeling process is performed on the privacy viewing angle half parallax angle, brightness half-decay width, and lateral stray light threshold to obtain a dynamic privacy viewing angle field distribution map representing the continuous distribution of the panel surface; According to the dynamic privacy viewing angle field distribution diagram, regional fine-tuning control is implemented on the laser pulse energy, imprinting pressure and curing dosage, and the optical phase grating elimination reaction is triggered to obtain the quality judgment result of the anti-peep light guide plate.

2. The method for detecting an anti-peep light guide plate according to claim 1, wherein: The method comprises the following steps: roll-coating a photosensitive phase change coating and writing a planar moiré fringe pattern under a low-energy mask light field, performing an optical phase grating writing process on the surface of the optical substrate before entering the micro-engraving station, and obtaining a phase reference data matrix calibrated in two dimensions of spatial coordinates and energy dose. The method comprises: Performing temperature gradient measurement processing on the surface of the optical substrate to obtain a temperature compensation parameter array distributed along the conveying direction; Performing roller coating thickness equalization processing on the photosensitive phase change coating according to the temperature compensation parameter array to obtain coating thickness distribution data; According to the coating thickness distribution data, the photosensitive phase change coating is exposed to a low-energy mask light field and a phase modulation writing process is performed to obtain a planar moiré fringe phase image containing biaxial periodic information; According to the plane moiré fringe phase image, performing phase difference calibration processing on the position coordinates of the substrate surface to obtain a spatial coordinate mapping matrix; According to the spatial coordinate mapping matrix and the energy absorption dose-phase response curve of the photosensitive phase change coating, the coating is subjected to energy dose calibration processing to obtain a phase reference data matrix.

3. The method for detecting an anti-peep light guide plate according to claim 1, wherein: The method of performing same-grid point superposition processing on the guided mode resonance spectrum sequence obtained during the laser etching process according to the phase reference data matrix to obtain a resonance spectrum feature data set characterizing the dynamic evolution of the microprism depth and tilt angle includes: Performing time window segmentation processing on the guided mode resonance spectrum sequence obtained during the laser etching process to obtain a spectrum frame set containing etching displacement indexes; According to the phase reference data matrix, performing grid point coordinate mapping and intensity superposition processing on the spectrum frame set to obtain an aligned spectrum grid frame set; According to the aligned spectral grid frame set, frequency shift vector extraction processing is performed on the cross-time spectrum line of the same grid point to obtain a depth-inclination change parameter matrix; A multi-order sequence fitting process is performed according to the depth-tilt variation parameter matrix and the corresponding time index to obtain a resonance spectrum feature data set.

4. The method for detecting an anti-peep light guide plate according to claim 1, wherein: According to the resonance spectrum characteristic data set, a time-space synchronous mapping process is performed on the Barkhausen transition signal sequence and the thermal radiation grayscale field generated by the iron-gallium alloy nanowire grid formed at the root of the microprism groove to obtain a multi-physics field coupling data set representing the stress-geometry-optical correlation, including: The Barkhausen transition signal sequence generated by the iron-gallium alloy nanowire grid formed at the root of the microprism groove is subjected to time segmentation processing to obtain a magnetoelastic signal frame set containing etching displacement index; performing grid point coordinate alignment and amplitude superposition processing on the magnetoelastic signal frame set according to the resonance spectrum feature data set to obtain an aligned magnetoelastic grid frame set; Perform position-time synchronous sampling processing on the thermal radiation grayscale field to obtain a temperature grayscale frame set; According to the aligned magnetoelastic grid frame set, the temperature grayscale frame set is subjected to same-grid point registration processing to obtain a thermal-magnetic combined grid frame set; According to the thermal-magnetic joint grid frame set, a co-correlation quantization process is performed on the magnetoelastic amplitude, temperature grayscale and depth-dip parameters to obtain a stress-geometry-optical coupling matrix; A multi-dimensional joint fusion process is performed based on the stress-geometry-optical coupling matrix and the spatial temperature gradient matrix calculated from the temperature grayscale frame set to obtain a multi-physics field coupling data set.

5. The method for detecting an anti-peep light guide plate according to claim 4, wherein: The method of performing co-correlation quantization processing on the magnetoelastic amplitude, temperature grayscale and depth-dip parameters according to the thermal-magnetic joint grid frame set to obtain a stress-geometry-optical coupling matrix includes: Normalization is performed on the magnetoelastic amplitude, temperature grayscale and depth-inclination parameters to obtain a set of standardized feature tensors; According to the standardized characteristic tensor set, performing directional wave packet coherence calculation processing on the magnetoelastic amplitude and depth-dip parameters to obtain an anisotropic coherence spectrum matrix; Performing multi-core information entropy mapping processing according to the anisotropic coherence spectrum matrix and the temperature grayscale parameter to obtain a cross-domain weight tensor; A Tucker decomposition and fusion process is performed on the cross-domain weight tensor and the standardized feature tensor set to obtain the stress-geometry-optical coupling matrix.

6. The method for detecting an anti-peep light guide plate according to claim 1, wherein: According to the multi-physics field coupling data set, topological persistent coherence dimensionality reduction modeling is performed on the privacy viewing angle half parallax angle, brightness half-decay width and lateral stray light threshold to obtain a dynamic privacy viewing angle field distribution map representing the continuous distribution of the panel surface, including: Performing privacy performance parameter extraction processing on the multi-physics field coupling data set to obtain a privacy parameter tensor set including a half-parallax angle, a brightness half-width, and a lateral stray light threshold; Performing multi-scale filter complex construction processing on the panel grid adjacency relationship according to the privacy parameter tensor set to obtain a filter complex sequence; According to the filter complex sequence, performing bar coding calculation processing on the persistence of the homology group to obtain a persistence bar matrix; According to the persistent strip matrix, performing a center-of-gravity embedding mapping process on the topological skeleton nodes to obtain a privacy perspective vector field; According to the privacy viewing angle vector field, continuous splicing processing is performed on the panel position coordinates to obtain a dynamic privacy viewing angle field distribution map.

7. The method for detecting an anti-peep light guide plate according to claim 6, wherein: The method of performing a gravity center embedding mapping process on the topology skeleton nodes according to the persistent strip matrix to obtain a privacy perspective vector field includes: Performing scale gating on the persistent bar matrix to obtain a threshold filtered bar set; Filter the bar set according to the threshold value, perform Laplace eigenmapping processing on the topological skeleton nodes, and obtain a skeleton spectral coordinate matrix; According to the skeleton spectral coordinate matrix, vector interpolation processing is performed on the panel grid coordinates to obtain a privacy perspective vector field.

8. The method for detecting an anti-peep light guide plate according to claim 1, wherein: According to the dynamic privacy viewing angle field distribution map, the laser pulse energy, imprinting pressure and curing dosage are fine-tuned in the region, and the optical phase grating elimination reaction is triggered to obtain the quality judgment result of the anti-peep light guide plate, including: Perform error vector field extraction processing on the dynamic privacy viewing angle field distribution map to obtain the fine-tuning instruction matrix of the coverage panel grid; According to the fine-tuning instruction matrix, a space-time solution process is performed on the laser pulse energy sequence, the imprint pressure sequence and the curing dose sequence to obtain a synchronous parameter modulation profile; According to the synchronization parameter modulation profile, regional modulation processing is performed on the roll-to-roll process time axis, and an optical phase grid elimination reaction is synchronously triggered at the corresponding plate grid to obtain a process execution record; The privacy viewing angle field snapshot is recalculated according to the process execution record, and a tolerance comparison process is performed with the original dynamic privacy viewing angle field distribution map to obtain a quality determination result.

9. A detection device for an anti-peep light guide plate, characterized in that: The detection device for the anti-peep light guide plate adopts the detection method for the anti-peep light guide plate according to any one of claims 1 to 8, and the detection device for the anti-peep light guide plate includes: The phase grating writing module is used to write a planar moiré fringe pattern by roll-coating a photosensitive phase change coating and writing it under a low-energy mask light field. This module performs an optical phase grating writing process on the surface of the optical substrate before entering the micro-engraving station, thereby obtaining a phase reference data matrix calibrated in two dimensions: spatial coordinates and energy dose. A resonance spectrum analysis module is used to perform a same-grid point superposition process on the guided mode resonance spectrum sequence obtained during the laser etching process based on the phase reference data matrix to obtain a resonance spectrum feature data set that characterizes the dynamic evolution of the microprism depth and tilt angle; a multi-physics field coupling module for performing spatiotemporal synchronous mapping processing on the Barkhausen transition signal sequence and the thermal radiation grayscale field generated by the embedded Fe-Ga alloy nanowire according to the resonance spectrum characteristic data set, thereby obtaining a multi-physics field coupling data set representing the stress-geometry-optical correlation; A topological dimensionality reduction modeling module is used to perform topological persistent coherence dimensionality reduction modeling on the privacy viewing angle half parallax angle, brightness half-width, and lateral stray light threshold based on the multi-physics field coupling data set to obtain a dynamic privacy viewing angle field distribution map representing the continuous distribution of the panel surface; The adaptive control module is used to implement regional fine-tuning control of laser pulse energy, imprinting pressure and curing dosage according to the dynamic privacy viewing angle field distribution map, and trigger the optical phase grating elimination reaction to obtain the quality judgment result of the anti-peep light guide plate.

10. A detection device for an anti-peep light guide plate, characterized in that: The detection device for the anti-peep light guide plate includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor calls the instruction in the memory to enable the anti-peeping light guide plate detection device to perform the steps of the anti-peeping light guide plate detection method according to any one of claims 1 to 8.

Images