A machine vision-based intelligent detection method for surface defects of a protective film

Abstract

This application provides a machine vision-based intelligent detection method for surface defects in protective films, comprising: extracting stress concentration areas and touch signal attenuation features from potential risks of touch response delay; performing group analysis on the stress concentration areas and touch signal attenuation features to obtain a joint attenuation trend of optical transmittance and touch signal attenuation; identifying the deformation degree of stress concentration areas and determining the touch signal attenuation degree of folding areas based on the joint attenuation trend of optical transmittance and touch signal attenuation to assess the durability of the folding screen; extracting early warning indicators for screen curvature radius and light scattering intensity defects from the attenuation trend prediction results, and generating a complete report including optical transmittance and touch signal attenuation defects.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to an intelligent detection method for surface defects of protective films based on machine vision. Background Technology

[0002] Foldable screen technology, as a core breakthrough in the field of smart devices, greatly expands the functionality and portability of mobile devices due to its foldable characteristics, playing a crucial role in improving user experience and driving industrial upgrading. However, existing performance analysis methods for foldable screen protectors have significant shortcomings in dynamic usage scenarios. These methods mostly rely on static testing, making it difficult to capture the real-time degradation trend of optical and touch performance during repeated folding, resulting in an inability to accurately predict the durability of the protector. Furthermore, existing solutions lack a comprehensive consideration of multi-physics coupling effects when analyzing the correlation between the complex stress distribution in the folding area and changes in polarized light reflectivity. Especially during the dynamic process of the foldable screen moving from flat to its extreme folding angle, the polarized light reflectivity of the screen display area gradually increases, directly leading to a decrease in display clarity. Areas of decreased clarity are often accompanied by stress concentration, and the unevenness of stress distribution further weakens the touch sensitivity of the protector. This interactive degradation of optical and touch performance constitutes the core challenge of performance evaluation for foldable screen protectors. The enhancement of polarized light reflectivity originates from the microstructural changes of the material in the folding area under repeated deformation, and these changes induce uneven stress distribution. For example, after a foldable screen is repeatedly bent 1000 times, the protective film near the folding axis may experience localized stress concentration, causing a delay in touch signal response. Therefore, how to analyze the degradation trend of optical and touch performance in the folding area by inputting continuous variation signals of polarized light reflectivity and stress distribution, and generate an early warning report on the performance degradation of the foldable screen protective film, has become a key issue in improving the durability of foldable screens and the user experience. Summary of the Invention

[0003] This invention provides a machine vision-based intelligent detection method for surface defects in protective films, mainly comprising:

[0004] By collecting polarized light reflectivity and stress distribution data of a foldable screen from its flat state to its extreme folding angle, a reflectivity and stress distribution dataset is constructed. Feature extraction is performed on this dataset to obtain an initial optical performance degradation index, which includes a reflectivity attenuation coefficient and a stress distribution non-uniformity value. Based on the polarized light reflectivity and stress distribution data, a display sharpness value is calculated. By analyzing the correspondence between the initial optical performance degradation index and the display sharpness value, a sharpness decline trend is determined. Touch response delay features are extracted based on this trend. From the touch response delay features, stress concentration areas and touch signal attenuation features are extracted. The touch signal attenuation features include signal delay time and attenuation amplitude. The attenuation characteristics of the central region and touch signal are grouped and analyzed to obtain the joint attenuation trend of optical transmittance and touch signal attenuation. Based on the joint attenuation trend of optical transmittance and touch signal attenuation, the degree of touch signal attenuation in the folded area is calculated. Combined with the degree of deformation in the stress concentration area, a durability index is obtained. Through the durability index and the degree of deformation, a comprehensive performance attenuation index is calculated, which includes optical transmittance attenuation, touch response delay, and material fatigue. Based on the comprehensive performance attenuation index, the relationship between the change of folding angle and optical refractive index is analyzed to determine the rate of change of folding angle and the cumulative value of material fatigue, and an attenuation trend prediction result is obtained. The screen curvature radius and light scattering intensity are extracted from the attenuation trend prediction result to obtain early warning index data.

[0005] Furthermore, by collecting polarized light reflectivity and stress distribution data of the foldable screen from its flat state to its extreme folding angle, a reflectivity and stress distribution dataset is constructed. Feature extraction is then performed on this dataset to obtain an initial optical performance degradation index, including:

[0006] Polarized light reflection signals are collected at preset angle intervals, and pressure values ​​at each measuring point at the corresponding angle are obtained simultaneously. A dataset of reflectivity and stress distribution is constructed. The difference between the reflectivity at each folding angle and the reflectivity in the initial flat state is calculated and divided by the initial reflectivity to obtain the attenuation ratio. The folding angle and attenuation ratio are linearly fitted using the least squares method, and the slope of the fitted line is used as the reflectivity attenuation coefficient. The square root of the sum of the squares of the differences between the stress values ​​at all measuring points and the average stress at each folding angle is divided by the number of measuring points to obtain the standard deviation. The ratio of the standard deviation to the average stress is used as the stress distribution non-uniformity value. The initial attenuation index of optical performance, which includes the reflectivity attenuation coefficient and the stress distribution non-uniformity value, is obtained.

[0007] Furthermore, the step of calculating the display sharpness value based on the polarized light reflectivity and stress distribution data, and determining the sharpness decline trend by analyzing the correspondence between the initial optical performance attenuation index and the display sharpness value, includes:

[0008] A reflectance distribution matrix is ​​generated based on the polarized light reflectance, a stress field distribution map is constructed using the stress distribution data, the spatial frequency response of the display area is calculated, and the display sharpness value is obtained. The display sharpness value and the initial optical performance attenuation index are processed by regression analysis to generate a sharpness decline trend curve.

[0009] Furthermore, the step of extracting stress concentration regions and touch signal attenuation features from the touch response delay features, wherein the touch signal attenuation features include signal delay time and attenuation amplitude, and performing group analysis on the stress concentration regions and touch signal attenuation features to obtain the joint attenuation trend of optical transmittance and touch signal attenuation, includes:

[0010] Read the spatial coordinates in the touch response delay feature, determine the boundary of the stress concentration area, collect the touch signal waveform within the boundary, calculate the signal delay time and the attenuation amplitude; group according to the signal delay time, construct a feature combination matrix, measure the optical transmittance at the position within the matrix, generate a regression equation between transmittance and attenuation amplitude, and fit to obtain a joint attenuation trend curve.

[0011] Furthermore, based on the combined attenuation trend of optical transmittance and touch signal attenuation, the degree of touch signal attenuation in the folded area is calculated, and combined with the degree of deformation in the stress concentration area, a durability index is obtained, including:

[0012] The surface profile of the stress concentration area is scanned, the local radius of curvature is calculated, and a deformation degree distribution map is generated; the touch signal attenuation degree is obtained according to the joint attenuation trend, the damage rate is calculated, the degradation curve is fitted, and the durability index is obtained.

[0013] Furthermore, the comprehensive performance degradation index is calculated using the durability index and the degree of deformation. This comprehensive performance degradation index includes optical transmittance degradation, touch response delay, and material fatigue, including:

[0014] The maximum values ​​of the durability index and the degree of deformation are obtained, and the comprehensive performance degradation index is calculated by weighted summation; the polarization state change of the folded region is measured, and the optical refractive index is calculated by inversion; the refractive index change rate is extracted according to the relationship between the optical refractive index and the folding angle; if the refractive index change rate exceeds the threshold, the components in the comprehensive performance degradation index are combined to generate an early warning signal.

[0015] Furthermore, the comprehensive performance degradation index is calculated using the durability index and the degree of deformation. This comprehensive performance degradation index includes optical transmittance degradation, touch response delay, and material fatigue, including:

[0016] Data on folding times and ambient temperature are collected, and the surface hardness and tensile strength of the protective film are measured to generate a dataset of material mechanical properties. The distribution of deformation degree is scanned to determine the boundary of the deformation region. The optical transmittance, touch response time and elastic modulus of the material within the boundary are measured to calculate the attenuation rate and fatigue life index, and to determine the contribution ratio of material aging and stress concentration.

[0017] Furthermore, based on the comprehensive performance degradation index, the relationship between the folding angle change and the optical refractive index is analyzed to determine the folding angle change rate and the material fatigue accumulation value, thereby obtaining the degradation trend prediction result, including:

[0018] The deformation depth of the stress concentration region is measured, the relative damage is calculated, and the cumulative fatigue value of the material is obtained by summing the results; the folding angle sequence is recorded, and the rate of change of the folding angle is calculated; according to the warning level of the comprehensive performance degradation index, the rate of change of the folding angle and the cumulative fatigue value of the material are adjusted to generate a predicted degradation value sequence.

[0019] Furthermore, the step of analyzing the relationship between the folding angle change and the optical refractive index based on the comprehensive performance degradation index, and determining the folding angle change rate, includes:

[0020] Measure the initial folding angle, record the dynamic offset, and generate the folding angle evolution trajectory; extract the stress distribution based on the evolution trajectory, calculate the material rigidity and the angular response to external loads, and determine the rate of change of the folding angle.

[0021] Furthermore, the step of extracting the screen curvature radius and light scattering intensity from the attenuation trend prediction results to obtain early warning index data includes:

[0022] Calculate the screen curvature radius and light scattering intensity based on the attenuation trend prediction results, and mark the warning status; extract the optical transmittance and touch signal response time at the corresponding positions, classify the defect types, and generate a data summary table containing risk levels.

[0023] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0024] This invention discloses a machine vision-based intelligent detection method for surface defects in protective films, addressing the challenges of optical performance, touch response, and durability degradation in foldable screens under repeated folding and various usage scenarios. By collecting data on polarized light reflectivity and stress distribution, this invention extracts initial optical performance degradation indicators and touch signal attenuation characteristics, analyzes their correlation with display clarity and touch response latency, and generates performance degradation early warning signals. Simultaneously, by monitoring changes in folding angle and the dynamic evolution of stress concentration areas, combined with material fatigue accumulation and environmental factors, the rate of change of folding angle is optimized to predict degradation trends. This invention integrates optical transmittance, touch latency, and material fatigue to generate a complete report including screen curvature radius and light scattering intensity defects, significantly improving the accuracy of foldable screen durability assessment and lifespan prediction capabilities, providing comprehensive technical support for the design optimization and performance maintenance of foldable screen devices. Attached Figure Description

[0025] Figure 1 This is a flowchart of a machine vision-based intelligent detection method for surface defects in protective films according to the present invention.

[0026] Figure 2 This is a schematic diagram of a machine vision-based intelligent detection method for surface defects in protective films according to the present invention. Detailed Implementation

[0027] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] like Figure 1-2 This embodiment of a machine vision-based intelligent detection method for surface defects in protective films may specifically include:

[0029] Step S101: By collecting the changes in polarized light reflectivity and stress distribution of the foldable screen from its flat state to its extreme folding angle, the polarized light reflectivity and the stress concentration area obtained based on the stress distribution changes are classified to obtain a reflectivity and stress distribution dataset. Feature extraction is performed on the reflectivity and stress distribution dataset to obtain the initial attenuation index of optical performance.

[0030] A foldable screen is continuously irradiated from its flat state to its extreme folded state using a polarized light irradiation device. Polarized light reflection signals are collected at preset angle intervals. Simultaneously, pressure sensors acquire pressure values ​​at corresponding measuring points at each angle. The ratio of the reflected signal intensity to the incident light intensity is used as reflectivity data. The pressure value is divided by the area of ​​force application to obtain the stress value, forming a raw dataset containing folding angles, reflectivity, and stress distribution. Differential operations are performed on the stress distribution at each folding angle in the raw dataset. The difference in stress values ​​between adjacent measuring points is calculated and divided by the distance between measuring points to obtain the stress change rate. When the stress change rate exceeds a preset threshold, the area is marked as a stress concentration area. The reflectivity change amplitude and stress concentration degree of each data point are extracted as feature vectors. A K-means clustering algorithm is used to cluster these feature vectors. The input is a two-dimensional feature vector, and the output is four cluster centers and a category label for each data point, constructing a classified and labeled reflectivity and stress distribution dataset. For the classified and labeled reflectivity and stress distribution dataset, the difference between the reflectivity at each folding angle and the reflectivity in the initial flat state is calculated and divided by the initial reflectivity to obtain the attenuation ratio. The folding angle and attenuation ratio are linearly fitted using the least squares method, and the slope of the fitted line is used as the reflectivity attenuation coefficient. The sum of the squares of the differences between the stress values ​​at all measuring points and the average stress at each folding angle is calculated and divided by the number of measuring points to obtain the standard deviation. The ratio of the standard deviation to the average stress is used as the stress distribution non-uniformity value, thus obtaining the initial attenuation index of optical performance that includes the reflectivity attenuation coefficient and the stress distribution non-uniformity value.

[0031] For example, polarized light irradiation devices play a crucial role in foldable screen testing. Their principle is based on the sensitivity of light polarization characteristics to changes in the microstructure of a material surface. When polarized light irradiates the surface of a foldable screen, the microscopic deformation of the screen material at different folding angles causes a change in the polarization state of the reflected light. This change directly reflects the alteration of the material's optical properties.

[0032] In one possible implementation, the polarized light source consists of LEDs and polarizers, emitting green light with a wavelength of 550nm, as the human eye is most sensitive to green light, and the foldable screen display material has a relatively stable response to this wavelength. The pressure sensors are arranged in a matrix with a spacing of 5mm, covering the entire folded area. The stress value at a point is obtained by dividing the pressure value measured by each sensor by its effective sensing area of ​​0.25 square centimeters. This dense arrangement can accurately capture detailed changes in stress distribution.

[0033] It should be noted that the stress distribution during the folding process exhibits significant non-uniformity. Stress values ​​near the folding axis can reach 50 MPa, while in areas far from the folding axis, the stress value is only around 5 MPa. This stress gradient is one of the main reasons for the performance degradation of foldable screens. Differential operations play a crucial role in highlighting local anomalies in the calculation of stress change rate.

[0034] Specifically, if the stress values ​​at two adjacent measuring points are 30 MPa and 35 MPa, and the distance between the measuring points is 5 mm, then the stress change rate is 1 MPa / mm. When this value exceeds a preset threshold of 0.8 MPa / mm, the area is identified as a stress concentration region. This identification method can effectively locate potential fatigue damage locations. The application of the K-means clustering algorithm makes data classification more scientific and reasonable. The two-dimensional feature vector input to the algorithm contains two dimensions: the magnitude of reflectivity change and the degree of stress concentration.

[0035] In one embodiment, the reflectivity decreased from an initial 0.92 to 0.85, a change of 7.6%. The stress concentration was represented by the ratio of the local maximum stress to the average stress, such as 2.5. Through iterative calculation, the algorithm categorized data points to the nearest cluster centers, ultimately forming four categories. The high-stress, high-attenuation category corresponds to the folding axis region, which experiences both maximum stress and significant optical attenuation; the low-stress, low-attenuation category corresponds to flat regions far from the folding area, which largely maintain their original performance. The least squares fitting process determines the best-fitting straight line by minimizing the sum of squared residuals. As the folding angle changes from 0 degrees to 180 degrees, the reflectivity attenuation ratio increases from 0 to 15%, and the slope of the fitted straight line is approximately 0.083% / degree. This slope value is the reflectivity attenuation coefficient. This coefficient directly reflects the rate of attenuation of the optical performance of the foldable screen with the degree of folding. The calculation of the standard deviation reflects the dispersion of the stress distribution. When the standard deviation is 12 MPa and the average stress is 20 MPa, the stress distribution non-uniformity is 0.6, indicating a significant non-uniformity in the stress distribution. These two indicators together constitute the initial degradation index of optical performance.

[0036] Step S102: Determine display clarity by polarized light reflectivity and stress distribution. By analyzing the correspondence between the initial attenuation index of optical performance and display clarity, obtain the decreasing trend of clarity when polarized light reflectivity and stress distribution increase. Determine the potential risk of touch response delay based on the decreasing trend of clarity.

[0037] A reflectance distribution matrix is ​​constructed based on polarized light reflectance data. A stress field distribution map is generated using the stress distribution data. The spatial frequency response of the display area is calculated using a modulation transfer function (MTF), which is a function of the ratio of output image contrast to input image contrast as a function of spatial frequency. This yields a quantitative evaluation value for display sharpness. The quantitative evaluation value for display sharpness is obtained, and combined with the reflectance attenuation coefficient and stress distribution unevenness in the initial optical performance attenuation index, a functional relationship between sharpness and attenuation index is established through multiple regression analysis to determine the sharpness decline trend curve. Based on the sharpness decline trend curve, the rate of change of the curve slope is calculated. If the rate of change exceeds a preset threshold, this moment is determined to be a point of sharp sharpness deterioration. The location of the folded area corresponding to the sharp deterioration point and its stress concentration degree are analyzed. Using the correspondence between the stress concentration degree and the sharp sharpness deterioration point, the degree of obstruction of the signal transmission path of the touch sensor in this area is assessed. By comparing the touch response time with that of the normal area, the response time extension ratio is calculated to obtain the potential risk level of touch response delay.

[0038] In one embodiment, the reflectivity distribution matrix is ​​constructed by dividing the foldable screen surface into 32×32 grid cells, with each cell recording the polarized light reflectivity value at that location. The stress field distribution map is characterized in the form of a heat map, with color depth corresponding to stress magnitude; red areas represent high stress concentration areas, and blue areas represent low stress areas. The modulation transfer function is measured using a sinusoidal grating test pattern. Grating patterns of different spatial frequencies are projected onto the foldable screen, and the contrast ratio between the output image and the input image is measured. The spatial frequency range is from 0.1 line pairs / mm to 10 line pairs / mm.

[0039] It should be noted that the quantitative evaluation value of display sharpness is obtained by calculating the integral area of ​​the modulation transfer function at a specific spatial frequency. When the foldable screen is in a flat state, the modulation transfer function curve is close to the ideal value of 1.0. As the folding angle increases, the curve gradually declines, especially in the high-frequency part, where the attenuation is more obvious, indicating a decrease in detail resolution.

[0040] Specifically, the multiple regression analysis employed the least squares method. Independent variables included the reflectivity attenuation coefficient, stress distribution heterogeneity, and folding angle, while the dependent variable was a clearly quantifiable value. The regression coefficients were obtained through iterative optimization, and the goodness-of-fit R² was [value missing]. 2 A value of 0.85 or higher indicates the model is valid. The resolution decline curve exhibits a non-linear characteristic, with the fastest rate of decline occurring within the folding angle range of 60 to 90 degrees.

[0041] Preferably, the rate of change of the curve slope is calculated using the three-point difference method, which approximates the second derivative values ​​of three adjacent data points to determine the changes in the curve's concavity and convexity. When the second derivative changes from negative to positive or its absolute value suddenly increases, it is marked as a potential point of sharp deterioration. The preset threshold is determined based on the characteristics of the protective film material and the usage requirements, and is typically set to twice the average rate of change.

[0042] In one possible implementation, the degree of obstruction to the touch sensor signal transmission path is assessed by measuring the change in mutual capacitance between the touch electrodes. Stress concentration in the folded area causes localized deformation of the protective film, altering the dielectric constant distribution between the electrodes and thus affecting the capacitive coupling effect. The touch response time in the normal area is approximately 10 milliseconds. When the response time in the stress-concentrated area exceeds 15 milliseconds, it is considered to pose a moderate risk, and exceeding 20 milliseconds is considered a high risk.

[0043] Step S103: Extract stress concentration areas and touch signal attenuation characteristics from the potential risks of touch response delay, and perform group analysis on the stress concentration areas and touch signal attenuation characteristics to obtain the joint attenuation trend of optical transmittance and touch signal attenuation.

[0044] Spatial coordinate information of each risk area is read from the potential risk level data of touch response delay. The boundary range of the stress concentration area is determined by coordinate mapping. The original waveform of the touch signal at each measurement point within the boundary range is collected, and the time difference from signal transmission to reception is calculated as the signal delay time. By comparing the reference signal strength with the actual received signal strength, the attenuation amplitude value of each measurement point is obtained. According to the signal delay time and attenuation amplitude values, the data points are grouped according to whether the delay time exceeds a preset threshold. Those exceeding the threshold are classified into the high delay group, and those not exceeding the threshold are classified into the low delay group. The data points in each group are sorted according to the attenuation amplitude, and a feature combination matrix containing two dimensions, delay time and attenuation amplitude, is constructed. Using the folding screen position corresponding to each data point in the feature combination matrix, the optical transmittance of the protective film at that position is measured. The optical transmittance is measured in the visible light band by a spectrophotometer to obtain a dataset of the correspondence between transmittance and attenuation amplitude. According to the dataset of the correspondence, the correlation coefficient between the percentage decrease in optical transmittance and the attenuation amplitude of the touch signal is calculated. If the correlation coefficient exceeds a preset value, it is determined that there is a strong correlation between the two, and a linear regression equation between transmittance and attenuation amplitude is established. The linear regression equation is used to predict the optical transmittance and touch signal attenuation at different folding angles. By performing time series analysis on the predicted values, a joint attenuation trend curve of optical transmittance and touch signal attenuation is obtained.

[0045] In one embodiment, the potential risk level data of touch response latency is stored in a two-dimensional array structure. Each array element contains three attributes: risk level value, X coordinate, and Y coordinate. The risk level is represented by a value from 0 to 10, where 0 represents no risk and 10 represents extremely high risk. The spatial coordinates are centered on the folding axis, with the X-axis parallel to the folding axis and the Y-axis perpendicular to the folding axis, and the coordinate unit is millimeters. By traversing the array to identify elements with a risk level greater than 5, their coordinate values ​​are extracted to form a set of boundary points for the stress concentration region. A convex hull algorithm is then used to connect these boundary points to obtain the closed boundary range.

[0046] Specifically, the raw waveform of the touch signal was acquired using an oscilloscope with a sampling rate of 1MHz and a recording duration of 100ms. The touch driving electrode emitted a 200kHz sine wave signal, and the receiving electrode captured the signal after it passed through the protective film. The signal delay time was calculated using a cross-correlation function, specifically by performing a sliding correlation operation between the transmitted and received signals; the time offset corresponding to the maximum correlation coefficient was the delay time. The reference signal strength was measured with the protective film folded, and the average of 10 measurements was used as a reference. The attenuation amplitude was calculated using the formula 20×log10(received signal amplitude / reference signal amplitude), in decibels.

[0047] It should be noted that the preset threshold for data grouping is determined based on the physiological characteristics of human touch perception. Research shows that when the touch latency exceeds 50 milliseconds, users begin to perceive a noticeable lag. Therefore, 50 milliseconds is set as the grouping threshold; data points exceeding this value indicate that the touch experience in that area has been affected. The feature combination matrix adopts an M×2 form, where M is the total number of measurement points. The first column stores the latency time, and the second column stores the attenuation magnitude. The data in the matrix are arranged in spatial order, with adjacent rows corresponding to measurement points that are physically adjacent.

[0048] Preferably, optical transmittance is measured using a dual-beam spectrophotometer, covering the visible spectrum from 380 nm to 780 nm. During measurement, the foldable screen protective film sample is fixed on a sample holder, with the reference beam passing through air and the sample beam passing through the protective film. Transmittance is defined as the ratio of transmitted light intensity to incident light intensity. After measurement at each wavelength point, the average transmittance across the entire wavelength band is calculated as the optical transmittance value at that location. The spatial resolution of the measurement points is 1 square millimeter to ensure a one-to-one correspondence with the touch signal measurement points. The correspondence dataset contains at least 100 data pairs, each recording the transmittance value and attenuation amplitude value at the same location.

[0049] In one possible implementation, the correlation coefficient is calculated using the Pearson correlation coefficient method. This method calculates the correlation coefficient by dividing the covariance of the two variables by the product of their respective standard deviations, with a value ranging from -1 to 1. When the absolute value of the correlation coefficient is greater than 0.7, the two variables are considered to be strongly correlated. The linear regression equation is fitted using the least squares method, with the form Y = aX + b, where Y represents the touch signal attenuation magnitude, X represents the percentage decrease in optical transmittance, a is the slope, and b is the intercept. The regression coefficient 'a' reflects the sensitivity of optical performance degradation to the impact on touch performance; a larger value of 'a' indicates a more significant impact.

[0050] For example, in actual testing, a protective film that underwent 5000 folding cycles showed a decrease in optical transmittance at the center of the folded area from an initial 92% to 78%, with the corresponding touch signal attenuation increasing from -3dB to -12dB. The equation obtained through linear regression analysis is: Attenuation Amplitude = 0.64 × Percentage Decrease in Transmittance - 1.8. This equation can be used to predict performance degradation at different folding cycles. Furthermore, time series analysis employed an autoregressive moving average model. This model considers the trend and periodicity of historical data, predicting future trends by analyzing data from the past N time points. Model parameters were determined using maximum likelihood estimation, with the prediction interval set at a 95% confidence level. The joint attenuation trend curve exhibits an S-shaped characteristic: slow initial degradation, accelerated mid-term degradation, and a later stabilization. The inflection point of the curve corresponds to the critical point where the protective film transitions from elastic deformation to plastic deformation, at which point the material's microstructure begins to undergo irreversible changes.

[0051] For example, as the folding angle increases from 0 degrees to 180 degrees, the combined attenuation trend curve shows that the attenuation rate is fastest in the range of 60 to 120 degrees, which is consistent with the physical characteristics of the smallest radius of curvature and the most severe stress concentration within this angle range. When the folding angle exceeds 150 degrees, the attenuation rate actually decreases because the protective film has adapted to the extreme folding state.

[0052] Understandably, by establishing a joint attenuation trend between optical transmittance and touch signal attenuation, a quantitative evaluation of the overall performance of foldable screen protectors can be achieved. This trend curve not only reflects changes in a single performance indicator but also reveals the intrinsic correlation and mutual influence mechanism between optical performance and touch performance.

[0053] Step S104: Identify the degree of deformation in the stress concentration area, and determine the degree of touch signal attenuation in the folding area based on the combined attenuation trend of optical transmittance and touch signal attenuation, thereby evaluating the durability of the folding screen.

[0054] The surface contour of the stress concentration area is scanned by a laser displacement sensor to obtain the deformation displacement value of each measurement point relative to the initial plane. The local radius of curvature is calculated based on the arc formed by three adjacent measurement points, where the radius of curvature refers to the radius of the circle determined by the three points, resulting in a deformation degree distribution map. Using the maximum deformation displacement value in the deformation degree distribution map, the corresponding attenuation value is found in the joint attenuation trend of optical transmittance and touch signal attenuation. The touch signal attenuation level of the folded area is determined by comparing the attenuation value with a preset level threshold. If the attenuation level exceeds a preset performance threshold, the number of folding cycles the folded screen has undergone is recorded, and the ratio of the attenuation level to the number of folding cycles is calculated as the damage rate. The change pattern of the damage rate with usage time is obtained. A degradation curve is established using an exponential fitting method based on this change pattern. The number of folds required to reach the failure threshold (the attenuation value at which the touch function completely fails) is predicted based on the degradation curve, thus evaluating the durability index of the folded screen under current usage conditions.

[0055] For example, based on the laser triangulation method, deformation displacement data is obtained by illuminating the surface of the foldable screen with a laser beam and receiving the reflected light signal.

[0056] In one possible implementation, the sensor scans point-by-point along the stress concentration region, recording the vertical displacement relative to the initial plane at each measurement point. This displacement directly reflects the degree of local deformation of the material during folding. When the folded screen bends from a flat state to a certain angle, the stress concentration region will produce obvious surface undulations, and the laser sensor can accurately capture these micrometer-level deformation differences. The method for calculating the local radius of curvature is based on the geometric principle that a circle is determined by three points, constructing the arc equation using the spatial coordinates of three adjacent measurement points.

[0057] Specifically, after obtaining displacement data from three measurement points, the radius of the circle passing through these three points is calculated using geometric relationships. This radius is the local radius of curvature. The smaller the radius of curvature, the more severe the curvature in that region, and the more concentrated the stress on the material.

[0058] For example, in the region near the folding axis, the radius of curvature is typically in the range of 2-5 mm, while in the region far from the folding axis, the radius of curvature can reach over 20 mm. This difference reflects the significant difference in stress at different locations. The deformation distribution map is constructed based on the displacement data and radius of curvature information of all measurement points, and the deformation status of the entire stress concentration area is visually displayed through color mapping or contour lines.

[0059] It should be noted that the maximum deformation displacement typically occurs at the center of the folding axis, where the greatest bending stress is experienced, and the deformation displacement can reach hundreds of micrometers. This maximum value becomes a key parameter for subsequent optical transmittance and touch signal attenuation analysis, as the maximum deformation of the material directly affects optical performance and the integrity of the capacitive touch layer. The joint attenuation trend of optical transmittance and touch signal attenuation establishes a quantitative relationship between the degree of deformation and functional performance.

[0060] In one embodiment, when the maximum deformation displacement reaches 200 micrometers, optical transmittance may decrease by 15%, while touch signal attenuation increases by 25%. This combined attenuation reflects the comprehensive impact of mechanical deformation on multiple functions. By finding the corresponding attenuation value and comparing it with a preset threshold, the continuous attenuation level can be divided into discrete levels such as mild, moderate, and severe, facilitating subsequent performance evaluation and early warning processing. Determining the touch signal attenuation level provides a standardized indicator for evaluating the usage status of foldable screens.

[0061] For example, when the degradation level exceeds the moderate threshold, the system begins recording the current number of folding cycles, which reflects the historical usage intensity of the foldable screen. The damage rate is calculated as the ratio of the degradation level to the number of folding cycles, characterizing the degree of performance degradation per unit of folds. The variation of the damage rate over usage time reveals the dynamic characteristics of foldable screen performance degradation.

[0062] Preferably, the damage rate is relatively low and stable in the initial stage of use, and then accelerates with the increase of folding cycles. This nonlinear change conforms to the basic law of material fatigue damage. The exponential fitting method can accurately describe this accelerated degradation process and predict future performance changes by establishing a degradation curve. The durability index assessment is based on the extrapolation prediction of the degradation curve, and the remaining service life of the foldable screen is quantified by calculating the number of folds required to reach the failure threshold.

[0063] Step S105: Integrate durability and surface deformation to obtain a comprehensive performance degradation index, analyze the change in optical refractive index, and determine whether the change in optical refractive index leads to accelerated touch signal degradation based on the change in folding angle. If so, generate a performance degradation warning signal including touch response delay based on the comprehensive performance degradation index.

[0064] The durability index and the maximum deformation value in the surface deformation distribution map are obtained. After normalization, they are weighted and summed according to preset weighting coefficients. These weighting coefficients are determined based on the impact of optical transmittance attenuation, touch response delay, and material fatigue on overall performance. The sum of the products of each value and its corresponding weight is calculated to obtain the comprehensive performance attenuation index. Using the folded region corresponding to the comprehensive performance attenuation index, the polarization state change of reflected light in this region at different incident angles is measured using an ellipsometer. Based on the relationship between the change in polarization state parameter and the material refractive index, the optical refractive index is calculated by inversion using the Fresnel equation. The change data of refractive index with the number of folds is recorded. Based on the refractive index change data, the difference in refractive index between adjacent folds is calculated as the rate of change. The continuous change process of the folding angle from zero degrees to the maximum folding angle is monitored, and the refractive index value corresponding to each angle position is recorded to establish a correspondence between refractive index change and folding angle. Using the refractive index change rate data in the correspondence, it is determined whether the rate of change exceeds a preset threshold. If it does, it is determined that the change in optical refractive index causes accelerated attenuation of the touch signal, and the values ​​of each component in the current comprehensive performance attenuation index are extracted. Based on the values ​​of each component, the touch response delay time, the percentage decrease in optical transmittance, and the material fatigue value are combined to set a warning level indicator and generate a performance degradation warning signal that includes touch response delay.

[0065] In one embodiment, the comprehensive performance degradation index is constructed using the analytic hierarchy process (AHP) to determine the weighting coefficients of each performance parameter. The specific process includes constructing a judgment matrix and determining the relative importance of the three factors—optical transmittance degradation, touch response latency, and material fatigue—by comparing them pairwise. Optical transmittance degradation directly affects the display effect, and its weighting coefficient is set to 0.4; touch response latency affects the user interaction experience, and its weighting coefficient is 0.35; material fatigue reflects long-term performance, and its weighting coefficient is 0.25. Normalization is performed using the maximum-minimum method, mapping each value to the interval between 0 and 1. The calculation formula is: the normalized value equals the current value minus the minimum value, divided by the difference between the maximum and minimum values.

[0066] It should be noted that during ellipsometer measurements, the incident light is linearly polarized, but becomes elliptically polarized after reflection by the protective film. Ellipsometry parameters include the amplitude ratio ψ and the phase difference Δ, which have a definite functional relationship with the complex refractive index of the material. The Fresnel equations describe the reflection and refraction behavior of light at interfaces. By measuring the ellipsoid parameters and combining them with the known incident angle and the refractive index of the surrounding medium, the real part of the refractive index n and the extinction coefficient k of the protective film material can be calculated. The real part of the refractive index reflects the change in the speed of light in the material, while the extinction coefficient reflects the degree of light absorption by the material. When the protective film undergoes repeated folding, the molecular arrangement inside the material changes, leading to a change in the refractive index.

[0067] Specifically, the refractive index variation data was collected at a frequency of once every 100 folds, forming a time-series data. The rate of change was calculated using the finite difference method, i.e., the refractive index of the i-th measurement minus the refractive index of the (i-1)-th measurement, then divided by the fold interval. The folding angle was monitored and recorded in real time using an angle sensor, starting from the 0-degree flat state, recording the corresponding refractive index value every 10 degrees. This intensive data acquisition accurately captures the nonlinear variation characteristics of the refractive index with the folding angle.

[0068] Preferably, the preset threshold is determined based on statistical analysis of a large amount of experimental data. Tests on protective films of different materials and thicknesses revealed that when the rate of refractive index change exceeds 0.005 per thousand folds, the attenuation rate of the touch signal significantly accelerates. This threshold serves as a criterion, providing timely warnings before severe performance degradation. The threshold setting also considers environmental factors, such as the impact of temperature and humidity on material performance, and is adjusted using correction coefficients.

[0069] In one possible implementation, the values ​​of each component of the overall performance degradation index are obtained through independent measurement channels. The percentage decrease in optical transmittance is measured using a spectrophotometer in the visible light band, averaging the values ​​from 380 nm to 780 nm. Touch response delay time is measured using an oscilloscope, measuring the rise time of the touch signal with microsecond-level accuracy. Material fatigue is measured using a dynamic mechanical analyzer, measuring changes in the storage modulus to reflect the degree of degradation of the material's elastic properties. These three parameters reflect the performance status of the protective film from different dimensions, complementing each other and improving the comprehensiveness and accuracy of the evaluation.

[0070] For example, in practical applications, after 8000 folding cycles, the overall performance degradation index of a foldable screen protector reaches 0.72. Specifically, the optical transmittance decreases from an initial 93% to 81%, a degradation percentage of 12.9%; the touch response latency increases from 8 milliseconds to 18 milliseconds; and the material fatigue index is 0.65. Ellipsometry measurements show the refractive index changes from 1.52 to 1.48, a change rate of 0.005 per thousand folds. The system determines that the accelerated degradation stage has been reached and immediately generates a warning signal. Furthermore, the warning signal generation employs a multi-level warning mechanism. Based on the numerical range of the overall performance degradation index, three warning levels are set: a yellow warning when the index is between 0.3 and 0.5, prompting the user to pay attention to usage; an orange warning when the index is between 0.5 and 0.7, suggesting a reduction in folding frequency; and a red warning when the index exceeds 0.7, indicating the need to replace the protector. The warning signal data packet uses a structured format, including fields such as timestamp, location information, current values ​​of various performance parameters, warning level, and recommended measures.

[0071] For example, the triggering conditions for warning signals consider not only the absolute value of a single parameter, but also the trend of parameter change. When a continuous increase in the rate of refractive index change is detected in three consecutive measurements, a trend warning will be issued even if the current comprehensive index has not reached the warning threshold, reminding the user to pay attention to the rapid deterioration of the protective film's condition.

[0072] Data on the number of folds and ambient temperature related to the durability of the device in actual use are collected. The changes in hardness, scratch resistance and tensile strength of the protective film surface under repeated folding or pressure are analyzed. By recording the deformation traces of the surface at different folding angles, the area and degree of deformation are obtained. The durability performance data are integrated with the area and degree of deformation to obtain optical transmittance, touch response delay and material fatigue. The optical transmittance attenuation rate, touch delay increase and fatigue life index of surface roughness and folding stress under different usage scenarios are analyzed to identify the contribution of material aging and stress concentration to the overall performance degradation.

[0073] The system collects cumulative folding counts and real-time ambient temperature data during device use using sensors. A hardness tester measures the Vickers hardness of the protective film surface at different folding counts. A scratch tester records the critical load for surface scratch resistance. A tensile testing machine obtains the tensile strength and elongation at break of the material, resulting in a material mechanical property degradation dataset. Based on the folding count information in this dataset, a 3D profilometer scans the deformation marks on the protective film surface at corresponding folding counts, acquiring mark depth and width distribution maps to identify the boundaries of areas with concentrated deformation and determine the deformation degree level and spatial distribution characteristics. Using the spatial distribution characteristics to define the regions, optical transmittance, touch response time, and material elastic modulus are measured in each region. The measured values ​​are compared with initial values ​​to calculate the changes, obtaining optical transmittance attenuation, touch response delay increment, and material fatigue value. Based on the optical transmittance attenuation, touch response delay increment, and material fatigue value, surface roughness parameters and folding stress peak values ​​for the corresponding regions are extracted. The attenuation rate is obtained by dividing the attenuation value by the usage time, the touch delay increase is obtained by dividing the delay increment by the folding count, and the fatigue life index is calculated using the fatigue value. The proportion of performance degradation caused by material aging and the proportion of performance degradation caused by stress concentration are calculated using the attenuation rate, touch delay increase and fatigue life index. If the proportion of material aging exceeds the preset threshold, it is identified as material aging dominating, otherwise it is identified as stress concentration dominating, and the contribution degree of overall performance degradation is identified.

[0074] In one embodiment, the number of folds is collected using a Hall sensor integrated into the hinge mechanism of the folding screen. Each time the folding screen transitions from an unfolded state to a folded state and back to an unfolded state, it is counted as one complete folding cycle. The sensor stores the accumulated number of folds in non-volatile memory, with the data updated immediately after each folding action. Ambient temperature data is collected using a distributed temperature sensor array, with sensors positioned at the four corners and center of the protective film. The sampling frequency is once per minute, and the recorded temperature range is -20°C to 60°C.

[0075] Specifically, the Vickers hardness test uses a diamond pyramidal indenter, applying a load of 0.98 N for 15 seconds. The hardness value is calculated by measuring the diagonal length of the indentation, with each test location measured five times and the average value taken. The scratch tester is equipped with a conical diamond indenter with a radius of 200 micrometers, which scratches across the protective film surface at a constant speed, with the load gradually increasing from 0.1 N to 10 N. The critical load is defined as the load value at which the first visible scratch appears. The tensile test is performed using a standard dumbbell-shaped specimen with a gauge length of 25 mm, a tensile speed of 5 mm / min, and the stress-strain curve is recorded until the specimen breaks.

[0076] It should be noted that the 3D profilometer uses the principle of white light interferometry for non-contact measurement. The scanning accuracy reaches 0.1 nanometers in the vertical direction and 0.5 micrometers in the horizontal resolution. Deformation traces are identified by comparing surface topography data before and after folding. When the height difference in a certain area exceeds 1 micrometer from the reference plane, it is determined to be a deformation trace. The trace depth is defined as the difference between the lowest point of the deformed area and the average height of the surrounding undeformed area, and the width is the span of the deformed area in the folding direction. The degree of deformation is divided into three levels: slight, moderate, and severe, corresponding to depths of less than 5 micrometers, 5-15 micrometers, and greater than 15 micrometers, respectively.

[0077] Preferably, optical transmittance is measured at nine uniformly distributed measurement points within the deformation region, each with a measurement area of ​​1 square millimeter. The measurement wavelength range covers the visible spectrum from 400 to 700 nanometers, with a spectral resolution of 2 nanometers. The transmittance attenuation value is calculated by subtracting the current transmittance from the initial transmittance and then dividing by the initial transmittance, expressed as a percentage. Touch response time is measured using a dedicated testing platform that simulates finger touch actions, recording the time interval from the issuance of the touch signal to the system response, with a measurement accuracy of 0.1 milliseconds. The material's elastic modulus is measured using nanoindentation technology, with the indentation depth controlled within 10% of the film thickness, and calculated from the slope of the unloading segment of the load-displacement curve.

[0078] In one possible implementation, surface roughness parameters include arithmetic mean roughness Ra and root mean square roughness Rq. These parameters are obtained through statistical analysis of three-dimensional profile data, with an evaluation length set at 2.5 mm and a sampling length of 0.8 mm. The peak folding stress is obtained through finite element simulation combined with measured strain data. During folding, stress concentration mainly occurs within a ±5 mm region near the folding axis. The attenuation rate is calculated considering time and usage intensity, specifically by dividing the change in performance parameters by the product of usage time and folding frequency. The increase in touch latency reflects the increase in response time per unit number of folds, with the slope value obtained through linear regression analysis.

[0079] For example, the fatigue life index is calculated based on the SN curve theory, extrapolating the measured fatigue value and historical fold count to the expected number of folds at material failure. When the material fatigue value reaches 0.8, it is considered close to failure. In actual testing, a protective film had a fatigue value of 0.45 after 10,000 folds. Based on the fatigue accumulation law, the expected fatigue life is estimated to be approximately 22,000 folds. Furthermore, the contribution of material aging and stress concentration is determined through multiple regression analysis. Material aging factors include molecular chain breakage, changes in crosslinking density, and additive migration. The performance degradation caused by these factors is time-dependent and has a relatively weak correlation with the number of folds. Stress concentration factors mainly manifest as local plastic deformation and microcrack propagation, which are strongly correlated with the number of folds. By separating the contributions of these two types of factors to the overall performance degradation, the formulation and structural design of the protective film can be optimized in a targeted manner.

[0080] For example, in one real-world case, the optical transmittance of the protective film decreased by 18% after 15,000 folds, with material aging contributing 7% and stress concentration contributing 11%. This indicates that the failure of the protective film was mainly caused by mechanical damage, and its fatigue resistance needs to be significantly improved.

[0081] Step S106: Analyze the degree of deformation in the stress concentration area to obtain material fatigue accumulation, and obtain the folding angle change rate by analyzing the relationship between the folding angle change and the stress concentration area. Optimize the folding angle change rate and material fatigue accumulation based on the performance degradation warning signal to obtain the degradation trend prediction result.

[0082] The deformation depth of each point in the stress concentration area is measured using a laser scanner. The relative damage per fold is calculated based on the ratio of deformation depth to the initial material thickness. This relative damage is accumulated over the number of folds to obtain the cumulative fatigue value. An angle sensor records the folding angle at each moment during the folding process, extracting the stress peak value in the stress concentration area at different folding angles. The rate of change of the folding angle is obtained by dividing the difference between folding angles at adjacent moments by the time interval. Based on the rate of change of the folding angle and the cumulative fatigue value, the warning level in the obtained performance degradation warning signal is read. If the warning level exceeds a preset threshold, the rate of change is multiplied by a first coefficient, and the cumulative fatigue value is multiplied by a second coefficient. The first coefficient is greater than the second coefficient, and the two are added together to obtain a comprehensive degradation index. Using this comprehensive degradation index as input data for the time series, the predicted value for future moments is calculated using an exponential smoothing method. This exponential smoothing method uses a weighted average of current and historical values, setting the prediction range to a predetermined number of future folds to obtain a predicted degradation value sequence. Based on the predicted decay value sequence, the difference value between adjacent data points is calculated to identify the rate of change. When the rate of change exceeds the critical value, it is determined as a decay acceleration point. The number of folds corresponding to this point is recorded to obtain the decay trend prediction result.

[0083] In one embodiment, the laser scanner uses the triangulation principle for non-contact measurement. A laser beam is projected onto the protective film surface at a 45-degree angle, and the reflected light is received by a CCD camera. The surface height is calculated based on the offset of the light spot position. The scanning accuracy reaches 0.1 micrometers, and the scanning speed is 1000 measurement points per second. The deformation depth value is defined as the vertical distance between the current surface height and the initial plane reference. The measurement range covers an area of ​​20 millimeters on each side of the folding axis. The calculation of the relative damage is based on Miner's linear cumulative damage theory, which posits that damage to materials under cyclic loading is linearly cumulative, and material failure occurs when the cumulative damage reaches 1.

[0084] Specifically, the calculation process for the relative damage amount of a single fold includes: first, measuring the deformation depth d, and then dividing it by the initial material thickness t0 to obtain the relative deformation rate ε = d / t0. Based on the material's SN fatigue curve, the fatigue life N corresponding to this deformation rate is found, and the single-fold damage amount is 1 / N. The cumulative fatigue value of the material is obtained by summing the single-fold damage amounts for all folds. When this value is close to 1, it indicates that the material is approaching fatigue failure. Actual testing revealed that the fatigue accumulation of the protective film exhibits a non-linear characteristic, with slow accumulation in the initial stage and accelerated accumulation in the middle and later stages.

[0085] It should be noted that the angle sensor uses a magnetic encoder principle, achieving a resolution of 0.01 degrees and a sampling frequency of 100Hz. The sensor is installed at the hinge axis of the folding screen, accurately recording the dynamic changes during the folding process. Stress peak values ​​are extracted by pausing at different folding angles and holding for 5 seconds, allowing the stress distribution to stabilize before reading the maximum value from the stress sensor array. The rate of change of the folding angle not only reflects the user's habits but is also closely related to the dynamic load borne by the protective film. Rapid folding generates greater inertial forces, leading to increased stress peaks and accelerated material fatigue.

[0086] Preferably, the warning level threshold is set using a three-level system: low risk corresponds to a warning value less than 0.3, medium risk corresponds to 0.3 to 0.7, and high risk corresponds to greater than 0.7. The determination of the first and second coefficients is based on statistical analysis of a large amount of experimental data. When the warning level is high risk, it indicates that the protective film has entered a rapid degradation stage. At this time, the rate of change of the folding angle has a more significant impact on performance degradation, so a larger weighting coefficient of 0.7 is assigned; while the impact of material fatigue accumulation is relatively stable, and the weighting coefficient is set to 0.3. The formula for calculating the comprehensive degradation index I is I = 0.7 × V + 0.3 × F, where V is the normalized rate of change and F is the normalized fatigue accumulation value.

[0087] In one possible implementation, the exponential smoothing method involves choosing a smoothing coefficient α, which determines the weighting of historical and current data. For predicting the degradation of foldable screen protectors, the α value is typically set between 0.2 and 0.4, a range that balances prediction stability and response speed. The prediction formula is S(t+1) = α × X(t) + (1-α) × S(t), where S(t+1) is the predicted value at the next moment, X(t) is the current actual value, and S(t) is the current smoothed value. Through recursive calculation, the degradation trend over a future period can be obtained. The prediction range is typically set to 1000 to 5000 future folds, a range that provides sufficient warning time without compromising accuracy due to excessively long prediction times.

[0088] For example, the analysis of the predicted decay value sequence uses a combination of first-order and second-order differencing. The first-order differencing ΔY(i) = Y(i+1) - Y(i) reflects the decay rate, and the second-order differencing Δ... 2Y(i) = ΔY(i+1) - ΔY(i) reflects the decay acceleration. When the second-order difference changes from negative to positive or suddenly increases, it indicates that the decay has entered the acceleration stage. The critical value is set based on statistical analysis of historical data, usually taking twice the standard deviation of the average rate of change as the judgment standard. Furthermore, in practical applications, after a protective film has undergone 12,000 folds, the material fatigue cumulative value is measured to be 0.52, and the folding angle change rate is 15 degrees / second. The comprehensive decay index is calculated. Using the exponential smoothing method, it is predicted that within the next 3,000 folds, the decay index will reach 0.75, the corresponding touch response delay will exceed 30 milliseconds, and the display transmittance will drop below 75%. The decay acceleration point is expected to occur around the 14,500th fold, at which point it is recommended that the user replace the protective film.

[0089] For example, by comparing the prediction results of protective films made of different materials, the degradation trend of PET material shows a typical S-shaped curve, which is flat in the early stage, steep in the middle stage, and slows down in the later stage; while the degradation of TPU material is closer to linear. This difference is mainly due to the different molecular structure and mechanical properties of the materials.

[0090] The initial state of the folding angle within the stress concentration region is measured to establish a benchmark for angle change. The dynamic shift of the folding angle with stress change is continuously monitored to obtain the evolution trajectory of the folding angle. The spatial correspondence between the evolution trajectory and the stress concentration region is analyzed to identify the driving factors of angle change, including material rigidity and external load. Using the driving factors, the response differences of the folding angle under different stress conditions are evaluated to obtain regularity data of angle change. Based on the regularity data, the rate of change of the folding angle is determined through time series analysis.

[0091] The initial angle is measured by a high-precision angle sensor in the stress concentration area of ​​the folding screen when no external force is applied. The stress sensor readings corresponding to this initial angle are recorded to establish a one-to-one correspondence between angle and stress. During the folding process, real-time data from the angle sensor is continuously collected, and the difference between the real-time angle and the initial angle is calculated to obtain the dynamic offset. The offsets are arranged in chronological order to form the evolution trajectory of the folding angle. Based on the angle values ​​and position information at each moment in the evolution trajectory, the stress distribution data of the stress concentration area at the corresponding moment is extracted. By correlating the angle change at each point on the trajectory with the stress value at that point, the degree of overlap between areas with drastic angle changes and high-stress areas is identified, obtaining the spatial correspondence between angle change and stress distribution. Using this spatial correspondence, the elastic modulus of the protective film is obtained through material mechanics testing as the material rigidity parameter. The external load value applied to the folding screen is measured by a force sensor array. The material rigidity parameter and the external load value are substituted into the mechanical equation to calculate the angle response, determining the angle change caused by material rigidity and external load respectively. Based on the respective angular changes caused by each stress level, the dynamic response of the folding angle is repeatedly measured under different stress levels. The time it takes for the angle to reach a steady state and the final angle value are recorded. The differences in response time and angle change amplitude under different stress conditions are calculated to obtain regularity data of angle change. Through this regularity data, the periodic and trend components of angle change over time are extracted. An autoregressive moving average method is used to establish a time series forecast. The difference between angle values ​​at adjacent time points is calculated and divided by the time interval to determine the rate of change of the folding angle.

[0092] In one embodiment, the high-precision angle sensor employs photoelectric encoder technology, achieving a resolution of 0.001 degrees. It is installed at three measurement points: the center point and both ends of the folding shaft. Initial angle measurements are performed under constant temperature and humidity conditions, with the temperature controlled at 25±0.5℃ and relative humidity at 50±5%, ensuring the material is in a standard state. The stress sensor uses a thin-film piezoresistive sensor array, distributed in a 5×5 matrix in the stress concentration area, with each sensor covering an area of ​​4 square millimeters. The baseline relationship is established by gradually increasing the folding angle, recording the corresponding stress value every 5 degrees to form discrete data points. Then, continuous angle-stress curves are obtained through cubic spline interpolation.

[0093] Specifically, the formation of the evolution trajectory involves high-frequency data acquisition and filtering. The sampling frequency of the angle sensor is set to 1000Hz to capture rapid changes during the folding process. The raw data is filtered to remove measurement noise; the process noise covariance of the filter is set to 0.001, and the measurement noise covariance is set to 0.01 according to the sensor specifications. The calculation of the dynamic offset considers not only the absolute change in angle but also introduces a rate-of-change factor. When the rate of change exceeds a threshold, the sampling density is increased to ensure the integrity of the trajectory. The evolution trajectory is stored as a two-dimensional time-angle curve, with each trajectory containing at least 10,000 data points.

[0094] It should be noted that the spatial correspondence was established using bilinear interpolation. Since the angle measurement points and stress measurement points do not completely overlap spatially, an interpolation algorithm is needed to map the two types of data onto a unified spatial grid. The grid resolution is set to 1 mm × 1 mm, covering the entire stress concentration area. The criterion for drastic angle changes is an angle difference exceeding 0.5 degrees between adjacent grid points; a high-stress region is defined as a region where the stress value exceeds 1.5 times the average value. The degree of overlap is obtained by calculating the ratio of the intersection area to the union area of ​​the two regions; the closer this ratio is to 1, the stronger the correlation.

[0095] Preferably, the material rigidity parameters are tested according to the standard tensile testing method. The specimen is dumbbell-shaped, with a gauge length of 50 mm, a width of 10 mm, and a thickness equal to the actual thickness of the protective film. The tensile speed is set to 1 mm / min, and the slope of the linear segment of the stress-strain curve is recorded as the elastic modulus. External loads are measured using distributed force sensors based on the capacitive principle, capable of measuring forces ranging from 0.01 N to 100 N. The mechanical equation uses the thin-plate bending theory, simplifying the protective film as an isotropic elastic thin plate, with the equation M = EI / (1-v) 2 The equation is )×K, where M is the bending moment, E is the elastic modulus, I is the moment of inertia of the section, v is Poisson's ratio, and K is curvature. This equation allows us to calculate the contributions of material stiffness and external load to the angle variation.

[0096] In one possible implementation, different stress levels are set using an arithmetic sequence, starting at 10 MPa and increasing in 10 MPa intervals until 80% of the material's yield stress. Measurements at each stress level are repeated 10 times to reduce random error. The criterion for determining that the angle has reached a stable state is that the angle change over 100 consecutive sampling points is less than 0.01 degrees. The response time is defined as the time required for the angle to stabilize after stress is applied, and is verified using a high-speed camera to ensure the accuracy of the time measurement. The angle change amplitude is calculated by the difference between the final stable angle and the initial angle.

[0097] For example, the analysis of regular data reveals two main characteristics of the folding angle variation. The periodic component is mainly caused by the viscoelasticity of the material, manifested as oscillations of the angle before reaching a stable value, with an oscillation period of approximately 0.5 seconds and an amplitude that decays exponentially over time. The trend component reflects the long-term creep behavior of the material, with the angle increasing slowly under continuous stress, and the rate of increase being positively correlated with the stress level. Furthermore, the construction of the autoregressive moving average model requires determining the model order. By calculating the autocorrelation function and partial autocorrelation function, it was found that the data exhibited AR(2) and MA(1) characteristics, therefore the ARMA(2,1) model was selected. The model parameters were determined by the maximum likelihood estimation method, with autoregressive coefficients φ1 = 0.6, φ2 = 0.3, and moving average coefficients θ1 = 0.4. The goodness of fit of the model was evaluated using the Akaike Information Criterion, with an AIC value of -2156, indicating that the model has a good fit.

[0098] For example, in actual tests, when the stress level is 30 MPa, the rate of angle change from 0 degrees to 90 degrees initially increases and then decreases as the folding angle changes. The initial rate is approximately 5 degrees / second, peaking at around 15 degrees / second near 45 degrees, and then gradually decreasing to 2 degrees / second. This variation is closely related to the nonlinear mechanical properties of the material. At small angles, the material is in the elastic deformation stage, with a slow response; at medium angles, the stress concentration effect is significant, and the change accelerates; at large angles, the material enters plastic deformation, and the accumulation of internal damage leads to a decrease in stiffness and a slower rate of change.

[0099] Step S107: Extract warning indicators about screen curvature radius and light scattering intensity defects from the attenuation trend prediction results, and generate a complete report including optical transmittance and touch signal attenuation defects.

[0100] The predicted attenuation value corresponding to the number of folds is read from the attenuation trend prediction results. The curvature radius of the screen in different folded states is calculated by using the folding angle and the protective film thickness. The curvature radius is equal to the protective film thickness divided by the folding angle in radians. The light scattering intensity is determined by the ratio of scattered light intensity to incident light intensity. If the curvature radius is less than a preset threshold or the scattering intensity exceeds a preset threshold, a warning state is marked, and warning index data is obtained. Based on the location information in the warning index data, the optical transmittance measurement value and touch signal response time of the corresponding location are extracted. The percentage decrease in transmittance relative to the initial value and the delay increment of the signal response time relative to the reference value are calculated. When the percentage decrease exceeds the threshold, it is identified as an optical defect; when the delay increment exceeds the threshold, it is identified as a touch defect, and a defect classification result is obtained. Using the defect classification result, the curvature radius value, light scattering intensity value, transmittance decrease percentage, and signal delay increment are arranged in chronological order and spatial location. According to the severity of the defect, it is divided into three risk levels: low, medium, and high, forming a data summary table containing various indicators and risk levels. By integrating the early warning indicators, defect classifications, and risk level information in the data summary table, and organizing the content into chapters according to defect type, a complete report including defects in optical transmittance and touch signal attenuation is generated.

[0101] In one embodiment, the radius of curvature is calculated based on the geometric deformation characteristics of the foldable screen. When the protective film is folded from a flat state to a specific angle, the folded area forms an arc shape. The radius of curvature R is calculated using the formula R = t / θ, where t is the thickness of the protective film and θ is the radian value of the folding angle. For a protective film with a thickness of 100 micrometers, the radius of curvature is approximately 63.7 micrometers when the folding angle is 90 degrees. The light scattering intensity is measured using an integrating sphere device. The incident light is a 632.8 nm helium-neon laser, and the scattered light is collected by a photodetector. An intensity ratio exceeding 0.1 indicates significant scattering.

[0102] Specifically, the warning thresholds are set based on statistical analysis of a large amount of experimental data. The warning threshold for the radius of curvature is typically set at 50 micrometers; below this value, the bending stress borne by the protective film approaches the material's yield limit. The warning threshold for light scattering intensity is set at 0.15; exceeding this value indicates the presence of microcracks or delamination within the material, affecting optical performance.

[0103] It should be noted that the defect classification uses a dual-index method. The criteria for optical defects are a transmittance decrease exceeding 10% or a haze increase exceeding 5%. The criteria for touch defects are a response time delay exceeding 20 milliseconds or a touch failure area exceeding 1 square millimeter. When both defect criteria are met simultaneously, it is defined as a composite defect and requires special attention.

[0104] Preferably, the risk level is divided into three levels. Low risk corresponds to a slight exceedance of a single indicator, which does not affect normal use; medium risk corresponds to multiple indicators exceeding the standard or a single indicator exceeding the standard severely, affecting user experience; high risk corresponds to a severe deterioration of key performance indicators, and it is recommended to replace the protective film immediately. The risk level is related to the expected remaining service life, with low, medium, and high risks corresponding to more than 5000 folds, 1000-5000 folds, and less than 1000 folds remaining, respectively.

[0105] In one possible implementation, the complete report adopts a structured format, including four parts: summary, detailed data, trend charts, and recommendations. The summary outlines the main problems and risk levels; the detailed data section lists all measurement parameters and calculation results; the trend charts show the performance indicators changing over time; and the recommendations section provides targeted improvement measures based on the type and severity of the defects, such as adjusting usage habits, replacing the protective film, or optimizing the folding mechanism design.

[0106] It should be noted that the above examples are merely some specific embodiments of the present invention. Obviously, the present invention is not limited to the above embodiments and many variations are possible. All variations that can be directly derived or conceived by those skilled in the art from the content disclosed in this invention should be considered within the scope of protection of this invention.

Claims

1. A machine vision-based intelligent detection method for surface defects in protective films, characterized in that, The method includes: By collecting polarized light reflectivity and stress distribution data of a foldable screen from its flat state to its extreme folding angle, a reflectivity and stress distribution dataset is constructed. Feature extraction is performed on this dataset to obtain an initial optical performance degradation index, which includes a reflectivity attenuation coefficient and a stress distribution non-uniformity value. Based on the polarized light reflectivity and stress distribution data, a display sharpness value is calculated. By analyzing the correspondence between the initial optical performance degradation index and the display sharpness value, a sharpness decline trend is determined. Touch response delay features are extracted based on this trend. From the touch response delay features, stress concentration areas and touch signal attenuation features are extracted. The touch signal attenuation features include signal delay time and attenuation amplitude. The attenuation characteristics of the central region and touch signal are grouped and analyzed to obtain the joint attenuation trend of optical transmittance and touch signal attenuation. Based on the joint attenuation trend of optical transmittance and touch signal attenuation, the degree of touch signal attenuation in the folded area is calculated. Combined with the degree of deformation in the stress concentration area, a durability index is obtained. Through the durability index and the degree of deformation, a comprehensive performance attenuation index is calculated, which includes optical transmittance attenuation, touch response delay, and material fatigue. Based on the comprehensive performance attenuation index, the relationship between the change of folding angle and optical refractive index is analyzed to determine the rate of change of folding angle and the cumulative value of material fatigue, and an attenuation trend prediction result is obtained. The screen curvature radius and light scattering intensity are extracted from the attenuation trend prediction result to obtain early warning index data.

2. The intelligent detection method for surface defects of protective films based on machine vision according to claim 1, characterized in that, The process involves collecting polarized light reflectivity and stress distribution data of the foldable screen from its flat state to its extreme folding angle to construct a reflectivity and stress distribution dataset. Feature extraction is then performed on this dataset to obtain initial optical performance degradation indices, including: Polarized light reflection signals are collected at preset angle intervals, and pressure values ​​at each measuring point at the corresponding angle are obtained simultaneously. A dataset of reflectivity and stress distribution is constructed. The difference between the reflectivity at each folding angle and the reflectivity in the initial flat state is calculated and divided by the initial reflectivity to obtain the attenuation ratio. The folding angle and attenuation ratio are linearly fitted using the least squares method, and the slope of the fitted line is used as the reflectivity attenuation coefficient. The square root of the sum of the squares of the differences between the stress values ​​at all measuring points and the average stress at each folding angle is divided by the number of measuring points to obtain the standard deviation. The ratio of the standard deviation to the average stress is used as the stress distribution non-uniformity value. The initial attenuation index of optical performance, which includes the reflectivity attenuation coefficient and the stress distribution non-uniformity value, is obtained.

3. The intelligent detection method for surface defects of protective films based on machine vision according to claim 1, characterized in that, The step of calculating the display sharpness value based on the polarized light reflectivity and stress distribution data, and determining the sharpness decline trend by analyzing the correspondence between the initial attenuation index of optical performance and the display sharpness value, includes: A reflectance distribution matrix is ​​generated based on the polarized light reflectance, a stress field distribution map is constructed using the stress distribution data, the spatial frequency response of the display area is calculated, and the display sharpness value is obtained. The display sharpness value and the initial optical performance attenuation index are processed by regression analysis to generate a sharpness decline trend curve.

4. The intelligent detection method for surface defects of protective films based on machine vision according to claim 1, characterized in that, The process involves extracting stress concentration regions and touch signal attenuation features from the touch response delay characteristics. The touch signal attenuation features include signal delay time and attenuation amplitude. Group analysis is then performed on the stress concentration regions and touch signal attenuation features to obtain the joint attenuation trend of optical transmittance and touch signal attenuation, including: Read the spatial coordinates in the touch response delay feature, determine the boundary of the stress concentration area, collect the touch signal waveform within the boundary, calculate the signal delay time and the attenuation amplitude; group according to the signal delay time, construct a feature combination matrix, measure the optical transmittance at the position within the matrix, generate a regression equation between transmittance and attenuation amplitude, and fit to obtain a joint attenuation trend curve.

5. The intelligent detection method for surface defects of protective films based on machine vision according to claim 1, characterized in that, The method involves calculating the degree of touch signal attenuation in the folded area based on the combined attenuation trend of optical transmittance and touch signal attenuation, and combining this with the degree of deformation in the stress concentration area to obtain a durability index, including: The surface profile of the stress concentration area is scanned, the local radius of curvature is calculated, and a deformation degree distribution map is generated; the touch signal attenuation degree is obtained according to the joint attenuation trend, the damage rate is calculated, the degradation curve is fitted, and the durability index is obtained.

6. The intelligent detection method for surface defects of protective films based on machine vision according to claim 1, characterized in that, The comprehensive performance degradation index is calculated based on the durability index and the degree of deformation. This comprehensive performance degradation index includes optical transmittance degradation, touch response delay, and material fatigue, including: The maximum values ​​of the durability index and the degree of deformation are obtained, and the comprehensive performance degradation index is calculated by weighted summation; the polarization state change of the folded region is measured, and the optical refractive index is calculated by inversion; the refractive index change rate is extracted according to the relationship between the optical refractive index and the folding angle; if the refractive index change rate exceeds the threshold, the components in the comprehensive performance degradation index are combined to generate an early warning signal.

7. The intelligent detection method for surface defects of protective films based on machine vision according to claim 1, characterized in that, The comprehensive performance degradation index is calculated based on the durability index and the degree of deformation. This comprehensive performance degradation index includes optical transmittance degradation, touch response delay, and material fatigue, including: Data on folding times and ambient temperature are collected, and the surface hardness and tensile strength of the protective film are measured to generate a dataset of material mechanical properties. The distribution of deformation degree is scanned to determine the boundary of the deformation region. The optical transmittance, touch response time and elastic modulus of the material within the boundary are measured to calculate the attenuation rate and fatigue life index, and to determine the contribution ratio of material aging and stress concentration.

8. The intelligent detection method for surface defects of protective films based on machine vision according to claim 1, characterized in that, The process involves analyzing the relationship between the folding angle change and the optical refractive index based on the comprehensive performance degradation index, determining the folding angle change rate and the material fatigue accumulation value, and obtaining the degradation trend prediction result, including: The deformation depth of the stress concentration region is measured, the relative damage is calculated, and the cumulative fatigue value of the material is obtained by summing the results; the folding angle sequence is recorded, and the rate of change of the folding angle is calculated; according to the warning level of the comprehensive performance degradation index, the rate of change of the folding angle and the cumulative fatigue value of the material are adjusted to generate a predicted degradation value sequence.

9. The intelligent detection method for surface defects of protective films based on machine vision according to claim 1, characterized in that, The step of analyzing the relationship between the change in folding angle and the optical refractive index based on the comprehensive performance degradation index, and determining the rate of change of folding angle, includes: Measure the initial folding angle, record the dynamic offset, and generate the folding angle evolution trajectory; extract the stress distribution based on the evolution trajectory, calculate the material rigidity and the angular response to external loads, and determine the rate of change of the folding angle.

10. The intelligent detection method for surface defects of protective films based on machine vision according to claim 1, characterized in that, The step of extracting the screen curvature radius and light scattering intensity from the attenuation trend prediction results to obtain early warning index data includes: Calculate the screen curvature radius and light scattering intensity based on the attenuation trend prediction results, and mark the warning status; extract the optical transmittance and touch signal response time at the corresponding positions, classify the defect types, and generate a data summary table containing risk levels.

Images