A PVA film stress detection method and system based on machine vision

Abstract

The present invention relates to a PVA film stress detection method and system based on machine vision, belonging to the technical field of film stress detection. The method comprises the following steps: building a test platform, the test platform comprising a curvature test component, a thickness test component and a stage; measuring the sample thickness by the thickness test component, testing the substrate curvature by the curvature test component, obtaining the sample thickness and the substrate curvature variation; based on the heat treatment of the PVA film during the test, inputting the tested substrate curvature variation into the modified Stoney formula to calculate the stress. The present invention improves the real-time and high efficiency of the detection by building a test platform for simultaneously performing film thickness test and substrate curvature test, and collects the test image and performs image processing by machine vision, thereby avoiding the physical damage to the sample that may be caused by the traditional contact measurement, realizing the accurate measurement of the sample thickness and substrate curvature, and improving the accuracy of the stress calculation.

Description

Technical Field

[0001] The present invention belongs to the technical field of film stress detection, and in particular relates to a PVA film stress detection method and system based on machine vision. Background Art

[0002] Polyvinyl alcohol (PVA) film is widely used in packaging, medical materials, optical applications and water treatment due to its water solubility, excellent mechanical properties and biodegradability. Thin film deposition is a non-equilibrium process, and the deposited atoms are not completely in a state of equilibrium, which means that the film is under stress. Usually, tensile stress will cause film cracking or limit the effective thickness of the film, and compressive stress will cause wrinkling, blistering and shedding of the film. It can be seen that film stress is an important cause of film failure.

[0003] At present, the commonly used methods for film stress testing include X-ray method, Raman spectroscopy method, substrate curvature method, etc. However, in actual measurement, it is necessary to consider the material properties of the film to select the test method. The film may be damaged during the stress detection process, and the real-time performance of stress detection is poor. At the same time, since the crystallinity of PVA film is different under different heat treatment conditions, and the crystallinity will directly affect the stress of the film, the crystallinity and other influencing factors cannot be associated during the test process, and the reliability of the test results is poor.

[0004] Therefore, there is an urgent need to propose a PVA film stress detection method and system based on machine vision, which can realize non-contact, efficient and accurate film stress detection. Summary of the invention

[0005] In order to solve the above problems existing in the prior art, the present invention provides a PVA film stress detection method and system based on machine vision. By building a test platform for simultaneously performing film thickness test and substrate curvature test, the real-time and efficiency of the detection are improved. The test images are collected and processed by machine vision, thereby avoiding the physical damage to the sample caused by traditional contact measurement, achieving accurate measurement of sample thickness and substrate curvature change, and improving the accuracy of stress calculation.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A first aspect of the present disclosure provides a PVA film stress detection method based on machine vision, comprising the following steps:

[0008] Build a test platform: The test platform includes a curvature test component, a thickness test component and a stage. Place the test sample on the stage and set the composition structure of the test component according to the test requirements;

[0009] Substrate curvature measurement: measure the sample thickness through the thickness test component, and test the substrate curvature through the curvature test component to obtain the sample thickness and substrate curvature change;

[0010] Stress calculation: Based on the heat treatment of the PVA film during the test, the Stoney formula is modified, and the change in the tested substrate curvature is input into the modified Stoney formula to calculate the stress;

[0011] The base curvature measurement comprises the following steps:

[0012] Measure sample thickness: Use the transmission optical density method to measure sample thickness through the thickness test component. Start the thickness test component to obtain the incident and output grayscale images before and after placing the sample, and calculate the transmission optical density value of the test sample; then use the Beer-Lambert law to convert the optical density value into sample thickness;

[0013] Test sample curvature: Use the base curvature method to test the base curvature through the curvature test component. Start the curvature test component to capture the reflected light signal, collect and record the reflected signal, obtain the distance between adjacent light spots, and convert it into the base curvature change.

[0014] Further, the curvature test assembly includes a laser, an etalon, a polarizer and a first camera, the first camera is aimed at the reflection area of ​​the laser, and the focal length and angle of the first camera are adjusted;

[0015] The thickness test assembly comprises a scattering surface light source, a milky white glass diffusion plate and a second camera. The film to be tested is arranged between the second camera and the scattering surface light source, and the film to be tested is formed on a glass substrate.

[0016] Furthermore, the PVA film stress detection method further comprises the steps of:

[0017] Prepare the test sample: select a glass substrate, take a certain amount of PVA solution on a clean glass substrate, control the low-speed coating machine to evenly coat the PVA solution on the surface of the glass substrate, and completely evaporate the water in the solution. After the film is formed, place it in a silica gel desiccant;

[0018] The temperature and time conditions required for the PVA film test were selected using a gradient heating furnace to heat treat the PVA film and obtain the final test sample.

[0019] Further, the measuring of the sample thickness comprises the following steps:

[0020] Determine the extinction coefficient of the PVA film: Use a contact measuring instrument to measure the thickness of different areas of the film sample to be tested, place film samples of different thicknesses on the test platform, use a CCD camera to collect the incident and output grayscale values ​​before and after placing the sample, calculate the optical density value, and the slope of the curve drawn based on the measured and calculated thickness-optical density value points is the extinction coefficient;

[0021] Calculate the transmitted optical density value: Calculate the optical density value by preprocessing the incident and outgoing grayscale images before and after placing the sample:

[0022] Without placing a sample, capture the distribution image of the incident light intensity, place the film sample to be tested between the light source and the camera, and capture the light intensity distribution image after passing through the sample again;

[0023] De-noising the collected images, making the incident and outgoing grayscale images have the same dynamic range through histogram equalization, calculating the grayscale value of each pixel in the incident and outgoing images, and normalizing the grayscale values;

[0024] Get sample thickness value: Calculate sample thickness according to Beer-Lambert law.

[0025] Furthermore, the optical density value D is calculated using the following formula:

[0026] ;

[0027] In the formula, is the gray value without film, is the gray value of the placed film;

[0028] The sample thickness d The calculation formula is:

[0029] ;

[0030] In the formula, a is the extinction coefficient.

[0031] Furthermore, the testing sample curvature comprises the following steps:

[0032] Collecting image data: first, before depositing the PVA film, collect the reflected light signal data image, and then use the prepared test sample to collect the reflected light signal data image using a CCD camera;

[0033] Image processing: Through image preprocessing, the light spots are detected to obtain the distances between adjacent light spots of the reflected light signals before and after the PVA film is deposited on the substrate;

[0034] Calculate the change in base curvature: The distance between adjacent light spots is represented by the distance between adjacent centroids, so as to obtain the initial value and change of the distance between adjacent light spots and calculate the change in base curvature. :

[0035] ;

[0036] In the formula, is the initial value of the distance between adjacent spots, is the change in the distance between adjacent light spots, is the incident angle of the light source, and L is the distance between the substrate and the camera.

[0037] Furthermore, the image processing comprises the following steps:

[0038] The image is converted into a grayscale image, the image is smoothed using a Gaussian filter, and the edge of the spot area is enhanced through morphological operations;

[0039] Select the Canny edge detection algorithm to calculate the gradient strength and direction of each pixel in the image and suppress the non-maximum values ​​in the gradient image; use double threshold detection to divide the edges into strong edges, weak edges and non-edges, and then connect all weak edges to strong edges to form a complete edge;

[0040] Traverse the binary graph to find all independent white pixel connected areas and represent them as contours. For each contour, calculate its spatial moment.

[0041] Using the calculated moment, the coordinates of the center of mass are obtained by the ratio of the first-order moment to the zero-order moment. Specifically, the x-coordinate of the center of mass is obtained by dividing the first-order moment in the x-direction of the contour by the zero-order moment of the contour, and the y-coordinate of the center of mass is obtained by dividing the first-order moment in the y-direction of the contour by the zero-order moment of the contour.

[0042] Furthermore, the stress calculation comprises the following steps:

[0043] Assume that the thermal expansion coefficient of the film is, the thermal expansion coefficient of the substrate is, and the temperature change is Δ T , then the thermal stress generated by the difference in thermal expansion coefficient between the film and the substrate is It is expressed as:

[0044] ;

[0045] In the formula, is the Young's modulus of the PVA film, and are the thermal expansion coefficients of the PVA film and substrate, is the temperature change, is the Poisson’s ratio of the PVA film;

[0046] The obtained base curvature change is input into the modified Stoney formula to calculate the stress. The modified Stoney formula is expressed as:

[0047] + ;

[0048] In the formula, is the stress of the film, is the Young's modulus of the substrate, is the thickness of the substrate, is the change in base curvature, is the Poisson's ratio of the base, is the thickness of the PVA film.

[0049] A second aspect of the present disclosure provides a PVA film stress detection system based on machine vision, which is applied to a PVA film stress detection method based on machine vision as described above, and includes a sample preparation module, a sample testing module, a machine vision module, and a stress calculation module;

[0050] The sample preparation module is used to prepare the PVA film, including spin coating, drying and heat treatment, to obtain the PVA film with a predetermined crystallinity for testing;

[0051] A PVA film of predetermined crystallinity is obtained, an XRD spectrum of the PVA film is measured by an X-ray powder diffractometer, and a corresponding crystallinity is calculated by an amorphous standard sample method, thereby obtaining a heat treatment temperature and time for obtaining the predetermined crystallinity;

[0052] The sample testing module is used to build a testing platform to obtain the transmission light density value of the PVA film and the change in substrate curvature, respectively, to measure the thickness of the PVA film and the change in substrate curvature before and after sample spin coating;

[0053] The test platform includes a curvature test component, a thickness test component and a stage, wherein the thickness test component calculates the transmitted light density value by obtaining the incident and output grayscale images before and after the sample is placed, and then converts the light density value into the sample thickness using the Beer-Lambert law; the curvature test component calculates the substrate curvature change by obtaining the distance between adjacent light spots before and after the sample is placed.

[0054] As a preferred technical solution of the present invention, the machine vision module is used to collect incident and outgoing grayscale images before and after the preparation of the PVA film and the reflected signal spot images before and after the sample is placed through a CCD camera, and pre-process the acquired image data;

[0055] The preprocessing of the incident and outgoing grayscale images is to perform denoising by averaging multiple images and calculate the grayscale value of each pixel in the image through histogram equalization;

[0056] Preprocessing of the reflected signal spot image: convert the image into a grayscale image, use a Gaussian filter to smooth the image, enhance the edge of the spot area through morphological operations, perform edge detection to extract the spot contour and obtain the spacing between adjacent spots;

[0057] The machine vision module is also used to calibrate the thickness measurement system before optical density measurement, including the following steps:

[0058] Dark noise acquisition and analysis: Test in completely dark conditions, and acquire multiple dark noise images under exactly the same conditions; average all acquired dark noise images to obtain a mean image of dark noise; calculate the variance of the dark noise image to obtain a variance map of dark noise;

[0059] Incident light grayscale calibration: prepare multiple standard light sources with known light densities, and collect multiple images for each standard light source; subtract the dark noise mean image from each standard light source image to obtain a net light grayscale image; average the multiple net light grayscale images of each standard light source to obtain an average light grayscale value; establish a calibration curve between light density and grayscale value based on the known light density value of the standard light source and the measured average light grayscale value;

[0060] System response correction: Use the optical density-grayscale curve to correct the nonlinear response of the system and evaluate the impact of dark noise on measurement accuracy under small signal conditions;

[0061] The stress calculation module is used to calculate the stress using the modified Stoney formula based on the PVA film thickness and the curvature change of the substrate obtained through the test, and to establish a stress thickness product versus film thickness change curve to obtain stress detection values ​​associated with the crystallinity and thickness of the PVA film.

[0062] The beneficial effects of the present invention are:

[0063] The present invention first prepares a test sample that meets the stress test requirements according to the characteristics of a PVA film, then builds a test platform for simultaneously performing film thickness tests and substrate curvature tests, collects test images through machine vision and performs image processing, obtains sample thickness and substrate curvature changes, and calculates film stress through the Stoney formula; during the test process, the film thickness is measured and the substrate curvature is measured simultaneously, which improves the efficiency of curvature and thickness measurement, ensures data consistency when measured under the same environmental conditions, and improves the accuracy of stress calculation.

[0064] When preparing a test sample, the present invention performs heat treatment on a PVA film to control the crystallinity of the film, thereby associating the crystallinity of the PVA film with a film stress test; correspondingly, due to the difference in thermal expansion coefficients between the film and a substrate, thermal stress is generated through heat treatment; the Stoney formula is corrected by considering the difference in thermal expansion coefficients, thereby reflecting a more realistic stress state and improving the accuracy of stress calculation; at the same time, the thickness measurement system is calibrated before optical density measurement, and each pixel is uniformly corrected to match it with a calibration value, thereby effectively reducing the influence of dark noise on optical density measurement and improving the accuracy and reliability of thickness measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0066] Figure 1 A schematic diagram of the steps of a PVA film stress detection method based on machine vision provided by an embodiment of the present invention;

[0067] Figure 2 A schematic diagram of the structure of a test platform provided in an embodiment of the present invention;

[0068] Figure 3 A schematic diagram of the steps of measuring substrate curvature provided in an embodiment of the present invention;

[0069] Figure 4 A structural schematic diagram of a PVA film stress detection system based on machine vision provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0070] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0071] This embodiment provides a PVA film stress detection method based on machine vision, such as Figure 1 As shown, the following steps are included:

[0072] S1. Prepare the test sample: select a glass substrate, take a certain amount of PVA solution on a clean glass substrate, control the low-speed coating machine to evenly coat the PVA solution on the surface of the glass substrate, and make the water in the solution evaporate completely. After the film is formed, place it in a silicone desiccant to prevent moisture absorption.

[0073] It should be noted that, since the film thickness is measured later, a glass substrate is selected as the substrate for the curvature test in this embodiment, which satisfies both the curvature test and the thickness test of the transmission optical density method. The selected glass substrate has known parameters such as thickness, Young's modulus and Poisson's ratio.

[0074] It is understood that the glass substrate is selected from optical grade glass with a smooth surface and no impurities. The substrate is cleaned with an appropriate detergent and a fiber-free cloth before coating to ensure that there is no dust and dirt and to dry it. In the process of coating the PVA solution on the glass substrate using a low-speed coating machine, the solution is evenly distributed on the surface of the substrate by rotating the substrate, reducing the thickness variation caused by uneven coating. By controlling the rotation speed and time, the drying process can be better controlled.

[0075] The temperature and time conditions required for the PVA film test were selected using a gradient heating furnace to heat treat the PVA film and obtain the final test sample.

[0076] It is understandable that the crystallinity of the PVA film is different under different heat treatment conditions. The higher the heat treatment temperature of the PVA film, the higher its crystallinity. In the test of film stress in this embodiment, in order to ensure the effectiveness of the test and the practicability of actual application, the corresponding heat treatment conditions are selected according to the predetermined crystallinity of the PVA film to obtain the test sample.

[0077] S2. Build the test platform: The test platform includes a curvature test component, a thickness test component and a stage, such as Figure 2 As shown, place the test sample on the stage and set the composition structure of the test components according to the test requirements:

[0078] The curvature test assembly includes a laser 1, an etalon 2, a polarizer 4 and a first camera 3. The first camera 3 is aimed at the reflection area of ​​the laser 1 to capture the reflected light spot. The focal length and angle of the first camera 3 are adjusted to ensure that the light spot can be clearly captured.

[0079] The thickness testing assembly includes a scattering surface light source 8 , a milky white glass diffusion plate 7 and a second camera 9 . The film 5 to be tested is arranged between the second camera 9 and the scattering surface light source 8 . The film 5 to be tested is formed on a glass substrate 6 .

[0080] It is understandable that both the first camera 3 and the second camera 9 use CCD (charge coupled device) cameras. In the curvature test component, the laser 1 is used to emit a monochromatic and highly directional light beam; the standard tool 2 is responsible for dividing the incident light into parallel light beams for projection; the polarizer 4 adjusts the polarization state of the laser beam to reduce light scattering and reflection interference and improve the accuracy of the measurement; the first camera 3 is used to capture the light spot image reflected from the substrate, convert the optical signal into an electrical signal, and form a digital image. In the thickness test component, the scattering surface light source 8 is used to provide uniform and stable illumination, the milky white glass diffuser plate 7 is used to further homogenize the light emitted from the light source, and the second camera 9 is used to capture the transmitted light passing through the film 5 to be tested, convert the optical signal into an electrical signal, and then convert it into a digital image.

[0081] It should be noted that in this embodiment, the test platform constructed simultaneously measures the film thickness and the substrate curvature, which is used for the calculation of film stress detection, thereby improving the efficiency of curvature and thickness measurement. Measurements under the same environmental conditions ensure data consistency and improve the accuracy of stress calculations.

[0082] S3, substrate curvature measurement: the thickness of the sample is measured by the thickness test component, and the substrate curvature is tested by the curvature test component, such as Figure 3 As shown, the following steps are included:

[0083] S31. Measure the sample thickness: Use the transmission optical density method to measure the sample thickness through the thickness test component, start the thickness test component to obtain the incident and output grayscale images before and after the sample is placed, and calculate the transmission optical density value of the test sample; then use the Beer-Lambert law to convert the optical density value into the sample thickness, which specifically includes the following steps:

[0084] S311. Determine the extinction coefficient of the PVA film: Use a contact measuring instrument to measure the thickness of different areas of a series of film samples to be tested, place film samples of different thicknesses on a test platform, use a camera to collect the incident and output grayscale values ​​before and after the samples are placed, calculate the optical density value, and the slope of the curve drawn based on the measured and calculated thickness-optical density value points is the extinction coefficient.

[0085] It is understood that for a particular film, its extinction coefficient is a constant, so the slope of a curve drawn through the measured and calculated thickness-optical density value points can be approximated as the extinction coefficient.

[0086] S312, calculate the transmission light density value: calculate the light density value by preprocessing the incident and outgoing grayscale images before and after placing the sample:

[0087] Without placing a sample, capture the distribution image of the incident light intensity, place the film sample to be tested between the light source and the camera, and capture the light intensity distribution image after passing through the sample again;

[0088] De-noising the collected images to reduce the impact of random noise in the images on the measurement. Usually, histogram equalization is used to make the incident and output grayscale images have the same dynamic range to improve the contrast and detail of the images.

[0089] For each pixel in the incident image and the outgoing image, its grayscale value is calculated. These grayscale values ​​are usually expressed as integers between [0, 255], and the grayscale values ​​are normalized to eliminate the influence of light source non-uniformity and camera response inconsistency;

[0090] Calculate the optical density value D, the formula is as follows:

[0091] ;

[0092] In the formula, is the gray value without film, is the gray value of the placed film.

[0093] S313, obtain the sample thickness value: According to Beer-Lambert law, the sample thickness d The calculation formula is:

[0094] ;

[0095] In the formula, a is the extinction coefficient.

[0096] It is understood that the extinction coefficient is a function of the film material and the illumination wavelength and is the key to establishing the optical density. D With thickness d The key parameter of the relationship between them, the extinction coefficient of different samples is obtained through experiments.

[0097] S32, testing the curvature of the sample: testing the base curvature by using the base curvature method through the curvature testing component, starting the curvature testing component to capture the reflected light signal, collecting and recording the reflected signal, obtaining the distance between adjacent light spots, and converting it into a base curvature change, including the following steps:

[0098] It can be understood that the curvature test assembly generates a set of two-dimensional parallel beam matrices, which are incident on the surface of the sample at a certain angle. These beams are usually evenly distributed in the form of a matrix to ensure that multiple locations on the sample surface can be covered for a wide range of stress measurements. When the two-dimensional parallel beam is incident on the sample surface, the beam is reflected from the surface. The shape or curvature of the sample surface causes the path of the reflected beam to change. The reflected light signal is captured by the CCD camera, and the position and intensity of the reflected light spot are accurately recorded.

[0099] S321, collecting image data: first, before depositing the PVA film, collecting a data image of a reflected light signal, and then using the prepared test sample to collect a data image of the reflected light signal using a CCD camera;

[0100] S322, image processing: by image preprocessing, the light spot is detected to obtain the distance between adjacent light spots of the reflected light signal before and after the PVA film is deposited on the substrate, including the following steps:

[0101] Convert the image to grayscale to simplify the processing process; a grayscale image is an image that only retains brightness information after removing color information, which is suitable for subsequent edge detection work.

[0102] The image is smoothed using a Gaussian filter to reduce the impact of noise on edge detection; morphological operations (such as dilation and erosion) are used to enhance the edges of the spot area to ensure edge continuity.

[0103] Select the Canny edge detection algorithm to calculate the gradient strength and direction of each pixel in the image. The gradient can reflect the change in pixel brightness. Suppress the non-maximum values ​​in the gradient image to refine the edges. Only the pixels with the local maximum gradient are retained, and the rest are suppressed to 0. Use dual threshold detection to divide the edges into strong edges, weak edges, and non-edges, and then connect all weak edges to strong edges to form complete edges.

[0104] It can be understood that a binary image is finally obtained through edge detection, in which white pixels represent detected edges and black pixels represent non-edge areas.

[0105] Traverse the binary graph to find all independent white pixel connected areas and represent them as contours. For each contour, calculate its spatial moment. The spatial moment contains a series of statistical information about the contour shape, including zero-order moment, first-order moment, etc.

[0106] Using the calculated moment, the coordinates of the center of mass are obtained by the ratio of the first-order moment to the zero-order moment. Specifically, the x-coordinate of the center of mass is obtained by dividing the first-order moment in the x-direction of the contour by the zero-order moment of the contour, and the y-coordinate of the center of mass is obtained by dividing the first-order moment in the y-direction of the contour by the zero-order moment of the contour.

[0107] S323, calculate the change of base curvature: the distance between adjacent light spots is represented by the distance between adjacent centroids, so as to obtain the initial value and change of the distance between adjacent light spots, and calculate the change of base curvature :

[0108] ;

[0109] In the formula, is the initial value of the distance between adjacent spots, is the change in the distance between adjacent light spots, is the incident angle of the light source, and L is the distance between the substrate and the camera.

[0110] S4. Stress calculation: Based on the heat treatment of the PVA film in the test, the Stoney formula is modified, and the change in the tested substrate curvature is input into the modified Stoney formula to calculate the stress, including the following steps:

[0111] S41, assuming that the thermal expansion coefficient of the film is, the thermal expansion coefficient of the substrate is, and the temperature change is Δ T , then the thermal stress generated by the difference in thermal expansion coefficient between the film and the substrate is It is expressed as:

[0112] ;

[0113] In the formula, is the Young's modulus of the PVA film, and are the thermal expansion coefficients of the PVA film and substrate, is the temperature change, is the Poisson's ratio of PVA film.

[0114] S42, inputting the obtained base curvature variation into the modified Stoney formula to calculate stress, the modified Stoney formula is expressed as:

[0115] + ;

[0116] In the formula, is the stress of the film, is the Young's modulus of the substrate, is the thickness of the substrate, is the change in base curvature, is the Poisson's ratio of the base, is the thickness of the PVA film.

[0117] In this embodiment, the Stoney formula is modified by calculating thermal stress, so that the stress calculation is better adapted to the actual needs of heat treatment and the accuracy of the stress calculation is improved.

[0118] S5. Establish stress variation curve: According to the calculated stress and the corresponding film thickness, establish a stress thickness product versus film thickness variation curve. The slope of the line connecting any point on the curve and the starting point represents the average stress of the thickness corresponding to the point.

[0119] This embodiment also provides a PVA film stress detection system based on machine vision, such as Figure 4 As shown, it includes a sample preparation module, a sample testing module, a machine vision module and a stress calculation module;

[0120] The sample preparation module is used to prepare a PVA film, including spin coating, drying and heat treatment, to obtain a PVA film with a predetermined crystallinity for testing.

[0121] The actual crystallinity control is to obtain a PVA film with a predetermined crystallinity for testing, measure the XRD spectrum of the PVA film by using an X-ray powder diffractometer, and calculate the corresponding crystallinity by using an amorphous standard sample method, thereby obtaining the heat treatment temperature and time for obtaining the predetermined crystallinity.

[0122] It should be noted that as a semi-crystalline polymer, the crystallinity of the PVA film increases with the increase of the heat treatment temperature. In this embodiment, the naturally dried PVA film is heat treated to test the stress of films with different crystallinity. The actual test heat treatment temperature should be lower than 180°C to prevent thermal degradation of the PVA film. By controlling the influencing factors such as the crystallinity and thickness of the PVA film, the test stress results are associated with the influencing factors, thereby enhancing the reference value of the results.

[0123] The sample testing module is used to build a testing platform to obtain the transmission light density value of the PVA film and the change in substrate curvature, respectively, to achieve measurement of the thickness of the PVA film and the change in substrate curvature before and after sample spin coating.

[0124] The test platform includes a curvature test component, a thickness test component and a stage, wherein the thickness test component calculates the transmitted light density value by obtaining the incident and output grayscale images before and after the sample is placed, and then converts the light density value into the sample thickness using the Beer-Lambert law; the curvature test component calculates the substrate curvature change by obtaining the distance between adjacent light spots before and after the sample is placed.

[0125] It should be noted that the stress of PVA film is greatly affected by environmental conditions (such as temperature and humidity). When conducting film stress testing, the temperature and humidity of the test environment should be kept under standardized conditions to ensure the consistency of environmental conditions.

[0126] The machine vision module is used to collect incident and output grayscale images of the PVA film before and after preparation and reflected signal spot images before and after sample placement through a CCD camera, and pre-process the acquired image data.

[0127] The preprocessing of the incident and outgoing grayscale images is to perform denoising by averaging multiple images and calculate the grayscale value of each pixel in the image through histogram equalization;

[0128] Preprocessing of the reflected signal spot image: convert the image into a grayscale image, use a Gaussian filter to smooth the image, enhance the edge of the spot area through morphological operations, perform edge detection to extract the spot contour and obtain the spacing between adjacent spots.

[0129] The machine vision module is also used to calibrate the thickness measurement system before optical density measurement, including the following steps:

[0130] Dark noise acquisition and analysis: Test in complete darkness, collect multiple (usually dozens to hundreds) dark noise images, each image is recorded under exactly the same conditions; average all collected dark noise images to obtain a mean image representing the dark noise; calculate the variance of all dark noise images to obtain a variance map of the dark noise.

[0131] It can be understood that the variance map describes the variation of each pixel in multiple measurements and is used for noise suppression in subsequent signal processing.

[0132] Incident light grayscale calibration: prepare multiple standard light sources with known optical density. The optical density of these light sources needs to cover multiple points within the measurement range to ensure the accuracy of the calibration. For each standard light source, collect multiple images, and ensure the stability and uniformity of the light source each time. Subtract the previously obtained dark noise mean image from each standard light source image to obtain a net light grayscale image. Average the multiple net light grayscale images of each standard light source to obtain the average light grayscale value. Based on the known light density value of the standard light source and the measured average light grayscale value, establish a calibration curve between light density and grayscale value.

[0133] System response correction: Use the optical density-grayscale curve to correct the nonlinear response of the system to ensure that the grayscale value output when measuring optical density can accurately reflect the actual optical density. Evaluate the impact of dark noise on measurement accuracy in small signal conditions (low optical density).

[0134] It is understandable that the thickness measurement system includes a thickness measurement component and a machine vision module, which are used to obtain the thickness of the PVA film. The thickness measurement component includes the thickness calculation process. By calibrating the thickness measurement system, each pixel is uniformly corrected and matched with the calibration value, thereby effectively reducing the impact of dark noise on optical density measurement and improving the accuracy and reliability of measurement.

[0135] The stress calculation module is used to calculate the stress using the modified Stoney formula based on the PVA film thickness and the curvature change of the substrate obtained through the test, and to establish a stress thickness product versus film thickness change curve to obtain stress detection values ​​associated with the crystallinity and thickness of the PVA film.

[0136] It should be noted that the effect of film thickness on film stress has a universal law. As the film thickness increases, the film stress will also change from compressive stress to tensile stress, and film stress will only occur when the film thickness reaches a certain critical thickness. Therefore, the thickness of the film is an indispensable influencing factor in stress analysis. For PVA film, its crystallization characteristics are related to heat treatment. The crystallinity has a significant effect on the mechanical properties, internal stress and stress distribution of PVA film. Therefore, in stress testing, films with different crystallinity have different stress test results. Therefore, in the process of sample preparation, it is necessary to pay attention to the conditions of heat treatment.

[0137] The present invention first prepares a test sample that meets the stress test requirements according to the characteristics of a PVA film, then builds a test platform for simultaneously performing film thickness tests and substrate curvature tests, collects test images through machine vision and performs image processing, obtains sample thickness and substrate curvature changes, and calculates film stress through the Stoney formula; during the test process, the film thickness is measured and the substrate curvature is measured simultaneously, which improves the efficiency of curvature and thickness measurement, ensures data consistency when measured under the same environmental conditions, and improves the accuracy of stress calculation.

[0138] When preparing a test sample, the present invention performs heat treatment on a PVA film to control the crystallinity of the film, thereby associating the crystallinity of the PVA film with a film stress test; correspondingly, due to the difference in thermal expansion coefficients between the film and a substrate, thermal stress is generated through heat treatment; the Stoney formula is corrected by considering the difference in thermal expansion coefficients, thereby reflecting a more realistic stress state and improving the accuracy of stress calculation; at the same time, the thickness measurement system is calibrated before optical density measurement, and each pixel is uniformly corrected to match it with a calibration value, thereby effectively reducing the influence of dark noise on optical density measurement and improving the accuracy and reliability of thickness measurement.

[0139] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A PVA film stress detection method based on machine vision, characterized in that: The following steps are involved: Prepare the test sample: select a glass substrate, take a certain amount of PVA solution on a clean glass substrate, control the low-speed coating machine to evenly coat the PVA solution on the surface of the glass substrate, and completely evaporate the water in the solution. After the film is formed, place it in a silica gel desiccant; Use a gradient heating furnace to select the temperature and time conditions required for the PVA film test to heat treat the PVA film to obtain the final test sample; Build a test platform: The test platform includes a curvature test component, a thickness test component and a stage. Place the test sample on the stage and set the composition structure of the test component according to the test requirements; Substrate curvature measurement: measure the sample thickness through the thickness test component, and test the substrate curvature through the curvature test component to obtain the sample thickness and substrate curvature change; Stress calculation: Based on the heat treatment of the PVA film during the test, the Stoney formula is modified, and the change in the tested substrate curvature is input into the modified Stoney formula to calculate the stress; The base curvature measurement comprises the following steps: Measure sample thickness: Use the transmission optical density method to measure sample thickness through the thickness test component. Start the thickness test component to obtain the incident and output grayscale images before and after placing the sample, and calculate the transmission optical density value of the test sample; then use the Beer-Lambert law to convert the optical density value into sample thickness; Test sample curvature: Use the base curvature method to test the base curvature through the curvature test component. Start the curvature test component to capture the reflected light signal, collect and record the reflected signal, obtain the distance between adjacent light spots, and convert it into the base curvature change; The method for measuring the thickness of the sample comprises the following steps: Determine the extinction coefficient of the PVA film: Use a contact measuring instrument to measure the thickness of different areas of the film sample to be tested, place film samples of different thicknesses on the test platform, use a CCD camera to collect the incident and output grayscale values ​​before and after placing the sample, calculate the optical density value, and the slope of the curve drawn based on the measured and calculated thickness-optical density value points is the extinction coefficient; Calculate the transmitted optical density value: Calculate the optical density value by preprocessing the incident and outgoing grayscale images before and after placing the sample: Without placing a sample, capture the distribution image of the incident light intensity, place the film sample to be tested between the light source and the camera, and capture the light intensity distribution image after passing through the sample again; De-noising the collected images, making the incident and outgoing grayscale images have the same dynamic range through histogram equalization, calculating the grayscale value of each pixel in the incident and outgoing images, and normalizing the grayscale values; Get the sample thickness value: Calculate the sample thickness according to Beer-Lambert law; The optical density value D is calculated using the following formula: ； In the formula, is the gray value without film, is the gray value of the placed film; The sample thickness d The calculation formula is: ； In the formula, a is the extinction coefficient; The test sample curvature comprises the following steps: Collecting image data: first, before depositing the PVA film, collect the reflected light signal data image, and then use the prepared test sample to collect the reflected light signal data image using a CCD camera; Image processing: Through image preprocessing, the light spots are detected to obtain the distances between adjacent light spots of the reflected light signals before and after the PVA film is deposited on the substrate; Calculate the change in base curvature: The distance between adjacent light spots is represented by the distance between adjacent centroids, so as to obtain the initial value and change of the distance between adjacent light spots and calculate the change in base curvature. : ； In the formula, is the initial value of the distance between adjacent spots, is the change in the distance between adjacent light spots, is the incident angle of the light source, L is the distance between the substrate and the camera; The image processing comprises the following steps: The image is converted into a grayscale image, the image is smoothed using a Gaussian filter, and the edge of the spot area is enhanced through morphological operations; Select the Canny edge detection algorithm to calculate the gradient strength and direction of each pixel in the image and suppress the non-maximum values ​​in the gradient image; use double threshold detection to divide the edges into strong edges, weak edges and non-edges, and then connect all weak edges to strong edges to form a complete edge; Traverse the binary graph to find all independent white pixel connected areas and represent them as contours. For each contour, calculate its spatial moment. Using the calculated moments, the coordinates of the center of mass are obtained by the ratio of the first-order moment to the zero-order moment. Specifically, the x-coordinate of the center of mass is obtained by dividing the first-order moment of the contour in the x-direction by the zero-order moment of the contour, and the y-coordinate of the center of mass is obtained by dividing the first-order moment of the contour in the y-direction by the zero-order moment of the contour. The stress calculation comprises the following steps: Assume the thermal expansion coefficient of the film is , the thermal expansion coefficient of the substrate is , the temperature change is Δ T , then the thermal stress generated by the difference in thermal expansion coefficient between the film and the substrate is It is expressed as: ； In the formula, is the Young's modulus of the PVA film, and are the thermal expansion coefficients of the PVA film and substrate, is the temperature change, is the Poisson’s ratio of the PVA film; The obtained base curvature change is input into the modified Stoney formula to calculate the stress. The modified Stoney formula is expressed as: + ； In the formula, is the stress of the film, is the Young's modulus of the substrate, is the thickness of the substrate, is the change in base curvature, is the Poisson's ratio of the base, is the thickness of the PVA film.

2. The PVA film stress detection method based on machine vision according to claim 1, characterized in that: The curvature test assembly includes a laser, an etalon, a polarizer and a first camera, the first camera is aligned with the reflection area of ​​the laser, and the focal length and angle of the first camera are adjusted; The thickness test assembly comprises a scattering surface light source, a milky white glass diffusion plate and a second camera. The film to be tested is arranged between the second camera and the scattering surface light source, and the film to be tested is formed on a glass substrate.

3. A PVA film stress detection system based on machine vision, using a PVA film stress detection method based on machine vision as described in any one of claims 1-2, characterized in that: It includes sample preparation module, sample testing module, machine vision module and stress calculation module; The sample preparation module is used to prepare the PVA film, including spin coating, drying and heat treatment, to obtain the PVA film with a predetermined crystallinity for testing; A PVA film of predetermined crystallinity is obtained, an XRD spectrum of the PVA film is measured by an X-ray powder diffractometer, and a corresponding crystallinity is calculated by an amorphous standard sample method, thereby obtaining a heat treatment temperature and time for obtaining the predetermined crystallinity; The sample testing module is used to build a testing platform to obtain the transmission light density value of the PVA film and the change in substrate curvature, respectively, to measure the thickness of the PVA film and the change in substrate curvature before and after sample spin coating; The test platform includes a curvature test component, a thickness test component and a stage, wherein the thickness test component calculates the transmitted light density value by obtaining the incident and output grayscale images before and after the sample is placed, and then converts the light density value into the sample thickness using the Beer-Lambert law; the curvature test component calculates the substrate curvature change by obtaining the distance between adjacent light spots before and after the sample is placed.

4. The PVA film stress detection system based on machine vision according to claim 3, characterized in that: The machine vision module is used to collect incident and outgoing grayscale images of the PVA film before and after preparation and reflected signal spot images before and after sample placement through a CCD camera, and pre-process the acquired image data; The preprocessing of the incident and outgoing grayscale images is to perform denoising by averaging multiple images and calculate the grayscale value of each pixel in the image through histogram equalization; Preprocessing of the reflected signal spot image: convert the image into a grayscale image, use a Gaussian filter to smooth the image, enhance the edge of the spot area through morphological operations, perform edge detection to extract the spot contour and obtain the spacing between adjacent spots; The machine vision module is also used to calibrate the thickness measurement system before optical density measurement, including the following steps: Dark noise acquisition and analysis: Test in completely dark conditions, and acquire multiple dark noise images under exactly the same conditions; average all acquired dark noise images to obtain a mean image of dark noise; calculate the variance of the dark noise image to obtain a variance map of dark noise; Incident light grayscale calibration: prepare multiple standard light sources with known light densities, and collect multiple images for each standard light source; subtract the dark noise mean image from each standard light source image to obtain a net light grayscale image; average the multiple net light grayscale images of each standard light source to obtain an average light grayscale value; establish a calibration curve between light density and grayscale value based on the known light density value of the standard light source and the measured average light grayscale value; System response correction: Use the optical density-grayscale curve to correct the nonlinear response of the system and evaluate the impact of dark noise on measurement accuracy under small signal conditions; The stress calculation module is used to calculate the stress using the modified Stoney formula based on the PVA film thickness and the curvature change of the substrate obtained through the test, and to establish a stress thickness product versus film thickness change curve to obtain stress detection values ​​associated with the crystallinity and thickness of the PVA film.

Images