Method, device and production system for detecting defects in a layer of transparent film material

Abstract

The application provides a transparent film material layer defect detection method, device and production system. The method comprises the following steps: placing a transparent film material between a light source emitter and a projection receiving surface; causing the light source emitter to emit a light beam, the light beam penetrating the transparent film material at an inclined angle with the surface of the transparent film material and forming a projection image on the projection receiving surface; collecting the projection image on the projection receiving surface by using an image collection device and performing image analysis on the collected projection image, so as to determine whether there is a layer defect in the transparent film material according to whether there is a shadow area caused by the internal refractive index difference of the transparent film material in the image. The application solves the problem that the existing technology cannot detect the layer defect of the transparent film material with no deformation, no chromatic aberration and only refractive index difference on the surface.

Description

Technical Field

[0001] This invention relates to the field of defect detection within transparent film materials, and more particularly to a method, apparatus, and production system for detecting defects within transparent film materials. Background Technology

[0002] Currently, defect detection in transparent film materials primarily employs optical inspection methods based on surface reflection or back-side transmission. Common reflection methods illuminate the film surface with reflective light sources of varying angles, flickering frequencies, and brightness, utilizing the differences in surface defect morphology to generate different reflected light. A camera captures this reflected light and identifies the defect's shape and size based on grayscale differences. Common transmission methods place a transmitted light source on the back of the product, acquiring dark-field images by adjusting exposure, and similarly determining defects based on grayscale differences. These methods typically fix the angles of the light source and camera, as well as the equipment parameters, adjusting filters or light source brightness during use to adapt to the identification of different defects.

[0003] However, defects of different morphologies behave differently under different detection methods at different angles. When the surface of the membrane material has no obvious deformation and no color difference, the above-mentioned detection methods based on grayscale differences cannot effectively identify it. Defects in transparent membrane materials (such as OCA adhesive, TPU film, etc.) are divided into exogenous defects and intrinsic defects: exogenous defects usually cause deformation or color difference on the surface and can be identified by conventional surface inspection; however, if intralayer defects (such as gel nodules) are formed due to uneven polymerization, their surface has no deformation or color difference, and only the refractive index differs from that of normal materials, which cannot be detected by existing methods.

[0004] Therefore, how to accurately detect intralayer defects in transparent film materials that have no surface abnormalities but only differences in refractive index is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] The purpose of this invention is to provide a method, apparatus and production system for detecting defects within a transparent film material layer, so as to solve the problem that the prior art cannot detect defects within a transparent film material layer that have no surface deformation, no color difference and only a difference in refractive index.

[0006] To achieve the above objectives, the present invention is implemented as follows: In a first aspect, the present invention provides a method for detecting defects within a transparent film material layer, comprising: A transparent film material is placed between the light source emitter and the projection receiving surface; The light source emitter emits a light beam, which penetrates the transparent film material at an angle to the surface of the transparent film material and forms a projected image on the projection receiving surface. The projected image on the projection receiving surface is acquired using an image acquisition device, and the acquired projected image is analyzed to determine whether there are intralayer defects in the transparent film material based on whether there are shadow areas caused by the difference in refractive index inside the transparent film material.

[0007] Secondly, a device for detecting defects within a transparent film material layer is provided, comprising: A light source emitter, used to emit a beam of light; The projection receiving surface is located on the other side of the transparent film material; An image acquisition device is used to acquire projected images on a projection receiving surface; And a controller configured to perform the method described in the first aspect.

[0008] Thirdly, a transparent film material production system is provided, comprising: A conveying device for transporting transparent film materials; And a detection device for defects within the transparent film material layer as described in the second aspect, the detection device being disposed beside the conveying device for online detection of defects within the transparent film material during conveyance.

[0009] Fourthly, the present invention also provides a computer-readable storage medium on which a computer program is stored, wherein the computer program, when executed by a processor, implements the steps of the method described in the first aspect.

[0010] The beneficial effects of this invention are as follows: The present invention provides a method for detecting intralayer defects in transparent film materials. This method places the transparent film material between a light source emitter and a projection receiving surface, and allows the light beam to penetrate the film material at an oblique angle. The refractive index difference between the intralayer defects and normal material causes the transmitted light path to deflect, forming a identifiable shadow area on the projection receiving surface. An image acquisition device automatically acquires the projected image and performs image analysis to accurately determine whether intralayer defects exist within the film material. Because this detection method does not rely on the surface morphology or color differences of the defects, but rather on the shadow changes caused by refractive index differences, it can effectively detect intralayer defects (such as gel nodules) that have no surface deformation or color difference, solving the technical problem that existing technologies cannot detect such defects. Furthermore, the use of an image acquisition device and image analysis enables automated detection, avoiding the subjectivity and risk of missed detection associated with manual observation. The method is fast, accurate, and can be integrated into online production lines, showing promising industrial application prospects.

[0011] Furthermore, by setting an adjustable angle between the beam and the film material surface and performing geometric transformations on the projected image based on the angle value, this invention can correct projection distortion caused by tilted illumination, restore the actual size of the defect, and thus improve the accuracy of defect size detection. Simultaneously, by using time-series switching or multiple fixed light sources to achieve illumination at least two different incident angles and registering and fusing multiple projected images, the projection differences of the same defect under different incident angles can effectively enhance the contrast of weak defects and even reconstruct the depth information of the defects, thereby achieving three-dimensional characterization of intralayer defects. The above techniques work together to solve the problems of low defect contrast, deformation distortion, and lack of depth information in single-angle detection, significantly improving the detection rate and accuracy of intralayer defects.

[0012] Finally, this invention estimates the relative difference in refractive index from the gray-level distribution characteristics of the shadow region using a pre-calibrated gray-level-refractive index mapping relationship, thus achieving a quantitative assessment of defect severity. Alternatively, this invention employs coded structured light illumination, and calculates the refractive index distribution changes within the film material by demodulating the local distortion information of the coded pattern, thereby obtaining the three-dimensional morphology and quantitative severity of the defect. Furthermore, by selectively using light beams of different wavelengths, the shadow contrast can be enhanced for specific defect types (such as gel nodules), further improving detection sensitivity. Thus, this invention overcomes the limitations of existing technologies that can only qualitatively determine the presence or absence of defects, achieving the identification of intralayer defect types, quantification of severity, and three-dimensional morphological reconstruction, providing richer and more reliable detection data for film material quality control. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of the transmitted light path and uniform projection of a transparent film material under normal conditions. Figure 2 This is a schematic diagram of optical path deflection and shadow projection under defect conditions within a transparent film material layer; Figure 3 This is a schematic flowchart of a method for detecting defects within a transparent film material layer according to an embodiment of the present invention; Figure 4 This is a schematic flowchart of a method for detecting defects within a transparent film material layer according to another embodiment of the present invention; Figure 5 This is a schematic flowchart of a method for detecting defects within a transparent film material layer according to another embodiment of the present invention; Figure 6 This is a schematic flowchart of a method for detecting defects within a transparent film material layer according to another embodiment of the present invention; Figure 7 The grayscale curve of a severe defect in the detection channel; Figure 8 The grayscale curve of a medium-sized defect in the detection channel; Figure 9 The grayscale curve of a minor defect in the detection channel; Figure 10 This is the grayscale curve of the invisible defect in the detection channel; Figure 11 This is a schematic structural diagram of a transparent film material production system according to an embodiment of the present invention; Figure 12 This is a schematic structural diagram of a transparent film material layer defect detection device according to an embodiment of the present invention; Figure 13 This is a topology diagram of a computer-readable storage medium disclosed in this invention. Detailed Implementation

[0014] The present invention will now be described in detail with reference to the embodiments shown in the accompanying drawings. However, it should be noted that these embodiments are not intended to limit the present invention. Equivalent changes or substitutions in function, method, or structure made by those skilled in the art based on these embodiments are all within the scope of protection of the present invention.

[0015] The technical solutions provided by the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0016] Example 1: As Figure 1 As shown, when the transparent film material 10 is internally uniform and free of defects, the parallel light beam OB emitted by the light source emitter penetrates the film material, and the transmitted light ray O'B' forms a uniformly bright projected image on the projection screen 11. Figure 2 As shown, when there are intralayer defects such as gel nodules formed due to uneven polymerization inside the membrane material, the refractive index of the defect region 101 is different from that of the normal material, causing the transmitted light path to be deflected (e.g., the light path path...). and optical path This generates a light spot shadow 11a at the corresponding position on the projection screen 11. The grayscale, shape, and position of this shadow are related to the difference in refractive index and size of the defect. By acquiring the projected image through an image acquisition device and performing image analysis, it can be determined whether there are intralayer defects inside the film material.

[0017] Based on the above principles, such as Figure 3 As shown, this embodiment provides a method for detecting defects within a transparent film material layer, including: Step 302. Place the transparent film material between the light source emitter and the projection receiving surface.

[0018] The transparent film material to be tested (e.g., OCA optical adhesive film, TPU film, etc.) is placed in the testing area between the light source emitter and the projection receiving surface. The transparent film material is continuously conveyed by a guide roller system to ensure it remains flat and runs smoothly. The light source emitter is located on one side of the film material, and the projection receiving surface (e.g., a white frosted screen) is located on the other side. The transparent film material to be tested and the projection receiving surface are generally arranged parallel to each other. The height and angle between the light source emitter and the transparent film material can be adjusted. The position of the projection receiving surface ensures that the projected image can be received. Simultaneously, controlling the distance between the projection receiving surface and the transparent film material allows for varying degrees of magnification or reduction of the projected image, creating an optimal arrangement. The distance between the light source emitter and the film material surface is adjusted to 100mm–2000mm, and the beam width on the film material surface is 10mm–200mm, with the beam length along the width of the film material exceeding the film material width to ensure full coverage. By rationally arranging the relative positions of the light source, film material, and projection receiving surface, the transmitted light can uniformly cover the entire width of the film material, laying the foundation for a clear projected image.

[0019] Step 304. The light source emitter emits a light beam, which penetrates the transparent film material at an angle to the surface of the transparent film material and forms a projected image on the projection receiving surface.

[0020] Activate the light source emitter to emit a ribbon-shaped focused beam (a white light source with an illuminance greater than 300 lm can be used in a dark environment, or a colored laser source can be used without a dark environment). Adjust the angle ∠A between the beam and the surface of the transparent film material to an acute angle within the range of 10° to 80° (e.g., 30°). The beam passes through the transparent film material at this angle: when there are no defects inside the film material, the light is uniformly transmitted, forming a projection area with uniform brightness on the projection receiving surface; when there are intralayer defects (such as gel nodules) formed due to uneven polymerization inside the film material, the light is deflected because the refractive index at the defect is different from that of the normal material, forming a shadow or light spot at the corresponding position on the projection receiving surface. The shape and grayscale of this shadow are related to the degree of refractive index difference of the defect. In this way, by utilizing the principle of light path deflection caused by refractive index difference, invisible internal defects are transformed into visible projected shadows, thereby achieving effective detection of intralayer defects with no surface deformation and no color difference. Meanwhile, by adjusting the angle of the light source emitter and the distance between the three (transparent film material, light source emitter, and projection receiving surface), the contrast and degree of shadow distortion can be changed, which facilitates subsequent image processing and analysis.

[0021] Step 306. Use an image acquisition device to acquire a projection image on the projection receiving surface and perform image analysis on the acquired projection image to determine whether there are intralayer defects inside the transparent film material based on whether there are shadow areas caused by the difference in refractive index inside the transparent film material.

[0022] Wherein, the image acquisition device (such as an industrial area scan camera or a linear array camera) is fixedly installed at a position directly facing the projection receiving surface, and its field of view covers the entire projection receiving surface. When the light source emitter emits light stably, the image acquisition device continuously collects projection images at a preset frame rate or synchronously collects projection images through an external trigger signal to obtain digital image data. The collected images are transmitted to an image processing unit, and the image analysis process first preprocesses the images (such as graying, contrast enhancement and noise filtering) to highlight potentially existing shadow areas. Then, an image segmentation algorithm (such as segmentation based on gray threshold, region growing method or edge detection operator) is used to divide the image into a background region and candidate shadow regions. Since the transmitted light of a normal membrane material is uniform, the projected image presents uniform brightness; when an intralayer defect exists, the difference in refractive index caused by the defect will cause the transmitted light to deflect, forming a shadow with a gray level significantly lower than the background or a light spot higher than the background at the corresponding position. The image analysis unit determines whether there is a shadow area by detecting these local gray abnormality areas.

[0023] If one or more shadow areas are detected in the projected image, it is determined that there are intralayer defects inside the transparent membrane material, and the position coordinates of the defect can be further recorded (for example, converting pixel coordinates into actual physical positions); if no shadow area is detected, the membrane material is determined to be qualified. Thus, automatically acquiring projected images by means of an image acquisition device and automatically identifying shadow areas generated by refractive index differences through an image analysis algorithm realizes automatic and objective detection of intralayer defects, and avoids the subjectivity, fatigue and low efficiency of manual visual observation. Meanwhile, digital detection results are convenient for recording, tracing and subsequent statistical analysis, and provide reliable data support for quality control. The entire acquisition and analysis process can be completed at the millisecond level, meeting the requirements of high-speed online detection.

[0024] In the above embodiment, the included angle ∠A between the light beam in step 304 and the surface of the transparent membrane material can be continuously adjusted within the range of 10° to 80° (for example, by an electric rotating table) or adjusted in steps (for example, preset 30°, 45°, 60°) according to actual detection requirements. The smaller the included angle, the longer the penetration path of the light beam, the more obvious the deflection effect caused by the defect, and the higher the shadow contrast, but at the same time the greater the projection deformation; the larger the included angle, the smaller the shadow deformation but the contrast may decrease. In actual operation, the optimal included angle value can be pre-calibrated for membrane materials of different thicknesses or materials. By presetting the included angle interval, the optimal included angle range is set for different degrees of defect (heavy-medium-light-invisible) control schemes. The optimal included angle range for serious defects can be between 60° and 80°, the optimal included angle range for moderate defects can be between 40° and 60°, and the optimal included angle range for slight defects can be between 20° and 40°. Meanwhile, the distance from the light source to the transparent membrane material and the distance from the transparent membrane material to the projection receiving surface are adjusted to obtain the optimal imaging effect.

[0025] After acquiring the projected image, a geometric transformation is performed on the acquired image according to the currently set angle value to correct the defect projection deformation caused by the tilted beam illumination and restore the actual size of the defect. That is, after acquiring the projected image in step 306, due to the tilted beam illumination, there is a geometric deformation relationship between the actual size of the defect and the projected shadow. The image analysis unit performs inverse perspective transformation or scaling correction on the extracted shadow area according to the currently set angle value, converting the pixel size of the shadow into the actual physical size on the film plane (when the angle is 30°, the correction factor is sin30°=0.5). This geometric transformation can be performed before or after image segmentation. Thus, by adjusting the angle, different film materials and defect types can be flexibly adapted, improving detection sensitivity; by correcting the projection deformation through geometric transformation, the true size of the defect is accurately restored, providing a quantitative basis for quality grading.

[0026] As an alternative implementation of step 304, the light source emitter emits light beams at different angles to the surface of the transparent film material at at least two different times in a time-sequence switching manner, so that the light beams penetrate the same area of ​​the transparent film material at at least two different incident angles, forming at least two corresponding projected images on the projection receiving surface; the image analysis process includes: registering and fusing at least two projected images to enhance the contrast of defects or reconstruct the depth information of defects.

[0027] In this embodiment, the light source emitter adopts a time-sequential switching method: at at least two different times (e.g., 1 millisecond interval), light beams are emitted at different angles to the film surface (e.g., first at 30°, then at 60°). The two beams penetrate the same area of ​​the transparent film material sequentially (spatial alignment can be ensured by precisely controlling the movement of the film material or by using synchronous triggering), forming two projected images on the projection receiving surface. The image acquisition device is synchronized with the light source and acquires the two images sequentially. Unlike the prior art that changes the illumination angle by physically rotating the light source, this embodiment does not require any relative rotation between the light source emitter and the transparent film material during the angle switching process. In the image analysis in step 306, the two images are first spatially registered (e.g., by using cross-correlation or feature point matching to eliminate small offsets), and then fusion processing is performed: the maximum gray value of the same pixel position in the two images can be selected to enhance the contrast, or the gray value ratio can be calculated to highlight the difference in refractive index, or the depth information of the defect can be reconstructed based on the shadow offset of the same defect at different angles using the principle of binocular stereo vision. With this setup, there is no need to add multiple light sources. The overall system structure is compact, and the registration and fusion can significantly improve the contrast of weak defects and reconstruct the depth and location of defects, distinguishing whether the defects are located on the upper, inner, or lower surface of the membrane material.

[0028] As an alternative implementation of step 304, the light source emitter includes at least two fixedly arranged light source units, each of which emits a light beam at a different angle to the surface of the transparent film material, so that the light beams penetrate the same area of ​​the transparent film material at at least two different incident angles, forming at least two corresponding projected images on the projection receiving surface; the image analysis specifically includes: registering and fusing the at least two projected images to enhance the contrast of the defect or reconstruct the depth information of the defect.

[0029] Specifically, in step 304, the light source emitter includes at least two fixedly installed light source units (e.g., one light source unit is fixedly installed at 30° and the other at 60°). The two light source units simultaneously emit light beams (which can be of different or the same color), penetrating the same area of ​​the transparent film material to form two superimposed projected images on the projection receiving surface. In this embodiment, all light source units of the light source emitter are fixedly installed, and there is no relative rotational movement between them and the transparent film material during the detection process. In this embodiment, each light source unit can be configured with an independent projection screen with an tilt angle matching its beam direction, and each can be received by a corresponding image acquisition device. This solves the problem that direct registration and fusion are difficult due to differences in the position and deformation of the projected images when light from different angles is projected onto the same planar screen. Separating the two images can be achieved using one of the following methods: 1. Wavelength differentiation method: Two light sources emit red and blue light respectively, which are captured by a color camera. Two independent images are obtained by separating the red and blue channels.

[0030] 2. Spatial separation method: Two light sources are slightly offset in space, so that the two projected images form different areas on the screen, and the camera captures the two areas respectively.

[0031] 3. Polarization differentiation method: Orthogonal polarizers are installed in front of the two light sources, and a switchable polarizer is installed in front of the camera for time-division acquisition.

[0032] In the image analysis of step 306, the two separated images are registered and fused, and the processing method is the same as that of the time-series switching scheme (registration, fusion, depth reconstruction). In this embodiment, the time delay of time-series switching is avoided by simultaneous illumination, which is particularly suitable for high-speed online detection applications. Furthermore, multispectral response features can assist in defect classification, and the spatial consistency of the acquired images is better, with smaller registration errors.

[0033] In another preferred embodiment, to further improve the detection capability of defects with weak refractive index differences (such as early gel nodules) and effectively distinguish gel nodules from bubbles, this embodiment introduces a polarization optical detection method in transmission projection detection. This embodiment introduces a polarization optical detection method based on the basic embodiment (single-source tilted transmission), the difference being the configuration of the light source emitter and image acquisition device, as well as the expansion of the image analysis method. Specifically: a polarizer (e.g., a linear polarizer) is set at the light outlet of the light source emitter, making the emitted light beam linearly polarized. This linearly polarized light penetrates the transparent film material at an tilted angle. When intralayer defects (such as gel nodules) exist within the film material, the uneven molecular arrangement in the defect region produces a weak birefringence effect, thereby changing the polarization state of the transmitted light (e.g., linearly polarized light becomes elliptically polarized light); while the normal region does not change its polarization state. Meanwhile, defects such as bubbles or impurities generally do not have birefringence characteristics.

[0034] A rotatable or switchable analyzer (e.g., an electrically driven rotating polarizer or a liquid crystal adjustable polarizer) is installed at the front of the lens of the image acquisition device. Within one detection cycle, the controller controls the analyzer to rotate sequentially to two orthogonal directions (e.g., 0° and 90°), and simultaneously acquires two projected images I0 and I90 in the corresponding directions. The grayscale of the normal area is basically the same in both images, while the grayscale of the defect area changes periodically with the analyzer angle. The image analysis unit calculates the difference image ΔI = |I0 - I90| between the two images. In the difference image, the grayscale of the normal area approaches zero, while the defect area exhibits significant bright spots due to the change in polarization state, thus greatly enhancing the contrast between the defect and the background. Furthermore, by analyzing the grayscale change curve of the defect area at different analyzer angles, polarization characteristic parameters (such as degree of polarization, ellipticity, etc.) can be extracted to determine whether the defect has birefringence characteristics—if it has birefringence characteristics, it is identified as a gel nodule; if it does not have birefringence characteristics, it may be a bubble or impurity. Thus, by introducing a polarization detection dimension, this implementation not only enhances the contrast of weak defects but also effectively distinguishes between gel nodules and bubbles, achieving accurate identification of defect types. Simultaneously, polarization differential imaging significantly suppresses background noise, improving the signal-to-noise ratio and reliability of the detection.

[0035] As an alternative or preferred implementation of step 304, "emitting a light beam from the light source emitter" (replacing a normal uniform light beam), the light beam emitted by the light source emitter is structured light modulated by a coded grating, and the structured light has a predetermined coded pattern. The image analysis content in step 306 is adjusted accordingly; specifically, the image analysis includes: calculating the refractive index distribution change inside the transparent film material by demodulating the local distortion information of the coded pattern in the projected image, in order to quantitatively assess the three-dimensional morphology and severity of defects within the layer.

[0036] In step 304, the light beam emitted by the light source emitter is not uniform light, but structured light modulated by an coded grating. Specifically, a replaceable or electrically adjustable grating (such as a sinusoidal grating or Roche grating) is set at the light outlet of the light source emitter to modulate the original light beam into a cosine fringe pattern with a predetermined period (e.g., 0.5 mm) and an initial phase. This structured light penetrates the transparent film material at an inclined angle. When there are intralayer defects with non-uniform refractive index within the film material, the defect region acts as a microlens, causing local phase distortion (fringe bending or spacing change) in the structured light fringes transmitted through that region. The image acquisition device captures the projected image of the deformed fringes.

[0037] In the image analysis of step 306, the phase distribution of the fringes is demodulated using either Fourier transform or phase-shifting methods. Specifically: For a single fringe image, Fourier transform can be used: a two-dimensional Fourier transform is performed on the image to extract the fundamental frequency component containing phase information in the frequency domain, followed by an inverse Fourier transform to obtain the wrapped phase map, and finally, a phase unwrapping algorithm is used to obtain the continuous phase distribution. For multiple phase-shifted images (obtainable by switching gratings or multiple exposures), the standard four-step phase-shifting method can be used: four fringe maps with a phase difference of π / 2 are acquired, and the principal phase value at each point is calculated.

[0038] Based on the obtained phase distribution φ(x,y), the relationship between the phase distortion Δφ caused by the defect region and the film thickness d, wavelength λ, and refractive index change Δn is: Δφ = (2π / λ)·Δn·d. Given the film thickness d and wavelength λ, the refractive index distribution change Δn(x,y) can be calculated. Furthermore, by combining the Δn distribution of the defect region, the three-dimensional morphology of the defect can be reconstructed (e.g., by integrating Δn to obtain the optical path difference, thereby estimating the thickness or depth of the defect along the optical path), and the severity of the defect can be quantitatively assessed (e.g., the maximum value, average value, or volume of Δn). Thus, the structured light method elevates the detection dimension from "brightness contrast" to "phase change," achieving sub-pixel-level sensitivity and a higher detection rate for gel nodules with minimal refractive index differences. Moreover, by calculating the refractive index distribution change, the three-dimensional morphology and severity of the defect can be quantitatively assessed, providing detailed data for product quality grading and overcoming the limitations of traditional methods that only offer qualitative judgments.

[0039] In the image analysis of step 306, after preprocessing and shadow region extraction, the extracted shadow regions are subjected to morphological parameter calculation and threshold comparison. For example... Figure 4 As shown, image analysis operations also include: Step 401. Extract the morphological parameters of the shadow region. These parameters include at least one of the following: area, aspect ratio, roundness, and edge gradient. Area refers to the total number of pixels within the region, converted to the actual physical area (mm²). Aspect ratio is the ratio of the longer side to the shorter side of the smallest bounding rectangle of the calculated region, used to distinguish between point defects (close to 1) and linear scratches (much greater than 1). Roundness is calculated as 4π × area / perimeter²; ​​a roundness closer to 1 indicates a shape closer to a circle. Gel nodules typically exhibit approximately circular shapes, while bubbles or impurity particles may have irregular shapes. Edge gradient is the average grayscale gradient of the pixels at the boundary of the calculated region. Defects with drastic changes in refractive index (such as bubbles) have clear edges and large gradients; while gel nodules have gradual changes in refractive index, resulting in blurred edges and small gradients.

[0040] Step 402. Compare the morphological parameters with preset defect feature thresholds to determine the presence of intralayer defects. The calculated morphological parameters are compared with thresholds pre-calibrated using a large number of standard defect samples. For example, if an area is within the range of 0.01–0.5 mm², has an aspect ratio <1.5, and a roundness >0.7, it is classified as a gel nodule defect; if the edge gradient is greater than the preset threshold, it is more likely to be classified as an air bubble. By comparing the comprehensive thresholds of multiple parameters, the presence and preliminary type of intralayer defects are determined. That is, through quantitative analysis of morphological parameters, real defects can be effectively distinguished from artifacts such as dust, scratches, and noise, reducing the false alarm rate. Simultaneously, morphological features can provide basic data for subsequent defect classification and severity assessment, enabling rapid initial screening.

[0041] After completing the morphological parameter extraction, such as Figure 5 As shown, as a further refinement of the image analysis in step 306, the image analysis steps also include: Step 501. Input the morphological parameters into a pre-trained classification model. The classification model outputs the defect type identification result, which includes at least one of gel nodules, bubbles, and impurity particles. In this step, a large number of labeled projection image samples (including gel nodules, bubbles, impurity particles, and defect-free images) are collected in advance. The morphological parameters (area, aspect ratio, roundness, edge gradient, etc.) of each sample are extracted as feature vectors. The model is trained using Support Vector Machine (SVM), Random Forest, or a lightweight neural network (such as MobileNet) to obtain the classification model. During online detection, the morphological parameters of the current shaded area are input into the trained classification model. The model outputs the defect type corresponding to the area (e.g., gel nodules with a confidence score of 0.85, bubbles with a confidence score of 0.10, impurities with a confidence score of 0.05, etc.). The highest confidence score is taken as the final identification result.

[0042] Step 502. Based on the grayscale distribution characteristics of the shaded area, estimate the relative difference in refractive index of the defective area using a pre-calibrated grayscale-refractive index mapping relationship as a quantitative indicator of defect severity. Analyze the grayscale distribution characteristics of each shaded area: the grayscale value of the normal area is G0, and the lowest grayscale value at the center of the shadow is Gmin. Then, define the contrast ratio C = (G0 - Gmin) / G0. Establish a calibration curve between the grayscale contrast ratio C and Δn in advance using a standard refractive index sample (with a known refractive index difference Δn). During online inspection, estimate the relative difference in refractive index Δn of the defective area based on the measured C value using the calibration curve (e.g., linear or polynomial fitting). This value directly reflects the severity of the defect; the smaller Δn is, the milder the defect; the larger Δn is, the more severe the defect.

[0043] Thus, by introducing a machine learning classification model, the accuracy and robustness of defect type identification can be improved. Different types such as gel nodules, bubbles, and impurities can be automatically distinguished. The severity of defects can be quantitatively assessed through gray-scale-refractive index mapping, providing a quantitative basis for quality grading and process improvement. This overcomes the limitation of traditional methods that can only qualitatively determine the presence or absence of defects.

[0044] During the inspection process, the transparent film material moves continuously along the running direction under the drive of guide rollers (speed, for example, 10–100 m / min). After step 306 determines that a defect exists and identifies its location and type, as follows... Figure 6 As shown, the method in this embodiment further includes: Step 601. Record the location coordinates and defect type of the detected defects on the transparent membrane material in real time. The location coordinates are obtained by recording the membrane material's running length as the X-coordinate using an encoder or grating ruler, and then converting this to the Y-coordinate along the membrane width using the camera pixel position, forming a two-dimensional coordinate system. The defect type is determined by the type output from the aforementioned classification model (gel nodules, bubbles, impurities, etc.).

[0045] Step 602. Based on the location coordinates, count the number of defects per unit area using a preset window size as the defect density, and compare the defect density with a preset defect density threshold.

[0046] In this step, using a preset window size (e.g., a 1m x 1m square area) as the unit, the number of defects per unit area (e.g., 1m²) within the window is counted to obtain the defect density D. This density is compared with a defect density threshold D0 pre-set according to product quality standards (e.g., gel nodules ≤ 5 / m²): if D ≤ D0, the window area is considered acceptable; if D > D0, an alarm is immediately triggered (audio-visual alarm or display screen prompt), and process adjustment instructions are sent to the upstream process control system (e.g., extruder, mixing tank, curing oven, etc.), such as adjusting the raw material ratio, increasing the mixing time, or reducing the extrusion speed, to suppress the generation of subsequent defects. All recorded data is stored in a database for generating quality reports and trend analysis.

[0047] Step 603. When the defect density exceeds the threshold, an alarm signal is automatically issued or a process adjustment command is sent to the upstream process control system.

[0048] Thus, this embodiment enables defect location tracing and density monitoring, timely detection of production line anomalies and triggering closed-loop feedback, elevating detection from "post-production rejection" to "online control," effectively reducing the scrap rate. Furthermore, the statistical results can provide data support for process optimization, contributing to intelligent manufacturing.

[0049] Before the detection begins, a beam of light of the corresponding wavelength is selected according to the type of intralayer defect to be detected. The light source emitter can selectively emit beams of at least two different wavelengths (e.g., by using multiple LED chips of different colors built in, or by using a tunable laser), or a switchable filter (such as an electric filter wheel equipped with red, green, blue, near-infrared, etc.) can be set in front of the image acquisition device. The operations prior to step 304 also include: Before detection, a beam of light of the corresponding wavelength is selected according to the type of intralayer defect to be detected in order to enhance the shadow contrast of that type of defect in the projected image.

[0050] Because transparent film materials have different refractive indices for different wavelengths (dispersion effect), the refractive index difference between defects and the substrate also varies with wavelength. Selecting the optimal wavelength for a specific defect type (such as gel nodules) can maximize shadow contrast. Specific operating steps: 1. Perform multi-wavelength scanning on standard defect samples in advance to establish a database of "defect type - wavelength - contrast".

[0051] 2. Before online detection, query the database according to the type of target defect (such as gel nodules), select the wavelength with the highest contrast, and set the light source wavelength or switch the filter accordingly.

[0052] 3. Perform the detection steps 304 and 306 at the selected wavelength.

[0053] Optional implementation methods include: Multiple LED switching: The light source emitter integrates red (630nm), green (532nm), and blue (450nm) LEDs, which can be quickly switched via an electronic switch to select the channel with the best contrast after imaging.

[0054] Tunable filter: Using a white light source, a liquid crystal tunable filter (LCTF) is installed in front of the camera. The center wavelength (400-700nm, resolution 10nm) is switched quickly by electronic control to acquire multispectral image sequences. Then, the spectral curve of each pixel is extracted, and the defect type is identified by spectral analysis.

[0055] In this embodiment, wavelength selection is used to enhance the shadow contrast of specific defects by utilizing the dispersive properties of the material, which can significantly improve the detection capability of defects with weak refractive index differences (such as gel nodules). At the same time, multispectral imaging can acquire the spectral characteristics of defects, assisting in the identification of defect types (for example, bubbles and gel nodules have different spectral responses), and achieving more accurate classification.

[0056] For indentations of varying severity (severe, medium, minor, and invisible) in the same transparent membrane material sample, imaging analysis was performed along the membrane material's running direction (longitudinal detection channel, MD) and perpendicular to the running direction (lateral detection channel, CD). The results are as follows: Figures 7 to 10 As shown in the figure, for severe and moderate defects, the detected grayscale curves all exhibit obvious peaks, with grayscale differences exceeding 60, making the defects clearly identifiable. For mild defects, the detected grayscale difference is approximately 20–30, still identifiable, but with blurred edges. For "invisible" defects, the detected grayscale differences are all less than 15, the curves show no significant fluctuations, and they cannot be effectively detected by conventional grayscale thresholding. The above comparison shows that this type of "invisible" defect is difficult to form significant grayscale differences in both longitudinal and lateral detection, and even using bidirectional joint interpretation, it still cannot be reliably identified.

[0057] This embodiment does not rely on a specific detection channel, but instead uses tilted transmission projection to directly utilize the difference in refractive index between the defect and the substrate to form a shadow. Figure 10 Taking the "invisible" defect as an example, the method of this embodiment (such as a red laser light source at a 30° angle) is used for transmission projection detection. A shadow area corresponding to the defect appears in the projected image, and the grayscale difference is increased to 18-22, reaching the threshold for stable detection. Further, by using fixed dual-light source multi-angle illumination (30° and 60°) and image registration and fusion, the grayscale difference of the fused image reaches 35 (vertical equivalent) and 38 (lateral equivalent), respectively, and the defect outline is clearly discernible. Thus, this embodiment can convert the weak refractive index difference (grayscale difference <15) that cannot be identified by traditional longitudinal / lateral detection into a detectable shadow signal, and further enhance the contrast through multi-angle fusion, breaking through the sensitivity limit of single-direction detection.

[0058] Example 2: Combination Figure 11 and Figure 12 This embodiment provides a transparent film material production system 1100, which includes a conveying device for transporting transparent film material and a detection device 112. The conveying device includes multiple guide rollers 111 (at least including unwinding rollers, rewinding rollers, and several guide rollers). The transparent film material moves continuously along the running direction under the drive of the guide rollers 111, and the moving speed can be set according to the production process (e.g., 10-100 m / min). At least one guide roller has a horizontal position adjustment device at both ends to adjust the stability of the film material's feed, preventing deviation or vibration from affecting detection accuracy. The detection device 112 is arranged beside the conveying device. Specifically, in the straight section 1111 of the film material's running path, a light source emitter 1121 is fixed to one side, a projection receiving surface 1122 is fixed to the other side, and an image acquisition device 1123 is fixed to the same side as the projection receiving surface. The detection device 112 does not contact the film material and does not affect normal production. The detection device 112 is located at any position after the unwinding process and before the rewinding process, preferably after key processes such as extrusion, coating or curing, so as to provide timely feedback on process problems.

[0059] like Figure 12As shown, the detection device 112 includes: a light source emitter 1121, a projection receiving surface 1122, an image acquisition device 1123, and a controller 1124. The light source emitter 1121 emits a strip-shaped focused light beam. The light source emitter 1121 can use a white light source with an illuminance greater than 300 lm and be used in a dark environment, or it can use a red laser light source without the need for a dark environment. The light source emitter is installed on one side of the transparent film material's running path, and the angle between its light emission direction and the film material surface can be continuously or incrementally adjusted within the range of 10° to 80° to accommodate film materials of different thicknesses and materials. The projection receiving surface 1122 is located on the other side of the transparent film material, i.e., the side opposite to the light source emitter 1121. The projection receiving surface 1122 is made of a white or milky white frosted non-transparent material (e.g., PVC, PP, ABS plastic sheets or metal sheets coated with frosted paint), with a frosted particle size between 2μm and 40μm. The screen length is greater than the width of the film material, and the width is 5mm to 300mm. The projection receiving surface 1122 serves to uniformly diffuse the light transmitted through the film material, forming a clear projected image. The image acquisition device 1123, an industrial area scan camera or line scan camera, is installed directly in front of the projection receiving surface, covering the entire surface. The camera is equipped with a fixed-focus lens and necessary filters (if required), and is synchronized with the light source emitter via an external trigger signal to ensure precise matching between the acquisition time and the light source emission time. The controller 1124 (such as an industrial computer, embedded system, or programmable logic controller) is electrically connected to the light source emitter 1121 and the image acquisition device 1123, respectively, and is configured to execute the detection method described in Example 1. The controller is used to: control the on / off state and angle adjustment of the light source; trigger the camera to acquire images; receive image data and perform preprocessing, segmentation, feature extraction, and classification; output defect detection results and statistical data; and communicate with other equipment on the production line.

[0060] When production system 1100 is running, transparent film material continuously passes through the detection area. Controller 1124 performs the following operations: 1. Parameter settings: Based on the thickness, refractive index, and expected defect type of the film material to be inspected, preset parameters such as the included angle of the light source emitter 1121 (e.g., 30°), the light source brightness, and the camera exposure time. If a multi-angle illumination mode is used, preset two or more included angle values.

[0061] 2. Online Detection: When the membrane material is running stably, the controller 1124 triggers the light source emitter 1121 to emit a light beam (if it is a time-switched multi-angle system, light beams with different angles are emitted sequentially according to the preset timing sequence; if it is a fixed multi-light source system, light is emitted simultaneously). After the light beam penetrates the membrane material, a projected image is formed on the projection receiving surface; if there are intralayer defects such as gel nodules inside the membrane material, the light is deflected due to the difference in refractive index, resulting in a shadow area on the projected image.

[0062] 3. Image Acquisition and Analysis: The controller 1124 triggers the image acquisition device 1123 to synchronously acquire projected images and performs real-time image analysis: preprocessing (grayscale normalization, filtering and noise reduction), segmentation and extraction of shadow areas, calculation of morphological parameters (area, aspect ratio, roundness, edge gradient), and identification of defect types and severity through threshold comparison or classification models. If multi-angle fusion is used, multiple images also need to be registered and fused to enhance contrast or reconstruct depth information.

[0063] 4. Recording and Feedback: The controller 1124 records the location coordinates of defects in real time (vertical position obtained through the encoder, and horizontal position converted from pixel coordinates) and defect type. It calculates the defect density per unit area using a preset window (e.g., 1m × 1m) and compares it with a threshold. When the defect density exceeds the limit, it automatically issues an audible and visual alarm and sends process adjustment commands to the upstream process control system (e.g., extruder temperature controller, coating head gap adjustment mechanism, curing oven power regulator) via industrial Ethernet or fieldbus. These commands can be used to improve mixing uniformity, adjust curing temperature, or reduce production speed.

[0064] 5. Data Management: All test data and alarm records are stored locally or uploaded to the cloud-based quality management system for generating quality reports, tracing defective batches, and trend analysis.

[0065] Regarding the detection device 112, through the coordinated operation of the light source emitter 1121, the projection receiving surface 1122, the image acquisition device 1123, and the controller 1124, the principle of light refraction caused by the difference in refractive index between internal defects in the film material and the substrate is utilized to form an identifiable shadow area on the projection receiving surface 1122, thereby achieving automatic detection of hidden defects such as intralayer gel nodules without surface deformation or color difference. This device employs non-contact optical detection, requiring no damage to the film material. Furthermore, the detection parameters can be flexibly configured by adjusting the light source angle, wavelength, or using structured light, multi-angle fusion, and other methods to adapt to different materials, thicknesses, and defect types, significantly improving the sensitivity and applicability of the detection. The controller performs automatic image analysis, outputting quantitative data such as defect location, type, size, and refractive index difference, avoiding the subjectivity and fatigue-induced omissions of manual visual inspection. It offers fast detection speed, high accuracy, and stable operation in industrial settings.

[0066] Regarding the production system 1100, by integrating the aforementioned detection device 112 alongside the transparent film material conveying device, continuous online full inspection is achieved without the need for downtime or offline sampling. The system 1100 records defect coordinates and types in real time and calculates defect density per unit area using a preset window. Once a threshold is exceeded, an automatic alarm is triggered or a process adjustment command is sent to the upstream process control system, forming a closed-loop quality control system of "detection-feedback-control." This system can promptly suppress defect generation and significantly reduce scrap rates. Simultaneously, the long-term accumulated defect data, correlated with process parameters, provides data support for process optimization and quality traceability, facilitating intelligent manufacturing. Furthermore, this system can work collaboratively with existing surface defect detection devices to match the spatial coordinates of surface defects with those within the layer, identify composite defects, and comprehensively assess film material quality. Therefore, the production system provided in this embodiment demonstrates significant economic benefits and technological advancements in improving product yield, reducing production costs, and ensuring process stability.

[0067] Furthermore, one or more external excitation sources (not shown in the figure), such as ultrasonic transducers or pulsed laser heaters, can be added to the detection device 112. The excitation source is installed near the detection station of the membrane material, either on the same side or opposite side of the membrane material, without interfering with the transmitted light path. The controller 1124 is electrically connected to the excitation source to achieve precise control of the excitation timing.

[0068] Ultrasonic excitation method: An ultrasonic transducer (e.g., a piezoelectric ceramic transducer with a frequency of 20kHz to 1MHz) is attached to the edge of the membrane material or radiates ultrasonic waves to the membrane material via air coupling. During the detection process, the controller first controls the light source emitter to emit light, and the image acquisition device acquires a first projection image without ultrasonic excitation as a reference. Then, the ultrasonic transducer is activated to induce high-frequency micro-vibrations in the membrane material (lasting from tens to hundreds of milliseconds), while a second projection image is acquired simultaneously. Because the elastic modulus of defective areas such as gel nodules differs from that of normal membrane materials, local vibration mode anomalies (such as abrupt amplitude changes or phase lag) will occur at the defective areas under ultrasonic vibration. These anomalies will cause slight changes in the transmitted light path, thus forming dynamic shadows in the projection image. The image analysis unit performs pixel-by-pixel difference between the second projection image and the first projection image to extract the abnormal areas that only appear under ultrasonic excitation, which are the candidate defect areas.

[0069] Pulsed thermal excitation method: A pulsed laser (e.g., infrared band, wavelength 1.5μm~10μm) is used as the heat source, which is expanded and then uniformly irradiated onto the film surface or focused into a linear spot. The laser pulse width is on the order of milliseconds, and the energy density is controlled within a range insufficient to damage the film material. The controller controls the laser to emit a single pulse to instantaneously heat the film material. Due to the difference in thermal expansion coefficient and thermal conductivity between the defective area and the normal film material, local thermal stress and thermal deformation will occur, causing transient deflection of the transmitted light path. A high-speed camera (frame rate ≥1kHz) is used to continuously acquire projection image sequences before, during, and after thermal excitation. The image analysis unit analyzes the grayscale change curve of each pixel in the image sequence over time, extracts characteristic parameters such as response peak, response time, and recovery time, and compares them with the pre-calibrated response curve of the normal material. Pixels with deviations exceeding the threshold are identified as defects.

[0070] This implementation upgrades passive optical detection to active composite detection by introducing ultrasonic or thermal excitation. This "amplifies" early defects with minimal refractive index differences into detectable optical signal changes, significantly improving the detection rate. Simultaneously, by comparing the differences between the presence and absence of excitation, fixed background noise can be effectively suppressed, making it suitable for high-precision detection scenarios. Furthermore, ultrasonic and thermal excitation methods can be used individually or in combination to adapt to the characteristics of different transparent film materials.

[0071] It should be noted that for the same or similar technical solutions as in Embodiment 1, please refer to Embodiment 1, and will not be repeated here.

[0072] This invention also provides a terminal device, which may include a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the above-described... Figure 1 The various processes of the transparent film material layer defect detection method embodiment shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0073] Combination Figure 13 As shown, this embodiment also discloses a specific implementation of a computer-readable storage medium 1300. This computer-readable storage medium 1300 can be configured wholly or partially in a physical computer, server, cluster server, or data center.

[0074] In this embodiment, the computer-readable storage medium 1300 stores computer program instructions 1301. The computer program instructions 1301 are read and executed by a processor 1302 to perform the steps in the method for detecting defects in the transparent film material layer as disclosed in Embodiment 1.

[0075] Optionally, the computer-readable storage medium 1300 can be configured as a server, and the server runs on a physical device used to build a private cloud, hybrid cloud, or public cloud. The computer-readable storage medium 1300 can also be configured as random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc.

[0076] The computer-readable storage medium 1300 is used to store a program, and the processor 1302, upon receiving an execution instruction, executes the method for detecting defects within a transparent film material layer disclosed in Embodiment 1.

[0077] Meanwhile, the processor 1302 disclosed in this embodiment may be an integrated circuit chip with signal processing capabilities. The processor 1302 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor.

[0078] The technical solution of the same part in the computer-readable storage medium 1300 disclosed in this embodiment as in Embodiment 1 and / or Embodiment 2 is described in Embodiment 1 and / or Embodiment 2, and will not be repeated here.

[0079] The detailed descriptions listed above are merely specific descriptions of feasible embodiments of the present invention, and are not intended to limit the scope of protection of the present invention. All equivalent embodiments or modifications made without departing from the spirit of the present invention should be included within the scope of protection of the present invention.

[0080] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0081] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method of detecting defects in a layer of a transparent film material, characterized by, include: A transparent film material is placed between the light source emitter and the projection receiving surface; The light source emitter emits a light beam, which penetrates the transparent film material at an angle to the surface of the transparent film material and forms a projected image on the projection receiving surface. The projected image on the projection receiving surface is acquired using an image acquisition device, and the acquired projected image is analyzed to determine whether there are intralayer defects in the transparent film material based on whether there are shadow areas caused by the difference in refractive index inside the transparent film material. The light source emitter emits light beams at different angles to the surface of the transparent film material at at least two different times in a time-sequence switching manner, so that the light beams penetrate the same area of ​​the transparent film material at at least two different incident angles, forming at least two corresponding projected images on the projection receiving surface. The image analysis includes: registering and fusing the at least two projected images to enhance the contrast of the defects or reconstruct the depth information of the defects.

2. The method according to claim 1, characterized in that, The angle between the light beam and the surface of the transparent film material is adjustable within the range of 10° to 80°. After acquiring the projected image, a geometric transformation is performed on the acquired projected image according to the currently set angle value to correct the defect projection deformation caused by the tilted beam illumination and restore the actual size of the defect.

3. The method according to claim 1, characterized in that, The light source emitter includes at least two fixedly arranged light source units, which respectively emit light beams at different angles to the surface of the transparent film material, so that the light beams penetrate the same area of ​​the transparent film material at at least two different incident angles, forming at least two corresponding projected images on the projection receiving surface. The image analysis includes: registering and fusing the at least two projected images to enhance the contrast of the defects or reconstruct the depth information of the defects.

4. The method according to claim 1, characterized in that, The image analysis also includes: Extract the shape parameters of the shadow region, including at least one of area, aspect ratio, roundness, and edge gradient; The morphological parameters are compared with a preset defect feature threshold to determine whether there are intralayer defects.

5. The method according to claim 4, characterized in that, The image analysis steps also include: The morphological parameters are input into a pre-trained classification model, and the classification model outputs the identification result of the defect type, wherein the defect type includes at least one of gel nodules, bubbles, and impurity particles. Based on the gray-scale distribution characteristics of the shaded area, the relative difference in refractive index of the defect area is estimated using a pre-calibrated gray-scale-refractive index mapping relationship, which serves as a quantitative indicator of the severity of the defect.

6. The method according to claim 4, characterized in that, The transparent membrane material moves continuously along the running direction during the detection process, and the method further includes: Record the location coordinates and defect type of detected defects on the transparent film material in real time; Based on the location coordinates, the number of defects per unit area is counted using a preset window size as the defect density, and the defect density is compared with a preset defect density threshold. When the defect density exceeds the threshold, an alarm signal is automatically issued or a process adjustment command is sent to the upstream process control system.

7. The method according to claim 1, characterized in that, The light beam emitted by the light source emitter is structured light modulated by a coded grating, and the structured light has a predetermined coded pattern. The image analysis also includes: calculating the change in refractive index distribution inside the transparent film material by demodulating the local distortion information of the coded pattern in the projected image, so as to quantitatively assess the three-dimensional morphology and severity of defects within the layer.

8. The method according to claim 1, characterized in that, The light source emitter can selectively emit light beams of at least two different wavelengths, or the image acquisition device is provided with a switchable filter. The method further includes: Before detection, a beam of light of the corresponding wavelength is selected according to the type of intralayer defect to be detected in order to enhance the shadow contrast of that type of defect in the projected image.

9. A device for detecting defects within a transparent film material layer, characterized in that, include: A light source emitter, used to emit a beam of light; The projection receiving surface is located on the other side of the transparent film material; An image acquisition device is used to acquire projected images on a projection receiving surface; And a controller configured to perform the method of any one of claims 1-8.

Images