Method for identifying undamaged area of optical transparent protective material

Abstract

The invention discloses a method for identifying an undamaged area of an optical transparent protective material, which effectively avoids transmission white edge errors and machine identification misleading caused by refractive index differences of transparent materials through multi-angle image acquisition and registration, and improves the identification accuracy of the undamaged area of the optical transparent protective material by combining morphological reconstruction and hole filling technologies. The accurate segmentation of the damaged area and the non-damaged area is realized; according to the method, the multi-view features are integrated by adopting the Poisson fusion algorithm, the accuracy and global consistency of damage identification are remarkably improved, meanwhile, the residual visible area of the transparent protective material can be quickly and reliably evaluated through pixel-level quantitative calculation, and objective maintenance and replacement basis is provided for a high-precision application scene of the optical transparent protective material.

Description

Technical Field

[0001] The present application relates to the technical field of optically transparent material testing, and in particular to a method for identifying undamaged areas of an optically transparent protective material. Background Art

[0002] Optically transparent protective materials, due to their high transparency and excellent protective properties, are widely used in key areas such as bank counter windows and transparent armor. These materials must provide physical protection while ensuring an unobstructed view for the user. However, after being subjected to external forces such as impact, friction, or strong shock, the material surface and interior are susceptible to damage such as cracks, holes, and localized refractive index changes, resulting in a decrease in light transmission, which directly affects both observation capabilities and protective effectiveness.

[0003] Currently, methods for detecting damaged areas on transparent materials have the following limitations:

[0004] Insufficient qualitative assessment: Existing technologies mostly rely on manual visual inspection or simple optical instruments to qualitatively describe the damaged area (such as marking the crack location or range), but cannot achieve accurate quantitative analysis of undamaged areas, making it difficult to evaluate the actual usability of materials in complex environments.

[0005] Poor algorithm adaptability: Traditional machine vision algorithms are susceptible to interference from light refraction when inspecting transparent materials. Due to the difference in refractive index between the transparent material and the air interface, a white edge effect is easily generated at the edges of damaged areas, leading to misjudgments or missed detections during image recognition.

[0006] Multi-angle collaborative defects: Single-angle image acquisition is difficult to fully capture the three-dimensional damage characteristics of transparent materials. If multi-angle images are not geometrically corrected and fused, registration errors will occur, further reducing detection accuracy.

[0007] In addition, the damaged areas of transparent materials often present irregular shapes (such as radial cracks, mesh-like micropores, etc.). Traditional threshold segmentation or edge detection algorithms are difficult to accurately distinguish the boundaries between damaged and undamaged areas, especially under low-contrast conditions, where the problem of missegmentation is particularly prominent. Summary of the Invention

[0008] The embodiments of the present application provide a method for identifying undamaged areas of an optically transparent protective material, thereby achieving a quantitative assessment of undamaged areas of a transparent protective material.

[0009] In order to achieve the above-mentioned purpose, the technical solution of the embodiment of the present invention is:

[0010] In a first aspect, an embodiment of the present invention provides a method for identifying an undamaged area of an optically transparent protective material, comprising:

[0011] Configure an optical contrast background for the transparent protective material sample to be tested, and enhance the visual difference between the damaged area and the undamaged area through preset optical properties;

[0012] Use high-resolution imaging equipment to capture images of the sample covered on an optical contrast background from multiple spatial orientations, and use the front view image as the spatial reference for coordinate calibration;

[0013] Based on the spatial reference, geometric correction and feature analysis are performed on multi-view images to identify abnormal transmittance areas caused by damage in the sample;

[0014] The damage features analyzed from multiple perspectives are used to generate a global damage distribution model through image fusion technology to eliminate edge errors caused by material refractive index gradients.

[0015] The global damage distribution model was morphologically segmented to separate the undamaged area from the damaged area, and the area ratio of the undamaged area was calculated by pixel density statistics.

[0016] In some possible implementations, the optical contrast background is a black backing plate whose size and shape are customized based on the size of the sample and whose surface has no reflective interfering texture.

[0017] In some possible implementations, the high-resolution imaging device captures images from at least nine angles: the front, top, bottom, left, right, upper left, lower left, upper right, and lower right of the sample.

[0018] In some possible implementations, the image fusion technology uses a Poisson fusion algorithm to achieve seamless stitching of multi-angle damage features through gradient domain optimization.

[0019] In some possible implementations, morphological segmentation is performed on the global damage distribution model to separate the undamaged area from the damaged area, and the area ratio of the undamaged area is calculated through pixel density statistics, including:

[0020] Morphological reconstruction of the global damage distribution model is performed using black and white contrasting colors to accurately segment the undamaged and damaged areas of the transparent protective material;

[0021] The segmented image is then expanded and eroded to fill holes in the reconstructed image and smooth edge burrs.

[0022] The padded image is divided into a preset number of pixels, and the number of pixels in the undamaged area is counted to determine the area ratio of the undamaged area.

[0023] In some possible implementations, the abnormal transmittance region includes at least one of a crack, a penetrating damage, and a surface friction scratch, and the abnormal transmittance region is determined by a grayscale value mutation threshold.

[0024] One or more technical solutions provided in the embodiments of the present invention have at least the following technical effects or advantages:

[0025] In the embodiment of the present invention, multi-angle image acquisition and registration are used to effectively avoid the transmission white edge error and machine recognition misleading caused by the refractive index difference of transparent materials. Combined with morphological reconstruction and hole filling technology, accurate segmentation of damaged and undamaged areas is achieved. The Poisson fusion algorithm is used to integrate multi-view features to significantly improve the accuracy and global consistency of damage identification. At the same time, through pixel-level quantitative calculation, the remaining visible area of the transparent protective material can be quickly and reliably evaluated, providing an objective maintenance and replacement basis for high-precision application scenarios of optically transparent protective materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments of the present invention. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0027] Figure 1 A schematic flow chart of an embodiment of a method for identifying undamaged areas of an optically transparent protective material provided for implementation of the present invention;

[0028] Figure 2 Schematic diagram of the collected sample information;

[0029] Figure 3 Schematic diagram of the sample with the edges marked;

[0030] Figure 4 is the front image of the sample used as a reference;

[0031] Figure 5 Schematic diagram of an image after acquisition and a corresponding image after perspective processing according to an embodiment of the present invention;

[0032] Figure 6 Schematic diagram of the damaged area of the transparent protective material sample obtained after feature extraction in an embodiment of the present invention;

[0033] Figure 7 Schematic diagram of the damaged areas of samples at different angles after geometric correction in an embodiment of the present invention;

[0034] Figure 8 Schematic diagram of global damage distribution after image fusion in an embodiment of the present invention;

[0035] Figure 9 A schematic diagram of a structure for reconstructing the fused image morphology according to an embodiment of the present invention;

[0036] Figure 10 Schematic diagram of the structure of the damaged area of the transparent protective material with morphological reconstruction in an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0038] In the relevant description of this embodiment, the terms "including, containing, having" and the like are open terms, and are generally understood to include but not be limited to; the term "at least one" is generally understood to mean one or more, where "plurality" refers to two or more; the term "at least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items, for example, "at least one of a, b or c", or "at least one of a, b and c", can all represent: a, b, c, ab (i.e., a and b), ac, bc, or abc, where a, b, c can be single or multiple respectively; the symbol "A / B" is used to describe the selection relationship of associated objects, generally indicating an "or" relationship before and after.

[0039] In the following description of the present embodiment, the terms used in the embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. The singular forms "a", "an" and "the" used in the embodiments of the present application and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise.

[0040] Those skilled in the art should understand that in the following description of the embodiments of the present application, the order of serial numbers does not mean the order of execution, some or all of the steps can be executed in parallel or sequentially, and the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0041] It will be understood by those skilled in the art that the numerical ranges in the examples of the present application are to be understood as also specifically disclosing each intermediate value between the upper and lower limits of the ranges. Each smaller range between the intermediate value in any stated value or stated range and any other stated value or intermediate value in the range is also included in the present invention. The upper and lower limits of these smaller ranges may be independently included or excluded in the scope.

[0042] Unless otherwise indicated, the technical / scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which this application belongs. Although this application describes only preferred methods and materials, any methods and materials similar or equivalent to those herein may also be used in the implementation or testing of this application. All documents mentioned in this specification are incorporated by reference to disclose and describe the methods and / or materials related to the documents. In the event of any conflict with any incorporated document, the content of this specification shall prevail.

[0043] In order to illustrate the technical solution of the present invention, specific embodiments are provided below.

[0044] Optically transparent protective materials, due to their high transparency and excellent protective properties, are widely used in key areas such as bank counter windows, police shield observation windows, and transparent armor. These materials must provide physical protection while ensuring an unobstructed view for the user. However, when subjected to external forces such as impact, friction, or strong shock, the material surface and interior are susceptible to damage such as cracks, holes, and localized refractive index changes, resulting in a decrease in light transmission, which directly affects both observation capabilities and protective effectiveness.

[0045] Currently, methods for detecting damaged areas on transparent materials have the following limitations:

[0046] Insufficient qualitative assessment: Existing technologies mostly rely on manual visual inspection or simple optical instruments to qualitatively describe the damaged area (such as marking the crack location or range), but cannot achieve accurate quantitative analysis of undamaged areas, making it difficult to evaluate the actual usability of materials in complex environments.

[0047] Poor algorithm adaptability: Traditional machine vision algorithms are susceptible to interference from light refraction when inspecting transparent materials. Due to the difference in refractive index between the transparent material and the air interface, a white edge effect is easily generated at the edges of damaged areas, leading to misjudgments or missed detections during image recognition.

[0048] Multi-angle collaborative defects: Single-angle image acquisition is difficult to fully capture the three-dimensional damage characteristics of transparent materials. If multi-angle images are not geometrically corrected and fused, registration errors will occur, further reducing detection accuracy.

[0049] In addition, the damaged areas of transparent materials often present irregular shapes (such as radial cracks, mesh-like micropores, etc.). Traditional threshold segmentation or edge detection algorithms are difficult to accurately distinguish the boundaries between damaged and undamaged areas, especially under low-contrast conditions, where the problem of missegmentation is particularly prominent.

[0050] Based on this, an embodiment of the present invention provides a method for identifying undamaged areas of an optically transparent protective material, thereby achieving a quantitative assessment of the undamaged areas of the transparent protective material.

[0051] Figure 1A schematic flow chart of an embodiment of a method for identifying undamaged areas of an optically transparent protective material provided for the implementation of the present invention is shown in FIG. Figure 1 As shown, the above method may include:

[0052] S101, configuring an optical contrast background for the transparent protective material sample to be tested, and enhancing the visual difference between the damaged area and the undamaged area through preset optical properties;

[0053] In some embodiments, the optically contrasting background is a black backing plate whose size and shape are customized based on the size of the sample and whose surface has no reflective interfering texture.

[0054] It should be noted that the black background can absorb light to the greatest extent, making the light reflected or scattered by the damaged area more prominent. In the embodiment of the present invention, the selection of a black backplane can enhance the visual contrast between the damaged area and the undamaged area. The optical contrast background can be a background designed with a specific pattern to further enhance the visual difference. For example, a uniformly distributed grid pattern can be used, and the line width and spacing of the grid are adjusted according to the damage characteristics of the material and the detection accuracy requirements. For the detection of minor damage, the grid spacing and line width can be appropriately reduced to make the interference of the damaged area on the grid pattern more obvious.

[0055] In some embodiments, to ensure optical contrast without the interference of reflective textures, the backplane can be made of a non-reflective composite material to suppress diffuse reflection of ambient light. For example, a carbon fiber-reinforced matte plastic can be used. Alternatively, a multi-layer light-absorbing structure can be used, such as a honeycomb light-absorbing layer with a black substrate. Furthermore, the backplane surface can be precision-polished or chemically etched to create a micron-level roughness, disrupting the specular reflection path and preventing light spots or streaks in the image.

[0056] In some embodiments, before collecting sample images, the surface of the damaged optical protective material sample can be cleaned using reagents such as anhydrous ethanol to remove wipeable contaminants such as stains, dust, and glue marks on the surface of the sample to prevent external factors from interfering with the accuracy of identification of undamaged areas.

[0057] In the process of enhancing the visual difference between damaged and undamaged areas through preset optical characteristics, a sample of optical protective material is covered on a visual identification card, and an optical flat panel light is used to fill in the light from the bottom of the sample. By selecting a suitable light source and lighting method, the optical characteristics of the damaged area are highlighted. Using an optical flat panel light to fill in the light from the bottom of the sample can provide uniform lighting and reduce the impact of shadows. At the same time, adjusting the brightness and angle of the light source can create a clear visual difference between the damaged and undamaged areas under lighting. For example, when illuminated from a specific angle, the damaged area may show an abnormal increase or decrease in reflected light. By adjusting the lighting angle, this difference can be made more obvious.

[0058] In some embodiments, a light shielding plate can be installed around the back panel to block the intrusion of lateral ambient light, ensuring that the light field in the detection area is only controlled by the bottom optical flat panel light to avoid errors caused by the influence of ambient light.

[0059] In some embodiments, the captured images can also be processed using contrast enhancement features in image processing software or equipment. By adjusting parameters such as contrast, brightness, and color saturation, the boundaries between damaged and undamaged areas can be made clearer, and the color differences more pronounced. For example, increasing contrast can make the colors of damaged areas more vivid, while undamaged areas appear relatively darker, thereby improving the recognition of damaged areas.

[0060] S102, using a high-resolution imaging device, capturing images of the sample covered on the optical contrast background from multiple spatial orientations, and performing coordinate calibration using the front view image as a spatial reference;

[0061] For example, high-resolution imaging equipment can use industrial-grade high-resolution cameras such as CCD or CMOS cameras with pixels greater than or equal to 12 million pixels to ensure clear image details and capture micron-level damage on the surface of transparent materials, such as fine cracks and scratches. The camera can be equipped with a fixed-focus or zoom lens, and the focal length of the lens is adjusted according to the size of the sample to ensure that the imaging distortion is small and avoid geometric distortion of the damaged area due to lens distortion. The camera can be installed on a three-dimensionally adjustable bracket or robotic arm, and the angle switching can be achieved through high-precision guide rails or turntables to ensure that the distance and inclination angle between the camera and the sample are consistent when collecting at different angles. Use a tripod or fixed tooling to stabilize the camera to avoid jitter caused by manual operation and ensure repeatability and consistency of image acquisition.

[0062] In some embodiments, the high-resolution imaging device captures images from at least nine angles of the sample, including the front, top, bottom, left, right, upper left, lower left, upper right, and lower right.

[0063] Specifically, the front view direction perpendicular to the sample surface is used as the reference reference image. The upper and lower edges of the image are parallel and there is no significant trapezoidal deformation. It is used to establish the origin of the coordinate system and the reference plane.

[0064] The top, bottom, left, and right parts (horizontally offset at 90°, 270°, 180°, and 0°, respectively) are photographed from the four sides of the sample at a 45° angle to capture damage in the edge area, such as edge cracks and chipping, to supplement the edge details not covered in the front view direction.

[0065] The upper left, lower left, upper right, and lower right (offsets of 45°, 225°, 315°, and 135°, respectively) are photographed from 30° to 45° above and below the four corners of the sample to focus on detecting stress concentration damage in the corner areas, such as star-shaped cracks and localized crushing, to avoid blind spots caused by occlusion from the front view angle.

[0066] For example, Figure 2 Schematic diagram of the collected sample information. Figure 3 Schematic diagram of the sample with the edges marked. Figure 4 This is the front image of the sample used as a reference. Figure 4 The middle green edges are basically parallel, and the width of the yellow edges is determined by the shooting height. The higher the height, the wider the edge width. The specific width is determined based on the needs of actual applications.

[0067] S103, based on the spatial reference, geometric correction and feature analysis are performed on the multi-view images to identify the abnormal transmittance areas caused by damage in the sample;

[0068] After capturing images of the sample at different angles, four positioning points are marked on the sample surface using the front image as a reference. The remaining eight angle images are then mapped to a reference coordinate system using methods such as perspective transformation algorithms. By extracting edge features or corner points such as the sample outline and damage boundary from each image, the coordinates of the non-frontal image are mapped to the reference coordinate system through rigid transformations such as translation and rotation, or affine transformations such as scaling and tilting, ensuring that the coordinates of the same physical point are consistent in images from different perspectives. A calibration plate can also be used to assist in calculating camera extrinsics (including rotation matrices and translation vectors) to achieve geometric alignment of multi-angle images.

[0069] In some embodiments, for large-sized samples, when the sample size exceeds the camera's field of view, regional stitching acquisition technology can also be used to achieve panoramic image synthesis through overlapping area feature matching to ensure that there are no blind spots in detection.

[0070] In other embodiments, for arc-shaped or curved samples, the recognition accuracy of curved surface damage can be improved by increasing image acquisition at side tilt angles and constructing a sample surface model through a three-dimensional reconstruction algorithm.

[0071] Through the above steps, full-view high-resolution imaging of transparent protective material samples can be achieved. Combined with the calibration of the orthographic reference coordinates, a unified and accurate spatial data basis is provided for the subsequent morphological processing of damaged areas, feature fusion, and quantitative analysis of undamaged areas, effectively solving the problem of blind spots in detection of transparent materials caused by differences in transmittance and refractive index.

[0072] In some embodiments, the abnormal transmittance area includes at least one of a crack, a penetrating damage, and a surface friction scratch, and the abnormal transmittance area can be determined by a gray value mutation threshold.

[0073] The above step S103 specifically includes: taking the reference image as a reference and the other angles as supplements, performing perspective processing on the collected image, performing feature extraction on the image after perspective extraction, and obtaining the damaged area of the transparent sample.

[0074] Perspective processing, based on the principle of perspective transformation, aims to eliminate image distortion caused by varying shooting angles. In actual photography, images captured from angles other than the reference image directly in front of you can exhibit trapezoidal distortion and other issues due to the tilted viewing angle. Perspective transformation uses a mathematical model to convert these distorted images into images with the same perspective and geometry as the reference image, ensuring spatial consistency between images from different angles.

[0075] To achieve perspective transformation, corresponding feature points must be found in the reference image and the other-angle images. These feature points can be edge corners of the sample or prominent pattern features. A feature point matching algorithm is used to accurately identify and match these feature points in the different images. Then, based on the matched feature point pairs, a perspective transformation matrix is calculated. This matrix contains the parameters such as translation, rotation, and scaling required to transform the other-angle images to the reference image's perspective.

[0076] The calculated perspective transformation matrix is used to correct images at other angles. Through matrix operations, each pixel in the image is mapped to the coordinate system of the reference image to achieve the image perspective conversion. During the conversion process, in order to ensure the quality and continuity of the image, a resampling operation can be performed. The resampling algorithm is used to estimate the pixel value of the new position based on the information of the surrounding pixels, thereby obtaining a high-quality corrected image. Figure 5 As shown, Figure 5 Schematic diagram of a captured image and a corresponding image after perspective processing in an embodiment of the present invention.

[0077] Feature extraction of the image after perspective extraction may include edge feature extraction, texture feature extraction, and feature fusion and screening.

[0078] Edges are regions in an image where grayscale values change dramatically, often corresponding to the boundaries and outlines of objects. Edge features are particularly prominent in damaged areas, such as crack edges and the boundaries of damaged areas. Edge detection algorithms, such as the Canny edge detector and the Sobel operator, can be used to extract edge information from the rectified image. By analyzing edge characteristics such as continuity, length, and shape, the extent and morphology of the damaged area can be further determined.

[0079] As you can understand, the texture of damaged areas differs from that of undamaged areas. For example, the texture of cracked areas is rough and irregular, while the texture of undamaged areas is relatively smooth. Texture analysis methods (such as gray-level co-occurrence matrix and local binary pattern) are used to extract the texture features of the image. By calculating parameters such as texture contrast, entropy, and correlation, the texture in the image is quantitatively described. Based on the differences in these texture features, damaged and undamaged areas can be further distinguished.

[0080] The extracted optical, edge, and texture features are fused to comprehensively consider multiple aspects of information and improve the accuracy of damage area identification. Feature fusion algorithms, such as weighted averaging and decision trees, can be used to combine and optimize different features. Furthermore, feature filtering is required to remove noise and false features. Feature thresholds and conditions are set, and the fused features are filtered to retain only areas that meet damage characteristics.

[0081] After feature extraction and screening, the possible damage areas are marked. Connectivity analysis can be used to merge adjacent pixels with similar features into a damage area. By analyzing the connectivity of the damage area, the scope and shape of the damage can be determined more accurately, avoiding misjudging scattered noise points as damage areas. Figure 6 As shown, Figure 6 Schematic diagram of the damaged area of the transparent protective material sample obtained after feature extraction in an embodiment of the present invention. Figure 7 Schematic diagram of the damaged areas of samples at different angles after geometric correction in an embodiment of the present invention. Figure 7 In the process, images collected at different angles have the same visual position as the reference image after geometric correction and registration.

[0082] S104, using image fusion technology to generate a global damage distribution model based on the damage features analyzed from multiple perspectives, eliminating edge errors caused by material refractive index gradients;

[0083] In some embodiments, the image fusion technology uses the Poisson fusion algorithm to achieve seamless splicing of multi-angle damage features through gradient domain optimization. Specifically, it includes:

[0084] S1041, performing gradient field calculation on multi-angle damage feature images;

[0085] The gradient field reflects the intensity and direction of brightness changes at each pixel in the image. For example, a crack edge will exhibit a high gradient due to a sudden change in transmittance, while a uniform area will have a lower gradient. By performing gradient analysis on damage feature maps collected from multiple angles, the system can accurately capture the contour details of damaged areas such as cracks and holes.

[0086] By performing gradient field calculation in step S1041, it is possible to ensure that the geometric features of the damage are completely preserved, avoid the loss of key information due to viewing angle differences, and provide high-precision input for global fusion.

[0087] S1042, optimizing the fusion boundary by the least square method based on the gradient field of the reference image;

[0088] Optimizing the fusion boundary mathematically resolves edge conflicts when stitching multi-angle images, particularly white edge errors caused by differences in material refractive index. Using the gradient field of the orthographic image as a benchmark, the least squares method is used to adjust the weights and directions of the gradient fields at other angles to ensure that the fused boundary aligns with the true damage morphology.

[0089] For example, false white edges caused by refraction at oblique viewing angles are automatically reduced, while the edges of real cracks are enhanced and aligned. This process effectively eliminates optical interference, ensures seamless spatial integration of multi-view data, and enhances the physical accuracy of the model.

[0090] S1043: Generate a global seamless damage feature distribution map.

[0091] In some embodiments, a global damage distribution map is generated by integrating the optimized gradient field, ultimately presenting a seamless, edge-free damage panorama. The Poisson equation can transform multi-angle gradient information into a coherent pixel distribution, reconstructing a high-fidelity damage model. Figure 8 As shown, Figure 8 Schematic diagram of global damage distribution after image fusion in an embodiment of the present invention.

[0092] S105, performing morphological segmentation on the global damage distribution model, separating the undamaged area from the damaged area, and calculating the area ratio of the undamaged area through pixel density statistics.

[0093] After constructing the global damage distribution model in step S104, accurate separation and quantitative calculation of undamaged areas are achieved through morphological segmentation and pixel statistics. In some embodiments, step S105 specifically includes:

[0094] S1051 uses black and white contrast to perform morphological reconstruction of the global damage distribution model and accurately segment the undamaged and damaged areas of the transparent protective material;

[0095] Morphological reconstruction of the global damage distribution model is performed using black and white contrast. A threshold can be set to perform binary segmentation between damaged areas (e.g., high brightness or specific grayscale values) and undamaged areas (e.g., low brightness or uniform grayscale). For example, damaged areas are marked as white pixels (value 1) and undamaged areas are marked as black pixels (value 0). This process can be combined with edge detection algorithms to optimize the segmentation boundaries, ensuring accurate extraction of damage contours such as cracks and holes, and avoiding mis-segmentation caused by gradual changes in transmittance. See [1]. Figure 9 As shown, Figure 9 Schematic diagram of the structure for reconstructing the fused image morphology in an embodiment of the present invention.

[0096] S1052, filling holes in the reconstructed image and smoothing edge burrs by dilation and erosion operations on the segmented image;

[0097] For example, the segmented binary image can be subjected to morphological closing operations (dilation followed by erosion) to fill tiny holes, and opening operations (erosion followed by dilation) to eliminate isolated noise points. The dilation operation expands the edges of the damaged area to close the gaps, and the erosion operation restores the original boundary shape, ultimately generating a continuous and smooth mask of the undamaged area. Figure 10 As shown, Figure 10 Schematic diagram of the structure of the damaged area of the transparent protective material with morphological reconstruction in an embodiment of the present invention.

[0098] S1053: Divide the padded image into a preset number of pixels, count the number of pixels in the undamaged area, and thus determine the area ratio of the undamaged area.

[0099] Specifically, the optimized image is divided into a preset number of pixels (e.g., 12 million pixels), with each pixel corresponding to an actual tiny area on the material surface. An automated script traverses all pixels and counts the total number of pixels in the undamaged area. For example, if the number of pixels in the undamaged area is 6.89 million, its area accounts for 57.37%. This process can leverage the efficient computing power of computers to quickly output quantitative results, avoiding the subjective errors of manual measurement and supporting batch processing.

[0100] Compared with the existing measurement of undamaged areas of optically transparent protective materials, the method for identifying undamaged areas of optically transparent protective materials provided by the embodiment of the present invention accurately describes the characteristic details such as damaged areas and extended cracks of transparent protective materials through a series of image processing methods such as visual transformation, feature fusion, and morphological reconstruction, thereby achieving accurate measurement of undamaged areas of transparent protective materials. The present invention avoids the system measurement error caused by the difference in sample acquisition image angles through image registration of damaged areas, reduces the workload of machine recognition calculations, and improves the accuracy of the test. The present invention is a method for measuring undamaged areas of transparent materials. By collecting image information from multiple angles, it avoids the transmission white edge error caused by transparent light passing through materials with different refractive indices, and avoids the machine's field of view being misled by the white edge area. The present invention's test of undamaged areas of transparent protective materials is achieved through physical calculations, without any interference factors. The algorithm has high accuracy and reliability, and can meet the measurement of the degree of damage of various transparent materials.

[0101] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other. Each embodiment focuses on the differences from other embodiments.

[0102] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit the present application. Although the present application has been described in detail with reference to the aforementioned embodiments, a person of ordinary skill in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some or all of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the present application.

Claims

1. A method for identifying undamaged areas of an optically transparent protective material, characterized in that: include: Configure an optical contrast background for the transparent protective material sample to be tested, and enhance the visual difference between the damaged area and the undamaged area through preset optical properties; Using high-resolution imaging equipment, collecting images of the sample covered on the optical contrast background from multiple spatial orientations, and performing coordinate calibration using the front view image as a spatial reference; Based on the spatial reference, geometric correction and feature analysis are performed on the multi-view images to identify areas of abnormal transmittance caused by damage in the sample; The damage features analyzed from multiple perspectives are used to generate a global damage distribution model through image fusion technology to eliminate edge errors caused by material refractive index gradients. The global damage distribution model is morphologically segmented to separate the undamaged area from the damaged area, and the area ratio of the undamaged area is calculated by pixel density statistics.

2. The method according to claim 1, characterized in that The optical contrast background is a black backing plate whose size and shape are customized based on the size of the sample and has no reflective interfering texture on the surface.

3. The method according to claim 2, characterized in that The high-resolution imaging device collects images from at least nine angles of the sample, namely, the front, top, bottom, left, right, upper left, lower left, upper right, and lower right.

4. The method according to claim 3, characterized in that The image fusion technology adopts the Poisson fusion algorithm and realizes seamless splicing of multi-angle damage features through gradient domain optimization.

5. The method according to claim 4, characterized in that Perform morphological segmentation on the global damage distribution model to separate the undamaged area from the damaged area, and calculate the area ratio of the undamaged area through pixel density statistics, including: Morphologically reconstructing the global damage distribution model using black and white contrasting colors to accurately segment the undamaged area and damaged area of the transparent protective material; The segmented image is then expanded and eroded to fill holes in the reconstructed image and smooth edge burrs. The padded image is divided into a preset number of pixels, and the number of pixels in the undamaged area is counted to determine the area ratio of the undamaged area.

6. The method according to claim 5, characterized in that The abnormal transmittance area includes at least one of a crack, a penetrating damage, and a surface friction scratch, and the abnormal transmittance area is determined by a gray value mutation threshold.

Images