Impregnated paper surface defect detection system and method

Abstract

The invention relates to the technical field of impregnated paper defect detection, in particular to an impregnated paper surface defect detection system and method. The method comprises the following steps: respectively acquiring reflection images of visible light and near-infrared bands by adopting a vertical illumination light source, calculating a reflection intensity difference index to separate a specular reflection dominant region from an impregnated paper body reflection region, and inhibiting the specular reflection dominant region through difference analysis and a binary mask. The problem that surface image acquisition of the impregnated paper is seriously interfered by mirror reflection is avoided; and combining the impregnated paper body reflection region and the suppressed specular reflection dominant region to obtain an impregnated paper integrated image, comparing the impregnated paper integrated image with an impregnated paper standard image, marking all suspicious regions, and identifying defect types according to image feature data of all suspicious regions. And the qualified state of the impregnated paper is judged based on the defect degree index corresponding to the defect type, so that automatic and standardized impregnated paper surface quality evaluation is realized, and the impregnated paper detection efficiency and detection reliability are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of impregnated paper defect detection, and in particular to a system and method for detecting surface defects of impregnated paper. Background Art

[0002] Impregnated paper is a key material widely used in furniture manufacturing, architectural decoration, and other fields. Its surface quality directly impacts the appearance and performance of end products. With the increasing automation of industry, machine vision-based inspection technology, with its advantages of non-contact, high speed, and quantifiable capabilities, has gradually become the mainstream method for impregnated paper surface defect detection.

[0003] Several patents exist for machine vision inspection of impregnated paper quality. For example, Chinese Patent Publication No. CN116721067A discloses a machine vision-based method for inspecting the impregnation quality of impregnated paper. This method uses the SIFT algorithm to perform key point matching, obtaining pixel correspondences between the base paper and impregnated paper images. It then calculates distance difference sequences and grayscale differences, yielding displacement and grayscale change values. Ultimately, the impregnation quality is inspected using warpage and impregnation uniformity. This approach replaces manual inspection with machine vision, significantly improving inspection efficiency and avoiding the randomness inherent in manual spot checks.

[0004] However, the existing technology has the following problems: 1. During the surface inspection of impregnated paper, since the paper surface may have a certain glossiness, it is easy to produce specular reflection under conditions such as vertical lighting. The existing technology does not fully consider the interference problem caused by specular reflection on the surface of impregnated paper.

[0005] This may result in a large number of highlight areas in the captured image, masking the actual surface defect information and affecting the extraction accuracy and detection accuracy of defect features.

[0006] 2. The existing technology mainly evaluates the impregnation quality from the perspective of overall warpage and impregnation uniformity. It lacks effective identification and classification methods for the various types of defects on the surface of the impregnated paper, and cannot distinguish the specific defect categories in detail. It is difficult to meet the needs of refined quality control and is not conducive to subsequent targeted quality assessment and treatment of different types of defects. Summary of the Invention

[0007] The present invention aims to overcome the deficiencies in the above-mentioned background technology and provide a system and method for detecting surface defects of impregnated paper, which can effectively solve the problems of mirror reflection interference and insufficient defect type recognition ability existing in the prior art, thereby realizing high-precision, rapid and comprehensive detection of surface defects of impregnated paper.

[0008] The present invention solves its technical problems by employing the following technical solutions: First, the present invention provides a surface defect detection system for impregnated paper, comprising a surface reflection image acquisition module, an image reflection characteristic separation module, a reflection interference dynamic suppression module, a suspicious image region marking module, and a defect identification and analysis module. The modules are connected as follows: the surface reflection image acquisition module is connected to the image reflection characteristic separation module, the reflection interference dynamic suppression module is connected to the image reflection characteristic separation module and the suspicious image region marking module, and the defect identification and analysis module is connected to the suspicious image region marking module.

[0009] The surface reflection image acquisition module is used to illuminate the surface of the impregnated paper using a vertical lighting source, and respectively acquire reflection images of the impregnated paper surface in the visible light band and the preset near-infrared band.

[0010] The image reflection characteristic separation module is used to calculate the reflection intensity difference index of the impregnated paper surface under the condition of vertical lighting source based on the collected reflection image, and separate the mirror reflection dominant area and the impregnated paper body reflection area based on the reflection intensity difference index.

[0011] The dynamic suppression module for reflective interference is used to control low-angle lighting sources to set timing. It combines multi-frame images collected synchronously by a near-infrared camera under different lighting conditions for image processing to dynamically suppress the mirror-dominated area.

[0012] The image suspicious area marking module is used to combine the impregnated paper body reflection area and the suppressed mirror reflection dominant area to obtain the impregnated paper integrated image, compare the impregnated paper integrated image with the impregnated paper standard image, and mark all suspicious areas.

[0013] The defect recognition and analysis module is used to identify the defect type based on the image feature data of all suspicious areas, and determine the qualified status of the impregnated paper based on the defect degree index corresponding to the defect type.

[0014] On the other hand, the present invention provides a method for detecting surface defects of impregnated paper. S1. A vertical lighting source is used to illuminate the surface of the impregnated paper, and reflection images of the impregnated paper surface in a visible light band and a preset near-infrared band are collected respectively.

[0015] S2. Based on the collected reflection image, calculate the reflection intensity difference index of the impregnated paper surface under the condition of vertical lighting source, and separate the specular reflection dominant area and the impregnated paper body reflection area based on the reflection intensity difference index.

[0016] S3. Control the low-angle lighting source to set the timing, combine the multi-frame images synchronously collected by the near-infrared camera under different lighting conditions to perform image processing, and dynamically suppress the mirror reflection dominant area.

[0017] S4. Combining the impregnated paper body reflection area and the suppressed specular reflection dominant area to obtain an impregnated paper integrated image, comparing the impregnated paper integrated image with the impregnated paper standard image, and marking all suspicious areas.

[0018] S5. Identify defect types based on the image feature data of all suspicious areas, and determine the qualified status of the impregnated paper based on the defect degree index corresponding to the defect type.

[0019] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention adopts a vertical illumination light source to collect reflection images of visible light and near-infrared bands respectively, calculates the reflection intensity difference index to separate the mirror reflection dominant area and the impregnated paper body reflection area, thereby avoiding the problem that the image collection of the impregnated paper surface is seriously interfered by the mirror reflection, improving the quality of the collected image, and providing a clear data source for subsequent defect detection.

[0020] (2) The present invention uses low-angle illumination light source timing control and near-infrared camera to synchronously capture multiple frames of images, and suppresses the mirror reflection dominant area through differential analysis and binary masking, thereby solving the problem of defects being masked due to local strong light interference in the mirror reflection dominant area, resulting in the effect of improving the defect recognition rate and ensuring that subtle defects are not missed.

[0021] (3) The present invention marks all suspicious areas, performs a fusion analysis on the color moments, texture features and shape features of all suspicious areas to form a comprehensive feature vector, and compares the comprehensive feature vector with the feature vector of the known defect type to determine the defect type, thereby distinguishing different types of surface defects in detail. It provides comprehensive and detailed information for the quality assessment and process improvement of impregnated paper, which is beneficial for manufacturers to take targeted quality control measures for different types of defects.

[0022] (4) The present invention performs deviation analysis on the comprehensive feature vectors of each suspicious area to obtain the feature deviation degree, adds the feature deviation degree to the corresponding weight product to obtain the defect degree index, and compares it with the threshold to determine the qualified state, thereby realizing automated and standardized impregnated paper surface quality assessment and improving the impregnated paper detection efficiency and detection reliability. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0024] Figure 1 This is a schematic diagram of system module connections of the present invention.

[0025] Figure 2It is a schematic diagram of the specific flow of the system content in the present invention.

[0026] Figure 3 This is a schematic diagram of the specific steps of the reflection interference dynamic suppression module in the present invention.

[0027] Figure 4 Schematic diagram of the process steps of the present invention. DETAILED DESCRIPTION

[0028] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that, unless otherwise specifically stated, the relative arrangement of components and steps, numerical expressions, and numerical values ​​described in these embodiments do not limit the scope of the present invention. It should also be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn to scale.

[0029] The following description of at least one exemplary embodiment is merely illustrative in nature and is not intended to limit the invention, its application, or uses. Technologies, methods, and apparatus known to those skilled in the art may not be discussed in detail, but where appropriate, such technologies, methods, and apparatus should be considered part of the specification.

[0030] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0031] See also Figure 1 and Figure 2 As shown, the present invention provides a surface defect detection system for impregnated paper, comprising a surface reflection image acquisition module, an image reflection characteristic separation module, a reflection interference dynamic suppression module, a suspicious image area marking module, and a defect identification and analysis module. The modules are connected as follows: the surface reflection image acquisition module is connected to the image reflection characteristic separation module, the reflection interference dynamic suppression module is connected to the image reflection characteristic separation module and the suspicious image area marking module, and the defect identification and analysis module is connected to the suspicious image area marking module.

[0032] The surface reflection image acquisition module uses a vertical illumination source to illuminate the impregnated paper surface and capture reflection images in the visible light band and a preset near-infrared band. Leveraging the differences in reflective characteristics across these bands, the module provides foundational data for more accurate separation of different reflective areas and defect detection. This enriches inspection information from the image acquisition source, enabling more detailed information on surface defect characteristics and improving detection accuracy.

[0033] It should be noted that the reflection images of the impregnated paper surface in the visible light band and the preset near-infrared band are collected by a high-resolution visible light camera and a high-sensitivity near-infrared camera respectively. The optical axes of the lenses of the two cameras are set in parallel and at a set angle to the plane of the conveyor belt that transports the impregnated paper to ensure the clarity and integrity of the collected images. At the same time, the two cameras are synchronously aligned with the impregnated paper detection area in space to ensure that the collected images have a one-to-one correspondence in position and have the same resolution and frame rate, so as to achieve simultaneous acquisition of impregnated paper surface information and provide basic data for subsequent separation of surface reflection characteristics. Both the high-resolution visible light camera and the high-sensitivity near-infrared camera are based on solid-state image sensors. The difference lies in the response ability of the solid-state image sensors used to light in different bands. The core sensor used by the high-resolution visible light camera is a solid-state image sensor optimized for the visible light band, such as a CMOS sensor or a CCD sensor. The core sensor used by the high-sensitivity near-infrared camera is a near-infrared enhanced CMOS sensor or a near-infrared enhanced CCD sensor.

[0034] The image reflection characteristics separation module calculates the reflection intensity difference index of the impregnated paper surface under vertical illumination conditions based on the collected reflection image. This reflection intensity difference index is then used to separate the specular reflection-dominant area from the impregnated paper's reflective area. This separation method effectively distinguishes areas prone to reflection interference from areas containing true surface information of the impregnated paper. This provides the prerequisite for subsequent targeted processing of reflection interference and defect detection in the reflective area of ​​the paper, helping to improve overall inspection results.

[0035] It should be noted that the specific content of the image reflection characteristic separation module is as follows: pixel alignment of the reflection images of the impregnated paper surface in the visible light band and the preset near-infrared band is performed to obtain the reflection intensity value of each pixel point of the visible light reflection image and the near-infrared reflection image.

[0036] The reflection intensity difference index of each pixel on the surface of the impregnated paper under the condition of vertical illumination is calculated according to the reflection intensity value of each pixel, and all pixels that meet the set specular reflection dominant area judgment conditions are screened based on the reflection intensity difference index.

[0037] The area encircled by all pixel positions that meet the set specular reflection dominant area determination conditions is taken as the specular reflection dominant area, and the remaining area in the reflected image except the specular reflection dominant area is taken as the impregnated paper body reflection area.

[0038] Furthermore, the calculation formula for the reflection intensity value of the visible light reflection image at each pixel point is: , where is the reflection intensity value of the visible light reflection image at the i-th pixel point, is the pixel value of the visible light reflection image at the i-th pixel point, are the pixel value of the standard whiteboard pixel and the calibrated reflectivity of the standard whiteboard respectively. At the same time, the reflection intensity value of the near-infrared reflection image at each pixel is calculated in the same way as the reflection intensity value of the visible light reflection image at each pixel.

[0039] The ratio of the reflection intensity value of the visible light reflection image at each pixel point to the reflection intensity value of the near-infrared reflection image at the corresponding pixel point is used as the reflection intensity difference index of each pixel point on the impregnated paper surface under vertical illumination conditions.

[0040] In a specific embodiment, the setting condition for determining the mirror reflection dominant area is: if the reflection intensity value of a pixel point in the visible light reflection image is greater than the set visible light high reflection intensity threshold and the reflection intensity value in the near-infrared reflection image is greater than the set near-infrared high reflection intensity threshold, and the reflection intensity difference index of the pixel point is within the preset mirror reflection intensity range, then the pixel point is a pixel point within the mirror reflection dominant area.

[0041] Specular reflection is the direct reflection of light from an object's surface, typically caused by smooth surfaces. Its reflection intensity is high and varies little across different wavelengths. Volume reflection from impregnated paper, on the other hand, occurs when light enters the paper, is scattered and absorbed, and then reflected back out. Impurities on the impregnated paper's surface and internal defects produce reflection characteristics in the visible and near-infrared bands that differ from specular reflection.

[0042] The present invention adopts a vertical illumination light source to respectively collect reflection images of visible light and near-infrared bands, calculates the reflection intensity difference index to separate the mirror reflection dominant area and the impregnated paper body reflection area, avoids the problem that the image collection of the impregnated paper surface is seriously interfered by the mirror reflection, improves the quality of the collected image, and provides a clear data source for subsequent defect detection.

[0043] The dynamic reflection interference suppression module controls low-angle illumination sources to set timings. Combined with image processing using multiple frames of images captured synchronously by a near-infrared camera under different lighting conditions, it dynamically suppresses areas dominated by specular reflections. Through specific timing control and image processing methods, it dynamically suppresses areas of reflection interference, preventing reflections from interfering with defect detection. This ensures better image quality and improves defect detection accuracy. This is particularly important for inspecting high-gloss impregnated paper surfaces, effectively resolving the key technical challenge of reflection interference.

[0044] like Figure 3As shown, the specific content and steps of the dynamic suppression module for reflective interference are as follows: W1. Control the low-angle lighting light source to work alternately in the on and off states according to the set timing, and synchronize the near-infrared camera with the low-angle lighting light source timing to capture the surface image each time the light source switches its state, and obtain a surface image group for each alternating state switching.

[0045] W2. Perform differential processing on the lighting-on state image and the lighting-off state image in each surface image group to obtain a differential image.

[0046] W3. Based on the pixel grayscale value distribution of the differential image, identify the local highlight interference area in the mirror reflection dominant area.

[0047] Furthermore, the identification of local highlight interference areas in the mirror reflection dominant area is specifically carried out by comparing the grayscale values ​​of different pixels in the lighting-on state image and the lighting-off state image in each surface image group to obtain the grayscale difference of different pixels, and constructing a differential image based on the grayscale difference of different pixels.

[0048] Pixels with grayscale differences greater than the set grayscale difference threshold are screened from the mirror reflection dominant area in the differential image corresponding to each surface image group. Pixel areas are circled according to the positions of the screened pixels. The pixel areas circled by each surface image group are subjected to overlapping area analysis, and the overlapping areas are used as local highlight interference areas in the mirror reflection dominant area.

[0049] W4. Generate a binary mask image with the local highlight interference area as white and the rest of the specular reflection dominant area as black. Perform the interference suppression operation on the binary mask image to generate the specular reflection dominant area after highlight interference suppression. The white value is set to 1 and the black value is set to 0.

[0050] Furthermore, the interference suppression operation performed on the binary mask image is: directly replacing the area where the mask mark is 1 in the binary mask image with the pixel value of the corresponding position of the lighting-on state image.

[0051] The area where the mask mark is 0 in the binary mask image continues to use the pixel value of the corresponding position in the binary mask image.

[0052] The above-mentioned binary mask processing can successfully suppress the local highlight interference area that was originally saturated under vertical illumination, and replace the information of the local highlight interference area with information captured under low-angle illumination that is closer to the actual surface features, avoiding information loss or distortion.

[0053] In a specific embodiment, the low-angle illumination light source is switched between on and off states between capturing two adjacent frames of images.

[0054] For example, frame 1: The near-infrared camera captures an image with the low-angle light source turned on.

[0055] Frame 2: The near-infrared camera captures an image with the low-angle light source turned off, relying solely on ambient light or another fixed light source, such as vertical illumination. This cycle is repeated, capturing an image of the surface each time the light source switches state.

[0056] The surface image set consists of images alternating between the illumination-on and illumination-off states. The illumination-on image features a low-angle light source illuminating the surface's microscopic undulations and any non-mirror-flat areas. Due to their smooth nature, mirror-flat areas reflect low-angle light away from the camera, appearing as relatively dark areas in the image.

[0057] The lighting-off state image shows the scene when the low-angle light source is not working. The scene lighting mainly comes from the vertical lighting source. The mirror-dominated area will strongly reflect the vertical incident light, appearing as a very bright white or highlighted area in the image, while the impregnated paper reflection area is in normal diffuse reflection with moderate brightness, containing surface texture and possible deep information.

[0058] The present invention uses low-angle lighting light source timing control and near-infrared camera to synchronously capture multi-frame images, and suppresses the mirror reflection-dominated area through differential analysis and binary masking, thereby solving the problem of defects being masked due to local strong light interference in the mirror reflection-dominated area, resulting in the effect of improving the defect recognition rate and ensuring that subtle defects are not missed.

[0059] The suspicious area marking module combines the reflective area of ​​the impregnated paper with the suppressed specular reflection dominant area to create an integrated image of the impregnated paper. This integrated image is then compared with a standard image of the impregnated paper to mark all suspicious areas. This module can comprehensively and meticulously identify areas with potential defects in the integrated image, enabling more detailed detection of potential surface defects, further improving detection accuracy and providing accurate regional positioning for subsequent defect type identification.

[0060] It should be noted that the specific method of marking all suspicious areas is: grayscale processing is performed on the integrated image of the impregnated paper to obtain the grayscale value of each pixel in the integrated image of the impregnated paper, the grayscale value deviation between each pixel and the pixel at the corresponding position in the standard image of the impregnated paper is calculated, and the pixel points whose grayscale value deviation exceeds the preset grayscale threshold are marked as potential suspicious points. The entire integrated image of the impregnated paper is traversed to form all suspicious areas according to the positions of all potential suspicious points.

[0061] The defect recognition and analysis module identifies defect types based on image feature data from all suspicious areas and determines the acceptable status of the impregnated paper based on the defect severity index corresponding to the defect type. This module analyzes the characteristics of suspicious areas from multiple dimensions, fully considering the differences in color, texture, and shape of surface defects. It effectively distinguishes different types of defects, enabling comprehensive and accurate classification and assessment of surface defects, providing a detailed and accurate basis for quality control.

[0062] It should be noted that the defect type is identified based on the image feature data of all suspicious areas. Specifically, the image of each suspicious area is obtained from the integrated image of the impregnated paper, the intensity value of each pixel in the RGB color space is extracted from the image of each suspicious area, and the color moment of each suspicious area in the RGB color space is calculated.

[0063] The grayscale images of each suspicious area are normalized, and the frequency of pixel pairs that meet the set distance and direction angle is counted based on the normalized grayscale images. The texture features of each suspicious area are calculated using the gray level co-occurrence matrix.

[0064] The area, perimeter and aspect ratio are obtained according to the contour of each suspicious area, which are used to form shape features. The shape features are fused with color moments and texture features to form a comprehensive feature vector of each suspicious area.

[0065] The comprehensive feature vector of each suspicious area is analyzed for similarity with the feature vector corresponding to each known defect type, and the defect type of each suspicious area is determined based on the similarity.

[0066] The defect type of each suspicious area is determined as follows: the comprehensive feature vector of each suspicious area is calculated with the feature vector corresponding to each known defect type using the cosine similarity analysis formula to obtain the similarity between the feature vectors of each suspicious area and each known defect type, and the known defect type with the largest similarity to the feature vector corresponding to each suspicious area is selected as the defect type of each suspicious area.

[0067] In one specific embodiment, the color moments are used to describe the color distribution characteristics of the suspicious area. These color moments include the color mean, variance, and skewness. The color mean reflects the central tendency of the primary colors, the variance reflects the degree of color dispersion, and the skewness measures the symmetry of the color distribution. For example, for a normal impregnated paper area, the color mean, variance, and skewness are relatively stable within a set range. However, if the color moments of a suspicious area significantly exceed the set range, a color defect may exist.

[0068] The gray-level co-occurrence matrix is ​​used to calculate the texture features of suspicious areas. The gray-level co-occurrence matrix describes the space of gray levels in an image. Based on the gray-level co-occurrence matrix, contrast, energy, entropy, and correlation are obtained. Contrast reflects the clarity of the texture and the depth of the texture grooves. Energy measures the uniformity of the image. Entropy reflects the complexity of the image. Correlation indicates the degree of linear correlation between elements in the gray-level co-occurrence matrix. For example, the scratch defect on the surface of impregnated paper will increase the contrast of the texture at the scratch, decrease the energy, increase the entropy, and the correlation may also change. The presence of texture defects such as scratches can be identified by changes in these texture features.

[0069] Shape features describe the shape characteristics of suspicious areas. For example, a hole defect typically appears as a closed region with a certain area and perimeter, and the aspect ratio may vary depending on the hole shape. On the other hand, a strip-shaped wrinkle defect has a relatively large aspect ratio. Shape features are used to initially screen out suspicious areas that may contain defects.

[0070] The present invention marks all suspicious areas, fuses and analyzes the color moments, texture features and shape features of all suspicious areas to form a comprehensive feature vector, and compares the comprehensive feature vector with the feature vector of known defect types to determine the defect type, thereby distinguishing different types of surface defects in detail. This provides comprehensive and detailed information for quality assessment and process improvement of impregnated paper, which is beneficial for manufacturers to take targeted quality control measures for different types of defects.

[0071] It should be noted that the defect type corresponding to the defect degree index analysis method is: extract the shape features, color moments and texture features of the corresponding defect type from the comprehensive feature vector of each suspicious area, and perform deviation analysis on them respectively with the normal range of shape features, color moments and texture features corresponding to the standard image of the impregnated paper to obtain the degree of deviation of the shape features, color moments and texture features, and multiply them with the corresponding feature weights of the corresponding defect type in the impregnated paper quality assessment process and then add them to obtain the defect degree index.

[0072] The normal ranges for shape, color, and texture features of standard images of impregnated paper refer to the numerical distribution intervals or statistical characteristic ranges of these features in standard images of impregnated paper used for quality inspection. The normal ranges are determined through statistical analysis of a large number of normal impregnated paper samples. By quantifying the degree of deviation of various image features in suspicious areas from the normal ranges, the severity of defects can be accurately assessed, providing an effective basis for quality inspection of impregnated paper.

[0073] In a specific embodiment, the analysis formula of the deviation degree is: , where is the degree of deviation, is the eigenvalue, are the maximum and minimum values ​​of the normal range of the corresponding features of the standard image of the impregnated paper, respectively.

[0074] The weight setting of the shape features, color moments and texture features of each defect type in the impregnated paper quality assessment process can be set according to industry experience, or obtained through a limited number of test data. For example, historical data of different defect types appearing on the surface of the impregnated paper is first collected, and then the correlation coefficient of the influence of the shape features, color moments and texture features in different defect types on the degree of impregnated paper surface defects is calculated. The contribution of the shape features, color moments and texture features is determined using a linear regression equation analysis. Finally, after normalization, the contribution is converted into the weights of the shape features, color moments and texture features, and the sum of them is 1.

[0075] For example, in impregnated paper surface defect detection, different shape features, color moments, and texture features have varying importance for identifying different defect types. Color moments dominate the identification of stains and color unevenness defects; texture features are most critical for identifying wrinkles and scratches; and shape features play a crucial role in identifying holes.

[0076] It should be noted that the qualified state of the impregnated paper is determined by comparing the defect degree index of the corresponding defect type of each suspicious area with the defect degree index threshold of the corresponding defect type. If the defect degree index of all suspicious areas is lower than the defect degree index threshold of the corresponding defect type, the impregnated paper is determined to be qualified. Conversely, if the defect degree index of one or more suspicious areas exceeds the defect degree index threshold of the corresponding defect type, the impregnated paper is determined to be unqualified.

[0077] The present invention performs deviation analysis on the comprehensive feature vectors of each suspicious area to obtain the feature deviation degree, adds the feature deviation degree to the corresponding weight product to obtain the defect degree index, and compares it with the threshold to determine the qualified status, thereby realizing automated and standardized impregnated paper surface quality assessment and improving impregnated paper detection efficiency and detection reliability.

[0078] like Figure 4 As shown, on the other hand, the present invention provides a method for detecting surface defects of impregnated paper, comprising: S1, using a vertical lighting source to irradiate the surface of the impregnated paper, and respectively collecting reflection images of the impregnated paper surface in a visible light band and a preset near-infrared band.

[0079] S2. Based on the collected reflection image, calculate the reflection intensity difference index of the impregnated paper surface under the condition of vertical lighting source, and separate the specular reflection dominant area and the impregnated paper body reflection area based on the reflection intensity difference index.

[0080] S3. Control the low-angle lighting source to set the timing, combine the multi-frame images synchronously collected by the near-infrared camera under different lighting conditions to perform image processing, and dynamically suppress the mirror reflection dominant area.

[0081] S4. Combining the impregnated paper body reflection area and the suppressed specular reflection dominant area to obtain an impregnated paper integrated image, comparing the impregnated paper integrated image with the impregnated paper standard image, and marking all suspicious areas.

[0082] S5. Identify defect types based on the image feature data of all suspicious areas, and determine the qualified status of the impregnated paper based on the defect degree index corresponding to the defect type.

[0083] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0084] The above embodiments may be implemented in whole or in part through software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0085] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0086] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0087] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0088] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A surface defect detection system for impregnated paper, characterized in that: include: A surface reflection image acquisition module is used to illuminate the surface of the impregnated paper using a vertical illumination light source, and respectively acquire reflection images of the impregnated paper surface in the visible light band and the preset near-infrared band; An image reflection characteristic separation module is used to calculate the reflection intensity difference index of the impregnated paper surface under the condition of vertical illumination light source based on the collected reflection image, and separate the specular reflection dominant area and the impregnated paper body reflection area based on the reflection intensity difference index; The dynamic suppression module for reflective interference is used to control low-angle lighting sources to set timing. It combines multi-frame images collected synchronously by a near-infrared camera under different lighting conditions for image processing to dynamically suppress the mirror-dominated area. The module for marking suspicious image areas is used to combine the impregnated paper body reflection area and the suppressed mirror reflection dominant area to obtain an integrated image of the impregnated paper, and compare the integrated image of the impregnated paper with the standard image of the impregnated paper to mark all suspicious areas; The defect recognition and analysis module is used to identify the defect type based on the image feature data of all suspicious areas, and determine the qualified status of the impregnated paper based on the defect degree index corresponding to the defect type.

2. The impregnated paper surface defect detection system according to claim 1, characterized in that: The specific contents of the image reflection characteristic separation module are as follows: Align the pixels of the reflection images of the impregnated paper surface in the visible light band and the preset near-infrared band to obtain the reflection intensity value of each pixel point of the visible light reflection image and the near-infrared reflection image; Calculate the reflection intensity difference index of each pixel on the surface of the impregnated paper under the condition of vertical illumination according to the reflection intensity value of each pixel, and screen all pixels that meet the set specular reflection dominant area judgment conditions based on the reflection intensity difference index; The area encircled by all pixel positions that meet the set specular reflection dominant area determination conditions is taken as the specular reflection dominant area, and the remaining area in the reflected image except the specular reflection dominant area is taken as the impregnated paper body reflection area.

3. The surface defect detection system for impregnated paper according to claim 2, characterized in that: The determination condition for setting the mirror reflection dominant area is: If the reflection intensity value of a pixel point in the visible light reflection image is greater than the set visible light high reflection intensity threshold and the reflection intensity value in the near-infrared reflection image is greater than the set near-infrared high reflection intensity threshold, and the reflection intensity difference index of the pixel point is within the preset specular reflection intensity range, then the pixel point is a pixel point in the specular reflection dominant area.

4. The impregnated paper surface defect detection system according to claim 1, characterized in that: The specific steps of the dynamic suppression module for reflective interference are as follows: The low-angle illumination light source is controlled to alternately operate in the on and off states according to a set timing, and a near-infrared camera is synchronized with the timing of the low-angle illumination light source to capture the surface image each time the light source switches between states, thereby obtaining a surface image group for each state switching; Performing differential processing on the illumination-on state image and the illumination-off state image in each surface image group to obtain a differential image; Based on the pixel gray value distribution of the differential image, the local highlight interference area in the mirror reflection dominant area is identified; A binary mask image is generated with the local highlight interference area as white and the other areas of the specular reflection dominant area as black. The interference suppression operation is performed on the binary mask image to generate the specular reflection dominant area after highlight interference suppression.

5. The impregnated paper surface defect detection system according to claim 4, characterized in that: The specific identification method of the local highlight interference area in the mirror reflection dominant area is as follows: Comparing the grayscale values ​​of different pixels in the lighting-on state image and the lighting-off state image in each surface image group to obtain grayscale differences of different pixels, and constructing a differential image based on the grayscale differences of different pixels; Pixels with grayscale differences greater than the set grayscale difference threshold are screened from the mirror reflection dominant area in the differential image corresponding to each surface image group. Pixel areas are circled according to the positions of the screened pixels. The pixel areas circled by each surface image group are subjected to overlapping area analysis, and the overlapping areas are used as local highlight interference areas in the mirror reflection dominant area.

6. The impregnated paper surface defect detection system according to claim 1, characterized in that: The specific method of marking all suspicious areas is: The integrated image of the impregnated paper is grayscaled to obtain the grayscale value of each pixel in the integrated image of the impregnated paper. The grayscale value deviation between each pixel and the pixel at the corresponding position in the standard image of the impregnated paper is calculated. The pixels whose grayscale value deviation exceeds the preset grayscale threshold are marked as potential suspicious points. The entire integrated image of the impregnated paper is traversed, and all suspicious areas are formed according to the positions of all potential suspicious points.

7. The impregnated paper surface defect detection system according to claim 1, characterized in that: The specific method of identifying the defect type based on the image feature data of all suspicious areas is as follows: Obtaining images of each suspicious area from the integrated image of the impregnated paper, extracting the intensity value of each pixel in the RGB color space from the image of each suspicious area, and calculating the color moment of each suspicious area in the RGB color space; Normalize the grayscale images of each suspicious area, count the frequency of pixel pairs that meet the set distance and direction angle based on the normalized grayscale images, and use the gray level co-occurrence matrix to calculate the texture features of each suspicious area; The area, perimeter and aspect ratio are obtained according to the contour of each suspicious area, which are used to form shape features. The shape features are fused with color moments and texture features to form a comprehensive feature vector of each suspicious area. The comprehensive feature vector of each suspicious area is analyzed for similarity with the feature vector corresponding to each known defect type, and the defect type of each suspicious area is determined based on the similarity.

8. The impregnated paper surface defect detection system according to claim 7, characterized in that: The defect severity index analysis method corresponding to the defect type is as follows: The shape features, color moments, and texture features of the corresponding defect type are extracted from the comprehensive feature vector of each suspicious area. The deviations are analyzed against the normal ranges of shape features, color moments, and texture features of the standard image of the impregnated paper, respectively. The deviation degrees of the shape features, color moments, and texture features are obtained. The deviation degrees are multiplied by the corresponding feature weights of the corresponding defect type in the impregnated paper quality assessment process, and then added together to obtain the defect degree index.

9. The impregnated paper surface defect detection system according to claim 8, characterized in that: The qualified state of the impregnated paper is determined as follows: The defect severity index of the defect type corresponding to each suspicious area is compared with the defect severity index threshold of the corresponding defect type. If the defect severity index of all suspicious areas is lower than the defect severity index threshold of the corresponding defect type, the impregnated paper is judged to be qualified. Conversely, if the defect severity index of one or more suspicious areas exceeds the defect severity index threshold of the corresponding defect type, the impregnated paper is judged to be unqualified.

10. A method for detecting surface defects of impregnated paper, characterized in that: S1, using a vertical lighting source to illuminate the surface of the impregnated paper, and collecting reflection images of the impregnated paper surface in the visible light band and the preset near-infrared band respectively; S2. calculating a reflection intensity difference index of the impregnated paper surface under vertical illumination conditions based on the collected reflection image, and separating the specular reflection dominant area and the impregnated paper body reflection area based on the reflection intensity difference index; S3, control the low-angle lighting source to set the timing, combine the near-infrared camera to synchronously capture multiple frames of images under different lighting conditions to perform image processing, and dynamically suppress the mirror reflection-dominated area; S4, combining the impregnated paper body reflection area and the suppressed specular reflection dominant area to obtain an impregnated paper integrated image, comparing the impregnated paper integrated image with the impregnated paper standard image, and marking all suspicious areas; S5. Identify defect types based on the image feature data of all suspicious areas, and determine the qualified status of the impregnated paper based on the defect degree index corresponding to the defect type.

Images