Glass fiber cloth production quality detection method and system

Abstract

The invention relates to the technical field of image processing, and discloses a glass fiber cloth production quality detection method and system.The method comprises the steps that a multi-scale texture decomposition algorithm and a dynamic light and shadow consistency analysis technology are used for separating texture signals of a fiber interweaving area, and hole candidate areas are extracted; dynamic change characteristics of the candidate areas are calculated in combination with a time sequence analysis model, and the holes are accurately judged through a reliability scoring model. In the prior art, the detection precision is insufficient due to the complex texture and light and shadow interference in the production of the glass fiber cloth, and especially in the high-speed production environment of the glass fiber cloth, the hole defect and the normal texture on the glass fiber cloth cannot be effectively distinguished. According to the method, the multi-scale texture decomposition algorithm and the reliability scoring model are used, so that the detection accuracy of the hole defects is remarkably improved, the problems of false detection and missing detection are avoided, and the quality control efficiency and the detection reliability of glass fiber cloth production are improved.

Description

Technical Field

[0001] The invention belongs to the technical field of image processing, and in particular relates to a method and system for detecting the production quality of glass fiber cloth. Background Art

[0002] At present, in the quality inspection of glass fiber cloth production, the detection of hole defects mainly relies on traditional static image analysis methods or single spectrum imaging technology. These methods have many shortcomings. For example, due to the complex surface texture of glass fiber cloth and the significant changes in light and shadow, traditional methods cannot effectively distinguish hole defects from normal fiber interweaving structures, especially when the hole size is small and the edge features are blurred, it is easy to misjudge or miss. In addition, in a high-speed production environment, the dynamic blur and reflected light interference caused by the movement of the cloth further reduce the detection accuracy. The existing detection technology also lacks dynamic analysis methods for multi-frame image sequences, and it is difficult to use information in the time dimension to make reliability judgments on hole defects. Therefore, there is an urgent need for a method that can combine multi-spectral imaging technology, multi-scale texture decomposition, light and shadow consistency analysis, and dynamic time series modeling to accurately identify hole defects and generate reliable quality assessment results under complex texture interference and high-speed production conditions, so as to improve the accuracy and real-time performance of glass fiber cloth production quality inspection. Summary of the invention

[0003] In view of the above-mentioned technical deficiencies, the purpose of the present invention is to propose a glass fiber cloth production quality inspection method, aiming to solve the technical problem in the prior art that the detection accuracy is insufficient due to the complex texture and light and shadow interference in the production of glass fiber cloth, especially in the high-speed production environment of glass fiber cloth, and it is impossible to effectively distinguish the hole defects on the glass fiber cloth from the normal texture.

[0004] In order to solve the above technical problems, the present invention adopts the following technical solutions: The present invention provides a method for detecting the production quality of glass fiber cloth.

[0005] The glass fiber cloth production quality detection method comprises:

[0006] Step S10: collecting multi-channel image data of the surface of the glass fiber cloth by multi-spectral imaging method;

[0007] The multi-channel image data includes:

[0008] Visible light image L v , used to obtain fiber surface texture characteristics;

[0009] Infrared image L ir , used to obtain thermal radiation characteristics;

[0010] UV image L uv , used to obtain the microscopic abnormal characteristics of materials;

[0011] Perform geometric calibration and registration on the multi-channel image data to generate a fused image L f (x, y), where x is the horizontal coordinate of the image pixel and y is the vertical coordinate of the image pixel;

[0012] Step S20: fusion image L f (x,y) performs optical property correction, including:

[0013] From the fused image L f (x, y) to obtain the reflected light intensity R(x, y), define the area where the reflected light intensity R(x, y) is greater than the preset reflected light intensity threshold as a high-reflection area, and perform brightness compensation on the high-reflection area to obtain the first optimized fusion image L c (x,y);

[0014] For the first optimized fusion image L c (x, y) is processed by grayscale equalization to obtain the second optimized fusion image L n (x,y);

[0015] Step S30: Using a multi-scale texture decomposition algorithm to decompose the second optimized fused image L n (x, y) is processed to decompose the texture signal of the fiber interweaving area into:

[0016] Low-frequency texture feature T1(x,y), used to represent local uniformity;

[0017] High frequency abnormal characteristics T h (x,y), used to represent the characteristics of the hole;

[0018] For high frequency abnormal features T h (x, y) applies the hole template matching algorithm to calculate the hole template matching value R(x, y), retains the pixel coordinate points whose hole template matching value R(x, y) is greater than the preset candidate threshold, and then generates the candidate area; calculates the light and shadow consistency score S for the extracted candidate area g (x, y), and remove the light and shadow consistency score S g The pixel points (x, y) below the preset light and shadow consistency score threshold are used to generate the optimized hole candidate area;

[0019] Step S40: Obtaining image sequence data L of cloth movement t (x, y), to optimize the hole candidate area, use the time series analysis model combined with the image sequence data of cloth movement L t (x,y), calculate the time variation variance of the optimized hole candidate area Combining time-varying variance The reliability score P of the hole is calculated using the reliability score model. defect (x,y), the formula is:

[0020]

[0021] Among them, μ is the time characteristic mean of the optimized hole candidate area, and exp is the exponential function;

[0022] Step S50: According to the reliability score P defect (x, y), determine whether the optimized hole candidate area is a hole defect and output the glass fiber cloth production quality inspection report, including the hole location, hole size and hole reliability score.

[0023] Preferably, in step S20, the first optimized fused image L c The calculation formula for (x,y) is:

[0024] L c (x,y)=L f (x,y)-αR(x,y)

[0025] Wherein, α is the light compensation coefficient of the high-reflective area, which is determined by the optical properties of the glass fiber cloth.

[0026] 3. A glass fiber cloth production quality detection method as claimed in claim 1, characterized in that in step S30, the low-frequency texture feature T l (x,y) and high-frequency abnormal features T h (x,y) satisfies the formula L n (x,y)=T l (x,y)+T h (x,y).

[0027] Preferably, in step S30, the low-frequency texture feature T l (x,y) and high-frequency abnormal features T h (x,y) satisfies the formula L n (x,y)=T l (x,y)+T h (x,y).

[0028] Preferably, in step S30, the calculation formula of the hole template matching value R(x, y) is:

[0029]

[0030] Among them, W(i,j) is the hole template weight coefficient matrix, m and n are the preset horizontal and vertical pixel coordinate sizes of the hole template, i is the horizontal offset of the pixel relative to the center point (x, y) in the filter window, in pixels, and j is the vertical offset of the pixel relative to the center point (x, y) in the filter window, in pixels.

[0031] Preferably, in step S30, the light and shadow consistency score S g The calculation formula for (x,y) is:

[0032]

[0033] Where i is the horizontal offset of the pixel relative to the center point (x, y) in the filter window, in pixels, j is the vertical offset of the pixel relative to the center point (x, y) in the filter window, in pixels, L n (x+i,y+j) is the grayscale value of the pixel coordinate (x+i,y+j), is the window mean gray value, N w is the number of pixels in the window.

[0034] Preferably, in step S20, the second optimized fused image L n The calculation formula for (x,y) is:

[0035]

[0036] Among them, L min and L max are the preset minimum grayscale value and the preset maximum grayscale value, L scale is the preset target grayscale range.

[0037] Preferably, in step S40, the time variation variance of the hole candidate region is optimized The calculation formula is:

[0038]

[0039] Where t is the time frame, T is the total number of frames in the time series of the image sequence data of the cloth movement obtained, is the time mean of the image sequence data of cloth motion.

[0040] The present invention also provides a glass fiber cloth production quality detection system comprising:

[0041] A multispectral imaging module, used for collecting multi-channel image data of the glass fiber cloth surface by multispectral imaging method;

[0042] The multi-channel image data includes:

[0043] Visible light image Lv , used to obtain fiber surface texture characteristics;

[0044] Infrared image L ir , used to obtain thermal radiation characteristics;

[0045] UV image L uv , used to obtain the microscopic abnormal characteristics of materials;

[0046] Perform geometric calibration and registration on the multi-channel image data to generate a fused image L f (x, y), where x is the horizontal coordinate of the image pixel and y is the vertical coordinate of the image pixel;

[0047] Optical property correction module, used to correct the fused image L f (x,y) performs optical property correction, including:

[0048] From the fused image L f (x, y) to obtain the reflected light intensity R(x, y), define the area where the reflected light intensity R(x, y) is greater than the preset reflected light intensity threshold as a high-reflection area, and perform brightness compensation on the high-reflection area to obtain the first optimized fusion image L c (x,y);

[0049] For the first optimized fusion image L c (x, y) is processed by grayscale equalization to obtain the second optimized fusion image L n (x,y);

[0050] The multi-scale texture analysis module is used to use a multi-scale texture decomposition algorithm to analyze the second optimized fusion image L n (x, y) is processed to decompose the texture signal of the fiber interweaving area into:

[0051] Low frequency texture feature T l (x,y), used to represent local uniformity;

[0052] High frequency abnormal characteristics T h (x,y), used to represent the characteristics of the hole;

[0053] For high frequency abnormal features T h (x, y) applies the hole template matching algorithm to calculate the hole template matching value R(x, y), retains the pixel coordinate points whose hole template matching value R(x, y) is greater than the preset candidate threshold, and then generates the candidate area; calculates the light and shadow consistency score S for the extracted candidate area g (x, y), and remove the light and shadow consistency score S g The pixel points (x, y) below the preset light and shadow consistency score threshold are used to generate the optimized hole candidate area;

[0054] Time series analysis module, used to obtain image sequence data L of cloth movement t (x, y), to optimize the hole candidate area, use the time series analysis model combined with the image sequence data of cloth movement L t (x,y), calculate the time variation variance of the optimized hole candidate area Combining time-varying variance The reliability score P of the hole is calculated using the reliability score model. defect (x,y0, the formula is:

[0055]

[0056] Among them, μ is the time characteristic mean of the optimized hole candidate area, and exp is the exponential function;

[0057] Report generation module, used to score P according to reliability defect (x, y), determine whether the optimized hole candidate area is a hole defect and output the glass fiber cloth production quality inspection report, including the hole location, hole size and hole reliability score.

[0058] The present invention also provides a glass fiber cloth production quality detection device, comprising a memory, a processor, and a glass fiber cloth production quality detection program stored in the memory and executable on the processor, wherein the glass fiber cloth production quality detection program implements the glass fiber cloth production quality detection method when executed by the processor.

[0059] The present invention also provides a computer program product, including a glass fiber cloth production quality detection program, and when the glass fiber cloth production quality detection program is executed by a processor, the glass fiber cloth production quality detection method is implemented.

[0060] The beneficial effects of the present invention are as follows: compared with the prior art, which has insufficient detection accuracy due to complex textures and light and shadow interference in the production of glass fiber cloth, especially in the high-speed production environment of glass fiber cloth, and cannot effectively distinguish between hole defects and normal textures on the glass fiber cloth, the present application uses a multi-scale texture decomposition algorithm and a reliability scoring model to significantly improve the detection accuracy of hole defects, thereby avoiding the problems of false detection and missed detection, and improving the quality control efficiency and detection reliability of glass fiber cloth production. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 paying creative work.

[0062] Figure 1 The figure is a flow chart of a first embodiment of a method for detecting the production quality of glass fiber cloth according to the present invention.

[0063] Figure 2 The present invention is a schematic diagram of equipment for a method for detecting the production quality of glass fiber cloth. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] Embodiment 1: Figure 1 , which is a schematic flow chart of a first embodiment of a method for detecting the quality of glass fiber cloth production according to the present invention, and provides a first embodiment of a method for detecting the quality of glass fiber cloth production according to the present invention.

[0066] In a first embodiment, the glass fiber cloth production quality detection method includes:

[0067] Step S10: collecting multi-channel image data of the surface of the glass fiber cloth by multi-spectral imaging method;

[0068] The multi-channel image data includes:

[0069] Visible light image L v , used to obtain fiber surface texture characteristics;

[0070] Infrared image L ir , used to obtain thermal radiation characteristics;

[0071] UV image L uv , used to obtain the microscopic abnormal characteristics of materials;

[0072] Perform geometric calibration and registration on the multi-channel image data to generate a fused image L f (x, y), where x is the horizontal coordinate of the image pixel and y is the vertical coordinate of the image pixel;

[0073] It should be noted that the multispectral imaging method obtains the visible light image L v , infrared image L ir and the UV image L uv , which can capture the texture characteristics, thermal radiation characteristics and microscopic abnormal characteristics of the glass fiber cloth surface in different spectral ranges. Compared with single spectral imaging, the technical advantage of multi-spectral imaging is that it can integrate the data of each spectral channel to provide more comprehensive and diverse feature information.

[0074] It can be understood that the fused image L f (x, y) is generated by geometric calibration and registration of multi-channel image data. During the calibration process, the spatial distortion of the image is effectively corrected, and the registration ensures the alignment of each spectral channel data in the same coordinate system. This fused image can provide a unified feature basis for subsequent texture decomposition and defect detection.

[0075] It should be understood that in practical applications, data from different spectral channels have specific functions: visible light image L v Provides a direct representation of surface texture, infrared image L ir It can reveal the material anomaly caused by the thermal conductivity of glass fiber cloth, while the ultraviolet image L uv It can amplify subtle abnormal reactions on the surface of the material. The diversity of this information effectively improves the sensitivity and accuracy of defect detection.

[0076] For example, when inspecting a batch of glass fiber cloth, the visible light image L v The local unevenness of the fiber interweaving texture is shown; infrared image L ir This area shows abnormal thermal radiation characteristics, indicating possible material damage; while the ultraviolet image L uv It further confirmed that there was a weak abnormal reflection characteristic at the edge of the hole in this area. By fusing the three spectral data into L f (x, y), the system can clearly identify the boundary and position of the hole, avoiding missed detection due to insufficient single spectrum

[0077] Step S20: fusion image L f (x,y) performs optical property correction, including:

[0078] From the fused image L f (x, y) to obtain the reflected light intensity R(x, y), and the reflected light intensity R(x, y) is greater than the preset reflected light intensity threshold T r The area is defined as a high-reflection area, and brightness compensation is performed on the high-reflection area to obtain the first optimized fusion image L c (x,y);

[0079] For the first optimized fusion image L c (x, y) is processed by grayscale equalization to obtain the second optimized fusion image L n (x,y);

[0080] It should be noted that for the fused image L f The purpose of optical property correction for (x, y) is to reduce the interference of high-reflective areas on subsequent texture decomposition and hole identification. High-value areas of reflected light intensity R(x, y) are prone to local overexposure, which masks the true texture features of the fiber surface. Therefore, for reflected light intensity R(x, y)>T r Brightness compensation for highly reflective areas can effectively balance the brightness distribution of the image and reduce the loss of image details.

[0081] It can be understood that the first optimized fusion image L c (x, y) is obtained by compensating for the brightness difference of the high-reflective area. In the brightness compensation process, the compensation coefficient α determines the degree of attenuation of the reflected light intensity R(x, y). Its value needs to be set according to the reflective characteristics of the material surface. In this way, L c (x,y) retains more fiber texture details and provides a more balanced image basis for subsequent equalization processing.

[0082] It should be understood that the main function of grayscale equalization is to enhance the contrast of the image and avoid the loss of important defect features caused by excessively high or low local brightness. c (x, y) is normalized to generate the second optimized fusion image L n (x, y) can ensure that texture features are evenly distributed throughout the image, providing high-quality data input for subsequent texture decomposition and feature extraction.

[0083] For example, in a certain glass fiber cloth area, the reflected light intensity R(x,y) exceeds the threshold value T r = 200. In the uncorrected fused image, the high reflective area causes the hole edge to be blurred. After brightness compensation, the generated L c (x,y) Adjust the grayscale value of the high reflective area to below T r , thus restoring the texture features of the region, and then c After grayscale equalization processing is performed on (x, y), the generated L n The brightness of (x,y) is uniform across the entire image, ensuring that the texture at the edge of the hole is clearly visible.

[0084] Step S30: Using a multi-scale texture decomposition algorithm to decompose the second optimized fused image L n(x, y) is processed to decompose the texture signal of the fiber interweaving area into:

[0085] Low frequency texture feature T l (x,y), used to represent local uniformity;

[0086] High frequency abnormal characteristics T h (x,y), used to represent the characteristics of the hole;

[0087] For high frequency abnormal features T h (x, y) applies the hole template matching algorithm to calculate the hole template matching value R(x, y), retains the pixel coordinate points whose hole template matching value R(x, y) is greater than the preset candidate threshold, and then generates the candidate area; calculates the light and shadow consistency score S for the extracted candidate area g (x, y), and remove the light and shadow consistency score S g The pixel points (x, y) below the preset light and shadow consistency score threshold are used to generate the optimized hole candidate area;

[0088] It should be noted that the texture signal of the fiber interlacing area is decomposed into low-frequency texture features T l (x,y) and high-frequency abnormal features T h (x, y), which helps to separate the uniform characteristics of the fiber surface from the abnormal characteristics of the holes, and the low-frequency characteristics T l (x, y) represents the uniform brightness distribution of the local area, mainly retaining the overall texture morphology of the glass fiber cloth surface, and the high-frequency feature T h (x, y) highlights the edge and detail characteristics of the hole area, providing an accurate feature basis for subsequent defect extraction.

[0089] It can be understood that the hole template matching algorithm is used to find the high-frequency abnormal features T h The template matching value R(x,y) is a quantitative indicator of the similarity between the high-frequency signal and the hole feature template. By retaining the pixels whose R(x,y) is greater than the preset candidate threshold and generating candidate areas, the interference of irrelevant textures on defect detection can be effectively reduced, providing preliminary hole feature segmentation for subsequent steps.

[0090] It should be understood that the light and shadow consistency score S g (x, y) is an indicator calculated based on the light and shadow changes in the candidate area, which is used to further optimize the boundary of the hole area. Pixels with low light and shadow consistency scores are usually caused by noise or pseudo defects. By eliminating these low-score points, the optimized hole candidate area can more accurately reflect the shape and position of the real hole defects.

[0091] For example, the high-frequency abnormal feature Th (x, y) shows the location of potential holes. During the template matching process, the hole template matching value R(x, y) exceeds the threshold T h = 0.8 pixels are retained as candidate regions. However, in the calculation of light and shadow consistency, the light and shadow consistency score S of some regions is g (x,y) is below the threshold T g =0.5, indicating that these areas may be affected by light and shadow noise. By eliminating these low-score points, the optimized hole candidate area finally generated accurately identifies the boundary and range of the hole.

[0092] Step S40: Obtaining image sequence data L of cloth movement t (x, y), to optimize the hole candidate area, use the time series analysis model combined with the image sequence data of cloth movement L t (x,y), calculate the time variation variance of the optimized hole candidate area Combining time-varying variance The reliability score P of the hole is calculated using the reliability score model. defect (x,y), the formula is:

[0093]

[0094] Among them, μ is the time characteristic mean of the optimized hole candidate area, and exp is the exponential function;

[0095] It should be noted that the image sequence data L of the cloth movement t (x,y), combined with the time series analysis model, the time variation variance of the optimized hole candidate area can be calculated The temporal variation variance reflects the dynamic stability of the candidate region in the image sequence. For the hole region, the temporal variation is small and shows a consistent feature; while the pseudo-defect region is usually affected by light or noise, resulting in a large temporal variation variance. Through this analysis, real holes and pseudo-defect regions can be effectively distinguished.

[0096] It is understandable that the reliability score P defect (x,y) is an indicator calculated based on the change of the time column, which is used to evaluate the possibility that a candidate area is a hole defect. In the scoring model, the mean μ and variance of the time-varying feature The impact on the scoring results is relatively large. Due to the stability of the dynamic characteristics of the real hole area, its reliability score P defect (x,y) approaches 1, while the pseudo-defect area has a score close to 0 due to its large variance.

[0097] For example, assuming that the mean of the temporal features of a candidate region is μ = 120, and its temporal variance is The pixel value of this area in a certain frame is L t (x,y)=122. According to the formula, the reliability score of this area is:

[0098]

[0099] A lower score indicates that the area may be a pseudo defect and will ultimately be rejected.

[0100] Step S50: According to the reliability score P defect (x, y), determine whether the optimized hole candidate area is a hole defect and output the glass fiber cloth production quality inspection report, including the hole location, hole size and hole reliability score.

[0101] It should be noted that the glass fiber cloth production quality inspection report contains key information on hole defects, including the hole location, which is accurately marked by image coordinates (x, y); the hole size, which is estimated based on the number of image points in the candidate area; and the hole reliability score, which indicates the detection credibility of each hole defect, which is convenient for quality analysis and grading.

[0102] In addition, the glass fiber cloth production quality detection system provided by the present invention adopts a glass fiber cloth production quality detection method in the above embodiment, which can solve a technical problem of glass fiber cloth production quality detection. Compared with the prior art, the beneficial effects of the glass fiber cloth production quality detection system provided by the present invention are the same as the beneficial effects of the glass fiber cloth production quality detection method provided by the above embodiment, and other technical features of the glass fiber cloth production quality detection system are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0103] The present invention provides a glass fiber cloth production quality inspection equipment, please refer to Figure 2, a glass fiber cloth production quality inspection device includes: at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a glass fiber cloth production quality inspection method in the above-mentioned embodiment 1. A glass fiber cloth production quality inspection device in the embodiment of the present invention may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. A glass fiber cloth production quality inspection device is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention. A glass fiber cloth production quality inspection device may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM) 1004. In RAM 1004, various programs and data required for the operation of a glass fiber cloth production quality inspection device are also stored. The processing device 1001, ROM 1002, and RAM 1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 can allow a glass fiber cloth production quality inspection device to communicate wirelessly or wired with other devices to exchange data. Although a glass fiber cloth production quality inspection device with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have alternatively.

[0104] The present invention also provides a computer program product, including a computer program, wherein when the computer program is executed by a processor, the steps of the above-mentioned method for detecting the production quality of glass fiber cloth are implemented. The computer program product provided by the present invention can solve the technical problem of detecting the production quality of glass fiber cloth. Compared with the prior art, the beneficial effects of the computer program product provided by the present invention are the same as the beneficial effects of the method for detecting the production quality of glass fiber cloth provided by the above-mentioned embodiment, and will not be described in detail here.

[0105] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through a communication device, or installed from the storage device 1003, or installed from the ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present invention are executed.

[0106] It should be understood that the various parts disclosed in the present invention can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0107] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for detecting the production quality of glass fiber cloth, characterized in that: Methods include: Step S10: collecting multi-channel image data of the surface of the glass fiber cloth by multi-spectral imaging method; The multi-channel image data includes: Visible light image L v , used to obtain fiber surface texture characteristics; Infrared image L ir , used to obtain thermal radiation characteristics; UV image L uv , used to obtain the microscopic abnormal characteristics of materials; Perform geometric calibration and registration on the multi-channel image data to generate a fused image L f (x, y), where x is the horizontal coordinate of the image pixel and y is the vertical coordinate of the image pixel; Step S20: fusion image L f (x,y) performs optical property correction, including: From the fused image L f (x, y) to obtain the reflected light intensity R(x, y), define the area where the reflected light intensity R(x, y) is greater than the preset reflected light intensity threshold as a high-reflection area, and perform brightness compensation on the high-reflection area to obtain the first optimized fusion image L c (x,y); For the first optimized fusion image L c (x, y) is processed by grayscale equalization to obtain the second optimized fusion image L n (x,y); Step S30: Using a multi-scale texture decomposition algorithm to decompose the second optimized fused image L n (x, y) is processed to decompose the texture signal of the fiber interweaving area into: Low frequency texture feature T l (x,y), used to represent local uniformity; High frequency abnormal characteristics T h (x,y), used to represent the characteristics of the hole; For high frequency abnormal features T h (x, y) applies the hole template matching algorithm to calculate the hole template matching value R(x, y), retains the pixel coordinate points whose hole template matching value R(x, y) is greater than the preset candidate threshold, and then generates the candidate area; calculates the light and shadow consistency score S for the extracted candidate area g (x, y), and remove the light and shadow consistency score S g The pixel points (x, y) below the preset light and shadow consistency score threshold are used to generate the optimized hole candidate area; Step S40: Obtaining image sequence data L of cloth movement t (x, y), to optimize the hole candidate area, use the time series analysis model combined with the image sequence data of cloth movement L t (x,y), calculate the time variation variance of the optimized hole candidate area Combining time-varying variance The reliability score P of the hole is calculated using the reliability score model. defect (x,y), the formula is: Among them, μ is the time characteristic mean of the optimized hole candidate area, and exp is the exponential function; Step S50: According to the reliability score P defect (x, y), determine whether the optimized hole candidate area is a hole defect and output the glass fiber cloth production quality inspection report, including the hole location, hole size and hole reliability score.

2. A glass fiber cloth production quality inspection method as claimed in claim 1, characterized in that: In step S20, the first optimized fusion image L c The calculation formula for (x,y) is: L c (x,y)=L f (x,y)-αR(x,y) Wherein, α is the light compensation coefficient of the high-reflective area, which is determined by the optical properties of the glass fiber cloth.

3. A glass fiber cloth production quality inspection method as claimed in claim 1, characterized in that: In step S30, the low-frequency texture feature T l (x,y) and high-frequency abnormal features T h (x,y) satisfies the formula L n (x,y)=T l (x,y)+T h (x,y).

4. A glass fiber cloth production quality inspection method as claimed in claim 1, characterized in that: In step S30, the calculation formula of the hole template matching value R(x,y) is: Among them, W(i,j) is the hole template weight coefficient matrix, m and n are the preset horizontal and vertical pixel coordinate sizes of the hole template, i is the horizontal offset of the pixel relative to the center point (x, y) in the filter window, in pixels, and j is the vertical offset of the pixel relative to the center point (x, y) in the filter window, in pixels.

5. A method for detecting the production quality of glass fiber cloth according to claim 1, characterized in that: In step S30, the light and shadow consistency score S g The calculation formula for (x,y) is: Where i is the horizontal offset of the pixel relative to the center point (x, y) in the filter window, in pixels, j is the vertical offset of the pixel relative to the center point (x, y) in the filter window, in pixels, L n (x+i,y+j) is the grayscale value of the pixel coordinate (x+i,y+j), is the window mean gray value, N w is the number of pixels in the window.

6. A method for detecting the production quality of glass fiber cloth according to claim 1, characterized in that: In step S20, the second optimized fused image L n The calculation formula for (x,y) is: Among them, L min and L max are the preset minimum grayscale value and the preset maximum grayscale value, L scale is the preset target grayscale range.

7. A glass fiber cloth production quality inspection method as claimed in claim 1, characterized in that: In step S40, the time variation variance of the hole candidate region is optimized The calculation formula is: Where t is the time frame, T is the total number of frames in the time series of the image sequence data of the cloth movement obtained, is the time mean of the image sequence data of cloth motion.

8. A glass fiber cloth production quality inspection system, characterized in that: The glass fiber cloth production quality detection system comprises: A multispectral imaging module, used for collecting multi-channel image data of the glass fiber cloth surface by multispectral imaging method; The multi-channel image data includes: Visible light image L v , used to obtain fiber surface texture characteristics; Infrared image L ir , used to obtain thermal radiation characteristics; UV image L uv , used to obtain the microscopic abnormal characteristics of materials; Perform geometric calibration and registration on the multi-channel image data to generate a fused image L f (x, y), where x is the horizontal coordinate of the image pixel and y is the vertical coordinate of the image pixel; Optical property correction module, used to correct the fused image L f (x,y) performs optical property correction, including: From the fused image L f (x, y) to obtain the reflected light intensity R(x, y), define the area where the reflected light intensity R(x, y) is greater than the preset reflected light intensity threshold as a high-reflection area, and perform brightness compensation on the high-reflection area to obtain the first optimized fusion image L c (x,y); For the first optimized fusion image L c (x, y) is processed by grayscale equalization to obtain the second optimized fusion image L n (x,y); The multi-scale texture analysis module is used to use a multi-scale texture decomposition algorithm to analyze the second optimized fusion image L n (x, y) is processed to decompose the texture signal of the fiber interweaving area into: Low frequency texture feature T l (x,y), used to represent local uniformity; High frequency abnormal characteristics T h (x,y), used to represent the characteristics of the hole; For high frequency abnormal features T h (x, y) applies the hole template matching algorithm to calculate the hole template matching value R(x, y), retains the pixel coordinate points whose hole template matching value R(x, y) is greater than the preset candidate threshold, and then generates the candidate area; calculates the light and shadow consistency score S for the extracted candidate area g (x, y), and remove the light and shadow consistency score S g The pixel points (x, y) below the preset light and shadow consistency score threshold are used to generate the optimized hole candidate area; Time series analysis module, used to obtain image sequence data L of cloth movement t (x, y), to optimize the hole candidate area, use the time series analysis model combined with the image sequence data of cloth movement L t (x,y), calculate the time variation variance of the optimized hole candidate area Combining time-varying variance The reliability score P of the hole is calculated using the reliability score model. defect (x,y), the formula is: Among them, μ is the time characteristic mean of the optimized hole candidate area, and exp is the exponential function; Report generation module, used to score P according to reliability defect (x, y), determine whether the optimized hole candidate area is a hole defect and output the glass fiber cloth production quality inspection report, including the hole location, hole size and hole reliability score.

9. A glass fiber cloth production quality inspection equipment, characterized in that: The glass fiber cloth production quality inspection device includes: a memory, a processor, and a glass fiber cloth production quality inspection program stored in the memory and executable on the processor. When the glass fiber cloth production quality inspection program is executed by the processor, the glass fiber cloth production quality inspection method described in any one of Claims 1 to 7 is implemented.

10. A computer program product, characterized in that The computer program product includes a glass fiber cloth production quality detection program, and when the glass fiber cloth production quality detection program is executed by a processor, the glass fiber cloth production quality detection method described in any one of Claims 1 to 7 is implemented.

Images