Method for detecting coating defects of color-coated steel sheet based on image enhancement

Abstract

The present application belongs to the technical field of image enhancement defect detection, and specifically discloses a color-coated steel plate coating defect detection method based on image enhancement, which comprises the following steps: acquiring an initial image of the color-coated steel plate coating, performing image enhancement processing on the initial image of the color-coated steel plate coating, determining a color-coated steel plate coating surface defect judgment compensation factor based on color-coated steel plate coating environment data, acquiring macro color-coated steel plate coating surface defect data, collecting micro color-coated steel plate coating surface defect data, obtaining a coating surface defect judgment signal, constructing a color-coated steel plate coating thickness-corrosion coupling detection model, performing ultrasonic detection on the color-coated steel plate coating, and outputting a color-coated steel plate coating defect grade. The present application realizes high-precision and multi-dimensional detection of color-coated steel plate coating defects, effectively improves the accuracy and reliability of the detection results, and solves the problems of limited detection precision and wide influence of environmental factors in traditional color-coated steel plate coating defect detection, and the application range is narrow.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of image enhancement defect detection, and specifically relates to a color-coated steel plate coating defect detection method based on image enhancement. BACKGROUND

[0002] Color-coated steel plates are widely used in many industries such as construction, home appliances, and automobiles. In the construction field, color-coated steel plates are used for roofing and walling, and if the coating has defects, it is easy to cause corrosion and rust, affecting the safety and aesthetics of the building structure. In the home appliance industry, color-coated steel plates are used to manufacture the housings of refrigerators, air conditioners, etc., and coating defects can reduce the appearance quality and durability of the products. In order to ensure product quality, industries are increasingly strict in controlling the quality of color-coated steel plate coatings, and accurate and effective defect detection methods are needed. With the development of science and technology, new breakthroughs have been made in non-destructive testing technologies such as ultrasonic, eddy current, and optical imaging. These technologies can detect coating defects without damaging the color-coated steel plate, and the precision and sensitivity are gradually improving, providing more technical means and possibilities for color-coated steel plate coating defect detection, prompting researchers to continuously explore how to better apply these advanced technologies to color-coated steel plate coating detection. Machine learning, deep learning algorithms, and image processing technologies in artificial intelligence are rapidly developing. If these technologies can be combined with traditional detection methods, automatic identification, classification, and quantitative analysis of color-coated steel plate coating defects can be achieved.

[0003] The current mainstream technology for detecting defects in the coating of color-coated steel plates includes:

[0004] Manual visual inspection: the characteristics are simple and direct, and the coating surface condition can be directly observed; the disadvantages are low detection efficiency, strong subjectivity, and difficulty in detecting small or internal defects.

[0005] Magnetic thickness gauge detection: the characteristics are fast measurement of coating thickness and simple operation; the disadvantages are that only thickness-related information can be detected, and other defects such as surface cracks and bubbles cannot be detected.

[0006] Traditional color-coated steel plate coating defect detection has limited precision and is greatly affected by environmental factors, with a narrow application range, making it difficult to ensure consistency and accuracy of detection. SUMMARY

[0007] In view of the shortcomings of the prior art, the present application provides a color-coated steel plate coating defect detection method based on image enhancement, which realizes high-precision, multi-dimensional detection of color-coated steel plate coating defects by enhancing the image, combining laser scanning, ultrasonic detection, and environmental factor compensation, and effectively improves the accuracy and reliability of the detection results.

[0008] To achieve the above purpose, the present application realizes the following technical scheme: a color-coated steel plate coating defect detection method based on image enhancement, comprising the following steps:

[0009] Obtain the initial image of the color-coated steel sheet coating, and perform image enhancement processing on the initial image of the color-coated steel sheet coating to obtain the enhanced image of the color-coated steel sheet coating.

[0010] Obtain environmental data on the application of color-coated steel sheet coatings and determine the compensation factor for surface defects in the color-coated steel sheet coatings.

[0011] Based on enhanced images of the color-coated steel sheet coating, macroscopic surface defect data of the color-coated steel sheet coating are obtained;

[0012] Laser scanning detection is performed on the coating of color-coated steel sheet to collect microscopic surface defect data of the coating. Combined with macroscopic surface defect data of the coating and the surface defect judgment compensation factor, the surface defect judgment signal of the coating is obtained.

[0013] Construct a coupled detection model for coating thickness and corrosion of color-coated steel sheets, and output a defect judgment signal for coupled coating thickness and corrosion.

[0014] Ultrasonic testing is performed on the coating of the pre-painted steel sheet to obtain the ultrasonic testing characteristics of the coating, including the ultrasonic velocity, ultrasonic amplitude, and ultrasonic transit time of the coating. The ultrasonic defect judgment signal of the coating is determined. Combining the coating surface defect judgment signal and the coating thickness-corrosion coupled defect judgment signal, the defect level of the pre-painted steel sheet coating is output.

[0015] Preferably, the process of obtaining the enhanced image of the color-coated steel sheet coating is as follows:

[0016] The initial image of the coating on the pre-painted steel sheet is acquired. Homomorphic filtering is then applied to this image to eliminate uneven illumination and enhance high-frequency details. After homomorphic filtering, the image is restored to the spatial domain, and its coordinate system is output. The pixel value at that location is used to perform histogram equalization on the initial image of the color-coated steel sheet coating, including:

[0017] Calculate the grayscale histogram:

[0018] ;

[0019] In the formula, Let k be the number of pixels at gray level k, N be the total number of pixels, and L be the number of gray levels. Let k be the probability of gray level k appearing.

[0020] Cumulative distribution function mapping:

[0021] ;

[0022] In the formula, o is the loop variable. Let be the probability of grayscale level 0. The cumulative distribution function reflects the proportion of pixels with gray values ​​less than or equal to k in the initial image of the color-coated steel sheet coating.

[0023] Gray-level mapping:

[0024] ;

[0025] In the formula, It is the new gray value corresponding to gray level k after gray level mapping, and round means rounding to the nearest integer.

[0026] Color image processing:

[0027] Convert to HSV or Lab color space, and only equalize the luminance channel to avoid color distortion;

[0028] Perform median filtering:

[0029] ;

[0030] In the formula, W is the neighborhood window. Represents the coordinates of pixels in the image of the color-coated steel sheet coating. After median filtering, the coordinates are The new grayscale value of the pixel, Represented by coordinates The current grayscale values ​​of each pixel within the neighborhood window centered on the center pixel, where i and j are relative to the center pixel. The offsets in the x and y directions, with median representing the midpoint;

[0031] Based on the median filtering process, for coordinates... The new grayscale value of the pixel is used to perform Laplacian sharpening to enhance edge and texture details:

[0032] ;

[0033] ;

[0034] In the formula, It is the Laplacian operator for image functions The result of the calculation Representing pixels The grayscale value of the adjacent pixel to the right in the horizontal direction. Representing pixels The grayscale value of the adjacent pixel to the left in the horizontal direction. Representing pixels The grayscale value of adjacent pixels in the vertical direction upwards. Representing pixels The grayscale value of adjacent pixels in the vertical direction downwards. The coordinates in the image are The output of the median filtering operation , After Laplacian sharpening, the coordinates are... The new grayscale value of the pixel, This represents the sharpening intensity factor;

[0035] Multi-scale Retinex enhances the color fidelity and detail of color-coated steel sheet images:

[0036] ;

[0037] In the formula, Coordinates are The One color channel, Coordinates are The m-th scale Gaussian kernel, Let M be the weight coefficient for the m-th scale, and M be the number of scales. This indicates that after multi-scale Retinex processing, the th Each color channel is located at coordinates The output value at the specified location, where m is the scale loop variable;

[0038] Output enhanced images of the color-coated steel sheet coating.

[0039] Preferably, the process of performing homomorphic filtering on the initial image of the color-coated steel sheet coating includes:

[0040] Logarithmic transformation converts the multiplicative model into an additive one:

[0041] ;

[0042] In the formula, The initial image of the coating of the color-coated steel sheet in coordinates Pixel value at that location, The initial image of the coating of the color-coated steel sheet in coordinates The amount of light at that location, The initial image of the coating of the color-coated steel sheet in coordinates The reflection component at that location;

[0043] Fourier transform to frequency domain:

[0044] ;

[0045] In the formula, Yes The frequency domain representation after Fourier transform. These are frequency domain coordinates. Represents the Fourier transform operation;

[0046] Frequency domain filtering: Design a Gaussian high-pass filter.

[0047] ;

[0048] In the formula, The designed Gaussian high-pass filter in the frequency domain The transfer function value at that point, The low-frequency gain coefficients are stored in the database. These are the high-frequency gain coefficients stored in the database. Represents frequency points in the frequency domain Distance to the frequency center The cutoff frequency is given by , and e is the natural constant.

[0049] Inverse Transform and Exponential Recovery:

[0050] ;

[0051] In the formula, The image of the color-coated steel sheet coating restored to the spatial domain after homomorphic filtering is in coordinates. Pixel value at that location, This is the inverse Fourier transform operation.

[0052] Preferably, the process of determining the compensation factor for surface defects in the color-coated steel sheet coating is as follows:

[0053] Obtain environmental data on the application of the pre-painted steel sheet coating, including the ambient temperature of the pre-painted steel sheet coating application environment. The ambient light intensity of the color-coated steel sheet coating ;

[0054] Based on the dimensionless processing of the environmental data of the pre-painted steel sheet coating, the environmental influencing factors of the pre-painted steel sheet coating were analyzed and obtained:

[0055] ;

[0056] In the formula, represents the environmental impact factor of the coating on color-coated steel sheets, where e is a natural constant.

[0057] Determine the compensation factor for surface defects in the color-coated steel sheet coating: Obtain the pre-stored mapping set of environmental impact factors and surface defect compensation factors for color-coated steel sheet coating in the database, and determine the matching compensation factor for surface defects in the color-coated steel sheet coating based on the current environmental impact factors.

[0058] Preferably, the process of obtaining macroscopic surface defect data of the color-coated steel sheet coating is as follows:

[0059] Based on enhanced images of the pre-coated steel sheet coating, macroscopic surface defect data of the pre-coated steel sheet coating is obtained, including the percentage of peeling area on the pre-coated steel sheet coating surface. Number of scratches on the surface of the color-coated steel sheet coating Percentage of color difference area on the surface of color-coated steel sheet .

[0060] Preferably, the process of obtaining the coating surface defect determination signal is as follows:

[0061] Laser scanning inspection was performed on the coating of the pre-painted steel sheet to collect microscopic surface defect data, including the number of bubbles on the coating surface. Pore ​​density on the surface of color-coated steel sheet coating ;

[0062] Based on microscopic surface defect data of color-coated steel sheet coating, macroscopic surface defect data of color-coated steel sheet coating, and compensation factor for surface defect judgment of color-coated steel sheet coating, the surface defect judgment signal of coating is obtained.

[0063] Preferably, after dimensionless processing of the microscopic and macroscopic color-coated steel sheet coating surface defect data, the method for obtaining the coating surface defect judgment signal is as follows:

[0064] ;

[0065] In the formula, e is the natural constant. This is a signal for determining defects on the coating surface. For macroscopic color-coated steel sheet coating signals, For microscopic signals of the coating on color-coated steel sheets, This is a compensation factor for judging surface defects in the coating of color-coated steel sheets. For setting Weighting factors For setting Weighting factors.

[0066] Preferably, the process of outputting the coating thickness-corrosion coupling defect determination signal is as follows:

[0067] Collect coating thickness characteristic data and coating corrosion characteristic data of pre-coated steel sheets, including:

[0068] The characteristic data of the coating thickness of pre-painted steel sheets include: the peak-to-valley difference in the coating thickness of pre-painted steel sheets. Average coating thickness of color-coated steel sheet The database stores the standard thickness of the coating on color-coated steel sheets. ;

[0069] Corrosion characteristic data of pre-painted steel sheet coatings include: corrosion area of ​​pre-painted steel sheet coatings. Maximum corrosion depth of color-coated steel sheet coating ;

[0070] Construct a coupled detection model for coating thickness and corrosion of color-coated steel sheets, and output a defect judgment signal for coupled coating thickness and corrosion.

[0071] After dimensionless processing of the coating thickness characteristic data and corrosion characteristic data of pre-coated steel sheets, the coating thickness-corrosion coupled detection model of pre-coated steel sheets is expressed as follows:

[0072] ;

[0073] In the formula, e is the natural constant. This is a signal for determining coating thickness-corrosion coupling defects. This is a signal for determining coating thickness defects. This is a signal for determining coating corrosion defects. For setting Weighting factors For setting Weighting factors.

[0074] Preferably, the process of outputting the coating defect level of the color-coated steel sheet is as follows:

[0075] Based on the ultrasonic testing characteristics of the coating of color-coated steel sheet, the ultrasonic analysis factors of the coating are analyzed and obtained.

[0076] Obtain the coating ultrasonic analysis factor-coating ultrasonic defect judgment signal mapping set stored in the database, and determine the matching coating ultrasonic defect judgment signal based on the current coating ultrasonic analysis factor;

[0077] Obtain the weighting factors of the coating ultrasonic defect judgment signal, the coating surface defect judgment signal, and the coating thickness-corrosion coupling defect judgment signal stored in the database.

[0078] Based on the ultrasonic defect determination signal of the coating, the surface defect determination signal of the coating, and the coating thickness-corrosion coupled defect determination signal, the weighted summation is used to obtain the coupling characteristic value of the coating defect of the color-coated steel sheet.

[0079] Obtain the threshold range of coating defect coupling characteristics of color-coated steel sheets stored in the database;

[0080] Compare the coupling characteristic value of the coating defect of the color-coated steel sheet with the threshold range of the coupling characteristic of the coating defect of the color-coated steel sheet;

[0081] If the coupling characteristic value of the coating defect of the pre-painted steel sheet is higher than the maximum value in the range of coupling characteristic thresholds for the coating defect of the pre-painted steel sheet, the defect level of the coating of the pre-painted steel sheet is level three, and a non-compliance warning is issued.

[0082] If the coupling characteristic value of the coating defect of the pre-painted steel sheet falls within the threshold range of the coupling characteristic value of the coating defect of the pre-painted steel sheet, then the defect level of the coating of the pre-painted steel sheet is level two, and an adverse warning is issued.

[0083] If the coupling characteristic value of the coating defect of the color-coated steel sheet is lower than the minimum value in the range of coupling characteristic thresholds for the coating defect of the color-coated steel sheet, then the ultrasonic defect judgment signal, the surface defect judgment signal, and the coating thickness-corrosion coupling defect judgment signal are compared with the ultrasonic defect judgment threshold, the surface defect judgment threshold, and the coating thickness-corrosion coupling defect judgment threshold stored in the database, respectively.

[0084] If the ultrasonic defect detection signal of the coating is higher than the ultrasonic defect detection threshold of the coating, it is marked as exceeding the limit;

[0085] If the coating surface defect judgment signal is higher than the coating surface defect judgment threshold, it is marked as exceeding the limit;

[0086] If the coating thickness-corrosion coupling defect judgment signal is higher than the coating thickness-corrosion coupling defect judgment threshold, it is marked as exceeding the limit;

[0087] When one or more exceedance marks are present, the coating defect level of the color-coated steel sheet is level two, and an adverse warning is issued;

[0088] If there is no over-limit mark, the coating defect level of the color-coated steel sheet is level one, and a qualified indication is issued.

[0089] Preferably, the method for obtaining the ultrasonic analysis factor of the coating is as follows:

[0090] ;

[0091] In the formula, e is the natural constant. For ultrasonic analysis factors of coatings, For the ultrasonic velocity of the coating, For the ultrasonic amplitude of the coating, The ultrasonic transit time of the coating. The standard coating ultrasonic velocity is stored in the database. The standard coating ultrasonic amplitude is stored in the database. The standard coating ultrasonic transit time is stored in the database, where, for , , , , , Dimensionless processing is performed.

[0092] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0093] This invention enhances the initial image, highlighting key features and making macroscopic defects more clearly identifiable. This helps to more accurately acquire surface defect data of macroscopic color-coated steel sheets, reducing missed detections or misjudgments due to poor image quality. Combining laser scanning to detect microscopic defects and ultrasonic testing to detect internal defect features supplements defect information at the microscopic and internal levels. Laser scanning can capture minute surface defects, while ultrasonic testing can detect internal defects such as delamination and bubbles within the coating. This multi-dimensional detection significantly improves the ability to identify various types of defects.

[0094] This invention obtains environmental data on the application of color-coated steel sheet coatings and determines a compensation factor to make the test results more consistent with actual conditions. Different application environments have different effects on the coating, which may lead to defects of varying degrees and types. By considering environmental factors, the actual quality of the coating can be assessed more accurately, avoiding misjudgments caused by environmental factors. Attached Figure Description

[0095] Figure 1 This is a schematic flowchart of the method of the present invention;

[0096] Figure 2 A flowchart illustrating the steps involved in image enhancement processing of the initial image of the coating on a color-coated steel sheet.

[0097] Figure 3 This is a detailed flowchart of the method of the present invention. Detailed Implementation

[0098] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0099] like Figure 1 , Figure 3 As shown, the image enhancement-based method for detecting coating defects in color-coated steel sheets includes the following steps:

[0100] Obtain the initial image of the color-coated steel sheet coating, and perform image enhancement processing on the initial image of the color-coated steel sheet coating to obtain the enhanced image of the color-coated steel sheet coating.

[0101] Obtain environmental data on the application of color-coated steel sheet coatings and determine the compensation factor for surface defects in the color-coated steel sheet coatings.

[0102] Based on enhanced images of the color-coated steel sheet coating, macroscopic surface defect data of the color-coated steel sheet coating are obtained;

[0103] Laser scanning detection is performed on the coating of color-coated steel sheet to collect microscopic surface defect data of the coating. Combined with macroscopic surface defect data of the coating and the surface defect judgment compensation factor, the surface defect judgment signal of the coating is obtained.

[0104] Construct a coupled detection model for coating thickness and corrosion of color-coated steel sheets, and output a defect judgment signal for coupled coating thickness and corrosion.

[0105] Ultrasonic testing is performed on the coating of the pre-painted steel sheet to obtain the ultrasonic testing characteristics of the coating, including the ultrasonic velocity, ultrasonic amplitude, and ultrasonic transit time of the coating. The ultrasonic defect judgment signal of the coating is determined. Combining the coating surface defect judgment signal and the coating thickness-corrosion coupled defect judgment signal, the defect level of the pre-painted steel sheet coating is output.

[0106] like Figure 2 As shown, the process of obtaining the enhanced image of the color-coated steel sheet coating is as follows:

[0107] The initial image of the coating on the pre-painted steel sheet is acquired. Homomorphic filtering is then applied to this image to eliminate uneven illumination and enhance high-frequency details. After homomorphic filtering, the image is restored to the spatial domain, and its coordinate system is output. The pixel value at that location is used to perform histogram equalization on the initial image of the color-coated steel sheet coating, including:

[0108] Calculate the grayscale histogram:

[0109] ;

[0110] In the formula, Let k be the number of pixels at gray level k, N be the total number of pixels, and L be the number of gray levels. Let k be the probability of gray level k appearing.

[0111] Cumulative distribution function mapping:

[0112] ;

[0113] In the formula, o is the loop variable. Let be the probability of grayscale level 0. The cumulative distribution function reflects the proportion of pixels with gray values ​​less than or equal to k in the initial image of the color-coated steel sheet coating.

[0114] Gray-level mapping:

[0115] ;

[0116] In the formula, It is the new gray value corresponding to gray level k after gray level mapping, and round means rounding to the nearest integer.

[0117] Color image processing:

[0118] Convert to HSV or Lab color space, and only equalize the luminance channel to avoid color distortion;

[0119] Perform median filtering:

[0120] ;

[0121] In the formula, W is the neighborhood window. Represents the coordinates of pixels in the image of the color-coated steel sheet coating. After median filtering, the coordinates are The new grayscale value of the pixel, Represented by coordinates The current grayscale values ​​of each pixel within the neighborhood window centered on the center pixel, where i and j are relative to the center pixel. The offsets in the x and y directions, with median representing the midpoint;

[0122] Based on the median filtering process, for coordinates... The new grayscale value of the pixel is used to perform Laplacian sharpening to enhance edge and texture details:

[0123] ;

[0124] ;

[0125] In the formula, It is the Laplacian operator for image functions The result of the calculation Representing pixels The grayscale value of the adjacent pixel to the right in the horizontal direction. Representing pixels The grayscale value of the adjacent pixel to the left in the horizontal direction. Representing pixels The grayscale value of adjacent pixels in the vertical direction upwards. Representing pixels The grayscale value of adjacent pixels in the vertical direction downwards. The coordinates in the image are The output of the median filtering operation , After Laplacian sharpening, the coordinates are... The new grayscale value of the pixel, This represents the sharpening intensity factor;

[0126] Multi-scale Retinex enhances the color fidelity and detail of color-coated steel sheet images:

[0127] ;

[0128] In the formula, Coordinates are The One color channel, Coordinates are The m-th scale Gaussian kernel, Let M be the weight coefficient for the m-th scale, and M be the number of scales. This indicates that after multi-scale Retinex processing, the th Each color channel is located at coordinates The output value at the specified location, where m is the scale loop variable;

[0129] Output enhanced images of the color-coated steel sheet coating.

[0130] The process of performing homomorphic filtering on the initial image of the coating of the pre-painted steel sheet includes:

[0131] Logarithmic transformation converts the multiplicative model into an additive one:

[0132] ;

[0133] In the formula, The initial image of the coating of the color-coated steel sheet in coordinates Pixel value at that location, The initial image of the coating of the color-coated steel sheet in coordinates The amount of light at that location, The initial image of the coating of the color-coated steel sheet in coordinates The reflection component at that location;

[0134] Fourier transform to frequency domain:

[0135] ;

[0136] In the formula, Yes The frequency domain representation after Fourier transform. These are frequency domain coordinates. Represents the Fourier transform operation;

[0137] Frequency domain filtering: Design a Gaussian high-pass filter.

[0138] ;

[0139] In the formula, The designed Gaussian high-pass filter in the frequency domain The transfer function value at that point, The low-frequency gain coefficients are stored in the database. These are the high-frequency gain coefficients stored in the database. Represents frequency points in the frequency domain Distance to the frequency center The cutoff frequency is given by , and e is the natural constant.

[0140] Inverse Transform and Exponential Recovery:

[0141] ;

[0142] In the formula, The image of the color-coated steel sheet coating restored to the spatial domain after homomorphic filtering is in coordinates. Pixel value at that location, This is the inverse Fourier transform operation.

[0143] Homomorphic filtering decomposes an image into illumination and reflection components through operations such as logarithmic transform and Fourier transform, effectively eliminating the problem of uneven illumination. It can avoid misjudgment or missed judgment of defects caused by differences in illumination; for example, even tiny defects in darker areas can be clearly presented.

[0144] Histogram equalization makes the image grayscale distribution more uniform and improves global contrast; Laplacian sharpening enhances edge and texture details; multi-scale Retinex enhances color fidelity and detail. These processes make defect features more prominent, facilitating subsequent detection and recognition, such as making minor scratches and holes more clearly visible.

[0145] Median filtering can effectively remove discrete noise such as salt-and-pepper noise while preserving image edges. Reducing noise interference can prevent noise from being mistaken for defects, improve detection accuracy, and lay the foundation for accurate subsequent analysis of defect features.

[0146] By comprehensively utilizing multiple image enhancement techniques, images are optimized from multiple dimensions such as lighting, contrast, detail, and noise. This enables a more comprehensive highlighting of defect features, accurate acquisition of macroscopic and microscopic defect data, and precise determination of defect type, location, size, and other information.

[0147] When processing color images, converting to HSV or Lab color space only applies equalization to the luminance channel to avoid color distortion. In actual inspection, ensuring undistorted colors helps to more accurately reflect the surface condition of the coating, preventing color changes from interfering with defect detection and making the inspection results more consistent with reality.

[0148] The process of determining the compensation factor for surface defects in color-coated steel sheets is as follows:

[0149] Obtain environmental data on the application of the pre-painted steel sheet coating, including the ambient temperature of the pre-painted steel sheet coating application environment. The ambient light intensity of the color-coated steel sheet coating ;

[0150] Based on the dimensionless processing of the environmental data of the pre-painted steel sheet coating, the environmental influencing factors of the pre-painted steel sheet coating were analyzed and obtained:

[0151] ;

[0152] In the formula, represents the environmental impact factor of the coating on color-coated steel sheets, where e is a natural constant.

[0153] Determine the compensation factor for surface defects in the color-coated steel sheet coating: Obtain the pre-stored mapping set of environmental impact factors and surface defect compensation factors for color-coated steel sheet coating in the database, and determine the matching compensation factor for surface defects in the color-coated steel sheet coating based on the current environmental impact factors.

[0154] By acquiring ambient temperature and light intensity and calculating environmental impact factors, the effect of environmental factors on coatings can be quantitatively assessed. Coating defects manifest and develop differently under different environments; for example, high temperatures and strong light can accelerate coating aging and discoloration. Considering these factors can avoid misjudgments of defects caused by environmental interference, making defect assessments more consistent with reality.

[0155] Compensation factors are determined from a pre-stored mapping set based on environmental impact factors. This method can specifically adjust the coating defect judgment for different environments, accurately compensating for the impact of environmental factors and making the defect judgment results more accurate. Color-coated steel sheets have a wide range of applications and significant environmental variations. This method can dynamically adjust the defect judgment criteria based on actual environmental data, enabling the detection method to adapt to various complex environments, effectively functioning in both high-temperature and high-humidity industrial environments and intensely lit outdoor environments.

[0156] The process of obtaining macroscopic surface defect data of color-coated steel sheet coating is as follows:

[0157] Based on enhanced images of the pre-coated steel sheet coating, macroscopic surface defect data of the pre-coated steel sheet coating is obtained, including the percentage of peeling area on the pre-coated steel sheet coating surface. Number of scratches on the surface of the color-coated steel sheet coating Percentage of color difference area on the surface of color-coated steel sheet .

[0158] The process of obtaining the coating surface defect determination signal is as follows:

[0159] Laser scanning inspection was performed on the coating of the pre-painted steel sheet to collect microscopic surface defect data, including the number of bubbles on the coating surface. Pore ​​density on the surface of color-coated steel sheet coating ;

[0160] Based on microscopic surface defect data of color-coated steel sheet coating, macroscopic surface defect data of color-coated steel sheet coating, and compensation factor for surface defect judgment of color-coated steel sheet coating, the surface defect judgment signal of coating is obtained.

[0161] After dimensionless processing of the microscopic and macroscopic surface defect data of color-coated steel sheet coatings, the method for obtaining the coating surface defect judgment signal is as follows:

[0162] ;

[0163] In the formula, e is the natural constant. This is a signal for determining defects on the coating surface. For macroscopic color-coated steel sheet coating signals, For microscopic signals of the coating on color-coated steel sheets, This is a compensation factor for judging surface defects in the coating of color-coated steel sheets. For setting Weighting factors For setting Weighting factors.

[0164] The above formula considers both macroscopic factors such as the percentage of peeling area, number of scratches, and percentage of color difference area on the surface of the color-coated steel sheet, and microscopic factors such as the number of bubbles and pore density. By integrating macroscopic and microscopic defect data, comprehensive information on coating surface defects can be obtained, avoiding omissions caused by focusing only on defects at a certain scale, and enabling a more complete assessment of coating quality.

[0165] By incorporating a compensation factor for surface defects in pre-painted steel sheets, the influence of the usage environment on the coating is integrated into the defect assessment system. The development and manifestation of coating defects differ under different environments; considering environmental factors makes the assessment more consistent with reality, enhancing the comprehensiveness and reliability of the inspection.

[0166] The process of outputting the coating thickness-corrosion coupling defect determination signal is as follows:

[0167] Collect coating thickness characteristic data and coating corrosion characteristic data of pre-coated steel sheets, including:

[0168] The characteristic data of the coating thickness of pre-painted steel sheets include: the peak-to-valley difference in the coating thickness of pre-painted steel sheets. Average coating thickness of color-coated steel sheet The database stores the standard thickness of the coating on color-coated steel sheets. ;

[0169] Corrosion characteristic data of pre-painted steel sheet coatings include: corrosion area of ​​pre-painted steel sheet coatings. Maximum corrosion depth of color-coated steel sheet coating ;

[0170] Construct a coupled detection model for coating thickness and corrosion of color-coated steel sheets, and output a defect judgment signal for coupled coating thickness and corrosion.

[0171] After dimensionless processing of the coating thickness characteristic data and corrosion characteristic data of pre-coated steel sheets, the coating thickness-corrosion coupled detection model of pre-coated steel sheets is expressed as follows:

[0172] ;

[0173] In the formula, e is the natural constant. This is a signal for determining coating thickness-corrosion coupling defects. This is a signal for determining coating thickness defects. This is a signal for determining coating corrosion defects. For setting Weighting factors For setting Weighting factors.

[0174] This model collects coating thickness characteristic data (such as thickness peak-to-valley difference, average thickness, and standard thickness) and corrosion characteristic data (such as corrosion area and maximum corrosion depth). By comprehensively considering the uniformity of coating thickness and corrosion status, two factors crucial to the performance of color-coated steel sheets, it can comprehensively evaluate the actual quality of the coating and avoid the limitations of single-factor evaluation.

[0175] Coating thickness and corrosion are interrelated; uneven thickness may make certain areas more susceptible to corrosion, while corrosion further affects coating thickness. By using a coupled detection model, the interaction and degree of influence between the two can be analyzed, allowing for a more accurate assessment of the coating's remaining life and potential risks.

[0176] The model calculates coating thickness defect judgment signals and coating corrosion defect judgment signals separately using specific formulas, and finally synthesizes them to obtain a coating thickness-corrosion coupled defect judgment signal. This transforms complex thickness and corrosion conditions into specific quantitative values, intuitively reflecting the severity of defects.

[0177] The process of outputting the coating defect level of color-coated steel sheet is as follows:

[0178] Based on the ultrasonic testing characteristics of the coating of color-coated steel sheet, the ultrasonic analysis factors of the coating are analyzed and obtained.

[0179] Obtain the coating ultrasonic analysis factor-coating ultrasonic defect judgment signal mapping set stored in the database, and determine the matching coating ultrasonic defect judgment signal based on the current coating ultrasonic analysis factor;

[0180] Obtain the weighting factors of the coating ultrasonic defect judgment signal, the coating surface defect judgment signal, and the coating thickness-corrosion coupling defect judgment signal stored in the database.

[0181] Based on the ultrasonic defect determination signal of the coating, the surface defect determination signal of the coating, and the coating thickness-corrosion coupled defect determination signal, the weighted summation is used to obtain the coupling characteristic value of the coating defect of the color-coated steel sheet.

[0182] Obtain the threshold range of coating defect coupling characteristics of color-coated steel sheets stored in the database;

[0183] Compare the coupling characteristic value of the coating defect of the color-coated steel sheet with the threshold range of the coupling characteristic of the coating defect of the color-coated steel sheet;

[0184] If the coupling characteristic value of the coating defect of the pre-painted steel sheet is higher than the maximum value in the range of coupling characteristic thresholds for the coating defect of the pre-painted steel sheet, the defect level of the coating of the pre-painted steel sheet is level three, and a non-compliance warning is issued.

[0185] If the coupling characteristic value of the coating defect of the pre-painted steel sheet falls within the threshold range of the coupling characteristic value of the coating defect of the pre-painted steel sheet, then the defect level of the coating of the pre-painted steel sheet is level two, and an adverse warning is issued.

[0186] If the coupling characteristic value of the coating defect of the color-coated steel sheet is lower than the minimum value in the range of coupling characteristic thresholds for the coating defect of the color-coated steel sheet, then the ultrasonic defect judgment signal, the surface defect judgment signal, and the coating thickness-corrosion coupling defect judgment signal are compared with the ultrasonic defect judgment threshold, the surface defect judgment threshold, and the coating thickness-corrosion coupling defect judgment threshold stored in the database, respectively.

[0187] If the ultrasonic defect detection signal of the coating is higher than the ultrasonic defect detection threshold of the coating, it is marked as exceeding the limit;

[0188] If the coating surface defect judgment signal is higher than the coating surface defect judgment threshold, it is marked as exceeding the limit;

[0189] If the coating thickness-corrosion coupling defect judgment signal is higher than the coating thickness-corrosion coupling defect judgment threshold, it is marked as exceeding the limit;

[0190] When one or more exceedance marks are present, the coating defect level of the color-coated steel sheet is level two, and an adverse warning is issued;

[0191] If there is no over-limit mark, the coating defect level of the color-coated steel sheet is level one, and a qualified indication is issued.

[0192] The method for obtaining the ultrasonic analysis factors of the coating is as follows:

[0193] ;

[0194] In the formula, e is the natural constant. For ultrasonic analysis factors of coatings, For the ultrasonic velocity of the coating, For the ultrasonic amplitude of the coating, The ultrasonic transit time of the coating. The standard coating ultrasonic velocity is stored in the database. The standard coating ultrasonic amplitude is stored in the database. The standard coating ultrasonic transit time is stored in the database, where, for , , , , , Dimensionless processing is performed.

[0195] This system integrates ultrasonic defect detection signals, surface defect detection signals, and thickness-corrosion coupled defect detection signals. By acquiring information from multiple dimensions, including ultrasonic characteristics, macroscopic and microscopic surface defects, and thickness and corrosion status, it avoids the limitations of single detection methods, comprehensively reflects coating quality, and improves the reliability of defect detection.

[0196] Each judgment signal is quantified by calculating ultrasonic analysis factors for the coating and then compared with thresholds stored in a database. This quantitative comparison and grading method makes the determination of defect levels more objective and accurate, reducing the subjectivity and uncertainty of human judgment.

[0197] The system clearly categorizes defects into three levels: Level 1 (Qualified), Level 2 (Defective), and Level 3 (Unqualified), and issues corresponding warnings. This allows production and quality management personnel to quickly understand the coating quality of the color-coated steel sheets, facilitating timely implementation of appropriate measures, such as handling unqualified products or adjusting production processes.

Claims

1. A method for detecting coating defects in color-coated steel sheets based on image enhancement, characterized in that, Includes the following steps: Obtain the initial image of the color-coated steel sheet coating, and perform image enhancement processing on the initial image of the color-coated steel sheet coating to obtain the enhanced image of the color-coated steel sheet coating. Obtain environmental data on the application of color-coated steel sheet coatings and determine the compensation factor for surface defects in the color-coated steel sheet coatings. Based on enhanced images of the color-coated steel sheet coating, macroscopic surface defect data of the color-coated steel sheet coating are obtained; Laser scanning detection is performed on the coating of color-coated steel sheet to collect microscopic surface defect data of the coating. Combined with macroscopic surface defect data of the coating and the surface defect judgment compensation factor, the surface defect judgment signal of the coating is obtained. A coupled detection model for coating thickness and corrosion of pre-coated steel sheets is constructed, and a defect judgment signal for coating thickness-corrosion coupling is output. The process is as follows: Collect coating thickness characteristic data and coating corrosion characteristic data of pre-coated steel sheets, including: The characteristic data of the coating thickness of pre-painted steel sheets include: the peak-to-valley difference in the coating thickness of pre-painted steel sheets. Average coating thickness of color-coated steel sheet The database stores the standard thickness of the coating on color-coated steel sheets. ; Corrosion characteristic data of pre-painted steel sheet coatings include: corrosion area of ​​pre-painted steel sheet coatings. Maximum corrosion depth of color-coated steel sheet coating ; Construct a coupled detection model for coating thickness and corrosion of color-coated steel sheets, and output a defect judgment signal for coupled coating thickness and corrosion. After dimensionless processing of the coating thickness characteristic data and corrosion characteristic data of pre-coated steel sheets, the coating thickness-corrosion coupled detection model of pre-coated steel sheets is expressed as follows: ； In the formula, e is the natural constant. This is a signal for determining coating thickness-corrosion coupling defects. This is a signal for determining coating thickness defects. This is a signal for determining coating corrosion defects. For setting Weighting factors For setting Weighting factors; Ultrasonic testing is performed on the coating of the pre-painted steel sheet to obtain the ultrasonic testing characteristics of the coating, including the ultrasonic velocity, ultrasonic amplitude, and ultrasonic transit time of the coating. The ultrasonic defect judgment signal of the coating is determined. Combining the coating surface defect judgment signal and the coating thickness-corrosion coupled defect judgment signal, the defect level of the pre-painted steel sheet coating is output.

2. The method for detecting coating defects in color-coated steel sheets based on image enhancement according to claim 1, characterized in that, The process of obtaining the enhanced image of the color-coated steel sheet coating is as follows: The initial image of the coating on the pre-painted steel sheet is acquired. Homomorphic filtering is then applied to this image to eliminate uneven illumination and enhance high-frequency details. After homomorphic filtering, the image is restored to the spatial domain, and its coordinate system is output. The pixel value at that location is used to perform histogram equalization on the initial image of the color-coated steel sheet coating, including: Calculate the grayscale histogram: ； In the formula, Let k be the number of pixels at gray level k, N be the total number of pixels, and L be the number of gray levels. Let k be the probability of gray level k appearing. Cumulative distribution function mapping: ； In the formula, o is the loop variable. Let be the probability of grayscale level 0. The cumulative distribution function reflects the proportion of pixels with gray values ​​less than or equal to k in the initial image of the color-coated steel sheet coating. Gray-level mapping: ； In the formula, It is the new gray value corresponding to gray level k after gray level mapping, and round means rounding to the nearest integer. Color image processing: Convert to HSV or Lab color space, and only equalize the luminance channel to avoid color distortion; Perform median filtering: ； In the formula, W is the neighborhood window. Represents the coordinates of pixels in the image of the color-coated steel sheet coating. After median filtering, the coordinates are The new grayscale value of the pixel, Represented by coordinates The current grayscale values ​​of each pixel within the neighborhood window centered on the center pixel, where i and j are relative to the center pixel. The offsets in the x and y directions, with median representing the midpoint; Based on the median filtering process, for coordinates... The new grayscale value of the pixel is used to perform Laplacian sharpening to enhance edge and texture details: ； ； In the formula, It is the Laplacian operator for image functions The result of the calculation Representing pixels The grayscale value of the adjacent pixel to the right in the horizontal direction. Representing pixels The grayscale value of the adjacent pixel to the left in the horizontal direction. Representing pixels The grayscale value of adjacent pixels in the vertical direction upwards. Representing pixels The grayscale value of adjacent pixels in the vertical direction downwards. The coordinates in the image are The output of the median filtering operation , After Laplacian sharpening, the coordinates are... The new grayscale value of the pixel, This represents the sharpening intensity factor; Multi-scale Retinex enhances the color fidelity and detail of color-coated steel sheet images: ； In the formula, Coordinates are The One color channel, Coordinates are The m-th scale Gaussian kernel, Let M be the weight coefficient for the m-th scale, and M be the number of scales. This indicates that after multi-scale Retinex processing, the th Each color channel is located at coordinates The output value at the specified location, where m is the scale loop variable; Output enhanced images of the color-coated steel sheet coating.

3. The image enhancement-based method for detecting coating defects in color-coated steel sheets according to claim 2, characterized in that, The process of performing homomorphic filtering on the initial image of the color-coated steel sheet coating includes: Logarithmic transformation converts the multiplicative model into an additive one: ； In the formula, The initial image of the coating of the color-coated steel sheet in coordinates Pixel value at that location, The initial image of the coating of the color-coated steel sheet in coordinates The amount of light at that location, The initial image of the coating of the color-coated steel sheet in coordinates The reflection component at that location; Fourier transform to frequency domain: ； In the formula, Yes The frequency domain representation after Fourier transform. These are frequency domain coordinates. Represents the Fourier transform operation; Frequency domain filtering: Design a Gaussian high-pass filter. ； In the formula, The designed Gaussian high-pass filter in the frequency domain The transfer function value at that point, The low-frequency gain coefficients are stored in the database. These are the high-frequency gain coefficients stored in the database. Represents frequency points in the frequency domain Distance to the frequency center The cutoff frequency is given by , and e is the natural constant. Inverse Transform and Exponential Recovery: ； In the formula, The image of the color-coated steel sheet coating restored to the spatial domain after homomorphic filtering is in coordinates. Pixel value at that location, This is the inverse Fourier transform operation.

4. The method for detecting coating defects in color-coated steel sheets based on image enhancement according to claim 1, characterized in that, The process for determining the compensation factor for surface defects in the color-coated steel sheet coating is as follows: Obtain environmental data on the application of the pre-painted steel sheet coating, including the ambient temperature of the pre-painted steel sheet coating application environment. The ambient light intensity of the color-coated steel sheet coating ; Based on the dimensionless processing of the environmental data of the pre-painted steel sheet coating, the environmental influencing factors of the pre-painted steel sheet coating were analyzed and obtained: ； In the formula, represents the environmental impact factor of the coating on color-coated steel sheets, where e is a natural constant. Determine the compensation factor for surface defects in the color-coated steel sheet coating: Obtain the pre-stored mapping set of environmental impact factors and surface defect compensation factors for color-coated steel sheet coating in the database, and determine the matching compensation factor for surface defects in the color-coated steel sheet coating based on the current environmental impact factors.

5. The method for detecting coating defects in color-coated steel sheets based on image enhancement according to claim 1, characterized in that, The process of obtaining macroscopic surface defect data of color-coated steel sheet coating is as follows: Based on enhanced images of the pre-coated steel sheet coating, macroscopic surface defect data of the pre-coated steel sheet coating is obtained, including the percentage of peeling area on the pre-coated steel sheet coating surface. Number of scratches on the surface of the color-coated steel sheet coating Percentage of color difference area on the surface of color-coated steel sheet .

6. The method for detecting coating defects in color-coated steel sheets based on image enhancement according to claim 5, characterized in that, The process of obtaining the coating surface defect determination signal is as follows: Laser scanning inspection was performed on the coating of the pre-painted steel sheet to collect microscopic surface defect data, including the number of bubbles on the coating surface. Pore ​​density on the surface of color-coated steel sheet coating ; Based on microscopic surface defect data of color-coated steel sheet coating, macroscopic surface defect data of color-coated steel sheet coating, and compensation factor for surface defect judgment of color-coated steel sheet coating, the surface defect judgment signal of coating is obtained.

7. The image enhancement-based method for detecting coating defects in color-coated steel sheets according to claim 6, characterized in that, After dimensionless processing of the microscopic and macroscopic surface defect data of color-coated steel sheet coatings, the method for obtaining the coating surface defect judgment signal is as follows: ； In the formula, e is the natural constant. This is a signal for determining defects on the coating surface. For macroscopic color-coated steel sheet coating signals, For microscopic signals of the coating on color-coated steel sheets, This is a compensation factor for judging surface defects in the coating of color-coated steel sheets. For setting Weighting factors For setting Weighting factors.

8. The method for detecting coating defects in color-coated steel sheets based on image enhancement according to claim 1, characterized in that, The process of outputting the defect level of the color-coated steel sheet coating is as follows: Based on the ultrasonic testing characteristics of the coating of color-coated steel sheet, the ultrasonic analysis factors of the coating are analyzed and obtained. Obtain the coating ultrasonic analysis factor-coating ultrasonic defect judgment signal mapping set stored in the database, and determine the matching coating ultrasonic defect judgment signal based on the current coating ultrasonic analysis factor; Obtain the weighting factors of the coating ultrasonic defect judgment signal, the coating surface defect judgment signal, and the coating thickness-corrosion coupling defect judgment signal stored in the database. Based on the ultrasonic defect determination signal of the coating, the surface defect determination signal of the coating, and the coating thickness-corrosion coupled defect determination signal, the weighted summation is used to obtain the coupling characteristic value of the coating defect of the color-coated steel sheet. Obtain the threshold range of coating defect coupling characteristics of color-coated steel sheets stored in the database; Compare the coupling characteristic value of the coating defect of the color-coated steel sheet with the threshold range of the coupling characteristic of the coating defect of the color-coated steel sheet; If the coupling characteristic value of the coating defect of the pre-painted steel sheet is higher than the maximum value in the range of coupling characteristic thresholds for the coating defect of the pre-painted steel sheet, the defect level of the coating of the pre-painted steel sheet is level three, and a non-compliance warning is issued. If the coupling characteristic value of the coating defect of the pre-painted steel sheet falls within the threshold range of the coupling characteristic value of the coating defect of the pre-painted steel sheet, then the defect level of the coating of the pre-painted steel sheet is level two, and an adverse warning is issued. If the coupling characteristic value of the coating defect of the color-coated steel sheet is lower than the minimum value in the range of coupling characteristic thresholds for the coating defect of the color-coated steel sheet, then the ultrasonic defect judgment signal, the surface defect judgment signal, and the coating thickness-corrosion coupling defect judgment signal are compared with the ultrasonic defect judgment threshold, the surface defect judgment threshold, and the coating thickness-corrosion coupling defect judgment threshold stored in the database, respectively. If the ultrasonic defect detection signal of the coating is higher than the ultrasonic defect detection threshold of the coating, it is marked as exceeding the limit; If the coating surface defect judgment signal is higher than the coating surface defect judgment threshold, it is marked as exceeding the limit; If the coating thickness-corrosion coupling defect judgment signal is higher than the coating thickness-corrosion coupling defect judgment threshold, it is marked as exceeding the limit; When one or more exceedance marks are present, the coating defect level of the color-coated steel sheet is level two, and an adverse warning is issued; If there is no over-limit mark, the coating defect level of the color-coated steel sheet is level one, and a qualified indication is issued.

9. The method for detecting coating defects in color-coated steel sheets based on image enhancement according to claim 8, characterized in that, The method for obtaining the ultrasonic analysis factor of the coating is as follows: ； In the formula, e is the natural constant. For ultrasonic analysis factors of coatings, For the ultrasonic velocity of the coating, For the ultrasonic amplitude of the coating, The ultrasonic transit time of the coating. The standard coating ultrasonic velocity is stored in the database. The standard coating ultrasonic amplitude is stored in the database. The standard coating ultrasonic transit time is stored in the database, where, for , , , , , Dimensionless processing is performed.

Images