A printing doctor blade wear state identification method and system

Abstract

This invention belongs to the field of image processing technology, specifically relating to a method and system for identifying the wear state of a printing squeegee. The method includes: performing edge detection on a grayscale image to obtain a set of pixel edge points; performing coordinate correction on the pixel edge point set based on virtual interpolation points and curve fitting along the gradient direction to obtain a set of contour coordinates; iteratively fitting the contour coordinate set to obtain an ideal baseline; obtaining a net loss residual sequence based on the vertical deviation distance of each point in the contour coordinate set from the ideal baseline and the ink tolerance threshold; calculating the micro-texture entropy, macro-deformation potential energy, and average wear depth based on the net loss residual sequence, and constructing a squeegee wear severity index accordingly; and generating a mechanical adjustment control signal based on the squeegee wear severity index. This invention can effectively shield the interference of ink residue on the detection baseline, achieving precise control for different wear stages.

Description

Technical Field

[0001] This invention relates to the field of image processing technology. More specifically, this invention relates to a method and system for identifying the wear condition of a printing squeegee. Background Technology

[0002] The doctor blade is a core component of the inking system in gravure and flexographic printing presses. Typically made of stainless steel or blued steel, its main function is to scrape away excess ink from the surface of the printing cylinder during printing, precisely controlling ink layer thickness and ensuring print clarity. Under high-speed, high-pressure printing conditions, the doctor blade edge rubs against the printing cylinder for extended periods, making it prone to wear, chipping, curling, or deformation. Defects in the doctor blade can lead to quality problems such as blade lines, color differences, or hazy imperfections in the print. In severe cases, it can even scratch the expensive printing cylinder, causing significant economic losses. Therefore, accurate inspection of the doctor blade's surface condition is crucial for ensuring printing quality and production safety.

[0003] Currently, the surface condition inspection of printing blades mainly relies on manual visual inspection or the operator's experience judgment. Typically, workers need to observe the blade edge while the machine is stopped, or infer the blade condition by observing the printing quality of the printed products during production.

[0004] However, printing doctor blades often have microscopic burrs or slight wavy deformations at the micrometer level. These minute defects are difficult to detect with the naked eye in the early stages. Differences in visual acuity and judgment standards among operators easily lead to missed detections or misjudgments. Often, the doctor blade failure is not discovered until a large batch of defective products appears, resulting in significant material waste. Furthermore, in actual operation, the blade edge often has randomly distributed ink residue, which can easily be mistaken for metal defects. Workers typically need to repeatedly wipe the doctor blade to confirm the diagnosis, which is not only time-consuming and labor-intensive but also increases downtime, severely impacting production efficiency. Summary of the Invention

[0005] To address the technical problem that existing printing blade inspection relies on manual experience, making it difficult to accurately distinguish between metal defects and surface stains in ink-interference environments, thus hindering the implementation of refined control for different wear stages, this invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides a method for identifying the wear state of a printing squeegee, comprising: acquiring a grayscale image of the cutting edge region of the printing squeegee; performing edge detection on the grayscale image to obtain a set of pixel edge points; taking any point in the set of pixel edge points as the center point, obtaining a virtual interpolation point according to the gradient direction of the center point, performing curve fitting between the center point and the virtual interpolation point, and performing coordinate correction on the center point according to the fitting result to obtain a correction point of the center point; combining the correction points and virtual interpolation points corresponding to all points in the set of pixel edge points into a contour coordinate set; performing curve fitting on the contour coordinate set to obtain an ideal baseline, and performing contour coordinate correction on the baseline. The net loss residual sequence is obtained by taking the vertical deviation distance of each point in the set from the ideal baseline and the preset ink tolerance threshold; the local texture entropy is obtained by traversing the net loss residual sequence using a sliding window, and the maximum value of the local texture entropy is taken as the micro texture entropy; the net loss residual sequence is smoothed to obtain the trend sequence, and the second difference of the trend sequence is calculated to obtain the macro deformation potential energy; the mean of the absolute values ​​of the elements less than zero in the net loss residual sequence is taken as the average wear depth; the doctor blade wear severity index is constructed based on the micro texture entropy, macro deformation potential energy and average wear depth; and the mechanical adjustment control signal is generated based on the doctor blade wear severity index.

[0007] This invention, after edge detection, corrects the edge position between pixels using virtual interpolation points based on gradient direction and utilizes grayscale distribution information to calculate the edge position between pixels. This enables the capture of minute edge chipping and improves the accuracy of blade edge positioning. The invention obtains an ideal baseline through asymmetric weighted iterative fitting, reducing the impact of outwardly protruding ink residue on the detection results while retaining information on inwardly recessed metal defects, thus solving the problem of baseline shift and false detection caused by ink adhesion. This invention evaluates the doctor blade condition by combining three dimensions: microtexture, macrodeformation, and average wear depth. This overcomes the problem that a single depth index cannot simultaneously reflect micro-serration and macro-wave deformation of the blade edge, improving the ability to identify complex wear patterns. This invention performs state maintenance, pressure compensation, and shutdown replacement operations according to different wear stages of the doctor blade, extending the doctor blade's service life while ensuring printing quality.

[0008] Preferably, the step of performing edge detection on the grayscale image to obtain a set of pixel edge points includes: performing Gaussian smoothing filtering on the grayscale image; calculating the gradient components of the image in the horizontal and vertical directions using an edge detection operator to obtain a gradient magnitude map; performing non-maximum suppression processing on the gradient magnitude map; and determining the final set of pixel edge points using a double threshold hysteresis algorithm.

[0009] Preferably, obtaining the virtual interpolation point based on the gradient direction of the center point includes: for each center point in the pixel edge point set, extending a preset unit distance along its gradient direction and anti-gradient direction respectively to determine the coordinate positions of two virtual interpolation points; and calculating the gradient magnitude of the two virtual interpolation points using a bilinear interpolation algorithm.

[0010] This invention constructs virtual interpolation points along the gradient direction and its opposite direction, explores the gradient change trend in the space between pixels, and uses the discrete gradient magnitude in the neighborhood to perform bilinear interpolation. Based on the grayscale information of the surrounding pixels, it calculates the gradient intensity of the virtual position, thereby providing high-density reference data for finding the real gradient extrema and ensuring the continuity of edge positioning.

[0011] Preferably, the step of performing curve fitting between the center point and the virtual interpolation points, and correcting the coordinates of the center point based on the fitting result to obtain the corrected center point includes: taking the coordinates and gradient magnitudes of the two virtual interpolation points, as well as the coordinates and gradient magnitudes of the center point, as inputs, and using a parabolic fitting algorithm to obtain a fitted curve; calculating the distance between the extreme point of the fitted curve and the center point, and decomposing the distance into a horizontal offset and a vertical offset; and superimposing the horizontal offset and the vertical offset onto the original coordinates of the center point to obtain the corrected center point.

[0012] This invention analyzes the gradient magnitude distribution of the center point and virtual interpolation points using a fitting algorithm, calculates the offset of the real edge peak relative to the center pixel, and superimposes the calculated offset onto the original coordinates. This enables the keen capture of minute chipping features and improves the detection sensitivity for minute defects.

[0013] Preferably, the step of curve fitting the contour coordinate set to obtain the ideal baseline includes: performing linear regression on the contour coordinate set as the initial dataset to determine the initial baseline; obtaining the vertical deviation distance based on the distance from each coordinate point in the set to the current baseline, wherein the vertical deviation distance can be positive or negative; when the vertical deviation distance is positive, substituting it into a preset decay function to obtain the weight value corresponding to the vertical deviation distance; when the vertical deviation distance is negative or zero, using the preset weight value as the weight value corresponding to the vertical deviation distance, constructing a weight matrix with all the obtained weight values, establishing a weighted least squares objective function in combination with the contour coordinate set, and recalculating and updating the baseline parameters by solving for the minimum value of the objective function; repeating the weight allocation and parameter update process until the baseline parameters converge to obtain the ideal baseline.

[0014] Preferably, the net loss residual sequence satisfies the following relationship: In the formula, The first in the net loss residual sequence The value of each element, The first in the set of contour coordinates Vertical deviation distance from the ideal baseline at each coordinate point The discrimination gain coefficient, It is a natural constant. This is the preset ink tolerance threshold.

[0015] This invention rapidly compresses the values ​​of positive deviations exceeding the tolerance threshold through the net loss residual sequence, thereby eliminating ink noise adhering to the doctor blade surface and maintaining the transmission of values ​​for negative wear deviations, thus fully preserving the depth information of metal defects and improving the anti-interference capability of the detection.

[0016] Preferably, the step of obtaining local texture entropy by traversing the net loss residual sequence using a sliding window includes: statistically analyzing the histogram distribution of the values ​​in the net loss residual sequence within the sliding window; and using the information entropy of the histogram distribution as the local texture entropy.

[0017] Preferably, the macroscopic deformation potential energy satisfies the following relationship: In the formula, This refers to macroscopic deformation potential energy; , and These are the first two terms in the trend sequence. The value, the first The value and the first One value, The length of the trend sequence.

[0018] Preferably, the wear severity index of the scraper satisfies the following relationship: In the formula, This refers to the severity index of scraper wear. For microscopic texture entropy, This is the macroscopic deformation potential energy. The average wear depth; The preset entropy feature scale parameter, The preset potential energy characteristic scale parameters, These are the preset depth feature scale parameters.

[0019] This invention processes evaluation indicators with different physical meanings through characteristic scale parameters, eliminating the synthesis obstacles caused by the differences in numerical ranges between indicators. It comprehensively describes micro-roughness, macro-wave deformation, and material loss, reflecting the health status of the scraper in different dimensions. This avoids omissions or misjudgments caused by the limitations of a single perspective, and provides a basis for subsequent control decisions.

[0020] Secondly, the present invention provides a printing blade wear condition identification system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned printing blade wear condition identification method is implemented.

[0021] By adopting the above technical solution, a computer program is generated from the above-mentioned method for identifying the wear state of a printing squeegee and stored in a memory so that it can be loaded and executed by a processor. This allows for the creation of a terminal device based on the memory and processor, making it convenient to use.

[0022] The beneficial effects of this invention are as follows: This invention can adapt to the complex working conditions of printing workshops filled with ink splatter and residue. Through anti-interference processing of image data, the detection system can still stably identify the true shape of the metal body even when there is a large amount of irregular ink adhering to the squeegee surface. This reduces the dependence of the detection process on the cleanliness of the squeegee and reduces the frequency of stopping the machine specifically to wipe the squeegee for detection during production, thereby improving the continuous operation efficiency of the printing production line. This invention alleviates the problem that traditional single-dimensional detection cannot cover multiple failure modes. Through comprehensive analysis of micro-roughness, macro-deformation trends, and physical defect depth, it achieves comprehensive monitoring of the squeegee's health status. Whether it is a tiny burr that causes foggy uncleanliness in the printed product or a wavy distortion that causes uneven ink layer thickness, both can be effectively identified, thereby preventing batch printing defects caused by hidden defects in the squeegee and improving the quality consistency of the final printed product. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating a method for identifying the wear state of a printing squeegee according to the present invention;

[0024] Figure 2 This is a grayscale image schematically showing the region of interest of the printing scraper blade in this invention;

[0025] Figure 3 This is a schematic diagram illustrating the ideal baseline comparison in this invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0028] This invention discloses a method for identifying the wear state of a printing squeegee, referring to... Figure 1 This includes steps S1-S5:

[0029] S1. Obtain the original image of the printing squeegee blade area, perform grayscale conversion and contrast stretching processing on it, and obtain the grayscale image of the region of interest of the printing squeegee blade.

[0030] It should be noted that printing squeegees are typically made of stainless steel or blued steel, and their blade surfaces have extremely strong specular reflective properties. Furthermore, during printing, the squeegee surface often has randomly distributed ink residue. These factors result in low contrast between the blade edge and the background under normal lighting, making it easy to misinterpret illuminated ink as a metal notch, or reflective metal as the background. To eliminate specular reflection and highlight the physical contours of the blade, a high-contrast imaging environment needs to be created, so that the squeegee blade appears as a clear dark or light boundary in the image, thus separating the physical edge from the complex background.

[0031] Specifically, a coaxial light source or a high-angle strip light source is controlled to illuminate the blade edge area at a specific angle, and a high-resolution raw image of the blade is captured by an industrial camera. The raw image is converted into single-channel data using a weighted average method, and a rectangular region containing the blade edge is extracted as the region of interest (ROI). A grayscale histogram is constructed, and the grayscale distribution of pixels within the ROI is statistically analyzed. Based on this pixel grayscale distribution, the ROI undergoes contrast stretching to obtain a grayscale image of the ROI of the printing blade edge.

[0032] S2. Perform edge detection on the grayscale image to obtain a set of pixel edge points, and perform coordinate correction based on virtual interpolation points in the gradient direction to obtain a set of contour coordinates.

[0033] It should be noted that ink flows on the surface of the doctor blade, and its edges typically exhibit a blurred, gradual transition, while the physical chipping or curling of the doctor blade's metal surface usually shows a step-like abrupt change in grayscale. To distinguish between ink edges and metal edges, the degree of abruptness in grayscale change should be considered. By locating the maximum point of the gradient modulus, the true physical boundary of the metal blade edge can be accurately captured. Furthermore, to mitigate image resolution limitations and detect minute chipping, further fine-tuning of the edge position using grayscale information from the edge's neighborhood is necessary.

[0034] Specifically, a Gaussian smoothing filter is constructed to convolve the grayscale image, and the Sobel operator is used to calculate the gradient components in the horizontal and vertical directions to obtain the gradient magnitude map. Each pixel in the gradient magnitude map is traversed, and its gradient magnitude is compared with two adjacent pixels along the gradient direction. If the pixel is not a maximum point, its grayscale value is set to zero to obtain the non-maximum suppression map. The non-maximum suppression map is then processed by the double threshold hysteresis algorithm to obtain the pixel edge point set.

[0035] Furthermore, for each edge pixel in the pixel edge point set, taking that pixel as the center point, extend it one pixel unit distance along the gradient direction and the anti-gradient direction respectively to determine the positions of two sub-pixel precision virtual interpolation points. Using the discrete gradient magnitude in the neighborhood of that pixel, calculate the gradient magnitude of these two virtual interpolation points using a bilinear interpolation algorithm. Use the coordinates of the two virtual interpolation points and the coordinates of the center point as input to the curve fitting algorithm to obtain the fitted curve. Obtain the distance between the extreme point in the fitted curve and the center point. Project this distance along the horizontal axis of the image to obtain the horizontal offset, and project this distance along the vertical axis of the image to obtain the vertical offset. Add the horizontal and vertical coordinates of the center point to the horizontal and vertical offsets respectively to obtain the correction point of the center point. Combine the virtual interpolation points and the correction points of the center point corresponding to all edge pixels into a contour coordinate set.

[0036] In one embodiment, the curve fitting algorithm is a parabolic fitting algorithm.

[0037] It's important to note that the horizontal and vertical offsets can be positive or negative. Their signs are determined by the relative magnitudes of the coordinates of the extreme points and the center point of the fitted curve. The horizontal offset is positive when the x-coordinate of the extreme point is greater than that of the center point; conversely, it is negative if the x-coordinate is less than that of the center point. Similarly, the vertical offset is positive when the y-coordinate of the extreme point is greater than that of the center point; otherwise, it is negative.

[0038] S3. Iteratively fit the contour coordinate set to obtain the ideal baseline, and calculate the net loss residual sequence in combination with the ink tolerance threshold.

[0039] It should be noted that in actual printing conditions, the doctor blade edge often exhibits both inward chipping and outward ink agglomeration. Traditional least squares fitting treats all deviations equally, causing the adhering ink to pull the fitting baseline outward, thus masking the true minute wear and reducing detection accuracy. To obtain the true wear depth, this invention constructs a fitting method that resists ink interference, ensuring the baseline closely adheres to the unworn portion of the metal body, rather than being deflected by the ink.

[0040] Specifically, linear regression fitting is performed using the contour coordinate set as the initial dataset to determine the initial baseline parameters. The vertical deviation distance is obtained based on the Euclidean distance from each coordinate point in the set to the initial baseline, and an initial distance sequence is constructed.

[0041] It should be added that the vertical deviation distance can be positive or negative; the ideal baseline is a zero potential surface, and the vertical deviation distance on the ink adhesion side of the ideal baseline is positive; the vertical deviation distance on the metal defect side of the ideal baseline is negative. This allows the net loss residual sequence to selectively attenuate the positive deviation representing ink based on the sign difference, while retaining the negative deviation representing wear.

[0042] Furthermore, each vertical deviation distance in the distance sequence is traversed. When the vertical deviation distance is positive, it is substituted into a preset attenuation function to obtain the weight value corresponding to that vertical deviation distance. When the vertical deviation distance is negative or zero, the preset weight value is used as the weight value corresponding to that vertical deviation distance. All the obtained weight values ​​are used to construct a weight matrix. A weighted least squares objective function is established by combining the contour coordinate set. The baseline parameters are recalculated and updated by solving for the minimum value of this objective function. The weight allocation and parameter update process is repeated until the baseline parameters converge, resulting in an ideal baseline that is resistant to interference. The vertical deviation of each point in the contour coordinate set to the ideal baseline is calculated, and the value of each element in the net loss residual sequence is obtained by combining it with the ink tolerance threshold.

[0043] For example, the preset decay function is an inverse proportional function, and the preset weight value is 1.

[0044] For example, when the absolute value of the difference between the baseline parameters obtained from two consecutive iterations is less than a preset convergence threshold, or when the number of iterations reaches a preset iteration limit, the baseline parameters are determined to have converged, and the iteration stops. For example, the convergence threshold is... The maximum number of iterations is 20.

[0045] Specifically, the net loss residual sequence satisfies the following relationship:

[0046] ;

[0047] In the formula, The first in the net loss residual sequence The value of each element, The first in the set of contour coordinates Vertical deviation distance from the ideal baseline at each coordinate point It is a natural constant. The discrimination gain coefficient, A preset ink tolerance threshold is used to control the steepness of the change in the net loss residual sequence, which determines the suppression rate of ink noise. In this embodiment... The value is set to 10. Implementers can adjust this coefficient based on actual implementation conditions. The empirical range is usually [5, 15]. When the ink used for printing has a high viscosity and the resulting edge transition area is wide, the value can be appropriately reduced. To achieve a smoother suppression effect; when extremely rapid truncation of tiny ink particles is required, the size can be appropriately increased. To enhance the binarization and discrimination capabilities of the signal; in this embodiment, we take =0.05. Implementers can determine this threshold based on actual implementation conditions; its empirical range is typically [0.02, 0.1]. Under high-speed printing conditions, if a thicker layer of lubricating ink is permissible on the doctor blade surface, the threshold can be appropriately increased. To avoid false alarms; if extremely high cleanliness of the scraper is required, the size should be appropriately reduced. .

[0048] in, This represents the degree of deviation of the original profile from the ideal baseline. When A negative value indicates that there is chipping or wear at that location. The denominator approaches 0, and the denominator approaches 1, making the th element in the net loss residual sequence... The value of each element retains information about the wear depth; when Positive value and exceeding the threshold When this occurs, it indicates that ink has accumulated in that area. The larger the value, the more likely it is to cause the first element in the net loss residual sequence to be affected. The value of each element rapidly decays to near 0, eliminating the noise from the attached ink and outputting only the net loss data representing the missing metal body.

[0049] For example, Figure 2 This is a grayscale image of the region of interest on the printing scraper blade in this invention. Figure 3 This is a schematic diagram comparing the ideal baseline in this invention. As can be seen from the figure, when faced with severe ink adhesion interference on the doctor blade surface, the traditional least squares method is severely affected by outlier noise, and the fitted line drifts significantly towards the ink area. However, this invention, through an asymmetric weighting mechanism, successfully identifies and eliminates the false edges generated by the ink, always closely adhering to the true metal contour at the bottom of the doctor blade, thus providing a good foundation for subsequent identification of doctor blade wear status.

[0050] S4. Extract features from the net loss residual sequence to obtain micro-texture entropy, macro-deformation potential energy and average wear depth, and construct a scraper wear severity index based on these.

[0051] It should be noted that the failure modes of doctor blades are usually not one-dimensional. Sometimes, although the doctor blade does not have deep chipping, high-frequency micro-serrations or burrs appear on the blade surface. This micro-roughness can lead to a hazy, unclean appearance in the printed product. Other times, although the blade surface is smooth, long-wavelength wavy deformation occurs overall. This macro-distortion can lead to uneven local pressure distribution. A single depth statistic cannot simultaneously cover these two distinct and unrelated wear modes. To construct a comprehensive evaluation standard, this invention extracts features from the net loss residual sequence to obtain micro-texture entropy, macro-deformation potential energy, and average wear depth, and constructs a doctor blade wear severity index based on these parameters.

[0052] Specifically, a sliding window is used to traverse the net loss residual sequence. Based on the histogram distribution of the residual values ​​within the sliding window, local texture entropy reflecting the density of micro-burrs on the cutting edge is obtained. Each sliding of the window yields one local texture entropy. All local texture entropies generated during the sliding window process are combined to form a local texture entropy sequence. The maximum value in the local texture entropy sequence is used as the micro-texture entropy reflecting the density of micro-burrs on the cutting edge. Simultaneously, the net loss residual sequence is smoothed to extract a trend sequence reflecting the overall direction of the scraper. The macroscopic deformation potential energy reflecting the degree of wave deformation of the scraper is obtained based on the sum of squares of the second difference of the trend sequence. The average wear depth is calculated as the mean of the absolute values ​​of all elements less than zero in the net loss residual sequence. The scraper wear severity index is obtained based on the micro-texture entropy, macroscopic deformation potential energy, and average wear depth.

[0053] For example, a method for smoothing the net loss residual sequence is low-pass filtering or a moving average algorithm.

[0054] Specifically, the scraper wear severity index satisfies the following relationship:

[0055] ;

[0056] ;

[0057] In the formula, This refers to the severity index of scraper wear. For microscopic texture entropy, This refers to macroscopic deformation potential energy; , and These are the first two terms in the trend sequence. The value, the first The value and the first One value, The length of the trend sequence. The average wear depth; The preset entropy feature scale parameter is used to eliminate the dimension of micro-texture entropy. These are preset potential energy characteristic scale parameters used to eliminate the dimensions of macroscopic deformation potential energy. The preset depth feature scale parameter is used to eliminate the dimension of the average wear depth. In this embodiment, It is 1.5. It is 100. It is 0.2. , and These are all normalized constants obtained from historical failure data statistics, used to balance the weights of different dimensional indicators in the total index. Implementers can adjust them according to actual conditions; for example, if the current printing process is extremely sensitive to hazy or unclean printed products, the weights can be appropriately reduced. This amplifies the proportion of microtexture entropy in the doctor blade wear severity index, thereby improving the system's sensitivity to burr detection. If the current process primarily focuses on ink layer thickness uniformity, then the size can be appropriately reduced. To highlight the influence weight of macroscopic deformation potential energy.

[0058] in, The roughness evaluation component represents the micro-dimensional aspect. The larger the value, the more chaotic the distribution of residual values ​​within a certain local area of ​​the scraper's cutting edge. Physically, this corresponds to the presence of dense, tiny burrs or serrations on the cutting edge. The larger the value, the greater the severity index of scraper wear. This represents the deformation evaluation component at the macroscopic level. The bending strain energy of the curve was simulated using the sum of squares of the second-order differences. The larger the value, the more drastically the overall trend line of the scraper blade deviates from a straight line, and the more likely the scraper has undergone wavy twisting, leading to... The larger the value, the greater the severity index of scraper wear. This represents the material loss assessment component. The larger the value, the more severe the loss of metal material at the blade edge, and the greater the blade wear severity index.

[0059] S5. Determine the graded status based on the scraper wear severity index and generate the corresponding mechanical adjustment control signal.

[0060] It should be noted that doctor blade wear is a gradual process from quantitative to qualitative change, and different wear stages correspond to drastically different maintenance strategies. In the early stages of wear, there are microscopic burrs or slight wear. Although the edge sharpness decreases, the contact tightness can be compensated by appropriately increasing the doctor blade pressure, thereby extending the doctor blade's service life. However, in the later stages of wear, there is severe chipping or deformation, and the physical defects have exceeded the compensation limit. If pressure is continued to be applied, it will not only fail to improve printing quality but will also scratch the expensive printing plate. To address doctor blade wear under various conditions, this invention determines the condition based on a doctor blade wear severity index and generates a mechanical adjustment control signal.

[0061] Specifically, a preset pressure compensation threshold and a shutdown alarm threshold exceeding the pressure compensation threshold are defined. When the doctor blade wear severity index is less than the pressure compensation threshold, the doctor blade is determined to be in a healthy state, a status hold signal is generated, and the doctor blade adjustment system of the printing press is controlled to maintain the current doctor blade pressure and angle parameters unchanged. When the doctor blade wear severity index is between the pressure compensation threshold and the shutdown alarm threshold, the doctor blade is determined to be in a compensable minor wear state, a pressure compensation control signal is generated, and the doctor blade pressure regulating cylinder is driven to increase the pressure by a preset pressure step based on the current pressure to compensate for the contact gap caused by blade wear. When the doctor blade wear severity index exceeds the shutdown alarm threshold, the doctor blade is determined to be in a severely failed state, a shutdown alarm signal and a reset control signal are generated, triggering an audible and visual alarm to prompt manual replacement of the doctor blade, while simultaneously driving the doctor blade angle adjustment motor to retract to the preset replacement position.

[0062] For example, the range of values ​​for the pressure compensation threshold is: The range of values ​​for the shutdown alarm threshold is: In this embodiment, the pressure compensation threshold is 0.4, the shutdown alarm threshold is 0.8, and the preset pressure step is 0.05MPa.

[0063] This invention also discloses a printing blade wear condition identification system, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a printing blade wear condition identification method according to the present invention.

[0064] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

Claims

1. A method for identifying the wear condition of a printing squeegee, characterized in that, include: Obtain a grayscale image of the printing squeegee blade area; perform edge detection on the grayscale image to obtain a set of pixel edge points; Taking any point in the pixel edge point set as the center point, a virtual interpolation point is obtained according to the gradient direction of the center point. Curve fitting is performed between the center point and the virtual interpolation point, and the coordinates of the center point are corrected according to the fitting result to obtain the correction point of the center point. The correction points and virtual interpolation points corresponding to all points in the pixel edge point set are combined into a contour coordinate set. The virtual interpolation points are obtained based on the gradient direction of the center point, including: for each center point in the pixel edge point set, extending a preset unit distance along its gradient direction and anti-gradient direction respectively to determine the coordinate positions of two virtual interpolation points; and calculating the gradient magnitude of the two virtual interpolation points using a bilinear interpolation algorithm. Curve fitting is performed between the center point and the virtual interpolation points, and the coordinates of the center point are corrected based on the fitting results to obtain the corrected center point. This includes: taking the coordinates and gradient magnitudes of the two virtual interpolation points, as well as the coordinates and gradient magnitude of the center point, as inputs, and using a parabolic fitting algorithm to obtain the fitted curve; calculating the distance between the extreme points of the fitted curve and the center point, and decomposing this distance into horizontal and vertical offsets; and superimposing the horizontal and vertical offsets onto the original coordinates of the center point to obtain the corrected center point. An ideal baseline is obtained by curve fitting of the contour coordinate set. The net loss residual sequence is obtained based on the distance of each point in the contour coordinate set to the ideal baseline and the preset ink tolerance threshold. The local texture entropy is obtained by traversing the net loss residual sequence using a sliding window, and the maximum value of the local texture entropy is taken as the micro texture entropy. The net loss residual sequence is smoothed to obtain a trend sequence, and the second difference of the trend sequence is calculated to obtain the macro deformation potential energy. The mean of the absolute values ​​of the elements less than zero in the net loss residual sequence is taken as the average wear depth. The blade wear severity index is constructed based on the micro texture entropy, macro deformation potential energy, and average wear depth. Mechanical adjustment control signals are generated based on the scraper wear severity index.

2. The method for identifying the wear state of a printing squeegee according to claim 1, characterized in that, The step of obtaining a set of pixel edge points by edge detection of a grayscale image includes: performing Gaussian smoothing filtering on the grayscale image; calculating the gradient components of the image in the horizontal and vertical directions using an edge detection operator to obtain a gradient magnitude map; performing non-maximum suppression processing on the gradient magnitude map; and determining the final set of pixel edge points using a double threshold hysteresis algorithm.

3. The method for identifying the wear state of a printing squeegee according to claim 1, characterized in that, The process of obtaining an ideal baseline by curve fitting of the contour coordinate set includes: performing linear regression on the contour coordinate set as the initial dataset to determine the initial baseline; obtaining the vertical deviation distance based on the distance from each coordinate point in the set to the current baseline, where the vertical deviation distance can be positive or negative; when the vertical deviation distance is positive, substituting it into a preset decay function to obtain the weight value corresponding to the vertical deviation distance; when the vertical deviation distance is negative or zero, using the preset weight value as the weight value corresponding to the vertical deviation distance, constructing a weight matrix with all the obtained weight values, establishing a weighted least squares objective function in combination with the contour coordinate set, and recalculating and updating the baseline parameters by solving for the minimum value of the objective function; repeating the weight allocation and parameter update process until the baseline parameters converge to obtain the ideal baseline.

4. The method for identifying the wear state of a printing squeegee according to claim 1, characterized in that, The net loss residual sequence satisfies the following relationship: ； In the formula, The first in the net loss residual sequence The value of each element, The first in the set of contour coordinates Vertical deviation distance from the ideal baseline at each coordinate point The discrimination gain coefficient, It is a natural constant. This is the preset ink tolerance threshold.

5. The method for identifying the wear state of a printing squeegee according to claim 1, characterized in that, The method of obtaining local texture entropy by traversing the net loss residual sequence using a sliding window includes: statistically analyzing the histogram distribution of the values ​​of the net loss residual sequence within the sliding window; and using the information entropy of the histogram distribution as the local texture entropy.

6. The method for identifying the wear state of a printing squeegee according to claim 1, characterized in that, The macroscopic deformation potential energy satisfies the following relationship: ； In the formula, This refers to macroscopic deformation potential energy; , and These are the first in the trend sequence. The value, the first The value and the first One value, The length of the trend sequence.

7. The method for identifying the wear state of a printing squeegee according to claim 6, characterized in that, The wear severity index of the scraper satisfies the following relationship: ； In the formula, This refers to the severity index of scraper wear. For microscopic texture entropy, This is the macroscopic deformation potential energy. The average wear depth; The preset entropy feature scale parameter, The preset potential energy characteristic scale parameters, These are the preset depth feature scale parameters.

8. A system for identifying the wear condition of a printing squeegee, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement a method for identifying the wear state of a printing scraper according to any one of claims 1-7.

Images