A machine vision-based doctor blade angle intelligent adjustment method and system

Abstract

The present application belongs to the technical field of image data processing, and particularly relates to a method and system for intelligently adjusting the angle of a doctor blade based on machine vision, which comprises the following steps: obtaining a doctor blade grayscale image; determining the local gradient directionality factor of a pixel point; determining the gradient consistency coefficient of the pixel point in the edge extension direction of the pixel point, and determining the edge continuity degree of the pixel point; determining the noise judgment threshold after the self-adaption of the pixel point; obtaining the doctor blade grayscale image after noise removal, extracting the doctor blade edge profile, and calculating the actual contact angle of the doctor blade to adjust the angle of the doctor blade. The present application analyzes the gradient amplitude characteristics and gradient direction characteristics of the pixel point, overcomes the problem that the traditional adaptive switch median filter algorithm is prone to confusing flying ink noise with the edge of the doctor blade, uses the noise judgment threshold after self-adaption to filter out noise and obtain the actual contact angle of the doctor blade, and improves the accuracy of the adjustment of the angle of the doctor blade.

Description

Technical Field

[0001] This invention relates to the field of image data processing technology. More specifically, this invention relates to a machine vision-based intelligent adjustment method and system for scraper angle. Background Technology

[0002] In the printing industry, the stability of print quality directly depends on the operating status of the doctor blade system. As a core component, the doctor blade's main function is to scrape off excess ink from the surface of the printing cylinder, leaving only the ink in the cells for transfer. The angle of contact between the doctor blade and the printing cylinder is a key parameter that determines the doctor blade effect. If the doctor blade angle is too large, it can easily damage the printing cylinder and shorten the doctor blade's life; if the doctor blade angle is too small, it cannot effectively remove ink, resulting in defects such as fogging, ink smudges, or color differences in the printed matter. Therefore, maintaining the optimal doctor blade angle is crucial for ensuring print quality.

[0003] Traditional doctor blade angle adjustment relies heavily on operator experience, involving manual adjustment of cylinder pressure or blade holder angle based on visual observation of print quality. This method is not only slow to respond but also highly susceptible to human error, making it difficult to meet the precision requirements of modern high-speed printing presses. With the development of machine vision technology, the method of using industrial cameras to monitor doctor blade status in real time has gradually gained popularity. However, printing presses typically operate at high speeds, generating a large amount of ink splatter at the contact area between the doctor blade and the printing roller. These tiny ink particles, suspended in the air, create high-density fog-like particle noise or salt-and-pepper noise-like interference in front of the lens during imaging. An existing algorithm for removing this type of noise is the adaptive switching median filtering algorithm. This algorithm determines whether to perform filtering replacement by judging whether a pixel is a noise point, demonstrating good performance in noise reduction.

[0004] Standard adaptive switching median filtering algorithms typically determine whether a pixel is noise based solely on the difference between its grayscale value and the median of its neighborhood, combined with a pre-defined, fixed noise threshold. This mechanism has significant limitations in high-speed inkjet printing environments. Specifically, the blade of a scraper appears as a straight edge with high-frequency characteristics in an image, with its grayscale value exhibiting abrupt changes compared to the surrounding background. Standard adaptive switching median filtering algorithms are prone to misclassifying these scraper edge pixels, which represent valid information, as high-amplitude impulse noise, and thus smoothing them. This misclassification leads to blurred, jagged, or even broken scraper edges in the denoised image, severely impacting the accuracy of subsequent scraper angle calculations based on edge extraction. Summary of the Invention

[0005] To address the technical problem that the aforementioned standard adaptive switching median filtering algorithm, based on a fixed noise threshold, easily treats the scraper edge as noise and removes it, thus affecting the accuracy of scraper angle calculation, this invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides a machine vision-based intelligent adjustment method for the angle of a scraper, comprising: acquiring a scraper image; preprocessing the scraper image to obtain a scraper grayscale image; determining a local gradient directionality factor of a pixel based on the gradient magnitude of the pixel; determining a gradient consistency coefficient of a pixel along its edge extension direction based on the gradient direction difference between the pixel and pixels along its edge extension direction; determining the edge continuity of a pixel based on the gradient consistency coefficient and the local gradient directionality factor; determining an adaptive noise threshold for the pixel based on the edge continuity; performing adaptive switching median filtering on the scraper grayscale image using the adaptive noise threshold to obtain a denoised scraper grayscale image; extracting the scraper blade contour and calculating the actual contact angle of the scraper based on the denoised scraper grayscale image; and adjusting the scraper angle based on the actual contact angle of the scraper.

[0007] This invention highlights the directional characteristics of the actual squeegee edge by calculating the local gradient directionality factor and utilizing the difference between the maximum gradient magnitude and the vertical magnitude. It quantifies the consistency and irregularity of the gradient direction in the edge extension direction by calculating the gradient consistency coefficient, thus enabling the identification of ink splatter noise. By combining the local gradient directionality factor and the gradient consistency coefficient to calculate the edge continuity and adaptively correcting the baseline noise threshold, the anti-interference capability of the switching median filtering algorithm in complex backgrounds is enhanced. Furthermore, by generating a denoised squeegee grayscale image based on the adaptive noise threshold and extracting the squeegee blade contour and calculating the actual contact angle, the squeegee angle is evaluated, accurately identifying printing defects caused by improper angles and improving the accuracy and robustness of adjustments.

[0008] Preferably, the acquisition of the doctor blade image includes: installing an industrial high-speed camera on the side of the doctor blade holder of the printing press to take a picture and obtain the doctor blade image.

[0009] Preferably, the local gradient directionality factor of the pixel satisfies the expression:

[0010] In the formula, For pixels The local gradient direction factor, For pixels The maximum gradient magnitude in four preset directions, For pixels The gradient magnitude in the direction perpendicular to the direction of the maximum gradient magnitude among the four preset directions. This represents the average gradient magnitude of all pixels within the grayscale image of the scraper. To prevent tiny constants with a denominator of 0, For the minimum normalization function, It is a natural exponential function.

[0011] This invention analyzes the gradient magnitude characteristics of pixels to evaluate the local gradient directionality factor. This results in a larger local gradient directionality factor being calculated for the real scraper edge region, while a smaller local gradient directionality factor is calculated for the ink noise region due to its chaotic direction. This provides a reliable basis for calculating the continuity of the edge.

[0012] Preferably, the method for obtaining the gradient consistency coefficient of the pixel in the edge extension direction is as follows: for any pixel in the edge extension direction, in response to the fact that the direction of the maximum gradient magnitude of the pixel is consistent with that of any pixel in the edge extension direction, 1 is taken as the gradient consistency coefficient of the pixel in the edge extension direction; otherwise, 0 is taken as the gradient consistency coefficient of the pixel in the edge extension direction.

[0013] This invention constructs a binary decision function to evaluate the gradient consistency coefficient. This decision term reflects the degree of continuity and consistency of the gradient direction in the edge extension direction of the pixel. This results in a larger gradient consistency coefficient being calculated for the real edge region, while a smaller gradient consistency coefficient is calculated for the ink noise region due to the disordered direction. This allows the continuity characteristics of the edge extension to be mapped, providing a reliable directional consistency benchmark for calculating the degree of edge continuity.

[0014] Preferably, the edge continuity of the pixel satisfies the expression:

[0015] In the formula, For pixels The degree of edge continuity, For pixels The local gradient direction factor, This represents the total number of pixels selected along the direction extending from the edge of the pixel. For pixels The first in the edge extension direction Local gradient directionality factor for each pixel For pixels The first in the edge extension direction Gradient consistency coefficient of each pixel It is a natural exponential function.

[0016] This invention achieves the evaluation of edge continuity by constructing a composite function containing a positive correlation term of local gradient directionality factor and a gradient consistency coefficient term. The local gradient directionality factor term enhances the continuous weight of the real scraper edge, while the gradient consistency coefficient term suppresses the continuous contribution of ink fly noise. This results in a smaller edge continuity in the ink fly noise region, while the real scraper edge region maintains a larger continuity, thereby effectively distinguishing the scraper edge from the background noise.

[0017] Preferably, the noise determination threshold after pixel adaptation satisfies the expression:

[0018] In the formula, For pixels Adaptive noise threshold The threshold for determining basic noise, Adjust the gain coefficient to the threshold. For pixels The degree of edge continuity.

[0019] This invention dynamically adjusts the basic noise judgment threshold by adjusting the edge continuity, achieving pixel-level adaptive noise judgment threshold. In areas with strong ink noise, the threshold is automatically lowered to enhance filtering, while in areas with obvious scraper blade features, a higher threshold is maintained to preserve the true response, ensuring the accuracy of the scraper grayscale image after denoising.

[0020] Preferably, obtaining the denoised scraper grayscale image includes: performing an adaptive switching median filtering algorithm based on a noise judgment threshold adapted to each pixel to determine and replace noisy pixels, and using the result as the denoised scraper grayscale image.

[0021] Preferably, the step of extracting the blade edge contour and calculating the actual contact angle of the scraper based on the denoised scraper grayscale image includes: extracting the edges of the denoised scraper grayscale image using the Canny operator, then fitting the extracted edges using the Hough transform, calculating the angle between the edge and the horizontal baseline, and obtaining the actual contact angle of the scraper.

[0022] Preferably, adjusting the scraper angle includes: calculating the difference between the actual contact angle of the scraper and the preset optimal scraper angle, determining the scraper angle deviation, and driving the scraper holder to make fine adjustments based on the scraper angle deviation, thereby achieving scraper angle adjustment.

[0023] Secondly, the present invention provides a machine vision-based intelligent adjustment system for scraper angle, comprising a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned machine vision-based intelligent adjustment method for scraper angle is implemented.

[0024] By adopting the above technical solution, a computer program for intelligent adjustment of scraper angle based on machine vision is generated 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.

[0025] The beneficial effects of this invention are as follows:

[0026] This invention solves the technical problem that traditional adaptive switching median filtering algorithms easily misjudge the edge of the scalpel as noise in the scalpel image environment, resulting in denoising distortion, by introducing an adaptive noise judgment threshold correction mechanism based on multi-feature fusion.

[0027] This invention establishes the intrinsic mapping relationship between the local gradient directionality factor and the real scraper structure by analyzing the gradient magnitude characteristics of pixels. On this basis, it further integrates the gradient consistency coefficient, tightly coupling the index reflecting gradient direction preference with the index reflecting edge continuity, thereby realizing the identification of ink background.

[0028] This invention calculates the edge continuity matching the local gradient features of each pixel. Through adaptive correction of the basic noise threshold, it achieves accurate matching between the noise threshold and local image features. For areas with ink splatter interference, the threshold is automatically lowered to enhance filtering; for the actual scraper area, the threshold is maintained to preserve its integrity. Ultimately, this invention improves the accuracy and anti-interference capability of scraper image processing, enhances the robustness of angle adjustment, and enables timely and accurate identification of angle deviations, providing a solid technical guarantee for printing quality control and equipment maintenance. Attached Figure Description

[0029] Figure 1 The flowchart illustrates a machine vision-based intelligent adjustment method for scraper angle in this invention.

[0030] Figure 2 A numerical comparison chart of the noise determination threshold fixed in the prior art and the adaptive noise determination threshold of the present invention. Detailed Implementation

[0031] 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.

[0032] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0033] This invention discloses a machine vision-based intelligent adjustment method for scraper angle, referring to... Figure 1 This includes steps S001 to S005, specifically:

[0034] S001: Acquire a scraper image, preprocess the scraper image to obtain a scraper grayscale image.

[0035] Specifically, an industrial high-speed camera is installed on the side of the doctor blade holder of the printing press to take pictures and obtain doctor blade images. The doctor blade images are then processed into grayscale to obtain doctor blade grayscale images.

[0036] S002: Determine the local gradient directionality factor of a pixel based on its gradient magnitude.

[0037] It should be noted that existing adaptive switching median filtering algorithms typically use a fixed preset noise threshold for noise detection. This makes them ill-suited to the complex conditions of high-density ink splatter noise and squeegee edge noise coexisting in high-speed printing environments. This leads to the algorithm easily confusing high-frequency ink splatter particles with high-frequency squeegee edges, mistakenly filtering out squeegee edge details and reducing the accuracy of squeegee angle detection. Based on image geometric feature analysis, the squeegee blade, as a linear mechanical structure, exhibits significant anisotropy in its gray-level gradient; that is, the gradient magnitude perpendicular to the edge direction is significantly greater than the gradient magnitude parallel to the edge direction. In contrast, ink splatter noise is usually distributed in a point-like pattern, and its gradient changes are isotropic. Therefore, this invention determines a local gradient directionality factor for each pixel to characterize the strength of anisotropy in the gray-level changes around that pixel.

[0038] Specifically, select the area around the pixel as the center. The range is used to construct its neighborhood window, and the Sobel operator is used to calculate the gradient magnitude of the pixel in the directions of four preset angles. In this embodiment, the four angles are set to 0°, 45°, 90° and 135° respectively. In other embodiments, the implementer can set them according to the actual implementation situation.

[0039] Specifically, the local gradient directionality factor of a pixel satisfies the expression:

[0040] ;

[0041] In the formula, For pixels The local gradient direction factor, For pixels The maximum gradient magnitude in four preset directions, For pixels The gradient magnitude in the direction perpendicular to the direction of the maximum gradient magnitude among the four preset directions. This represents the average gradient magnitude of all pixels within the grayscale image of the scraper. To prevent tiny constants with a denominator of 0, For the minimum normalization function, It is a natural exponential function.

[0042] In the formula, The larger the value, the more pixels there are. The stronger the gradient change in the direction with the largest gradient magnitude change among its four preset directions, the better the pixel. The stronger the anisotropy of the surrounding grayscale changes, the more important the pixel is. The greater the likelihood that a pixel belongs to the edge of a scraper, the more likely it is to be a pixel. The larger the local gradient directionality factor, the better. The larger the value, the more pixels there are. The greater the gradient strength at a given point, the stronger the gradient at that point. The more drastic the grayscale change at a point, the more likely it is to be a pixel. The higher the probability that a pixel belongs to the edge of a scraper, the greater its credibility. The larger the local gradient directionality factor, the better.

[0043] S003: Determine the gradient consistency coefficient of pixels along the edge extension direction based on the gradient direction difference between a pixel and pixels along its edge extension direction. Determine the edge continuity of a pixel based on the gradient consistency coefficient and the local gradient directionality factor.

[0044] It should be noted that under high-speed operation, high-density ink droplets can easily aggregate into short, linear ink trails, forming pseudo-edges with a certain directionality. These pseudo-edges also exhibit significant gradient directionality in their local neighborhoods, making it difficult to accurately distinguish between real scraper edges and pseudo-edges. This can easily leave noise in the denoised image. Based on the physical geometry of the scraper blade, a real scraper edge macroscopically appears as a long, continuous straight line traversing the field of view, with relatively stable grayscale brightness on both sides. In contrast, pseudo-edges formed by ink droplet aggregation are usually shorter and their grayscale polarity is prone to flipping or interruption during extension. Therefore, this invention determines the edge continuity of a pixel to characterize the continuity and stability of the edge structure at a global scale.

[0045] Specifically, the direction perpendicular to the direction corresponding to the maximum gradient magnitude of a pixel in four preset directions is taken as the edge extension direction of the pixel. For any pixel in the edge extension direction, the gradient consistency coefficient of the pixel in the edge extension direction is 1 if the direction of the maximum gradient magnitude of the pixel is consistent with that of any pixel in the edge extension direction; otherwise, the gradient consistency coefficient of the pixel in the edge extension direction is 0.

[0046] Specifically, the edge continuity of a pixel satisfies the expression:

[0047] ;

[0048] In the formula, For pixels The degree of edge continuity, For pixels The local gradient direction factor, The total number of pixels selected along the edge extension direction is 10 in this embodiment. In other embodiments, the implementer can set the number of pixels according to the actual implementation situation. For example, when ink splatter interference is severe in the printing workshop and long chain-like false edges are easily formed, the total number of pixels selected can be appropriately increased to enhance the algorithm's ability to identify long straight line features and accurately remove short line-like ink splatter false edges. When the response speed requirement for the doctor blade angle adjustment is high, the total number of pixels selected can be appropriately reduced to reduce the computational load and improve the algorithm efficiency. For pixels The first in the edge extension direction Local gradient directionality factor for each pixel For pixels The first in the edge extension direction Gradient consistency coefficient of each pixel It is a natural exponential function.

[0049] In the formula, The larger the value, the more pixels there are. The greater the likelihood that a pixel belongs to the edge of a scraper, the more likely it is to be a pixel. The greater the degree of edge continuity. The larger the value, the more pixels there are. The more pixels that have a height that is consistent with their own gradient direction along the edge extension direction, the more pixels there are. The greater the probability that a pixel is located on a long and stable edge, the better. The greater the likelihood that it belongs to the edge of the scraper rather than ink artifacts, the more likely it is to be a pixel. The greater the degree of edge continuity.

[0050] S004: Determine the noise threshold after pixel adaptation based on the degree of edge continuity.

[0051] It should be noted that after obtaining the edge continuity of the pixel, the present invention will modify the basic noise judgment threshold of the pixel according to the edge continuity of the pixel. The traditional adaptive switching median filtering algorithm usually presets the same basic noise judgment threshold for all pixels. The present invention, however, uses the edge continuity to modify the basic noise judgment threshold of the pixel to obtain the noise judgment threshold after the pixel is adapted, so that the adaptive switching median filtering algorithm can remove noise more accurately.

[0052] Specifically, the noise threshold after pixel adaptation satisfies the following expression:

[0053] ;

[0054] In the formula, For pixels Adaptive noise threshold In this embodiment, the basic noise threshold is set to 20. In other embodiments, the implementer can set it according to the actual implementation situation. For example, when the ink splatter noise density generated by high-speed printing is large, the basic noise threshold can be appropriately reduced to improve the algorithm's ability to filter out tiny ink droplets. When it is necessary to prevent micro-cracks and other fine features of the doctor blade from being over-smoothed, the basic noise threshold can be appropriately increased to protect effective high-frequency information. The threshold adjustment gain coefficient is set to 5 in this embodiment. In other embodiments, the implementer can set it according to the actual implementation situation. For example, when the clarity of the scraper edge is poor due to wear or lighting, and edge breakage needs to be prevented, the threshold adjustment gain coefficient can be appropriately increased to significantly improve the noise judgment threshold of the edge area. When it is necessary to avoid misjudging high-density ink flow as an edge for protection, the threshold adjustment gain coefficient can be appropriately decreased to reduce the tolerance for false edges. For pixels The degree of edge continuity.

[0055] In the formula, The larger the value, the more pixels there are. The greater the likelihood that the pixel belongs to the edge of the scraper rather than ink artifact, the more likely it is to be a pixel. The larger the adaptive noise threshold, the better to prevent pixel-level noise discrimination. It was misdiagnosed as noise.

[0056] S005: Adaptive switching median filtering is used to denoise the scraper grayscale image using the adaptive noise threshold to obtain the denoised scraper grayscale image. The scraper blade contour is extracted based on the denoised scraper grayscale image and the actual contact angle of the scraper is calculated. The scraper angle is adjusted according to the actual contact angle of the scraper.

[0057] Specifically, the scraper angle is adjusted, including:

[0058] For each pixel, the median grayscale value within its 8-neighborhood is calculated, and then the absolute value of the difference between the pixel and the median grayscale value within its 8-neighborhood is calculated. If the absolute value of the difference between the pixel and the median grayscale value within its 8-neighborhood is greater than the noise threshold after adaptive evaluation, the pixel is considered a noise pixel, and its grayscale value is replaced with the median grayscale value within its 8-neighborhood. If the absolute value of the difference between the pixel and the median grayscale value within its 8-neighborhood is less than or equal to the noise threshold after adaptive evaluation, the pixel is considered normal, and its grayscale value is retained unchanged. This process is repeated for all pixels in the scraper grayscale image to obtain the denoised scraper grayscale image.

[0059] The Canny operator is used to extract the edges of the denoised scraper grayscale image. Then, the Hough transform is used to fit the extracted edges and calculate the angle between the edge and the horizontal baseline to obtain the actual contact angle of the scraper.

[0060] The difference between the actual contact angle of the scraper and the preset optimal scraper angle is calculated to determine the scraper angle deviation. Based on the scraper angle deviation, the scraper holder is driven to make fine adjustments to achieve scraper angle adjustment. In this embodiment, the optimal scraper angle is set to 45°. In other embodiments, the implementer can set it according to the actual implementation situation.

[0061] like Figure 2 As shown in the figure, the spatial distribution of noise judgment thresholds across the scraper edge region is illustrated. The dashed line represents the fixed threshold benchmark used in the prior art, which is always maintained at a constant low level. In contrast, the solid line represents the adaptive threshold generated in real time by the embodiment of the present invention. Observing the trend of the adaptive noise judgment threshold, it can be seen that in the non-edge background area, the adaptive noise judgment threshold only shows a small dynamic oscillation near the benchmark line, while in the scraper edge region, the adaptive noise judgment threshold is significantly improved, effectively preventing the scraper edge from being misjudged as noise and achieving accurate noise reduction.

[0062] This invention also discloses a machine vision-based intelligent adjustment system for scraper angle, including a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a machine vision-based intelligent adjustment method for scraper angle according to the present invention.

[0063] 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 intelligent adjustment of scraper angle based on machine vision, characterized in that, include: Acquire a scraper image, preprocess the scraper image to obtain a scraper grayscale image; Determine the local gradient directionality factor of a pixel based on its gradient magnitude. In the formula, For pixels The local gradient direction factor, For pixels The maximum gradient magnitude in four preset directions, For pixels The gradient magnitude in the direction perpendicular to the direction of the maximum gradient magnitude among the four preset directions. This represents the average gradient magnitude of all pixels within the grayscale image of the scraper. To prevent tiny constants with a denominator of 0, For the minimum normalization function, It is a natural exponential function; Based on the difference in gradient direction between a pixel and pixels along its edge extension direction, a gradient consistency coefficient for pixels along the edge extension direction is determined. This includes: for any pixel along the edge extension direction, if the direction of the maximum gradient magnitude of the pixel is consistent with that of any pixel along its edge extension direction, a coefficient of 1 is used as the gradient consistency coefficient for the pixel along the edge extension direction; otherwise, a coefficient of 0 is used. Based on the gradient consistency coefficient and the local gradient directionality factor, the edge continuity of the pixel is determined. In the formula, For pixels The degree of edge continuity, This represents the total number of pixels selected along the direction extending from the edge of the pixel. For pixels The first in the edge extension direction Local gradient directionality factor for each pixel For pixels The first in the edge extension direction Gradient consistency coefficient of each pixel; Based on the degree of edge continuity, determine the noise judgment threshold after pixel adaptation. In the formula, For pixels Adaptive noise threshold The threshold for determining basic noise, Adjust the gain coefficient for the threshold. The grayscale image of the scraper is denoised by adaptive switching median filtering using the adaptive noise threshold to obtain a denoised grayscale image of the scraper. The scraper blade contour is extracted based on the denoised grayscale image of the scraper and the actual contact angle of the scraper is calculated. The scraper angle is adjusted according to the actual contact angle of the scraper.

2. The intelligent adjustment method for scraper angle based on machine vision according to claim 1, characterized in that, The acquisition of the doctor blade image includes: installing an industrial high-speed camera on the side of the doctor blade holder of the printing press to take pictures and obtain the doctor blade image.

3. The intelligent adjustment method for scraper angle based on machine vision according to claim 1, characterized in that, The process of obtaining the denoised scraper grayscale image includes: performing an adaptive switching median filtering algorithm based on the noise judgment threshold adapted to the pixel points to realize the judgment and replacement of noise pixels, and using the result as the denoised scraper grayscale image.

4. The intelligent adjustment method for scraper angle based on machine vision according to claim 1, characterized in that, The step of extracting the blade edge contour and calculating the actual contact angle of the scraper based on the denoised scraper grayscale image includes: using the Canny operator to extract the edge of the denoised scraper grayscale image, then using the Hough transform to fit the extracted edge, calculating the angle between the edge and the horizontal baseline, and obtaining the actual contact angle of the scraper.

5. The intelligent adjustment method for scraper angle based on machine vision according to claim 1, characterized in that, The adjustment of the scraper angle includes: calculating the difference between the actual contact angle of the scraper and the preset optimal scraper angle, determining the scraper angle deviation, and driving the scraper holder to make fine adjustments based on the scraper angle deviation, thereby achieving scraper angle adjustment.

6. A machine vision-based intelligent adjustment system for scraper angle, characterized in that, include: The processor and memory, wherein the memory stores computer program instructions that, when executed by the processor, implement a machine vision-based intelligent adjustment method for scraper angle according to any one of claims 1-5.

Images