Visual monitoring control method for circular knife machine

Abstract

The application provides a kind of circular cutter machine vision monitoring control method, belongs to monitoring control technical field, by building for circular cutter machine vision monitoring control vision acquisition system, namely in the position of circular cutter corresponding to the cutter holder of circular cutter machine installs industrial camera;After starting circular cutter machine, the vision acquisition system for circular cutter machine vision monitoring control carries out image acquisition and pretreatment;The key operating parameters of circular cutter are extracted to the circular cutter profile image obtained after pretreatment;The control system compares the key operating parameters of circular cutter extracted with the preset standard parameter threshold, judges whether the running state of circular cutter is abnormal, and carries out corresponding regulation and control;Realize the automation, real-time, high-precision monitoring of circular cutter machine running state and complete abnormal regulation and control at the same time, guarantee the stable operation of circular cutter machine, improve the processing quality and production efficiency of circular cutter machine.

Description

Technical Field

[0001] This invention belongs to the field of monitoring and control technology, and specifically relates to a visual monitoring and control method for a circular knife machine. Background Technology

[0002] Circular die-cutting machines are widely used in industries such as textiles, die-cutting, and printing. Their main working component is the circular die, and the operating status of the circular die directly determines the precision and quality of the processed products. It also affects the operating stability and service life of the circular die-cutting machine. Currently, the operation monitoring of circular die-cutting machines mainly relies on manual inspection. Operators observe the surface condition and operating posture of the circular die with the naked eye, or use manual tools such as calipers and micrometers to measure key parameters such as the diameter and thickness of the circular die, thereby determining whether there are any abnormalities in the circular die.

[0003] However, this manual monitoring method has many insurmountable drawbacks: low detection accuracy, as the naked eye cannot detect hidden defects such as minute wear on the cutting edge of the circular cutter or fine surface cracks; manual measurement is limited by the operator's technique and tool precision, resulting in large errors and making it difficult to meet the needs of high-precision machining; it also heavily relies on the operator's experience and judgment, as different operators have varying levels of expertise and responsibility, leading to inconsistent standards for judging abnormal states, which can easily result in missed or false detections, causing abnormal circular cutters to continue operating, leading to product scrap, equipment failure, or even safety accidents; furthermore, the detection efficiency is low, as manual inspection requires machine shutdown or low-speed operation, consuming a significant amount of production time and affecting production efficiency, and it cannot achieve real-time monitoring of the circular cutter's operation, making it difficult to detect and handle sudden anomalies in a timely manner; finally, the labor intensity is high, as operators need to focus on observing the circular cutter's operating status and manually measuring for extended periods, which can easily lead to fatigue and further reduce the reliability of the detection.

[0004] Therefore, there is an urgent need for a method that can overcome the shortcomings of manual monitoring, match the operating characteristics of the circular knife machine, and achieve high-precision, real-time, and automated monitoring and control to solve the problems existing in the current technology. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of the embodiments of the present invention and to briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section, as well as in the abstract and title of the present application, to avoid obscuring the purpose of this section, the abstract and title of the invention. Such simplifications or omissions shall not be used to limit the scope of the present invention.

[0006] In view of the aforementioned existing problems, the present invention is proposed.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0008] This invention provides a visual monitoring and control method for a circular knife machine, comprising:

[0009] Build a vision acquisition system for vision monitoring and control of the rotary cutter machine, that is, install an industrial camera next to the tool holder of the rotary cutter machine at the position corresponding to the rotary cutter.

[0010] After starting the circular cutter machine, image acquisition and preprocessing are performed on the vision acquisition system for the visual monitoring and control of the circular cutter machine.

[0011] Key operating parameters of the circular cutter are extracted from the preprocessed circular cutter contour image.

[0012] The control system compares the extracted key operating parameters of the circular cutter with preset standard parameter thresholds to determine whether there are any abnormalities in the operating status of the circular cutter and makes corresponding adjustments.

[0013] The beneficial effects of the present invention are as follows, compared with the prior art:

[0014] This invention employs an industrial camera to construct a visual acquisition system, combining an improved sub-pixel edge fitting algorithm and an improved method for extracting the outer circular contour coordinates of a circular cutter. This eliminates image noise and interference, improves the detection accuracy of key operating parameters, and enables the identification of defects such as minute wear and micro-cracks on the cutting edge of the circular cutter, solving the problem of low accuracy in manual inspection. The improved outer circular contour coordinate extraction method eliminates irrelevant areas and noise interference, improving coordinate extraction accuracy and avoiding parameter calculation errors caused by coordinate deviations. The improved sub-pixel edge fitting algorithm offers excellent fitting accuracy. The method for calculating the radial offset of the center of rotation can capture the offset of the center of rotation during the circular cutter's rotation, improving the accuracy of runout detection and enabling timely detection of minute runout anomalies. Through preset... The system employs standardized parameter thresholds and automated judgment methods to objectively assess the operating status of the circular cutter, eliminating the need for operator experience and effectively preventing missed or false detections caused by insufficient experience and fatigue during manual inspections. An industrial camera captures real-time images of the circular cutter's operation, eliminating the need for machine downtime and ensuring uninterrupted production. Simultaneously, it automatically outputs control messages for abnormal states, enabling wear repair and position correction, thus improving production efficiency. Compared to manual inspections, monitoring efficiency is significantly enhanced. Furthermore, the improved sub-pixel edge fitting algorithm reduces computational redundancy through positioning, further improving fitting efficiency. By monitoring parameters such as the circular cutter's runout and wear in real-time, anomalies are promptly detected and addressed, preventing equipment failures caused by continued operation of abnormal circular cutters and reducing equipment maintenance costs. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0016] Figure 1 This is an overall flowchart of the visual monitoring and control method for the circular knife machine described in this invention;

[0017] Figure 2 This is a flowchart of the method for extracting the actual diameter of a circular knife as described in this invention. Detailed Implementation

[0018] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0019] Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without inventive effort should fall within the scope of protection of this invention.

[0020] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0021] According to an embodiment of the present invention, in combination Figure 1 The flowchart shown illustrates a visual monitoring and control method for a circular knife machine according to the present invention, comprising:

[0022] S1. Build a vision acquisition system for the visual monitoring and control of the rotary cutter machine, that is, install an industrial camera next to the tool holder of the rotary cutter machine at the position corresponding to the rotary cutter. Note the following in this step:

[0023] In a preferred embodiment, the industrial camera is electrically connected to the control system of the rotary cutting machine; the field of view of the industrial camera covers the cutting edge, outer surface, and tool holder connection of the rotary cutting machine.

[0024] It should be noted that the vision acquisition system includes an industrial camera and the control system of the rotary cutting machine.

[0025] For example, the control system of a circular knife machine can be a PLC control system.

[0026] S2. After starting the circular saw machine, perform image acquisition and preprocessing on the vision acquisition system for the visual monitoring and control of the circular saw machine. Note the following in this step:

[0027] In a preferred embodiment, after the rotary cutter is started, the industrial camera acquires continuous color images of the rotary cutter during its operation at a preset frequency, and transmits the acquired color images to the control system of the rotary cutter. The control system of the rotary cutter preprocesses the color images, that is, it sequentially performs grayscale conversion, Gaussian filtering, and contrast enhancement using the CLAHE algorithm to eliminate image noise, improve the contrast between the rotary cutter outline and the background, and obtain a clear rotary cutter outline image.

[0028] It should be noted that the grayscale algorithm can use a weighted average, and S2 to S4 are all processed by the control system.

[0029] For example, the preset frequency is set according to specific requirements, such as 10 frames to 30 frames per second.

[0030] S3. Extract key operating parameters of the circular cutter from the preprocessed circular cutter contour image. Note that the following points should be noted in this step:

[0031] In a preferred embodiment, a method for extracting key operating parameters of a circular cutter from a preprocessed circular cutter contour image includes:

[0032] Based on the preprocessed circular cutter contour image, an improved connected component labeling algorithm and sub-pixel edge fitting algorithm are used to extract the key operating parameters of the circular cutter. The key operating parameters of the circular cutter include the actual diameter of the circular cutter, the wear of the cutting edge, the runout of the circular cutter, and the area of ​​surface defects.

[0033] like Figure 2 As shown, in a preferred embodiment, the method for extracting the actual diameter of the circular cutter specifically includes:

[0034] Traditional connected component labeling and contour tracking methods can only simply traverse edge pixels without filtering out irrelevant connected components such as the tool holder, and they do not optimize the extracted coordinates, which easily leads to coordinate repetition, resulting in discontinuities and large deviations in contour coordinates, affecting the fitting accuracy of subsequent diameter and center coordinates.

[0035] To address the aforementioned issues, this invention, based on a preprocessed circular knife contour image, employs an improved combination of connected component filtering, contour tracking, and coordinate optimization to extract the pixel coordinates of the outer circular contour of the circular knife. This overcomes the shortcomings of traditional methods, such as low extraction accuracy, susceptibility to noise interference, and coordinate discontinuities. Then, the least squares method is used to fit the circle equation to obtain the pixel diameter of the circular knife contour. The actual diameter of the circular knife is calculated using pixel equivalents. The improved connected component labeling algorithm includes an improved combination of connected component filtering, contour tracking, and coordinate optimization, the specific methods of which are shown below:

[0036] Connected component filtering is performed by using a four-neighbor labeling algorithm to traverse the circular knife contour image, label all connected components, and combine the geometric features of the circular knife to filter out the connected components corresponding to the outer circle of the circular knife, while removing irrelevant connected components such as the knife holder and image noise.

[0037] It should be noted that the method for filtering out the connected components of the corresponding outer circle of the circular cutter and removing irrelevant connected components such as the cutter holder and image noise includes:

[0038] First, the criteria for determining the geometric features, including the circular outline and preset size range of the circular blade, are clarified. The circular outline is determined based on the outline approximation of the connected components, with a preset outline approximation threshold of 0.8. The closer the outline approximation is to 1, the closer it is to a standard circle. The preset size range includes the pixel size range corresponding to the diameter of the new circular blade, combined with pixel equivalents. The preset ranges for the pixel area and pixel diameter of the outer circle of the circular blade are calculated. Then, each marked connected component is extracted one by one, and its key geometric parameters are calculated. These key geometric parameters include the contour approximation, pixel area, and circumscribed circle diameter of the connected component. The contour approximation is calculated by dividing the perimeter² of the connected component by (4π × area of ​​the connected component). The closer this ratio is to 1, the closer the connected component is to a circle. The pixel area is obtained by counting the total number of pixels within the connected component; the circumscribed circle diameter is obtained using the minimum circumscribed circle algorithm. Then, each... The key geometric parameters of the connected components are compared with preset judgment criteria to select connected components with a contour approximation of ≥0.8, a pixel area within a preset range of pixel area, and a circumscribed circle diameter within a preset range. This connected component is the connected component corresponding to the outer circle of the circular cutter. Irrelevant connected components are eliminated. Among them, the connected component corresponding to the cutter holder has the characteristics of a contour approximation of <0.5 and a pixel area much larger than the outer circle of the circular cutter. The connected component corresponding to image noise has the characteristics of a pixel area of ​​<10 pixels². All such irrelevant connected components that do not meet the selection criteria are eliminated to ensure that subsequent contour tracking is only for the outer circle region of the circular cutter.

[0039] Furthermore, a new circular blade is a circular blade that has never been used; based on the pixel size range corresponding to the diameter of the new circular blade, combined with pixel equivalent... The methods for calculating the preset range of pixel area and pixel diameter of the connected region of the outer circle of the circular knife include:

[0040] Based on the actual diameter of the new round knife To calculate the fluctuation range of the actual diameter, the actual diameter fluctuation range of a new round cutter is usually calculated as follows: [ ×(1-1%), [×(1+1%)], ×(1-1%) is the minimum actual diameter. ×(1+1%) is the maximum actual diameter; this gives the preset range of pixel diameter. , Minimum pixel diameter =Minimum actual diameter ÷ Maximum pixel diameter =Maximum actual diameter ÷ The preset range for pixel area is [ , Minimum pixel area =π×( ÷2)², maximum pixel area =π×( ÷2)².

[0041] Contour tracking is performed, which involves using a chain code tracking algorithm based on the connected components of the filtered outer circle of the circular cutter. Starting from the edge starting point of the connected components of the outer circle of the circular cutter, the algorithm sequentially traverses the edge pixels of the connected components and records the x-coordinate of each edge pixel. and ordinate The set of pixel coordinates that form the outer circle contour of the circular knife. , This represents the total number of pixels on the outer circular outline of the circular knife.

[0042] It should be noted that the starting point of the edge of the connected region of the outer circle of the circular knife is selected as the leftmost pixel of the connected region of the outer circle of the circular knife.

[0043] Coordinate optimization is performed, which involves deduplicating the pixel coordinates of the extracted outer circle contour of the circular knife and removing duplicate pixels.

[0044] The actual diameter of the circular cutter is calculated based on the pixel coordinates of the outer circular contour.

[0045] In a preferred embodiment, the mathematical expression for calculating the actual diameter of the circular cutter based on the pixel coordinates of its outer circular contour is as follows:

[0046] ;

[0047] In this mathematical expression, This is the actual diameter of the circular cutter; The pixel diameter of the outer circle of the circular cutter is the number of pixels corresponding to the diameter of the fitted circle of the outer circle of the circular cutter. In pixel equivalents The unit is mm / pixel, which is the actual physical length corresponding to each pixel.

[0048] It should be noted that the pixel diameter of the outer circle contour of the circular cutter is obtained by using the least squares method to fit the circle equation to the coordinates of the contour pixel points after coordinate optimization, and then obtaining the standard equation of the fitted circle. , ( , () represents the pixel coordinates of the center of the fitted circle. The fourth step is to calculate the pixel radius of the fitted circle. Based on the relationship between the diameter and radius of a circle, the pixel diameter of the outer circle contour of the circular cutter is thus determined. =2 ; The method of obtaining the calibration block can be to take an image of a standard calibration block of known size using an industrial camera, measure the pixel size of the standard calibration block in the calibration block image, and calculate the calibration block size. = Actual size of the standard calibration block / pixel size of the calibration block, where pixel size is the pixel length.

[0049] Furthermore, the improved combination of connected component filtering, contour tracking, and coordinate optimization can filter out irrelevant connected components such as tool holders, and optimize the extracted coordinates to avoid the problem of discontinuous contour coordinates and large deviations caused by coordinate repetition, which affects the fitting accuracy of subsequent diameter and center coordinates.

[0050] In a preferred embodiment, the method for extracting the wear amount of the cutting edge specifically includes:

[0051] The wear amount of the cutting edge is defined as the maximum distance between the actual contour of the circular cutter's cutting edge and the standard cutting edge contour. This is calculated by comparing the sub-pixel coordinates of the current cutting edge contour of the circular cutter with the sub-pixel coordinates of the standard cutting edge contour. The maximum pixel distance between the two is then calculated, and combined with the pixel equivalent, to obtain the actual wear amount. This is essentially the mathematical expression for the wear amount of the cutting edge, which is:

[0052] ;

[0053] In this mathematical expression, This represents the actual wear of the cutting edge; The maximum pixel distance between the current cutting edge contour and the standard cutting edge contour is determined by an improved sub-pixel edge fitting algorithm. This algorithm extracts the sub-pixel coordinates of both the current and standard cutting edge contours, calculates the pixel distance between these sub-pixel coordinates using the Euclidean distance formula, and takes the maximum value of this distance as the maximum pixel distance. ; In pixel equivalents.

[0054] It should be noted that traditional subpixel edge fitting algorithms, such as the Zernike moment algorithm or the Sobel subpixel algorithm, are weak in their anti-interference ability when applied to the extraction of the cutting edge contour of a circular knife. They are easily affected by image blurring and residual reflections caused by the high-speed rotation of the circular knife, resulting in misjudgment of the cutting edge edge. Furthermore, their fitting accuracy is insufficient, as they use quadratic polynomial fitting, which cannot adapt to the contour irregularities caused by minor wear on the cutting edge.

[0055] Therefore, improvements are made, and the improved sub-pixel edge fitting algorithm includes:

[0056] Coarse localization of the cutting edge region is performed by using the Sobel operator to detect edges in the connected components of the outer circle of the circular knife. The Sobel operator smooths noise based on the weighted difference of the gray levels of the pixels above and below and left and right neighbors of the connected components of the outer circle of the circular knife, while providing clear edge direction information. This accurately locates the cutting edge points of the circular knife's cutting edge contour, reducing the amount of subsequent fitting calculations. Specifically, when extracting the sub-pixel coordinates of the current cutting edge contour, the connected components of the outer circle of the circular knife are the connected components of the outer circle of the current circular knife; when extracting the sub-pixel coordinates of the standard cutting edge contour, the connected components of the outer circle of the circular knife are the connected components of the outer circle of the new circular knife. In addition, the Sobel operator is mainly used to detect areas with obvious gray level changes. Because the cutting edge is sharp and thin, the gray level change between its gray value and the circular knife body is more drastic than that of the outer circle contour. Therefore, the Sobel operator has an advantage in locating the cutting edge points of the circular knife's cutting edge contour.

[0057] Edge point expansion is performed, that is, the edge points of the cutting edge obtained by coarse positioning of the cutting edge area are expanded by 5 pixels along the edge normal direction.

[0058] In a preferred embodiment, the method for edge point expansion specifically includes:

[0059] First, based on the coarse localization of the cutting edge region, the edge direction vector of each edge point is calculated. The method for calculating the edge direction vector of each edge point is as follows:

[0060] The Sobel operator includes a horizontal direction operator. and vertical direction operator ,use and Calculate the horizontal gradient value at each point on the cutting edge. and vertical gradient value The gradient direction vector is determined by the gradient value; the gradient direction vector is a vector... The expression is The direction of the gradient direction vector is the edge gradient direction, and the edge direction is perpendicular to the gradient direction. Therefore, the edge direction vector is obtained by rotating the gradient vector by 90°. If the gradient vector is... Then the edge direction vector is or These two correspond to the two perpendicular directions of the gradient vector, respectively.

[0061] Then, the edge normal direction of the edge point is determined, which is perpendicular to the edge direction vector of that edge point. That is, the edge normal direction is consistent with the edge gradient direction. The edge normal direction is divided into positive normal and negative normal. The positive normal is the direction of gradient increase, and the negative normal is the direction of gradient decrease. Using this edge point as a reference point, pixels are uniformly expanded along the positive and negative normal directions. The specific methods for uniformly expanding pixels include:

[0062] First, determine the coordinates of this edge point, which will serve as the reference point. Combined with the already determined normal direction vector , That is , That is , That is Next, calculate the unit normal vector, which is the normal direction vector. Normalization is performed to obtain the unit normal vector. The normalization formula is Ensure the expansion step size is uniform; then determine the coordinates of the expanded pixels, using the reference point as an example. Centered on the x-coordinate, with its index set to 0, along the positive unit normal, the coordinates of two extended pixels are calculated sequentially with a step size of 1 pixel. The coordinates of these two extended pixels are respectively... and , The index of the x-coordinate is set to 1. The x-coordinate index is set to 2; along the negative unit normal, with a step size of 1 pixel, the coordinates of the two extended pixels are calculated sequentially. The coordinates of these two extended pixels are respectively... and , The index of the x-coordinate is set to -1. The index of the horizontal coordinate is set to -2; then coordinate calibration is performed, which involves rounding the calculated coordinates of the extended pixels. Since pixel coordinates must be integers, the specific rounding method is to round to the nearest integer. This method can preserve the actual positional accuracy of the extended pixels to the greatest extent and avoid inaccurate grayscale value acquisition due to rounding deviation; finally, 5 consecutive pixels are formed, covering the edge of the blade and the small areas on both sides; the grayscale values ​​of these 5 pixels are obtained to provide a complete grayscale distribution data basis for subsequent accurate fitting, and the coordinates of the reference point are also regarded as extended pixels.

[0063] Next, a fifth-order orthogonal polynomial fitting is performed, that is, based on the gray-level distribution features of the five extended pixels, a gray-level distribution fitting function is obtained. ;

[0064] In a preferred embodiment, compared to a traditional quadratic polynomial, a quintic polynomial can better adapt to irregular changes in the cutting edge profile, improving fitting accuracy; the fitted function obtained is the gray-level distribution fitting function of the quintic polynomial. The methods specifically include:

[0065] Determine the basic fitting parameters, that is, based on the indices of the horizontal coordinates of the 5 extended pixels (-2, -1, 0, 1, 2), determine the gray value corresponding to each horizontal coordinate index. =[ , , , , ],in The x-axis index is The grayscale values ​​of the extended pixels, =-2,-1,0,1,2, the fitting objective is to construct... This minimizes the sum of squared errors between the function and the actual grayscale value;

[0066] Next, we select the basis {1,} of the ordinary quintic polynomial. , , , , }, which is then transformed into an orthogonal polynomial basis using the Schmidt orthogonalization method { , , , , , The orthogonalization process satisfies The orthogonal polynomials are shown below:

[0067] Zero-order orthogonal polynomial basis: =1;

[0068] First-order orthogonal polynomial basis Substituting into the calculation, we get = ;

[0069] Quadratic orthogonal polynomial basis Substituting into the calculation, we get ;

[0070] Cubic orthogonal polynomial basis Substituting into the calculation, we get ;

[0071] Quadratic orthogonal polynomial base Substituting into the calculation, we get ;

[0072] Fifth-order orthogonal polynomial basis Substituting into the calculation, we get ;

[0073] Fitting function to grayscale distribution It is represented as a linear combination of orthogonal polynomial bases, i.e. , , , , , , All are fitting coefficients, and therefore, based on the least squares method, the sum of squared errors is calculated. To minimize this, and considering the orthogonality of orthogonal polynomials, the fitting coefficients can be calculated using the following formula:

[0074] , =0,1,2,3,4,5 , , , and These are the indices of the x-coordinates of the extended pixels: -2, -1, 0, 1, and 2. The x-axis index is The grayscale value of the extended pixels;

[0075] The obtained fitting coefficients , , , , , Substitution This function can accurately characterize the gray-level distribution pattern in the normal direction of the cutting edge, providing a basis for subsequent sub-pixel coordinate extraction.

[0076] Based on the extreme value solution conditions, the fitted function Find the second differential and set it equal to zero. Solve for the result. The value represents the sub-pixel offset of the cutting edge. Therefore, the sub-pixel coordinates of the cutting edge contour are determined as follows: , , All the sub-pixel coordinates of the cutting edge contour constitute its sub-pixel coordinate set, thus achieving sub-pixel level precise positioning of the cutting edge contour.

[0077] Extract the sub-pixel coordinate sets of the current cutting edge contour and the standard cutting edge contour respectively. Then, take the sub-pixel coordinate set with the larger number of sub-pixel coordinates from both sets, and use the center pixel coordinates of the fitted circle (the center of the circular tool) as the reference point. , Based on the first subpixel coordinate, sort the subpixel coordinates in ascending order of the angle between the subpixel coordinate and the center of the circle, resulting in an ordered coordinate set. , The first sub-pixel coordinate in the set with the most sub-pixel coordinates Sub-pixel coordinates, , Let be the number of sub-pixel coordinates in this set of sub-pixel coordinates; then, for another set of sub-pixel coordinates, sort the sub-pixel coordinates in ascending order of the angle between the sub-pixel coordinates and the center of the circle, resulting in a second ordered set of coordinates. , For the first sub-pixel coordinate in this other set of sub-pixel coordinates Sub-pixel coordinates, , This represents the number of subpixel coordinates in the other set of subpixel coordinates, where the angle between the subpixel coordinate and the center of the circle is the subpixel coordinate. , ) and the center pixel coordinates of the fitted circle ( , The angle between the line connecting (x, y) and the positive x-axis. The sub-pixel coordinates are sorted in ascending order of the angle between the sub-pixel coordinates and the center of the circle, ensuring that the sub-pixel coordinate points are evenly distributed along the cutting edge contour and conform to the circular characteristics of the cutting edge contour. The filtering interval is calculated based on the number of sub-pixel coordinates in the two sets of sub-pixel coordinates. , =round( / `round()` is a rounding function that rounds the coordinates according to an ordered set of coordinates. The sorted subpixel coordinates are ordered, starting from the first subpixel coordinate, every... One sub-pixel is deleted at each point until the number of sub-pixel coordinates in the current cutting edge contour set is the same as the number in the standard cutting edge contour set. Then, the resulting ordered coordinate set is sorted again according to the ascending order of the angles between the sub-pixel coordinates and the center of the circle. This ordered coordinate set is the result of the deletion process. ,in For the first ordered set of coordinates after deletion Sub-pixel coordinates, , and For the one-to-one correspondence of subpixel coordinates, the Iterative Closest Point (ICP) algorithm is used to rotate and translate the deleted ordered coordinate set one and ordered coordinate set two to form the aligned subpixel coordinate set of the current cutting edge contour and the subpixel coordinate set of the standard cutting edge contour. The order of the subpixel coordinates after the rotation and translation of the deleted ordered coordinate set one and ordered coordinate set two remains unchanged from the order of the subpixel coordinates before the rotation and translation. In addition, the one-to-one correspondence of the corresponding subpixel coordinates avoids distance calculation errors caused by contour offset.

[0078] It should be noted that the improved subpixel edge fitting algorithm can effectively resist image blurring and residual reflection interference caused by the high-speed rotation of the circular blade, and will not misjudge the edge of the blade, ensuring the accuracy of the blade contour extraction; high fitting accuracy: abandoning the traditional quadratic polynomial fitting, a fifth-order orthogonal polynomial fitting is adopted, which can better adapt to the contour irregularity caused by the slight wear of the blade, improve the fitting accuracy, and improve the extraction accuracy of the coordinates of the blade contour to meet the needs of high-precision monitoring.

[0079] In a preferred embodiment, the pixel distance between corresponding sub-pixel coordinates is calculated using the Euclidean distance formula, and the maximum value of this pixel distance is taken as... The methods specifically include:

[0080] Based on the sub-pixel coordinate sets of the aligned current cutting edge contour and the standard cutting edge contour, the pixel distance between the corresponding sub-pixel coordinates is calculated using the Euclidean distance formula, and the maximum value is taken as the maximum pixel distance between the current circular cutter's cutting edge contour and the standard cutting edge contour. , The mathematical expression is:

[0081] ;

[0082] In this mathematical expression, The set of sub-pixel coordinates of the aligned current cutting edge profile and the set of sub-pixel coordinates of the standard cutting edge profile. The pixel distance between corresponding sub-pixel coordinates =1,2,..., ; The first subpixel coordinate in the current cutting edge profile set Sub-pixel coordinates; The first in the sub-pixel coordinate set Subpixel coordinates, max() is the MAX function.

[0083] In a preferred embodiment, the method for extracting the runout of a circular cutter specifically includes:

[0084] The runout of a circular cutter refers to the maximum radial displacement of a point on the outer surface of the cutter during high-speed rotation, reflecting the stability of the cutter's rotation. It is determined by acquiring consecutive frame images within one revolution of the cutter, extracting the center pixel coordinates of the cutter's outer contour in each frame, calculating the maximum radial offset of the center pixel coordinates, and combining this with pixel equivalents to obtain the runout. The mathematical expression for the runout of the circular cutter is as follows:

[0085] ;

[0086] This represents the runout of the circular cutter; This represents the maximum radial offset of the center pixel coordinates within one revolution of the circular cutter. This is achieved by acquiring consecutive frame images of the circular cutter rotating in one revolution, extracting the center pixel coordinates of the outer circle contour of the cutter in each frame, combining them with the theoretical center coordinates, and using the Euclidean distance formula to calculate the radial offset of the center in each frame. The maximum radial offset is then taken as the value of the offset. ;

[0087] The methods for obtaining it specifically include:

[0088] The preset frequency of the industrial camera is determined based on the rotation speed of the circular cutter to ensure that several consecutive frame images within one revolution of the circular cutter are captured.

[0089] It should be noted that the higher the rotary cutter speed, the higher the preset frequency. The unit of rotary cutter speed here is r / min. The rotation period is calculated based on the rotary cutter speed. The unit of rotation period is seconds. The rotation period is the time required for the rotary cutter to rotate one revolution. Rotation period = 60 / rotary cutter speed. The preset frequency is ≥ preset frame number / rotation period.

[0090] For each frame of the continuous frame image, after preprocessing using the aforementioned preprocessing method, the improved combination of connected component filtering, contour tracking, and coordinate optimization is then applied to each preprocessed frame image to extract the pixel coordinates of the outer circle contour of the circular knife.

[0091] For the pixel coordinates of the outer circular contour extracted from each frame of the image, the least squares method is used to fit the circle equation to the pixel coordinates of the outer circular contour of each frame of the image after coordinate optimization, and the standard equation of the fitted circle for each frame of the image is obtained. , For the first The pixel coordinates of the center of the outer circle of the circular knife in the frame image. For the first The pixel radius of the outer circle contour of the circular knife in the frame image;

[0092] The pixel coordinates corresponding to the center of the cutter axis of the circular cutter are set as the theoretical center coordinates. The theoretical center coordinates are obtained by calibration using a circular standard calibration block. It is preset in the control system as a reference for calculating radial offset;

[0093] It should be noted that the theoretical center coordinates are obtained by calibration using a circular standard calibration block. The methods may include:

[0094] A circular standard calibration block is fixed on the cutter shaft of the circular cutter machine and concentric with the center of the cutter shaft, ensuring that the center of the circular calibration block is completely coincident with the center of the cutter shaft. An industrial camera captures an image of the circular standard calibration block and transmits it to the control system. The control system preprocesses the image of the circular standard calibration block using the aforementioned preprocessing method. Then, it applies the aforementioned improved combination method of connected component filtering, contour tracking, and coordinate optimization to extract the pixel coordinates of the outer circular contour of the circular standard calibration block. The pixel coordinates of the outer circular contour are then fitted with a circle equation using the least squares method to obtain the standard equation of the fitted circle for the image of the circular standard calibration block. , The coordinates of the center pixel of the outer circle of the circular standard calibration block in the image. The pixel radius of the outer circle of the circular standard calibration block.

[0095] The Euclidean distance formula is used to calculate the center pixel coordinates of the outer circular contour of the circular knife in each frame of the image. coordinates of the theoretical center The distance is the first... Radial pixel offset of the center pixel coordinates of the frame image , ;

[0096] Based on the radial pixel offset of the center pixel coordinates of all frames within one revolution of the circular cutter. ,Pick The maximum value in the value is taken as the maximum radial offset of the center pixel coordinates during one revolution of the circular cutter. , , This refers to the number of consecutive frames in the acquired image within one revolution of the circular cutter.

[0097] In a preferred embodiment, the method for extracting the area of ​​surface defects specifically includes:

[0098] The surface defect area refers to the area of ​​defects on the end face of a circular cutter, such as cracks, scratches, and rust. It is calculated by extracting the number of pixels in the defect region within the connected domain of the outer circle of the cutter, and combining this with the area conversion relationship corresponding to the pixel equivalent. The mathematical expression for the surface defect area is:

[0099] ;

[0100] In this mathematical expression, The area of ​​the surface defect. The number of pixels in the defective area. The method for obtaining the defect includes: in the connected region corresponding to the outer circle of the circular cutter, removing the connected regions corresponding to the outer circle contour of the circular cutter and the current cutting edge contour of the circular cutter to avoid misjudging the normal contour of the circular cutter as a defect; then, taking the remaining connected regions in the connected region corresponding to the outer circle of the circular cutter as the defect region; counting the total number of pixels contained in all defect regions to obtain the total number of pixels in the defect region. .

[0101] It should be noted that the pixel coordinates of the standard cutting edge contour are the same as the pixel coordinates of the cutting edge contour of the new circular knife. The industrial system pre-captures color images of the new circular knife and transmits them to the control system. The control system also preprocesses the color images of the new circular knife to obtain a clear image of the circular knife contour. The four-neighbor labeling algorithm is also used on the clear image of the new circular knife contour to traverse the image and label all connected components. Combined with the geometric features of the circular knife, the connected components of the outer circle of the new circular knife are selected.

[0102] S4. The control system compares the extracted key operating parameters of the circular cutter with preset standard parameter thresholds to determine if there are any abnormalities in the operating status of the circular cutter and makes corresponding adjustments. It should be noted that in this step:

[0103] In a preferred embodiment, the method for the control system to compare the extracted key operating parameters of the circular cutter with preset standard parameter thresholds to determine whether there is an abnormality in the operating state of the circular cutter and to perform corresponding adjustments specifically includes:

[0104] The standard thresholds for key operating parameters of the circular cutter are preset, including the threshold for the actual diameter of the circular cutter. Threshold for the amount of wear on the cutting edge Threshold for runout of the circular cutter and the threshold of surface defect area The threshold of the actual diameter of the circular cutter The cut-off diameter of the circular cutter, determined based on specific machining accuracy requirements, and the threshold for edge wear. The maximum allowable wear of the circular cutter is determined based on the specific material being processed and the required precision; the threshold value for the runout of the circular cutter is also specified. The maximum allowable runout of the circular cutter, determined based on the specific stability requirements of the circular cutter machine operation, is the threshold value for the surface defect area. The maximum allowable defect area for a circular cutter, determined based on specific machining quality requirements;

[0105] exist At that time, the operating status of the circular cutter was determined to be abnormal in diameter; At that time, the operating state of the circular cutter was determined to be abnormal edge wear; At that time, the running state of the circular cutter was determined to be abnormal, indicating an abnormal runout. At that time, the operating status of the circular cutter is determined to be abnormal due to surface defects;

[0106] If any of the following abnormalities are present: abnormal diameter, abnormal cutting edge wear, abnormal runout, or abnormal surface defects, the operating state of the circular cutter is determined to be abnormal.

[0107] When there is an abnormality in the operation of the circular cutter, the control system makes corresponding adjustments. After the adjustment is completed, the circular cutter machine is restarted. After returning to the start of the circular cutter machine in S2, the visual acquisition system for visual monitoring and control of the circular cutter machine performs image acquisition and preprocessing.

[0108] It should be noted that the control system can perform the following adjustments: stop the operation of the circular cutter machine, and remind the operator to perform maintenance or repair by sounding a buzzer and displaying corresponding inspection or maintenance messages on a connected display screen. For example, under abnormal diameter conditions, the corresponding inspection or maintenance message may be to instruct the operator to replace the circular cutter; under abnormal cutting edge wear conditions, the corresponding inspection or maintenance message may be to instruct the operator to grind and repair the cutting edge; under abnormal runout conditions, the corresponding inspection or maintenance message may be to instruct the operator to adjust the tightness of the tool holder fixing bolts and correct the installation position of the circular cutter; under abnormal surface defects conditions, the corresponding inspection or maintenance message may be to instruct the operator to grind and repair the end face of the circular cutter or replace the circular cutter.

[0109] The purpose of this invention is to provide a visual monitoring and control method for a circular knives machine, which overcomes the shortcomings of manual inspection of circular knives machines in the prior art, such as low detection accuracy, reliance on experience, easy omissions and false detections, and low efficiency. This method enables automated, real-time, and high-precision monitoring of the operating status of the circular knives machine and simultaneously completes abnormal control, ensuring stable operation of the circular knives machine and improving the processing quality and production efficiency of the circular knives machine.

[0110] It should be recognized that embodiments of the present invention may be implemented or carried out by computer hardware, a combination of hardware and software, or by computer instructions stored in a non-transitory computer-readable storage medium.

[0111] The method can be implemented using standard programming techniques, including a non-transitory computer-readable storage medium configured with a computer program in the computer program, wherein the storage medium is configured such that the computer operates in a specific and predefined manner.

[0112] Each program can be implemented in a high-level procedural or object-oriented programming language to communicate with the computer system; however, if necessary, the program can be implemented in assembly or machine language.

[0113] In any case, the language can be either compiled or interpreted.

[0114] Furthermore, for this purpose, the program can run on a programmed application-specific integrated circuit.

[0115] The processes described herein (or variations and / or combinations thereof) can be executed under the control of one or more computer systems configured with executable instructions, and can be implemented by hardware or a combination thereof as code (e.g., executable instructions, one or more computer programs, or one or more applications) that commonly executes on one or more processors. The computer program includes a plurality of instructions executable by one or more processors.

[0116] Furthermore, the method can be implemented in any suitable computing platform, including but not limited to personal computers, minicomputers, mainframes, workstations, networked or distributed computing environments, standalone or integrated computer platforms, or in communication with charged particle tools or other imaging devices.

[0117] Various aspects of the present invention can be implemented in machine-readable code stored on a non-transitory storage medium or device, whether portable or integrated into a computing platform, such as a hard disk, optical read and / or write storage medium, RAM, ROM, etc., such that it can be read by a programmable computer, and when the storage medium or device is read by the computer, it can be used to configure and operate the computer to perform the processes described herein.

[0118] Furthermore, machine-readable code, or parts thereof, can be transmitted via wired or wireless networks.

[0119] When such media includes instructions or programs that combine with a microprocessor or other data processor to implement the steps described above, the invention described herein includes these and other different types of non-transitory computer-readable storage media.

[0120] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A visual monitoring and control method for a circular knife machine, characterized in that, include: Build a vision acquisition system for vision monitoring and control of the rotary cutter machine, that is, install an industrial camera next to the tool holder of the rotary cutter machine at the position corresponding to the rotary cutter. After the rotary cutter is started, the vision acquisition system for the visual monitoring and control of the rotary cutter performs image acquisition and preprocessing. That is, after the rotary cutter is started, the industrial camera acquires continuous color images of the rotary cutter during its operation at a preset frequency, and transmits the acquired color images to the control system of the rotary cutter. The control system of the rotary cutter performs preprocessing on the color images, that is, it sequentially performs grayscale conversion, Gaussian filtering, and contrast enhancement by applying the CLAHE algorithm to obtain the outline image of the rotary cutter. The key operating parameters of the circular cutter are extracted from the preprocessed circular cutter contour image. Specifically, based on the preprocessed circular cutter contour image, an improved connected component labeling algorithm and a sub-pixel edge fitting algorithm are used to extract the key operating parameters of the circular cutter. These key operating parameters include the actual diameter of the circular cutter, the wear amount of the cutting edge, the runout of the circular cutter, and the area of ​​surface defects. The method for extracting the actual diameter of the circular cutter specifically includes: Connected component filtering is performed by using a four-neighbor labeling algorithm to traverse the circular knife contour image, label all connected components, and combine the geometric features of the circular knife to filter out the connected components of the corresponding outer circle of the circular knife and remove irrelevant connected components. Contour tracking is performed, which involves using a chain code tracking algorithm based on the connected components of the filtered outer circle of the circular cutter. Starting from the edge starting point of the connected components of the outer circle of the circular cutter, the algorithm sequentially traverses the edge pixels of the connected components and records the x-coordinate of each edge pixel. and ordinate The set of pixel coordinates that form the outer circle contour of the circular knife. , This represents the total number of pixels on the outer circular outline of the circular knife. Coordinate optimization is performed, which involves deduplicating the pixel coordinates of the extracted outer circle contour of the circular knife and removing duplicate pixels. The actual diameter of the circular cutter is calculated based on the pixel coordinates of the outer circular contour of the cutter. Methods for determining the wear of a cutting edge include: The mathematical expression for the amount of wear on the cutting edge is as follows: ； In this mathematical expression, This represents the actual wear of the cutting edge; The maximum pixel distance between the current cutting edge contour and the standard cutting edge contour is determined by an improved sub-pixel edge fitting algorithm. This algorithm extracts the sub-pixel coordinates of both the current and standard cutting edge contours, calculates the pixel distance between these sub-pixel coordinates using the Euclidean distance formula, and takes the maximum value of this pixel distance as the maximum pixel distance. ; In pixel equivalents; The control system compares the extracted key operating parameters of the circular cutter with preset standard parameter thresholds to determine whether there are any abnormalities in the operating status of the circular cutter and makes corresponding adjustments.

2. The visual monitoring and control method for a circular knife machine according to claim 1, characterized in that, The industrial camera is electrically connected to the control system of the rotary cutter; the field of view of the industrial camera covers the cutting edge, outer surface, and tool holder connection of the rotary cutter.

3. The visual monitoring and control method for a circular knife machine according to claim 2, characterized in that, The mathematical expression for calculating the actual diameter of the circular cutter based on the pixel coordinates of its outer circular contour is as follows: ； In this mathematical expression, This is the actual diameter of the circular cutter; The pixel diameter of the outer circle of the circular knife; In pixel equivalents.

4. The visual monitoring and control method for a circular knife machine according to claim 3, characterized in that, The improved sub-pixel edge fitting algorithm includes: Coarse localization of the cutting edge region is performed by using the Sobel operator to perform edge detection on the connected region of the outer circle of the circular cutter, thereby locating the cutting edge points of the cutting edge contour of the circular cutter. Edge point expansion is performed, that is, the edge points of the cutting edge obtained by coarse positioning of the cutting edge area are expanded by 5 pixels along the edge normal direction. Next, a fifth-order orthogonal polynomial fitting is performed, that is, based on the gray-level distribution features of the five extended pixels, a gray-level distribution fitting function is obtained. ; For the fitted function Find the second differential and set it equal to zero. Solve for the result. The value represents the sub-pixel offset of the cutting edge. Therefore, the sub-pixel coordinates of the cutting edge contour are determined as follows: , , The sub-pixel coordinates of the blade edge contour constitute its sub-pixel coordinate set. Extract the sub-pixel coordinate sets of the current cutting edge contour and the standard cutting edge contour respectively. Then, use the sub-pixel coordinate set with the larger number of sub-pixel coordinates from both sets to fit the center pixel coordinates of the circle. , Based on the first subpixel coordinate, sort the subpixel coordinates in ascending order of the angle between the subpixel coordinate and the center of the circle, resulting in an ordered coordinate set. , The first sub-pixel coordinate in the set with the larger number of sub-pixel coordinates Sub-pixel coordinates, , Let be the number of sub-pixel coordinates in this set of sub-pixel coordinates; then, for another set of sub-pixel coordinates, sort the sub-pixel coordinates in ascending order of the angle between the sub-pixel coordinates and the center of the circle, resulting in a second ordered set of coordinates. , For the first sub-pixel coordinate in this other set of sub-pixel coordinates Sub-pixel coordinates, , This represents the number of subpixel coordinates in the other set of subpixel coordinates, where the angle between the subpixel coordinate and the center of the circle is the subpixel coordinate. , ) and the center pixel coordinates of the fitted circle ( , The angle between the line connecting (x, y) and the positive x-axis. ; Calculate the filtering interval based on the number of sub-pixel coordinates in the two sub-pixel coordinate sets. , =round( / `round()` is a rounding function that rounds the coordinates according to an ordered set of coordinates. The sorted subpixel coordinates are ordered, starting from the first subpixel coordinate, every... One sub-pixel is deleted at each point until the number of sub-pixel coordinates in the current cutting edge contour set is the same as the number in the standard cutting edge contour set. Then, the resulting ordered coordinate set is sorted again according to the ascending order of the angles between the sub-pixel coordinates and the center of the circle. This ordered coordinate set is the result of the deletion process. ,in For the first ordered set of coordinates after deletion Sub-pixel coordinates, , and For subpixel coordinates that correspond one-to-one, the Iterative Nearest Point (ICP) algorithm is used to rotate and translate the ordered coordinate set one after deletion and the ordered coordinate set two to align them, thereby forming the aligned subpixel coordinate set of the current cutting edge contour and the subpixel coordinate set of the standard cutting edge contour.

5. The visual monitoring and control method for a circular knife machine according to claim 4, characterized in that, Methods for edge point expansion include: First, based on the coarse localization of the cutting edge region, the edge direction vector of each edge point is calculated. The method for calculating the edge direction vector of each edge point is as follows: The Sobel operator includes a horizontal direction operator. and vertical direction operator ,use and Calculate the horizontal gradient value at each point on the cutting edge. and vertical gradient value The vector of the gradient direction vector The expression is The edge direction vector is obtained by rotating the gradient vector by 90°; Then, the edge normal direction of the edge point is determined, that is, the edge normal direction is consistent with the edge gradient direction, and the edge normal direction is divided into positive normal and negative normal. Taking this edge point as the reference point, pixels are uniformly expanded along the positive normal and negative normal directions. The specific methods for uniformly expanding pixels include: First, determine the coordinates of the edge point that will serve as the reference point. Combined with the already determined normal direction vector , That is , That is , That is Next, calculate the unit normal vector, which is the normal direction vector. Normalization is performed to obtain the unit normal vector. The normalization formula is Then determine the coordinates of the extended pixels, using the reference point. Centered on the x-coordinate, with its index set to 0, along the positive unit normal, the coordinates of two extended pixels are calculated sequentially with a step size of 1 pixel. The coordinates of these two extended pixels are respectively... and , The index of the x-coordinate is set to 1. The x-coordinate index is set to 2; along the negative unit normal, with a step size of 1 pixel, the coordinates of the two extended pixels are calculated sequentially. The coordinates of these two extended pixels are respectively... and , The index of the x-coordinate is set to -1. The index of the horizontal coordinate is set to -2; then coordinate calibration is performed, which involves rounding the calculated coordinates of the extended pixels. The grayscale distribution fitting function is obtained by fitting. The methods specifically include: Determine the basic fitting parameters, that is, based on the indices of the horizontal coordinates of the 5 extended pixels (-2, -1, 0, 1, 2), determine the gray value corresponding to each horizontal coordinate index. =[ , , , , ],in The x-axis index is The grayscale value of the extended pixels, =-2,-1,0,1,2; Next, we select the basis {1,} of the ordinary quintic polynomial. , , , , }, which is then transformed into an orthogonal polynomial basis using the Schmidt orthogonalization method { , , , , , The orthogonalization process satisfies The orthogonal polynomials are shown below: Zero-order orthogonal polynomial basis: =1; First-order orthogonal polynomial base Substituting into the calculation, we get = ; Quadratic orthogonal polynomial basis Substituting into the calculation, we get ; Cubic orthogonal polynomial basis Substituting into the calculation, we get ; Quadratic orthogonal polynomial base Substituting into the calculation, we get ; Fifth-order orthogonal polynomial basis Substituting into the calculation, we get ; Fitting function to grayscale distribution It is represented as a linear combination of orthogonal polynomial bases, i.e. , , , , , , These are all fitting coefficients, which are calculated using the following formula: , =0,1,2,3,4,5 , , , and These are the indices of the x-coordinates of the extended pixels: -2, -1, 0, 1, and 2. The x-axis index is The grayscale value of the extended pixels; The obtained fitting coefficients , , , , , Substitution ; The pixel distance between the corresponding sub-pixel coordinates is calculated using the Euclidean distance formula, and the maximum value of this pixel distance is taken as the mean. The methods specifically include: Based on the sub-pixel coordinate sets of the aligned current cutting edge contour and the standard cutting edge contour, the pixel distance between the corresponding sub-pixel coordinates is calculated using the Euclidean distance formula, and the maximum value is taken as the maximum pixel distance between the current circular cutter's cutting edge contour and the standard cutting edge contour. , The mathematical expression is: ； In this mathematical expression, The set of sub-pixel coordinates of the aligned current cutting edge profile and the set of sub-pixel coordinates of the standard cutting edge profile. The pixel distance between corresponding sub-pixel coordinates =1,2,..., ; The first subpixel coordinate in the current cutting edge profile set Sub-pixel coordinates; The first in the sub-pixel coordinate set Subpixel coordinates.

6. The visual monitoring and control method for a circular knife machine according to claim 5, characterized in that, Methods for extracting the runout of a circular cutter include: By acquiring consecutive frame images of the circular cutter during one revolution, the center pixel coordinates of the outer circle contour of the circular cutter in each frame are extracted. The maximum radial offset of the center pixel coordinates is calculated, and the runout of the circular cutter is obtained by combining the pixel equivalent. The mathematical expression for the runout of the circular cutter is as follows: ； This represents the runout of the circular cutter; This represents the maximum radial offset of the center pixel coordinates within one revolution of the circular cutter. The methods for obtaining it specifically include: The preset frequency of the industrial camera is determined based on the rotation speed of the circular cutter to ensure that several consecutive frame images within one revolution of the circular cutter are captured. For each frame in a series of images, a preprocessing method is used to preprocess the image. Then, an improved combination of connected component filtering, contour tracking, and coordinate optimization is applied to each preprocessed frame to extract the pixel coordinates of the outer circle contour of the circular knife. For the pixel coordinates of the outer circular contour extracted from each frame of the image, the least squares method is used to fit the circle equation to the pixel coordinates of the outer circular contour of each frame of the image after coordinate optimization, and the standard equation of the fitted circle for each frame of the image is obtained. , For the first The pixel coordinates of the center of the outer circle of the circular knife in the frame image. For the first The pixel radius of the outer circle contour of the circular knife in the frame image; The pixel coordinates corresponding to the center of the cutter axis of the circular cutter are set as the theoretical center coordinates. The theoretical center coordinates are obtained by calibration using a circular standard calibration block. And it is pre-installed in the control system; The Euclidean distance formula is used to calculate the center pixel coordinates of the outer circular contour of the circular knife in each frame of the image. coordinates of the theoretical center The distance is the first... Radial pixel offset of the center pixel coordinates of the frame image , ; Based on the radial pixel offset of the center pixel coordinates of all frames within one revolution of the circular cutter. ,Pick The maximum value in the value is taken as the maximum radial offset of the center pixel coordinates during one revolution of the circular cutter. , , The number of consecutive frames in the acquired circular cutter during one revolution; Methods for extracting the area of ​​surface defects include: By extracting the number of pixels in the defect region within the connected domain of the outer circle of the circular knife, and combining this with the area conversion relationship corresponding to the pixel equivalent, the surface defect area is calculated. The mathematical expression for the surface defect area is: ； In this mathematical expression, The area of ​​the surface defect. This represents the number of pixels in the defective area.

7. The visual monitoring and control method for a circular knife machine according to claim 6, characterized in that, The control system compares the extracted key operating parameters of the circular cutter with preset standard parameter thresholds to determine whether there are any abnormalities in the operating status of the circular cutter and performs corresponding adjustments. Specifically, this includes: The standard thresholds for key operating parameters of the circular cutter are preset, including the threshold for the actual diameter of the circular cutter. Threshold for the amount of wear on the cutting edge Threshold for runout of the circular cutter and the threshold of surface defect area The threshold of the actual diameter of the circular knife The cut-off diameter of the circular cutter, and the threshold for the amount of wear on the cutting edge. The maximum allowable wear of the circular cutter, and the threshold for the runout of the circular cutter. The maximum allowable runout of the circular cutter, and the threshold value for the surface defect area. This represents the maximum allowable defect area for a circular cutter. exist At that time, the operating status of the circular cutter was determined to be abnormal in diameter; At that time, the operating state of the circular cutter was determined to be abnormal edge wear; At that time, the running state of the circular cutter was determined to be abnormal, indicating an abnormal runout. At that time, the operating status of the circular cutter is determined to be abnormal due to surface defects; If any of the following abnormalities are present: abnormal diameter, abnormal cutting edge wear, abnormal runout, or abnormal surface defects, the operating state of the circular cutter is determined to be abnormal. When there is an abnormality in the operation of the circular cutter, the control system makes corresponding adjustments. After the adjustment is completed, the circular cutter machine is restarted. After the circular cutter machine is restarted, the visual acquisition system for visual monitoring and control of the circular cutter machine performs image acquisition and preprocessing.

Images