Intelligent monitoring method for a spherical tank connecting shell drawing forming process

Abstract

The present application relates to the technical field of image processing, and more particularly to a kind of intelligent monitoring method for drawing forming process of spherical tank connecting shell, the method includes, collecting drawing forming video under each angle, frame processing is handled to drawing forming video under each angle, obtain each frame image of each angle, carry out edge detection to current frame image under each angle, obtain several edges in current frame image under each angle, obtain the probability that each edge in current frame image under each angle is pipe orifice edge, based on the probability, obtain pipe orifice edge of current frame image under each angle;Obtain pipe orifice drawing deformation degree of current frame image under each angle, according to pipe orifice drawing deformation degree of current frame image under each angle, obtain drawing deformation degree of current frame image, and then judge whether deformation occurs to current drawing, the present application can identify initial deformation in real time, guarantee spherical tank shell forming quality.

Description

Technical Field

[0001] This invention relates to the field of image processing technology. More specifically, this invention relates to an intelligent monitoring method for the drawing and forming process of a spherical tank connecting shell. Background Technology

[0002] The drawing process for the shell of the spherical tank involves using a press to drive a specially designed mold to gradually and repeatedly stretch the pre-formed holes in the tank, thereby plastically deforming them into the predetermined nozzle shape. Each stretching stage in this process requires precise control and meticulous operation, because any minute deformation or forming defect could become a potential structural failure point in the high-pressure environment of future product service, threatening the product's safety and reliability.

[0003] In the drawing process, equipment stability is crucial. If the press drives a specific die with even a slight deviation due to continuous vibration during operation, it will cause uneven plastic flow at the nozzle of the metal material during drawing. This uneven flow will initially manifest as nascent deformation at the nozzle edge, such as minor deformation or wrinkling at the local nozzle. If this nascent deformation is not addressed promptly, it will be rapidly amplified in subsequent drawing passes, ultimately leading to severe deformation and quality problems in the casing. Therefore, a smart method is needed to monitor the drawing process in real time and issue early warnings at the "nascent deformation" stage, facilitating timely parameter adjustments to ensure product quality.

[0004] Under uniform drawing conditions, the deformation of the tube edge caused by each drawing is consistent in height and expands outward uniformly in the radial direction. Therefore, by comparing the deformation of the tube edge in two frames of images before and after drawing, the initial deformation can be identified. However, when the actual mold contacts the pre-made hole and gradually applies force to stretch it, the outline of the mold equipment in the collected image affects the identification of the tube edge, which makes the monitoring system unable to accurately warn of the initial deformation, making the monitoring and warning unreliable. Summary of the Invention

[0005] To address the problem that the outline of the mold equipment in the image affects the recognition of the pipe opening edge, thus causing the monitoring system to fail to accurately warn of initial deformation and making the monitoring and warning unreliable, this invention proposes an intelligent monitoring method for the drawing and forming process of the spherical tank connecting shell. The method includes the following steps:

[0006] The drawing forming video at each angle is acquired, and the drawing forming video at each angle is processed into frames to obtain each frame image at each angle.

[0007] Edge detection is performed on the current frame image at each angle to obtain several edges in the current frame image at each angle; based on the difference in gray values ​​on both sides of the edge and the slope of the edge, the probability that each edge in the current frame image at each angle is the pipe opening edge is obtained; based on the probability, the pipe opening edge of the current frame image at each angle is obtained.

[0008] Based on the changes in the pipe opening edge between the current frame image and the previous frame image at each angle, the degree of pipe opening pull-out deformation in the current frame image at each angle is obtained.

[0009] Based on the degree of tube pull deformation in the current frame image at each angle, obtain the degree of pull deformation in the current frame image; based on the difference between the degree of pull deformation in the current frame image and the degree of pull deformation in the previous frame image, determine whether deformation has occurred during the current pull.

[0010] The innovation of this invention lies in first performing edge detection on each frame of the image, and then using corner points to break up the adhered edge segments. This effectively avoids the adhesion between the tube opening edge and the contour edge of the mold equipment, which would affect subsequent analysis. Furthermore, based on the difference in gray values ​​and slope on both sides of the tube opening edge, the probability of each edge in the image being the tube opening edge is obtained, which can accurately determine which edges are the real tube opening edges, thereby avoiding identification errors. Furthermore, by comparing the changes in the tube opening edge between two consecutive frames of images, the degree of deformation during drawing can be clearly identified, thereby effectively detecting initial deformation and issuing timely warnings.

[0011] Preferably, the step of performing edge detection on the current frame image at each angle to obtain several edges in the current frame image at each angle includes:

[0012] An edge detection algorithm is used to perform edge detection on the current frame image at the i-th angle to obtain the current frame edge image at the i-th angle. The Shi-Tomasi corner detection algorithm is used to detect the corners in the current frame edge image at the i-th angle to obtain the corners in the current frame edge image at the i-th angle. Based on the corners in the current frame edge image at the i-th angle, the edges in the current frame edge image at the i-th angle are broken to obtain several edges in the current frame image at the i-th angle.

[0013] This facilitates subsequent identification of the pipe opening edge.

[0014] Preferably, obtaining the probability that each edge in the current frame image at each angle is the edge of the pipe opening includes:

[0015] , This represents the probability that the j-th edge in the current frame image at the i-th angle is the edge of the pipe opening; This represents the number of pixels on the j-th edge in the current frame image at the i-th angle. This represents the absolute value of the grayscale difference between the two grayscale lines at the z-th pixel on the j-th edge in the current frame image at the i-th angle. This represents the number of pixel pairs along the j-th edge in the current frame image at the i-th angle; It represents the absolute value of the slope between the m-th pixel pairs of the j-th edge in the current frame image at the i-th angle; norm() represents the normalization function.

[0016] Based on the characteristics of the pipe opening edge, the probability of each edge being a pipe opening edge is obtained, which can accurately identify the pipe opening edge in the image.

[0017] Preferably, obtaining the absolute value of the grayscale difference between the two grayscale lines of the z-th pixel on the j-th edge in the current frame image at the i-th angle includes:

[0018] Given a preset number of pixels N, for the j-th edge in the current frame image at the i-th angle, obtain the normal direction of each pixel on the edge. Taking the z-th pixel on the edge as the center, extend N pixels on both sides of its normal direction to obtain two grayscale lines of the z-th pixel on the edge. Obtain the absolute value of the difference between the grayscale mean values ​​of the pixels on the two grayscale lines of the z-th pixel on the edge, and use it as the absolute value of the grayscale difference between the two grayscale lines of the z-th pixel on the j-th edge in the current frame image at the i-th angle.

[0019] Preferably, obtaining the degree of pipe end stretching deformation of the current frame image at each angle includes:

[0020] Obtain the fitting curve of the pipe opening edge from the previous frame image of the current frame image at the i-th angle;

[0021] ;

[0022] In the formula, This represents the degree of pipe pull-out deformation in the current frame image at the i-th angle; The standard deviation represents the shortest distance from the fitted curve of the pipe edge of the current frame image to the pipe edge of the previous frame image at the i-th angle. The standard deviation of the gradient direction of all pixels on the pipe edge of the current frame image at the i-th angle is represented by norm(), which represents the normalization function.

[0023] It can accurately reflect whether deformation occurs during the drawing process.

[0024] Preferably, obtaining the fitting curve of the pipe edge of the previous frame image at the i-th angle includes:

[0025] Obtain the position coordinates of each pixel on the pipe edge of the previous frame image of the current frame image at the i-th angle. Use the least squares method to perform curve fitting on the position coordinates of each pixel on the pipe edge of the previous frame image of the current frame image at the i-th angle to obtain the pipe edge fitting curve of the previous frame image of the current frame image at the i-th angle.

[0026] Preferably, obtaining the degree of stretching deformation of the current frame image includes:

[0027] The average value of the pipe end stretching deformation of the current frame image at all angles is taken as the stretching deformation degree of the current frame image.

[0028] The degree of pipe end pull-out deformation under different angles was considered, which improved the accuracy of primary deformation identification.

[0029] Preferably, determining whether deformation has occurred during the current pull based on the difference between the degree of pull deformation in the current frame image and the degree of pull deformation in the previous frame image includes:

[0030] A preset deformation threshold T is set. If the difference between the degree of stretching deformation in the current frame and the degree of stretching deformation in the previous frame exceeds the deformation threshold T, a device malfunction will cause stretching deformation at the pipe opening, and the system will immediately issue a warning notification.

[0031] Preferably, obtaining the pipe opening edge of the current frame image at each angle includes:

[0032] The edge corresponding to the maximum probability among all edges in the current frame image at the i-th angle that are pipe opening edges is taken as the pipe opening edge of the current frame image at the i-th angle.

[0033] Preferably, obtaining the pixel pair of the j-th edge in the current frame image at the i-th angle includes:

[0034] Take any two adjacent pixels on the j-th edge of the current frame image at the i-th angle as the pixel pair of the j-th edge of the current frame image at the i-th angle. Similarly, obtain all pixel pairs of the j-th edge of the current frame image at the i-th angle.

[0035] The present invention has the following beneficial effects: The purpose of the present invention is to perform edge detection on each frame of image, and then use corner points to break the adhered edge segments, which can effectively avoid the adhesion of the tube opening edge and the contour edge of the mold equipment in the image, affecting subsequent analysis. Furthermore, based on the difference in gray values ​​and slope on both sides of the tube opening edge, the probability of each edge in the image being the tube opening edge can be obtained, which can accurately determine which edges belong to the true tube opening edge, thereby avoiding the identification of the mold equipment edge as the tube opening edge. Furthermore, by comparing the changes in the tube opening edge of two consecutive frames of images, the degree of deformation of the drawing process can be clearly identified, thereby effectively detecting the initial deformation and issuing an early warning for the initial deformation in a timely manner. Attached Figure Description

[0036] Figure 1 This is a flowchart illustrating the steps of an intelligent monitoring method for the drawing and forming process of a spherical tank connecting shell according to an embodiment of the present invention. Detailed Implementation

[0037] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0038] Please see Figure 1 The diagram illustrates a flowchart of a method for intelligent monitoring of the drawing process of a spherical tank connecting shell according to an embodiment of the present invention. The method includes the following steps:

[0039] S001. Acquire drawing forming video at each angle, process the drawing forming video at each angle frame by frame, and obtain each frame image at each angle.

[0040] In this embodiment of the invention, a camera is placed in front of, to the left of, to the right of and behind the drawing forming area of ​​the spherical tank connecting shell, and the height of each camera is the same. During the drawing forming process, the camera at different angles is used to record the video of the drawing forming process to obtain the drawing forming video at each angle.

[0041] The drawing forming video at each angle is processed into frames to obtain each frame image at each angle.

[0042] S002. Perform edge detection on the current frame image at each angle, obtain several edges in the current frame image at each angle, and obtain the probability that each edge in the current frame image at each angle is the pipe opening edge based on the difference in gray values ​​on both sides of the edge and the slope of the edge. Based on the probability, obtain the pipe opening edge of the current frame image at each angle.

[0043] It should be noted that the drawing process of the spherical tank connecting shell involves using a press to drive a specific die to stretch the pre-made holes of the spherical tank multiple times and gradually, causing it to plastically deform into a preset pipe or flange shape. However, during the drawing process, if the press-driven die experiences slight deviation due to continuous vibration of the equipment, it will cause uneven plastic flow in the spherical tank connecting shell during drawing. This uneven flow will cause localized minor deformation and wrinkling of the pipe opening of the spherical tank connecting shell. If not intervened in time, it will be rapidly amplified in subsequent drawing processes, ultimately leading to serious deformation and quality problems of the shell. Therefore, this invention needs to analyze the deformation of the pipe opening edge of the spherical tank connecting shell during the drawing process. Thus, it is necessary to extract the pipe opening edge to facilitate the analysis of the pipe opening deformation during the drawing process.

[0044] Drawing is a process in which a press drives a specific die to stretch the pre-made holes of a spherical tank multiple times and step by step. Therefore, one side of the image is the die driven by the press. It is known that the actual shell of the spherical tank is made of aluminum, while the die driven by the press is generally made of iron. Therefore, the grayscale difference between the two is significant. Thus, the grayscale difference between the two sides of the pipe opening connecting the shell of the spherical tank in the image is large. Therefore, an edge detection algorithm is used to extract the edges of each frame of the image.

[0045] The edge of the canister's connecting shell will appear as an independent planar curve. However, during the drawing process, the edge will also be detected due to the large difference in grayscale values ​​between the two sides of the pull rod and the background. Therefore, the edge of the canister's connecting shell may be connected with the edges generated on both sides of the pull rod, resulting in inaccurate extraction of the edge. Known corner points usually appear at the edges, intersections, or turning points of objects. Therefore, detecting corner points in the image can help segment the edges and prevent the edge of the canister's connecting shell from being connected with the edge of a specific mold driven by the press.

[0046] In this embodiment of the invention, an edge detection algorithm is used to perform edge detection on the current frame image at the i-th angle to obtain the current frame edge image at the i-th angle. The Shi-Tomasi corner detection algorithm is used to detect the corners in the current frame edge image at the i-th angle to obtain the corners in the current frame edge image at the i-th angle. Based on the corners in the current frame edge image at the i-th angle, the edges in the current frame edge image at the i-th angle are broken to obtain several edges in the current frame image at the i-th angle.

[0047] It should be further explained that, given the known large difference in grayscale between the two sides of the pipe opening edge connecting the spherical tank to the shell in the image, this invention can obtain the pipe opening edge based on the difference in grayscale values ​​between the two sides of the edge. However, since the mold driven by the press is generally made of iron and has a large difference in grayscale compared to the background area, the difference in grayscale values ​​between the two sides of the pull rod edge is also large. If the probability of each edge in the image being the pipe opening edge is obtained solely based on the difference in grayscale values ​​between the two sides of the edge, it will be inaccurate and will lead to the inability to find the accurate pipe opening edge. Furthermore, since the geometric features of the pipe opening edge exhibit a certain curvature change, while the pull rod edge is a straight line, this invention combines the difference in grayscale values ​​between the two sides of each edge and the slope difference between the pixels on each edge to obtain the probability of each edge being the pipe opening edge, thereby obtaining the pipe opening edge.

[0048] In this embodiment of the invention, the preset number of pixels N=5. For the j-th edge in the current frame image at the i-th angle, the normal direction of each pixel on the edge is obtained. Taking the z-th pixel on the edge as the center, N pixels are extended on both sides of its normal direction to obtain two gray lines of the z-th pixel on the edge. The absolute value of the difference between the gray mean values ​​of the pixels on the two gray lines of the z-th pixel on the edge is obtained as the absolute value of the gray difference between the two gray lines of the z-th pixel on the j-th edge in the current frame image at the i-th angle.

[0049] Take any two adjacent pixels on the j-th edge of the current frame image at the i-th angle as the pixel pair of the j-th edge of the current frame image at the i-th angle. Similarly, obtain all pixel pairs of the j-th edge of the current frame image at the i-th angle.

[0050] Obtain the probability that each edge in the current frame image at the i-th angle is the edge of the pipe opening:

[0051] ;

[0052] In the formula, This represents the probability that the j-th edge in the current frame image at the i-th angle is the edge of the pipe opening; This represents the number of pixels on the j-th edge in the current frame image at the i-th angle. This represents the absolute value of the grayscale difference between the two grayscale lines at the z-th pixel on the j-th edge in the current frame image at the i-th angle. This represents the number of pixel pairs along the j-th edge in the current frame image at the i-th angle; This represents the absolute value of the slope between the m-th pixel pairs of the j-th edge in the current frame image at the i-th angle; norm() represents the normalization function.

[0053] The larger the value, the more significant the grayscale difference between the two sides of the j-th edge in the current frame image at the i-th angle. This means that the j-th edge in the current frame image at the i-th angle may be the boundary line of two different regions. Therefore, the probability that the j-th edge in the current frame image at the i-th angle is the edge of the pipe opening is greater.

[0054] The larger the value, the larger the absolute value of the slope between all pixel pairs of the j-th edge in the current frame image at the i-th angle. This means that the geometry of the j-th edge is more complex and more in line with the curve characteristic of the pipe opening edge. Therefore, the probability that the j-th edge in the current frame image at the i-th angle is the pipe opening edge is greater.

[0055] Obtain the probability that each edge in the current frame image at the i-th angle is the edge of the pipe opening. Take the edge corresponding to the maximum probability among all edges in the current frame image at the i-th angle that is the edge of the pipe opening as the edge of the pipe opening in the current frame image at the i-th angle.

[0056] S003. Based on the comparison of the pipe opening edge between the current frame image at each angle and the previous frame image, obtain the degree of pipe opening pull-out deformation of the current frame image at each angle.

[0057] It should be noted that under normal uniform drawing conditions, the deformation of the spherical tank connecting shell should be equal after each stamping and stretching, and the outline of the nozzle (flared mouth) should expand uniformly along the stretching direction. Therefore, after each stretching, the shortest distance from each pixel on the nozzle edge to the nozzle edge after the previous stretching should be equal, and the gradient direction of the pixels on the nozzle edge after stretching should be consistent. However, during the drawing process, if the press equipment malfunctions and vibrates or the parameters are not set correctly, the regularity of the drawing process will be broken, resulting in changes in the shape of the nozzle edge and the gradient direction of the pixels on the nozzle edge. Therefore, this invention obtains the degree of nozzle drawing deformation in the current frame image by measuring the consistency of the shortest distance from each pixel on the nozzle edge in the current frame image to the nozzle edge in the previous frame image and the consistency of the gradient direction of each pixel on the nozzle edge in the current frame image, which is convenient for subsequent measurement of whether the nozzle has deformed during the drawing process.

[0058] In this embodiment of the invention, the position coordinates of each pixel on the pipe edge of the previous frame image of the current frame image at the i-th angle are obtained, and the least squares method is used to perform curve fitting on the position coordinates of each pixel on the pipe edge of the previous frame image of the current frame image at the i-th angle to obtain the pipe edge fitting curve of the previous frame image of the current frame image at the i-th angle.

[0059] Obtain the degree of tube end stretching deformation in the current frame image at each angle:

[0060] ;

[0061] In the formula, This represents the degree of pipe pull-out deformation in the current frame image at the i-th angle; The standard deviation represents the shortest distance from the fitted curve of the pipe edge of the current frame image to the pipe edge of the previous frame image at the i-th angle. The standard deviation of the gradient direction of all pixels on the pipe edge of the current frame image at the i-th angle; norm() represents the normalization function;

[0062] It reflects the consistency of the shortest distance from all pixels on the tube edge of the current frame image to the tube edge of the previous frame image at the i-th angle. The larger the value, the more the tube edge line is deformed during the stretching process. Therefore, the greater the degree of tube stretching deformation in the current frame image at the i-th angle.

[0063] It reflects the consistency of the gradient direction of the pixels on the edge of the tube in the current frame image at the i-th angle. The larger the value, the more inconsistent the gradient direction of the pixels on the edge of the tube in the current frame image at the i-th angle, indicating that the edge line of the tube has been deformed during the stretching process. At this time, the degree of tube stretching deformation in the current frame image at the i-th angle is greater.

[0064] S004. Based on the degree of pipe pulling deformation in the current frame image at each angle, obtain the degree of pulling deformation in the current frame image; based on the difference between the degree of pulling deformation in the current frame image and the degree of pulling deformation in the previous frame image, determine whether the current pulling has caused deformation.

[0065] It should be noted that uneven stretching caused by equipment problems usually manifests as initial deformation. If this initial deformation is not addressed in time, it will rapidly intensify during subsequent stretching, ultimately leading to severe deformation and quality issues in the casing. Therefore, it is necessary to determine whether deformation has occurred based on the degree of tube end stretching deformation in the current frame image at each angle, thereby providing early warning of initial deformation during the stretching process and adjusting the equipment status in a timely manner to avoid more serious quality problems.

[0066] First, this invention considers the degree of tube end stretching deformation in the current frame image at different angles to obtain the degree of stretching deformation in the current frame image, which can comprehensively reflect the abnormal deformation situation that occurs during the current stretching process. Then, if the degree of stretching deformation in the current frame image is greater than the degree of stretching deformation in the previous frame image, it indicates that there may be a problem with the equipment during the current stretching process, causing tube end stretching deformation, and the system should immediately issue an early warning notification.

[0067] In this embodiment of the invention, the average value of the pipe opening stretching deformation degree of the current frame image at all angles is taken as the stretching deformation degree of the current frame image; the stretching deformation degree of each frame image is obtained.

[0068] A preset deformation threshold T is set. If the difference between the degree of stretching deformation in the current frame and the degree of stretching deformation in the previous frame is greater than the deformation threshold T, it indicates that there may be a problem with the equipment during the stretching process at the corresponding moment in the current frame, which may cause stretching deformation of the tube opening. The system should immediately issue an early warning notification to remind the operator to adjust the equipment in time to ensure that the product production quality meets the standard requirements.

[0069] It should be noted that the preset deformation threshold T=0.1. In other embodiments, the implementer may preset the value of the deformation threshold T according to the specific implementation situation.

[0070] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method of intelligent monitoring of a draw forming process for a spherical tank connecting shell, characterized in that, The method comprises the following steps: Collecting the drawing forming video at each angle, performing frame processing on the drawing forming video at each angle, and obtaining each frame image at each angle; Performing edge detection on the current frame image at each angle to obtain the current frame edge image at each angle, detecting the corner points in the current frame edge image at each angle using a Shi-Tomasi corner detection algorithm to obtain the corner points, and disconnecting the edges in the corresponding current frame edge image according to the corner points to obtain a plurality of edges in the current frame image at each angle; According to the gray value difference on both sides of the edge and the slope of the edge, a probability that each edge in the current frame image under each angle is a nozzle edge is obtained, including: , represents the probability that the jth edge in the current frame image under the ith angle is a nozzle edge, , respectively represent the number of pixel points on the jth edge in the current frame image under the ith angle, and the absolute value of the gray value difference of the two gray lines of the zth pixel point, , respectively represent the number of pixel point pairs of the jth edge in the current frame image under the ith angle, and the absolute value of the slope between the mth pixel point pair, and norm() represents a normalization function; based on the probability, a nozzle edge of the current frame image under each angle is obtained. According to the change of the nozzle edge of the current frame image and the previous frame image under each angle, the nozzle drawing deformation degree of the current frame image under each angle is obtained, including: obtaining the nozzle edge fitting curve of the previous frame image of the current frame image under the i-th angle; ; representing the nozzle drawing deformation degree of the current frame image under the i-th angle, representing the standard deviation of the shortest distance from all pixel points on the nozzle edge of the current frame image under the i-th angle to the nozzle edge fitting curve of the previous frame image, representing the standard deviation of the gradient direction of all pixel points on the nozzle edge of the current frame image under the i-th angle; According to the pipe mouth drawing deformation degree of the current frame image at each angle, the drawing deformation degree of the current frame image is obtained; and according to the difference between the drawing deformation degree of the current frame image and the drawing deformation degree of the previous frame image, it is judged whether the current drawing has deformed.

2. The method of claim 1, wherein, The edge detection on the current frame image at each angle and the obtaining of a plurality of edges in the current frame image at each angle comprise: Using an edge detection algorithm, the current frame image at the i-th angle is subjected to edge detection to obtain the current frame edge image at the i-th angle, the corner points in the current frame edge image at the i-th angle are detected using a Shi-Tomasi corner detection algorithm to obtain the corner points in the current frame edge image at the i-th angle, and the edges in the current frame edge image at the i-th angle are disconnected according to the corner points in the current frame edge image at the i-th angle to obtain a plurality of edges in the current frame image at the i-th angle.

3. The method of claim 1, wherein, The obtaining of the absolute value of the gray value difference of the two gray lines of the z-th pixel point on the j-th edge in the current frame image at the i-th angle comprises: A preset number of pixel points N, for the j-th edge in the current frame image at the i-th angle, the normal direction of each pixel point on the edge is obtained, and the z-th pixel point on the edge is taken as the center to extend N pixel points on both sides along the normal direction to obtain two gray lines of the z-th pixel point on the edge; and the absolute value of the difference between the average gray values of the pixel points on the two gray lines of the z-th pixel point on the edge is obtained as the absolute value of the gray value difference of the two gray lines of the z-th pixel point on the j-th edge in the current frame image at the i-th angle.

4. The method of claim 1, wherein, The obtaining of the pipe edge fitting curve of the previous frame image of the current frame image at the i-th angle comprises: Obtaining the position coordinates of each pixel point on the pipe edge of the previous frame image of the current frame image at the i-th angle, and using the least square method to perform curve fitting on the position coordinates of each pixel point on the pipe edge of the previous frame image of the current frame image at the i-th angle to obtain the pipe edge fitting curve of the previous frame image of the current frame image at the i-th angle.

5. The method of claim 1, wherein, The obtaining of the drawing deformation degree of the current frame image comprises: Taking the average of the pipe mouth drawing deformation degrees of the current frame images at all angles as the drawing deformation degree of the current frame image.

6. The method of claim 1, wherein, The judgment of whether the current drawing has deformed according to the difference between the drawing deformation degree of the current frame image and the drawing deformation degree of the previous frame image comprises: A preset deformation threshold T is set, if the value of the drawing deformation degree of the current frame image minus the drawing deformation degree of the previous frame image is greater than the deformation threshold T, the device is malfunctioning to cause the pipe mouth drawing deformation, and the system immediately issues a warning notification.

7. The method of claim 1, wherein, The pipe mouth edge of the current frame image under each angle is obtained, including: The maximum value of the probabilities of all edges being the pipe mouth edge in the current frame image under the i th angle is taken as the pipe mouth edge of the current frame image under the i th angle.

8. The method of claim 1, wherein, The pixel point pair of the j th edge in the current frame image under the i th angle is obtained, including: Any two adjacent pixel points on the j th edge in the current frame image under the i th angle are taken as the pixel point pair of the j th edge in the current frame image under the i th angle, and all pixel point pairs of the j th edge in the current frame image under the i th angle are obtained in the same way.

Images