A sewer pipe shearing quality detection method and system based on shape recognition

Abstract

The present application relates to the technical field of industrial detection, in particular to a kind of drainage pipe shearing quality detection method and system based on shape identification.The method comprises: obtaining the gray image of the shearing section of the drainage pipe and extracting sub-pixel edge points, constructing polar coordinate sub-pixel edge sequence containing original radial distance and shearing phase angle and calculating mechanical collapse index;Based on the shearing phase angle, calculate the elastic modulus weight and plastic collapse weight;Local curvature and local gradient distribution standard deviation of sub-pixel edge points are obtained, combined with original radial distance, mechanical collapse index, elastic modulus weight, plastic collapse weight to calculate theoretical radial distance, and then calculate the true diameter and roundness error of the drainage pipe, generate equipment adjustment feedback signal.The present application effectively suppresses the influence of reflection interference and shear stress deformation through light shape decoupling and rheological compensation mechanism, and realizes high-precision detection of the shearing quality of the drainage pipe.

Description

Technical Field

[0001] This invention relates to the field of industrial testing technology, specifically to a method and system for detecting the shear quality of drainage pipes based on shape recognition. Background Technology

[0002] Drainage pipes, especially those made of polymer materials such as PVC-U and PE, are widely used in building and municipal engineering. In the production process of drainage pipes, fixed-length cutting is one of the key processes. However, because drainage pipes have a certain degree of plasticity and elasticity, and the cutting process is a strong destructive process, the quality of the sheared section is often seriously affected by shear stress.

[0003] Traditional methods for inspecting the quality of shear sections of drainage pipes primarily rely on manual caliper measurements or basic machine vision techniques. However, manual measurement is inefficient and prone to subjective errors, while existing machine vision methods often employ pixel-level edge detection operators such as Canny or Sobel to extract the contour of the shear section of the drainage pipe and then directly fit the roundness. However, this approach faces two major challenges in practical applications:

[0004] First, optical calibration is difficult. Under industrial light source illumination, the chamfer and arc surface of the drain pipe opening are prone to strong specular reflection highlights. The traditional gradient amplitude detection method cannot distinguish between the edge produced by physical collapse and the edge of the light spot produced by optical reflection. It is easy to misjudge the reflection as a serious edge collapse or gap, resulting in a very high false alarm rate.

[0005] Second, the deformation mechanism is complex. The shearing process introduces a non-uniform stress field into the shear cross-section of the drain pipe, causing complex nonlinear deformations such as elastic flattening at the top and plastic involute at the bottom. Traditional testing methods usually assume that the drain pipe undergoes uniform geometric deformation and lack a compensation mechanism for the physical rebound process. As a result, the measured dimensions cannot reflect the true geometric properties of the drain pipe under stress-free conditions, making it difficult to guide the optimization of the production process. Summary of the Invention

[0006] To address the problems of low detection accuracy and high false positive rate in traditional methods due to the inability to distinguish between optical reflection and physical collapse, and the lack of compensation mechanisms for non-uniform shear stress, this invention provides a drainage pipe shear quality detection method and system based on shape recognition.

[0007] In a first aspect, the present invention provides a method for detecting the shear quality of drainage pipes based on shape recognition, employing the following technical solution:

[0008] A method for detecting the shear quality of a drainage pipe based on shape recognition includes: acquiring a grayscale image of the shear cross-section of the drainage pipe and extracting sub-pixel edge points; constructing a polar coordinate sub-pixel edge sequence based on the sub-pixel edge points, the polar coordinate sub-pixel edge sequence including the original radial distance and shear phase angle of the sub-pixel edge points; extracting a grayscale profile along the normal direction of the sub-pixel edge points, with the side away from the center of the drainage pipe wall as the outer side and the side pointing towards the center of the drainage pipe wall as the inner side; calculating the mechanical collapse index of the sub-pixel edge points based on the inner gradient exponential decay rate and the outer gradient exponential decay rate obtained by exponential fitting of the grayscale profile; calculating the elastic mode weight and plastic collapse weight of the sub-pixel edge points based on the shear phase angle; acquiring the local curvature and local gradient distribution standard deviation of the sub-pixel edge points, and calculating the theoretical radial distance of the sub-pixel edge points by combining the original radial distance, the mechanical collapse index, the elastic mode weight, and the plastic collapse weight; calculating the true diameter and roundness error of the drainage pipe based on the theoretical radial distance; and generating a quality judgment result and equipment adjustment feedback signal based on the roundness error.

[0009] This invention introduces an optical decoupling mechanism based on the difference in bilateral gradient attenuation. By utilizing the difference between the Gaussian symmetry of optical reflection and the asymmetry of physical collapse, a mechanical collapse index is constructed, which can accurately remove optical artifacts and lock in the real mechanical deformation, thus solving the problem of misjudgment in strong reflective environments.

[0010] Furthermore, the mechanical collapse index satisfies the following relationship:

[0011]

[0012] The polar coordinate sub-pixel edge sequence contains several sub-pixel edge points, wherein, For the first The mechanical collapse index of each sub-pixel edge point. , The first The inner gradient exponential decay rate and the outer gradient exponential decay rate of the sub-pixel edge points. To prevent extremely small positive numbers with a denominator of zero, Noise tolerance threshold, It is a step function.

[0013] This invention determines the accuracy of real-world collapse by judging the symmetry of the gradients on both sides of a sub-pixel edge point. When there is high-light reflection, the attenuation rates on both sides are similar, and the exponent approaches 0. When there is physical collapse, one side is steep and the other side is gentle, with a huge difference and the exponent approaches 1. This invention achieves accurate judgment of real-world collapse from a physical logic level.

[0014] Furthermore, the elastic modal weights satisfy the following relationship:

[0015]

[0016] The plastic collapse weights satisfy the following relationship:

[0017]

[0018] The polar coordinate sub-pixel edge sequence contains several sub-pixel edge points, wherein, , The first The elastic mode weights and plastic collapse weights of the sub-pixel edge points For the first The shear phase angle of the sub-pixel edge points. As a sidewall resilience enhancement factor, This represents the focusing coefficient of the plastic region.

[0019] This invention establishes a spatial mapping mechanism based on shear phase angle, which divides sub-pixel edge points into different stress modes, effectively solving the mode mismatch problem of insufficient compensation at the top and overcompensation at the bottom on the shear cross-section by a single compensation model, which is consistent with the physical facts of vertical shearing process.

[0020] Furthermore, the theoretical radial distance satisfies the following relationship:

[0021]

[0022] The polar coordinate sub-pixel edge sequence contains several sub-pixel edge points, wherein, For the first The theoretical radial distance of each of the sub-pixel edge points For the first The original radial distance of each of the sub-pixel edge points , , The first The elastic mode weight, plastic collapse weight, and mechanical collapse index of each sub-pixel edge point. , The first The local curvature and the standard deviation of the local gradient distribution of each of the sub-pixel edge points As the reference curvature, The standard deviation of the baseline gradient distribution. The correlation coefficient is the elastic modulus. This is the plastic collapse compensation coefficient. It is the natural logarithm function. It is a very small positive number.

[0023] Based on the comparison between local curvature and reference curvature, this invention drives the original radial distance to contract inward or expand outward, accurately restoring the springback process of the drain pipe after the shear force is removed, thereby obtaining the true geometric contour of the drain pipe's shear cross-section under stress-free conditions.

[0024] Furthermore, the noise tolerance threshold was determined through experimental comprehensive calibration.

[0025] Furthermore, the equipment adjustment feedback signal includes: calibrating the 0-degree phase based on the physical installation position of the tool for cutting the drain pipe, the tool including a cutting tool and a clamp; if the quality judgment result is unqualified, then statistically analyzing the distribution of the sub-pixel edge points whose roundness error exceeds a preset threshold; if they are concentrated in the 180-degree phase region, generating a cutting tool wear warning signal; if they are concentrated in the 90-degree phase region, generating a clamp pressure adjustment signal.

[0026] This invention introduces a phase decoupling feedback mechanism to transform detection results into specific process optimization instructions, thus achieving a leap from passive detection to active control.

[0027] Furthermore, the extraction of sub-pixel edge points is achieved using an edge detection algorithm.

[0028] Furthermore, the sidewall resilience enhancement factor was determined through experimental comprehensive calibration.

[0029] Furthermore, the grayscale image is acquired using a high-frequency industrial camera in conjunction with a low-angle ring light source.

[0030] Secondly, the present invention provides a drainage pipe shear quality detection system based on shape recognition, which adopts the following technical solution:

[0031] A shape recognition-based drainage pipe shear quality detection system includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned shape recognition-based drainage pipe shear quality detection method is implemented.

[0032] The present invention has the following technical effects:

[0033] This invention constructs an optical decoupling model based on the difference in bilateral gradient attenuation, and utilizes the symmetric difference between physical collapse and optical reflection in grayscale distribution to effectively suppress the interference of specular artifacts and ensure the purity of edge feature extraction.

[0034] Secondly, by introducing an anisotropic rheological compensation model, this invention nonlinearly couples the microscopic mechanical collapse index with the macroscopic stress mode weights, and mathematically reconstructs the rebound process of the drainage pipe. It can infer the actual shear section size from the deformed shear section, greatly improving the detection accuracy.

[0035] Furthermore, based on the feedback mechanism of the shear phase angle, this invention can automatically diagnose problems such as tool wear or fixture abnormalities according to the error distribution, providing strong support for the intelligent operation and maintenance of the production line. Attached Figure Description

[0036] Figure 1 This is a flowchart of a drainage pipe shear quality detection method based on shape recognition provided in an embodiment of the present invention;

[0037] Figure 2 This is a comparison diagram of the effects of the traditional method provided in the embodiments of the present invention and the present invention. Detailed Implementation

[0038] This invention provides a method for detecting the shear quality of drainage pipes based on shape recognition, referring to... Figure 1 This includes steps S1-S5:

[0039] S1: Subpixel edge contour acquisition and polar coordinate construction.

[0040] To achieve high-precision quality inspection of the shear section of drainage pipes, it is necessary to establish a polar coordinate reference based on the geometric center of the shear section.

[0041] Specifically, a standardized analysis space with the center of the shear section as the origin is established by acquiring high-precision images of the shear section of the drainage pipe, extracting sub-pixel edge features, and performing a benchmark transformation from Cartesian coordinates to polar coordinates.

[0042] First, a high-frequency industrial camera is used in conjunction with a low-angle ring light source to photograph the sheared section of the drain pipe, obtaining a grayscale image of the sheared section. In this embodiment, a 5-megapixel monochrome camera equipped with a telecentric lens is selected, along with an LED ring shadowless light source with an illumination angle of 30 degrees, placed 50 cm above the drain pipe. When the trigger sensor detects that the drain pipe is in place, the high-frequency industrial camera takes a picture to obtain a grayscale image of the sheared section of the drain pipe, which is then transmitted to the processor via a gigabit network port.

[0043] It should be noted that although a low-angle ring light source helps to highlight the edge contour of the shear section of the drain pipe, it may also have the effect of reflection. Therefore, it is necessary to use an edge detection algorithm to extract features from the grayscale image of the shear section of the drain pipe. Compared with the traditional Canny operator, this embodiment selects the Zernike moment sub-pixel edge detection algorithm, which can achieve sub-pixel level positioning accuracy by utilizing the integral characteristics of the image. The specific operation is as follows:

[0044] A 7×7 Zernike moment template is constructed, and convolution operations are performed on the edge regions of the acquired grayscale image. Edge parameters such as rotation angle and distance of each pixel are calculated. By determining the zero-crossing points of the second derivative, sub-pixel edge points with an accuracy of 0.1 pixels are located, resulting in a dataset containing... sub-pixel edge point set Each sub-pixel edge point contains precise pixel coordinates, such as the first... Sub-pixel edge points The pixel coordinates are .

[0045] Since the center of the acquired grayscale image is not necessarily the geometric center of the shear section of the drain pipe, this embodiment uses a least squares circle fitting algorithm to... By performing fitting and iterative processing, the coordinates of the geometric center of the shear section of the drainage pipe are calculated. .

[0046] Finally, to visually describe the circumferential deformation of the drain pipe during shearing, this embodiment uses a Cartesian-polar coordinate transformation matrix to transform the pixel coordinates of each sub-pixel edge point:

[0047] by Using the origin as the reference point, the physical installation vertical position of the tool for cutting the drain pipe at the image acquisition site is obtained. The direction on the grayscale image corresponding to this position is defined as the phase angle reference, and the counterclockwise direction is set as positive. This yields the relative position of each sub-pixel edge point. original radial distance and shear phase angle This constructs a polar coordinate subpixel edge sequence arranged in phase angle order, providing standardized data input for subsequent operations.

[0048] S2: Light-shaped decoupling feature extraction based on bilateral gradient decay difference.

[0049] Specifically, optical decoupling feature extraction eliminates optical reflection interference to extract the true shape features of the shear cross-section of the drainage pipe. This step aims to complete the in-depth analysis of the gray-level distribution features of the neighborhood of the sub-pixel edge point. By evaluating the symmetry difference of gradient attenuation on both sides of the sub-pixel edge point, a mechanical collapse index is constructed to lock the physical deformation features of the true shear cross-section of the drainage pipe.

[0050] During the shearing process of a drainage pipe, the edges of its sheared cross-section exhibit both optical reflection and physical collapse. Optical reflection is determined by the smoothness of the drainage pipe surface, and the edges of its reflected light spots usually show a Gaussian symmetrical distribution. Physical collapse, such as involute and tearing, is caused by the plastic flow of the drainage pipe. Due to the presence of fracture surfaces and rheological surfaces at its edges, the grayscale distribution inevitably exhibits an asymmetrical characteristic with one side steep and the other side gentle.

[0051] High-precision grayscale sampling in the normal direction: Since traditional operators only focus on gradient magnitude and ignore direction information, this invention adopts a bilateral grayscale sampling method in the normal direction, as follows:

[0052] For each sub-pixel edge point, the normal vector of the point is first calculated using the local differential method. Then, with the point as the center, along the normal vector, the direction pointing to the outside of the drain pipe wall is taken as the positive direction, and the direction pointing to the inside of the drain pipe wall is taken as the negative direction. To ensure complete coverage of the spot area and the collapsed area, the sampling length is set according to the chamfer radius of the drain pipe wall. In this embodiment, the sampling length is set to 20 pixels. Pixel-level sampling is performed along the positive and negative directions respectively. If the sampling coordinates are not integers, the accurate gray value is obtained by bilinear interpolation, thereby constructing the inner gray-scale profile sequence and the outer gray-scale profile sequence of each sub-pixel edge point.

[0053] Gradient exponential decay fitting: Based on physical optics and material mechanics properties, the specular reflection spot at the chamfer of the drain pipe opening typically exhibits a Gaussian or quasi-Gaussian distribution with bilateral symmetrical decay; while the physical collapse caused by shearing typically exhibits an asymmetrical distribution with a steep fracture surface on one side and a gentle rheological surface on the other. Based on this, this invention uses an exponential decay model to perform feature fitting on the inner grayscale profile sequence and the outer grayscale profile sequence of each sub-pixel edge point. The specific operation is as follows:

[0054] The last three data points of the inner grayscale profile sequence are selected, and their mean is calculated and denoted as the background light noise constant. Then, the background light noise constant is subtracted from each data point in the inner grayscale profile sequence to obtain the net attenuation data sequence. The net attenuation data sequence is then subjected to a natural logarithmic transformation to map it to linear space data. Finally, the transformed linear space data is fitted with a straight line using the least squares method. The absolute value of the slope of this fitted line represents the inner gradient exponential attenuation rate of that sub-pixel edge point. The exponential decay rate of the inner gradient at each sub-pixel edge point is denoted as... ;No. The calculation method of the outer gradient exponential decay rate of each sub-pixel edge point and Similarly, denoted as .

[0055] It should be noted that, because the reflected light spots at the chamfered corner of the drain pipe opening typically exhibit a Gaussian or Gaussian distribution, the light diffuses in a generally consistent manner to both sides. The value and The values ​​are very close; however, physical collapse caused by shearing often exhibits asymmetry, i.e. The value and The values ​​differ greatly; based on this logic, the first... Mechanical collapse index of subpixel edge points This is used to determine the probability that it is a physical collapse.

[0056] Mechanical collapse index The relationship is as follows:

[0057]

[0058] in, For the first The mechanical collapse index of a sub-pixel edge point; the larger the value, the greater the probability that the point has physically collapsed. To prevent extremely small positive numbers with a denominator of zero, this embodiment takes... , This is the noise tolerance threshold; It is a step function.

[0059] When optical reflection is present The term approaches 1. Approaching 0, the first The probability of a sub-pixel edge point being physically collapsed is low; when involution or tearing exists, The term approaches 0. Approaching 1, the first The probability of a sub-pixel edge point being physically collapsed is relatively high.

[0060] It should be noted that in actual drainage pipe inspection scenarios, interference from non-physical factors such as camera thermal noise, minor light source vibrations, and the natural texture of the drainage pipe wall can all lead to... The value and The values ​​differ, therefore it is necessary to introduce... To filter out these interferences, when Item greater than hour Take 1, when The term is less than or equal to hour Take 0.

[0061] It should be noted that the noise tolerance threshold The calibration is based on the statistical distribution of the background noise of the image acquisition system and the optical properties of the drainage pipe surface. The specific steps are as follows:

[0062] (1) Select a section with a surface roughness of less than 0.8 Using a standard circular drainage pipe without shear defects as a calibration component, 50 frames of shear section images of the calibration component are continuously acquired under the same conditions as in step S1, using a high-frequency industrial camera and a low-angle ring light source. All sub-pixel edge points are extracted from each frame image, and the absolute value of the difference between the inner gradient exponential decay rate and the outer gradient exponential decay rate of each sub-pixel edge point is calculated and denoted as the baseline gradient difference. The baseline gradient differences of all sub-pixel edge points are combined into a baseline gradient difference set, denoted as B.

[0063] (2) Perform histogram statistical analysis on set B. Since, under defect-free conditions, the difference between the inner gradient exponential decay rate and the outer gradient exponential decay rate mainly originates from non-physical collapses such as camera thermal noise, slight light source jitter, and natural texture of the drainage pipe wall, and this difference follows a normal distribution, calculate the mean and standard deviation of set B. According to the statistical principle of 3... In principle, the upper bound of the coverage area of ​​99.7% is defined as the noise tolerance threshold. The embodiment is calibrated It is 0.043.

[0064] By comprehensively considering the inner and outer gradient exponential decay rates of sub-pixel edge points, and simultaneously introducing... This feature can filter out interference caused by non-physical collapse, ensuring that the calculation of the mechanical collapse index is activated only when the difference between the inner gradient exponent decay rate and the outer gradient exponent decay rate exceeds the noise tolerance threshold, thus effectively enhancing robustness and anti-interference ability.

[0065] S3: Stress mode weighting based on shear phase.

[0066] Specifically, this step aims to establish a spatial mapping relationship between the shear phase angle and the stress mode. Because a non-uniform stress field is generated on the shear cross-section of the drain pipe during shearing—for example, the top of the shear cross-section at 0 degrees is under compression, the bottom at 180 degrees is under tension, and the sidewalls at 90 or 270 degrees are under bending—this invention evaluates the elastic rebound trend and plastic tearing risk of different regions of the shear cross-section of the drain pipe by calculating the elastic mode weight and plastic collapse weight of each sub-pixel edge point. Simultaneously, a correction factor calibrated experimentally is used to ensure the physical authenticity of the results.

[0067] Specifically, the shear phase angle is determined using a preset stress mode distribution function. Perform mapping calculations:

[0068] The relationship between the elastic modal weights is as follows:

[0069]

[0070] The relationship for the plastic collapse weight is as follows:

[0071]

[0072] in, It is the first Shear phase angle of sub-pixel edge points It is a sine function. It is a cosine function. As a sidewall resilience enhancement factor, This represents the focusing coefficient of the plastic region.

[0073] for Relationship, The term corresponds to the vertical direction of the shear section, that is, the principal compressive stress zone at 0 degrees and 180 degrees; The term corresponds to the horizontal direction of the shear section, namely the sidewall bulging areas at 90 degrees and 270 degrees. Used to correct the lateral deformation of the sidewalls of the shear section caused by the Poisson effect;

[0074] for The relationship is that, since the plastic tearing caused by shearing on the shear section of the drainage pipe, such as involutes and burrs, only exists in the top and bottom regions, while the sidewalls only undergo elastic bending, a filter for each angle on the circumference of the shear section is constructed using the power property of the cosine function. When it approaches 90 degrees or 270 degrees, Approaching 0, at this point, a focusing coefficient for the plastic zone is introduced. This allows for further compression of tiny bases; and when When approaching 0 degrees or 180 degrees, Approaching 1, it ultimately achieves effective weight compensation.

[0075] It should be noted that the greater the brittleness of the drain pipe material, the more concentrated the area affected by shearing. The value is in the range [2.5, 3]; the better the toughness of the drain pipe material, the more dispersed the area affected by shear. The value is in the range [1.5, 2]. This embodiment is for PVC-U material drainage pipes. The value is 2, but this value is only an example based on typical working conditions. In practical applications, it can be adjusted or recalibrated within a reasonable engineering range according to the diameter, wall thickness and material characteristics of the drainage pipe, and all of these are within the protection scope of this invention.

[0076] It should be noted that, due to the sidewall resilience enhancement factor This characterizes the proportional relationship between the lateral bulging of the sidewalls and the vertical compression of the top of the drain pipe when it is subjected to compression. At the top and bottom, that is The value is 1 when the angle is 0 or 180 degrees, and it is on the side wall, i.e. The value is when it is 90 degrees or 270 degrees. ,therefore, Directly related to the physical deformation rate of the drain pipe, this invention uses an elastic field calibration experiment to... The comprehensive calibration process involves the following steps:

[0077] 1. Select a standard circular drain pipe. In this embodiment, a 50cm long PVC-U solid wall drain pipe with a DN110 specification conforming to the international standard GB / T5836.1 and an outer diameter of 110mm±0.2mm is selected as the test sample to ensure that the sample is in a stress-free relaxation state and the initial roundness error is less than 0.2mm.

[0078] 2. Place the sample on a flat plate compression tester and apply a vertical radial load at a speed of 5 mm / min so that the compression in the vertical direction of the sample reaches 3% to 5% of the nominal outer diameter. This range can ensure that the deformation characteristics are obvious to reduce measurement error, and also ensure that the drainage pipe is completely within the range governed by the elastic Hooke's law.

[0079] 3. Keep the load stable, and use a vernier caliper with an accuracy of 0.02mm to measure the minimum vertical outer diameter of the sample at this time, which is the distance between the line connecting 0 degrees and 180 degrees, and record it as . And the maximum horizontal outer diameter, i.e., the distance between the lines connecting 90 degrees and 270 degrees, is denoted as... ;

[0080] 4. Calculate the vertical compression of the sample. and horizontal bulging Sidewall resilience enhancement factor .

[0081] It can be seen that the elastic field calibration experiment... Comprehensive calibration effectively addresses the problem of traditional methods that simply equate the shear cross-section of a drainage pipe to an ideal ellipse, i.e., a longitudinal-to-transverse deformation ratio of 1:1. Due to the complex ring stiffness characteristics of polymer materials, the sidewall bulging amplitude is often greater than the top compression amplitude. This invention, by comprehensively considering the material's true mechanical response properties, constructs elastic mode weights that can accurately reproduce the true rebound of the drainage pipe. This effectively improves the geometric accuracy of contour reconstruction of the shear section of the drainage pipe.

[0082] S4: Establish an anisotropic rheological compensation model and perform coordinate correction.

[0083] Specifically, this step is the core step in realizing the deduction of the true size from the deformation image of the shear section of the drain pipe. By constructing an anisotropic rheological compensation model, the springback process of the drain pipe after the shear force is removed is restored by mathematical means, thereby obtaining the true geometric contour of the shear section of the drain pipe in the stress-free state.

[0084] First, perform a second-order central difference operation on the polar coordinate sub-pixel edge sequence constructed in step S1 to obtain the local curvature of each sub-pixel edge point; at the same time, process the inner grayscale profile sequence extracted in step S2 using the statistical standard deviation algorithm to obtain the local gradient distribution standard deviation of each sub-pixel edge point.

[0085] The theoretical radial distance satisfies the following relationship:

[0086]

[0087] in, For the first The theoretical radial distance of sub-pixel edge points As the reference curvature, The standard deviation of the baseline gradient distribution. The correlation coefficient is the elastic modulus. This is the plastic collapse compensation coefficient. For elastic modal weights, As the weight for plastic collapse, For the first Local curvature of sub-pixel edge points For the first Standard deviation of local gradient distribution at sub-pixel edge points For the first The mechanical collapse index of sub-pixel edge points For the first The original radial distance of each sub-pixel edge point To prevent extremely small positive numbers where the logarithmic term is zero, this embodiment takes... ; It is the natural logarithm function.

[0088] Statistical analysis of experimental data revealed that and The logarithm of the ratio is significantly correlated with the radial rebound of the drain pipe. Therefore, this invention constructs an empirical compensation model based on the natural logarithm function. When the value is positive, it drives... Inward contraction; when When the value is negative, the driving force is... Outward expansion; this process is affected Weighted control, by applying a restoring force at the mathematical level that is opposite to the direction of the actual stress, effectively reproduces the rebound process of the drain pipe;

[0089] Tearing caused by shearing inevitably increases the disorder of the micro-texture at the edge of the shear cross-section of the drain pipe, specifically manifested as follows: By utilizing this characteristic, through the introduction and This achieves the effect of directional compensation only at the tear point and no interference at the un-torn point, thus restoring its true theoretical radial distance without contacting the drain pipe.

[0090] It should be noted that, and The calibration process is as follows:

[0091] A DN110 PVC-U solid-wall drainage pipe conforming to international standard GB / T 5836.1 was selected as the experimental sample. CMM measurements confirmed that the roundness error of the experimental sample was less than 0.05 mm and the surface roughness was less than 0.8 μm. Grayscale images of the shear section of the experimental sample were acquired using a high-frequency industrial camera and a low-angle ring light source, similar to step S1. The arithmetic mean of the local curvature at each sub-pixel edge point was calculated and recorded as the reference curvature. Calculate the arithmetic mean of the standard deviations of the local gradient distribution at each sub-pixel edge point, and denote it as the baseline gradient distribution standard deviation. .

[0092] It should be noted that the elastic modulus correlation coefficient Characterizing the ability of the drain pipe to resist and recover bending deformation, this embodiment uses -0.16;

[0093] When a drain pipe tears due to shearing at the bottom, the deeper the tear, the more chaotic the microstructure of its edges, specifically manifested as... ;However, Unable to be directly associated Therefore, a plastic collapse compensation coefficient is required. To establish a linear mapping relationship between the two, this embodiment... Take 0.8;

[0094] It should be noted that the and in the present invention are only examples based on typical working conditions, obtained by comparing the offline CMM measurement values and online image calculation values of standard drain pipes and using the least squares method for regression. For drain pipes made of common polymer materials such as PVC-U and PE, usually lies in the interval [-0.1, -0.3], usually lies in the interval [0.5, 1.2]. Although the in this embodiment takes -0.16, and the

[0095] S5: Comprehensive judgment and feedback of cross-sectional quality.

[0096] Specifically, in this step, based on the theoretical radial distance of each sub-pixel edge point obtained in S4, geometric fitting and error analysis are performed on the grayscale image of the shear cross-section of the drain pipe, which can not only determine whether the drain pipe is qualified, but also reverse-diagnose the operating state of the production equipment by analyzing the spatial distribution characteristics of the error, and finally achieve the leap from passive error reduction to active process optimization.

[0097] First, combine the theoretical radial distance of each sub-pixel edge point obtained in step S4 with its corresponding shear phase angle to construct a theoretical contour point set, perform Gaussian smoothing on this point set, and then use the least squares circle fitting algorithm to perform regression analysis on this point set. The diameter of the circle obtained by fitting is the true diameter of the drain pipe after eliminating the influence of shear stress, denoted as ; Calculate the distance deviation from the theoretical radial distance of each sub-pixel edge point in the theoretical contour point set to the fitted circumference, and take the sum of the maximum positive deviation and the maximum negative deviation as the roundness error, denoted as .

[0098] Compare with the preset quality threshold . If , determine that the drain pipe is qualified, and store in the production database; if , determine that the drain pipe is unqualified;

[0099] When it is determined that the drain pipe is unqualified, not only send an alarm message to the production line, but also further analyze the aggregation position of the roundness error, and based on this, reverse-deduce which production component has a problem and generate an adjustment instruction. The specific steps are as follows:

[0100] 1. Using the center of the fitted circle as the origin, establish an angle index from 0 degrees to 360 degrees; calculate the difference between the radius of each sub-pixel edge point in the theoretical contour point set and the radius of the fitted circle, denoted as . ,structure Follow The changing error distribution curve; find the error distribution curve in... For the interval, calculate the geometric center angle of that interval, and denot it as the fault phase angle. ;

[0101] 2. If The value is between 165 degrees and 195 degrees, corresponding to the position where the shearing tool leaves the drain pipe, generating a tool wear warning signal and increasing the feed rate of the shearing tool; if The value is between 75 degrees and 105 degrees or between 255 degrees and 295 degrees. This corresponds to the clamping contact surface of the cutting machine, generating a clamping pressure adjustment signal to drive the pneumatic proportional valve to reduce the pressure of the clamping cylinder.

[0102] Figure 2 This is a comparison diagram of the effects of the traditional method and the present invention provided in the embodiments of the present invention. It can be seen that, since the traditional method assigns the same weight to all sub-pixel edge points, the result is affected by the bottom collapse, the center of the circle shifts downward and the radius is smaller than the true radius of the drain pipe; while the present invention can effectively identify and automatically reduce the weight of sub-pixel edge points in the bottom collapse area, successfully restore the rebound process of the drain pipe, and demonstrate the robustness of the present invention.

[0103] It should be noted that the preset quality threshold It was calibrated comprehensively through assembly performance correlation experiments, and the specific steps are as follows:

[0104] 1. Define the stable process state: Select a PVC-U drainage pipe production line equipped with an SJZ-80 / 156 conical twin screw extruder, with a screw speed of 18 r / min, a constant vacuum degree of -0.04 MPa in the vacuum sizing box, and a barrel heating zone temperature controlled at 190℃±1℃. Run under these conditions for 30 minutes, and confirm with an online diameter measuring instrument that the outer diameter fluctuation of the drainage pipe is less than 0.05 mm and the roundness error is less than 0.1 mm.

[0105] 2. Set the clamping air pressure of the traction machine to 0.25 MPa, which will not cause significant deformation of the drain pipe. Gradually increase the clamping air pressure to 0.65 MPa in increments of 0.05 MPa. After running stably for 2 minutes at each air pressure setting, cut 5 sections of drain pipe with a length of 100 mm as experimental samples.

[0106] 3. Calculate the roundness error of all experimental samples, and group the experimental samples into several gradient test groups with a gradient range of 0.1 mm.

[0107] 4. A universal testing machine with a range of 50KN is used, equipped with a special drain pipe compression clamp. In this embodiment, INSTRON 5960 is selected. A DN110 straight socket conforming to GB / T 5836.1 standard is selected as the standard socket. Lubricant is evenly applied to the spigot end of each experimental sample in the gradient test group and the inner wall of the standard socket. In this embodiment, potassium soap solution is selected.

[0108] 5. Fix the standard socket to the bottom of the testing machine, and fix the test sample to the upper moving crossbeam. Apply downward pressure at a constant speed of 20 mm / min until the test sample is inserted into the standard socket mark. According to the ergonomically set single-person operation physical limit, when the detected peak insertion resistance exceeds 800 N, record the average roundness error of the test samples in this gradient test group, which is the preset quality threshold. .

[0109] It can be seen that comprehensive calibration is achieved through assembly performance correlation experiments. This effectively avoids the blindness of setting qualification standards based solely on experience or general tolerances in traditional testing, transforming abstract geometric error indicators into clearly defined engineering functional boundaries, thus achieving a balance between product quality reliability and manufacturing costs.

Claims

1. A method for detecting the shear quality of drainage pipes based on shape recognition, characterized in that, include: Obtain a grayscale image of the shear section of the drainage pipe and extract sub-pixel edge points. Construct a polar coordinate sub-pixel edge sequence based on the sub-pixel edge points. The polar coordinate sub-pixel edge sequence contains the original radial distance and shear phase angle of the sub-pixel edge points. Along the normal direction of the sub-pixel edge point, a grayscale profile is extracted with the outer side facing away from the center of the drainage pipe wall as the outer side and the inner side facing the center of the drainage pipe wall as the inner side. The mechanical collapse index of the sub-pixel edge point is calculated based on the inner gradient exponential decay rate and the outer gradient exponential decay rate obtained by exponential fitting of the grayscale profile. The elastic mode weight and plastic collapse weight of the sub-pixel edge point are calculated based on the shear phase angle. The local curvature and local gradient distribution standard deviation of sub-pixel edge points are obtained. The theoretical radial distance of the sub-pixel edge points is calculated by combining the original radial distance, mechanical collapse index, elastic mode weight, and plastic collapse weight. The actual diameter and roundness error of the drainage pipe are calculated based on the theoretical radial distance. The quality judgment result and equipment adjustment feedback signal are generated based on the roundness error. The mechanical collapse index satisfies the following relationship: The polar coordinate subpixel edge sequence contains several subpixel edge points, among which... For the first The mechanical collapse index of sub-pixel edge points , The first The inner gradient exponential decay rate and the outer gradient exponential decay rate of each sub-pixel edge point To prevent extremely small positive numbers with a denominator of zero, Noise tolerance threshold, It is a step function; The elastic modal weights satisfy the following relationship: The plastic collapse weights satisfy the following relationship: The polar coordinate subpixel edge sequence contains several subpixel edge points, among which... , The first Elastic mode weights and plastic collapse weights for sub-pixel edge points For the first Shear phase angle of sub-pixel edge points As a sidewall resilience enhancement factor, This is the focusing coefficient of the plastic region; The theoretical radial distance satisfies the following relationship: The polar coordinate subpixel edge sequence contains several subpixel edge points, among which... For the first The theoretical radial distance of sub-pixel edge points For the first The original radial distance of each sub-pixel edge point , The first Local curvature and standard deviation of local gradient distribution at sub-pixel edge points As the reference curvature, The standard deviation of the baseline gradient distribution. The correlation coefficient is the elastic modulus. This is the plastic collapse compensation coefficient. It is the natural logarithm function. It is a very small positive number.

2. The method for detecting the shear quality of drainage pipes based on shape recognition according to claim 1, characterized in that, The noise tolerance threshold was determined through experimental comprehensive calibration.

3. The method for detecting the shear quality of drainage pipes based on shape recognition according to claim 1, characterized in that, The equipment adjustment feedback signal includes: calibrating the 0-degree phase based on the physical installation position of the tool for shearing the drain pipe. The tool includes a cutting tool and a clamp. If the quality judgment result is unqualified, the distribution of sub-pixel edge points with roundness error exceeding a preset threshold is statistically analyzed. If they are concentrated in the 180-degree phase area, a cutting tool wear warning signal is generated; if they are concentrated in the 90-degree phase area, a clamp pressure adjustment signal is generated.

4. The method for detecting the shear quality of drainage pipes based on shape recognition according to claim 1, characterized in that, The extraction of sub-pixel edge points is achieved using an edge detection algorithm.

5. The method for detecting the shear quality of drainage pipes based on shape recognition according to claim 1, characterized in that, The sidewall resilience enhancement factor was determined through experimental comprehensive calibration.

6. The method for detecting the shear quality of drainage pipes based on shape recognition according to claim 1, characterized in that, The grayscale image was acquired using a high-frequency industrial camera in conjunction with a low-angle ring light source.

7. A drainage pipe shear quality detection system based on shape recognition, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement a drainage pipe shear quality detection method based on shape recognition according to any one of claims 1-6.

Images