Automatic welding identification and repair method and system for pipeline crevasse, medium and terminal

Abstract

The invention provides an automatic welding identification and repair method and system for a pipeline crevasse, a medium and a terminal. The method comprises the following steps that pipeline images, collected by a plurality of cameras, of different visual angles are obtained; the pipeline image is preprocessed; carrying out crevasse identification and positioning based on the preprocessed pipeline image; carrying out welding repair on the identified and positioned crevasses; and carrying out welding detection evaluation on the repaired crevasse. According to the pipeline crevasse automatic welding identification and repair method and system, the medium and the terminal, through automatic pipeline crevasse welding identification and repair, the pipeline crevasse repair efficiency is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of pipeline welding, and in particular to a pipeline rupture automatic welding identification and repair method, system, medium and terminal. Background Art

[0002] In industrial production and infrastructure construction, pipeline systems are widely used to transport various fluid media, such as oil, natural gas, water, etc. However, due to the long-term influence of internal medium pressure, external environmental corrosion and other mechanical stress factors, pipelines are prone to defects such as cracks. These cracks will not only cause medium leakage, environmental pollution and waste of resources, but may also cause serious safety accidents, threatening the safety of life and property.

[0003] Traditional pipeline breach repair methods mainly rely on manual inspection and repair, which have problems such as low efficiency, poor precision, and high labor intensity. With the development of industrial automation technology, automated welding technology has been applied to a certain extent in the field of pipeline repair, but the existing methods still need to be improved in terms of breach recognition accuracy, welding path planning rationality, welding parameter adaptability, and welding quality detection reliability. For example, in terms of breach recognition, a single image feature extraction method is difficult to fully and accurately identify breach features; when planning the welding path, the complex shape of the breach and the robot kinematic constraints are not fully considered; the welding parameter adjustment is not intelligent enough and cannot be optimized in real time according to the breach and material characteristics; the post-welding detection method is not comprehensive and accurate enough, and it is difficult to effectively ensure the repair quality.

[0004] Therefore, there is an urgent need for an efficient, accurate and highly automated pipeline rupture automatic welding identification and repair method. Summary of the invention

[0005] In view of the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a pipeline rupture automatic welding identification and repair method, system, medium and terminal, which effectively improves the efficiency of pipeline rupture repair through automated pipeline rupture welding identification and repair.

[0006] In a first aspect, the present invention provides a method for automatic welding identification and repair of pipeline ruptures, the method comprising the following steps: acquiring pipeline images of different viewing angles acquired by multiple cameras; preprocessing the pipeline images; identifying and locating the rupture based on the preprocessed pipeline images; performing welding repair on the identified and located ruptures; and performing welding detection and evaluation on the repaired ruptures.

[0007] In an implementation of the first aspect, acquiring pipeline images at different angles captured by multiple cameras includes the following steps:

[0008] Setting the shooting angles and shooting distances of the multiple cameras;

[0009] according to Adjust the camera acquisition parameters, which include exposure time T, gain G and aperture size A, L represents the ambient light intensity, C represents the complexity of the pipeline surface texture, k 1 ,k 2 ,k 3 ,k 4 ,k 5 ,k 6 represents the empirical coefficient;

[0010] The pipeline image I is acquired based on the shooting angle, the shooting distance and the camera acquisition parameters. i =f(θ i ,d i ,P), where I i represents the pipeline image captured by the i-th camera, P represents the pipeline surface features, θ i represents the shooting angle of the i-th camera, d i Represents the shooting distance of the i-th camera.

[0011] In an implementation of the first aspect, preprocessing the pipeline image includes the following steps:

[0012] According to LoG new =w(x,y)×▽ 2 G filters the pipeline image, where LoG new represents the filtered pipeline image, w(x,y) represents the adaptive weight factor, ▽ 2 G represents LoG filtering, (x, y) represents the pixel of the pipeline image;

[0013] Perform image enhancement on the filtered pipeline image based on deep learning algorithm;

[0014] Perform adaptive threshold binarization on the pipeline image after image enhancement;

[0015] Perform morphological gradient edge detection and optimization on the binarized pipeline image.

[0016] In an implementation of the first aspect, the breach identification and location based on the preprocessed pipeline image includes the following steps:

[0017] Based on the preprocessed pipeline image, an image pyramid is constructed to obtain pipeline images of different scales;

[0018] For each scale of pipeline image, extract local binary pattern features and oriented gradient histogram features;

[0019] Fusing the local binary pattern feature and the oriented gradient histogram feature to obtain a fused feature;

[0020] Classifying the fused features based on a support vector machine to identify breaches in the pipeline image, and using pixel points corresponding to the identified breaches as breach feature points;

[0021] The identified breach feature points under different perspectives are matched and fused to obtain the fused breach position.

[0022] In an implementation of the first aspect, for the identified and located breach, performing welding repair comprises the following steps:

[0023] Constructing a three-dimensional model of the breach based on the shape and position information of the breach;

[0024] Constructing a spiral welding planning path based on the three-dimensional model of the breach, wherein the spiral welding planning path is x=r cos(θ), y=r sin(θ), z=hθ, wherein r represents the spiral radius, θ represents the angle parameter, and h represents the pitch;

[0025] Calculating the welding robot joint angle corresponding to each path point in the spiral wire welding planning path;

[0026] Inputting the rupture depth, rupture width and pipeline material type into a welding parameter adjustment model to obtain welding parameters, wherein the welding parameters include welding current, welding voltage and welding speed;

[0027] Fine-tuning the welding parameters based on the welding temperature field distribution and the molten pool shape;

[0028] Inputting the adjusted welding parameters into the welding prediction model to obtain welding state parameters, wherein the welding dynamic parameters include weld penetration, weld width and welding deformation;

[0029] The welding parameters are adjusted based on the welding state parameters and a welding quality target.

[0030] In an implementation of the first aspect, performing welding detection and evaluation on the repaired breach includes the following steps:

[0031] Construct welding quality assessment model;

[0032] The image of the pipeline after the breach is repaired is input into the welding quality assessment model to obtain the welding quality assessment result.

[0033] In an implementation of the first aspect, performing welding detection and evaluation on the repaired breach further includes the following steps:

[0034] Obtain the internal structure information of the weld collected by the ultrasonic sensor;

[0035] Obtain the surface flatness and deformation information of the pipe after welding collected by the laser displacement sensor;

[0036] Based on the pipeline image after welding, the internal structure information of the weld, and the surface flatness and deformation information of the pipeline, a welding quality assessment result is obtained through a welding quality assessment model.

[0037] In a second aspect, the present invention provides a pipeline breach automatic welding identification and repair system, the system comprising an acquisition module, a preprocessing module, an identification and positioning module, a repair module and an evaluation module;

[0038] The acquisition module is used to acquire pipeline images of different viewing angles collected by multiple cameras;

[0039] The preprocessing module is used to preprocess the pipeline image;

[0040] The identification and positioning module is used to identify and locate the breach based on the preprocessed pipeline image;

[0041] The repair module is used to perform welding repair on the identified and located breach;

[0042] The evaluation module is used to perform welding detection evaluation on the repaired breach.

[0043] In a third aspect, the present invention provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned pipeline rupture automatic welding identification and repair method.

[0044] In a fourth aspect, the present invention provides a terminal, comprising: a processor and a memory;

[0045] The memory is used to store computer programs;

[0046] The processor is used to execute the computer program stored in the memory, so that the terminal executes the above-mentioned pipeline rupture automatic welding identification and repair method.

[0047] As described above, the pipeline breach automatic welding identification and repair method, system, medium and terminal of the present invention have the following features:

[0048] Beneficial effects:

[0049] (1) A multi-scale feature fusion breach recognition method is adopted. By constructing an image pyramid to obtain images of different scales, the local texture, edge and shape information of the image can be effectively combined to improve the accuracy of breach recognition. The breach positioning optimization based on spatial geometric constraints matches and fuses the breach feature points under different viewing angles, further improving the accuracy of breach positioning.

[0050] (2) A three-dimensional model is constructed according to the shape and position of the rupture, and a spiral path planning strategy is adopted to ensure that the welding path covers the entire rupture area with uniform spacing. At the same time, the robot kinematic constraints and workspace limitations are considered, which improves the rationality and feasibility of the welding path;

[0051] (3) The depth, width and material type of the breach are introduced as input parameters, and a neural network model is established to adjust the welding current, voltage and welding speed in real time, so that the welding parameters can be dynamically optimized according to the actual situation of the breach to ensure the welding quality; the temperature field distribution and the shape of the molten pool are monitored in real time during the welding process, and the welding parameters are further fine-tuned to effectively deal with the unstable factors in the welding process and improve the stability of the welding process;

[0052] (4) A welding quality assessment model is constructed based on deep learning. The model is trained using appropriate loss functions for classification and regression tasks, which can accurately assess the welding quality. Multimodal data fusion detection uses ultrasonic sensors and laser displacement sensors to obtain the internal structure information of the weld and the surface flatness and deformation of the pipeline, and comprehensively detects the welding quality from multiple angles. Dynamic threshold adaptive adjustment detection adjusts the detection threshold according to the welding parameters and pipeline material characteristics to avoid misjudgment and improve the accuracy of defect detection.

[0053] (5) The whole process from image acquisition and preprocessing, breach identification and positioning, welding repair to post-welding inspection is automated, which reduces manual intervention, greatly improves the efficiency of pipeline breach repair, reduces labor intensity, and is suitable for large-scale pipeline repair operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 Shown is a flow chart of an embodiment of the pipeline breach automatic welding identification and repair method of the present invention;

[0055] Figure 2 Shown is a schematic structural diagram of an automatic welding identification and repair system for pipeline breaches according to an embodiment of the present invention;

[0056] Figure 3 Shown is a schematic structural diagram of a terminal in one embodiment of the present invention. DETAILED DESCRIPTION

[0057] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0058] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0059] like Figure 1 As shown, in one embodiment, the pipeline breach automatic welding identification and repair method of the present invention includes steps S1 to S5.

[0060] Step S1: Acquire pipeline images of different viewing angles captured by multiple cameras.

[0061] Specifically, pipeline images of different viewing angles are collected based on multiple cameras, and the pipeline images are sent to the pipeline breach automatic welding identification and repair terminal of the present invention.

[0062] When collecting pipeline images from different perspectives based on multiple cameras, the shooting angles and shooting distances of the multiple cameras are first set to ensure that all areas on the pipeline surface can be clearly photographed. Then, based on the lighting conditions and texture characteristics of the pipeline surface, Adjust the camera acquisition parameters, which include exposure time T, gain G and aperture size A, L represents the ambient light intensity, C represents the complexity of the pipeline surface texture, k 1 ,k 2 ,k 3 ,k 4 ,k 5 ,k 6 represents the empirical coefficient, which is optimized and determined through a large amount of experimental data. Among them, the ambient light intensity L is detected based on the light sensor, and the pipeline surface texture complexity C is calculated based on the texture analysis algorithm. Finally, the pipeline image I is obtained based on the shooting angle, the shooting distance and the camera acquisition parameters. i =f(θ i ,d i ,P), where I i represents the pipeline image captured by the i-th camera, P represents the pipeline surface features, θ i represents the shooting angle of the i-th camera, d iRepresents the shooting distance of the i-th camera, i=1,2,…,n, n represents the number of cameras. The surface features of the pipeline include texture features, color features and geometric features. The texture features include roughness, uniformity, etc. They can be obtained through a texture analysis algorithm. For the color features, if there are color differences on the pipeline surface, a quantitative analysis can be performed through a color space model (such as RGB, HSV, etc.). The geometric features include flatness, curvature, etc. Flatness can be determined by using laser measurement technology to measure the time or phase difference of laser reflected light to calculate the surface height change, and then judge the flatness. The curvature can be determined by performing curve fitting on the sampling points of the pipeline surface profile (such as least squares fitting of a circle or a polynomial curve) and calculating the curvature parameters of the fitting curve.

[0063] Step S2: preprocessing the pipeline image.

[0064] Specifically, preprocessing the pipeline image includes the following steps:

[0065] 21) According to LoG new =w(x,y)×▽ 2 G filters the pipeline image, where LoG new represents the filtered pipeline image, w(x,y) represents the adaptive weight factor, ▽ 2 G represents LoG filtering, and (x, y) represents the pixel point of the pipeline image.

[0066] The pipeline image is filtered by using an improved LoG filter. Based on this, an adaptive weight factor w(x,y) is introduced to adjust the weight factor according to the grayscale change rate of the pixels around the pixel point (x,y). The formula is improved to LoG new =w(x,y)×▽ 2 G. Grayscale change rate calculation method is: Where I(x,y) is the gray value of the pixel (x,y), m,n is the neighborhood window size, and σ represents the standard deviation in the Gaussian function. Therefore, the weight factor w(x,y) = 1 + k × R(x,y), and k is the weight adjustment coefficient.

[0067] 22) Perform image enhancement on the filtered pipeline image based on deep learning algorithm.

[0068] Among them, a convolutional neural network (CNN) model is constructed for image enhancement. The convolutional neural network model includes multiple convolutional layers, pooling layers and activation function layers. The input image (filtered pipeline image) is extracted features by the convolutional layer, the pooling layer is downsampled, and the activation function layer increases the nonlinear mapping capability. During the training process of the convolutional neural network model, a large number of image samples containing pipeline breaches are used for supervised learning, and the loss function adopts the mean square error (MSE) function, and the formula is: Where N is the number of samples, Y i To truly enhance the image, The enhanced image predicted by the model is enhanced by using the trained convolutional neural network model to enhance the filtered pipeline image, thereby improving the clarity and contrast of the filtered pipeline image and making the breach feature more obvious.

[0069] 23) Perform adaptive threshold binarization processing on the pipeline image after image enhancement.

[0070] In the process of image binarization, an adaptive threshold segmentation method is used. The local average grayscale value μ(x,y) and the local grayscale standard deviation σ(x,y) of the pipeline image after image enhancement are calculated, and T(x,y)=μ(x,y)+k×σ(x,y) is used as the adaptive threshold, where k is the threshold adjustment parameter. For each pixel point (x,y), if its grayscale value is greater than or equal to the threshold T(x,y), it is assigned as the foreground, that is, the breach part, otherwise it is assigned as the background. This method can automatically select the appropriate threshold according to the grayscale distribution characteristics of different areas of the image, and effectively segment the breach and the background.

[0071] 24) Perform morphological gradient edge detection and optimization on the binarized pipeline image.

[0072] In the morphological operation, the morphological gradient image of the binarized pipeline image is first calculated, and the formula is: Where I is the original image, B is the structural element, represents the expansion operation, Represents the erosion operation. The morphological gradient can highlight the edge of the object in the image. Then the gradient image is optimized, and the non-maximum suppression algorithm is used to remove the non-maximum points in the edge response, and only the local maximum value is retained as the final edge point, so as to further refine the edge of the rupture and improve the edge positioning accuracy.

[0073] Step S3: identifying and locating the breach based on the preprocessed pipeline image.

[0074] Specifically, breach identification and location based on the preprocessed pipeline image includes the following steps:

[0075] 31) Based on the preprocessed pipeline image, an image pyramid is constructed to obtain pipeline images of different scales.

[0076] For the preprocessed pipeline image, an image pyramid is constructed to obtain pipeline images at different scales. Assume that the number of layers of the image pyramid is l, and the resolution of the i-th layer image is r i (i=1,2,…l), pipeline images of different scales are obtained through downsampling operations: I i+1 =domnsample(I i ), where domnsample represents the downsampling function.

[0077] 32) For each scale of the pipeline image, extract local binary pattern features and oriented gradient histogram features.

[0078] For each scale of image, local binary pattern (LBP) features and directional gradient histogram (HOG) features are extracted respectively. LBP features can effectively describe the local texture information of the image, while HOG features are sensitive to the edge and shape information of the image.

[0079] 33) Fusing the local binary pattern feature and the oriented gradient histogram feature to obtain a fused feature.

[0080] Among them, the LBP features and HOG features at different scales are fused to obtain the fused feature F. The fusion method adopts the weighted summation strategy, and the formula is: where w i is the weight of the i-th layer feature, which is determined through experiments based on the importance of different scale features to breach identification.

[0081] 34) Classifying the fused features based on a support vector machine to identify breaches in the pipeline image, and using pixel points corresponding to the identified breaches as breach feature points.

[0082] The support vector machine (SVM) classifier is used to classify the fusion features to determine whether there is a breach in the image, thereby identifying the breach in the pipeline image. The pixel points corresponding to the fusion features that can highlight the key information such as the breach position, shape, edge and texture in the image are used as the breach feature points. When training the support vector machine, the decision function of the SVM is: Where n is the number of training samples, α i is the Lagrange multiplier, y is the sample label, and b is the bias term. K(x i ,x) is the kernel function, and the radial basis function K(xi ,x)=exp(-γx i -x 2 ), γ is the kernel parameter. A large number of image samples containing cracks and without cracks are used to train the SVM classifier, so that it can accurately identify cracks.

[0083] 35) Match and fuse the identified breach feature points under different perspectives to obtain the fused breach position.

[0084] When optimizing breach location based on spatial geometric constraints, the position information and calibration parameters of the camera during multi-view pipeline image acquisition are used to establish a spatial geometric model. Assume that the external parameter matrix of the i-th camera is R i and the translation vector t i , the internal parameter matrix is ​​K i In the image coordinate system, for the breach feature point p identified in the pipeline image i , and transform it to the world coordinate system using the following formula:

[0085] According to the spatial geometric constraints, the breach feature points located under different perspectives are matched and fused. Assume that the breach feature points identified under two perspectives are P w1 and P w2 , calculate the spatial distance between two points If the spatial distance d is less than the set threshold T d , then the two breach feature points are considered to belong to the same breach, and they are fused to obtain a more accurate breach location estimate. The fusion method uses a weighted average strategy, and the breach location after fusion is where w 1 and w 2 is the weight, which is determined according to the positioning accuracy and reliability of the breach feature points, where the breach feature points with high positioning accuracy have larger weights.

[0086] Preferably, for the identified breach, the upper boundary point and the lower boundary point of the laser stripe are fitted based on a cubic spline curve; and the coordinates of the center point of the laser stripe are calculated based on the cubic spline curve.

[0087] Among them, laser stripes play an auxiliary role in the process of breach identification and positioning. By projecting laser stripes onto the surface of the pipe, the reflection and refraction characteristics at the breach are used to obtain relevant information about the breach. When the laser stripes are irradiated to the edge of the breach, the laser stripes will be deformed due to the discontinuity of the surface at the breach. At the raised edge of the breach, the laser stripes may bend or interrupted; at the concave part, the brightness and width of the laser stripes may change. By analyzing these deformation characteristics, the approximate position and outline of the breach can be preliminarily determined. Among them, the direction of the laser stripes is related to the shape and direction of the breach. For long strip breaches, the deformation of the laser stripes when passing through the breach will extend along the length of the breach; for circular or elliptical breaches, the deformation of the laser stripes will show the characteristics of surrounding the breach. Therefore, the laser stripes are used to assist in the positioning of the breach.

[0088] On the basis of determining the boundary points of the laser stripes and calculating the geometric center method, the curve fitting technology is introduced to model the laser stripes more accurately. The present invention adopts a cubic spline curve to fit the upper boundary point P of the laser stripes. u (x u ,y u ) and the lower boundary point P 1 (x 1 ,y 1 ). The cubic spline curve formula is: S(x)=a i +b i (xx i )+c i (xx i ) 2 +d i (xx i ) 3 , where x i ≤x≤x i+1 ,i=1,2,…,n-1,n is the number of boundary points. Determine the curve coefficient a by solving the equations i ,b i ,c i ,d i , so that the cubic spline curve can smoothly pass through the boundary points and more accurately describe the shape of the laser stripes.

[0089] Based on the fitted cubic spline curve, recalculate the coordinates of the center point of the laser stripe. Assume that the coordinates of the point on the fitted cubic spline curve are (x, S(x)), then the horizontal coordinate of the center point is: where x min and x max is the value range of the fitted cubic spline curve on the x-axis, and the ordinate

[0090] Preferably, the present invention optimizes the traditional positioning method in the improved sub-pixel breach positioning algorithm. The gradient information of the grayscale values ​​of neighboring pixels is introduced when calculating the correction value. The traditional grayscale centroid method may have certain errors when determining the breach position because it does not fully consider the trend of pixel grayscale changes. Therefore, the present invention calculates the correction value based on the gradient information of the grayscale values ​​of neighboring pixels. Among them, for sub-pixel positioning, the grayscale centroid method is further optimized. When calculating the correction value Δ, the gradient information of the grayscale values ​​of neighboring pixels is introduced, and the formula is improved to: Where G(x,y+k) is the grayscale gradient value at the pixel point (x,y+k), which is obtained by calculating the difference between the grayscale values ​​of adjacent pixels. This improvement can more accurately consider the trend of pixel grayscale changes and improve sub-pixel positioning accuracy.

[0091] Through the above spatial geometric constraint optimization, the accuracy and stability of breach positioning can be effectively improved, the deviation caused by single-view positioning error can be reduced, and more accurate position information can be provided for subsequent welding repair. Through the above innovative breach identification and positioning method, combined with image acquisition and preprocessing steps, breaches on the pipeline can be more accurately and reliably identified and their position and shape can be precisely located, providing strong support for subsequent welding repair work and improving the performance and repair quality of the entire system.

[0092] Step S4: Perform welding repair on the identified and located breaches.

[0093] Specifically, for the identified and located breach, welding repair includes the following steps:

[0094] 41) Constructing a three-dimensional model of the breach based on the shape and position information of the breach.

[0095] According to the shape and position information of the breach, a three-dimensional model of the breach is constructed. Assume that the coordinates of the boundary point of the breach in the world coordinate system are P b (x b ,y b ,z b ), where b = 1, 2, …, m, where m is the number of boundary points. The surface curve of the breach is fitted using the cubic spline interpolation method, and the surface equation of the breach is obtained as S(u, v) = a(u) + b(u)v + c(u)v 2 +d(u)v 3 , where u and v are parameter variables, a(u), b(u), c(u), and d(u) are coefficient functions related to u, which are solved by the coordinates of the boundary points.

[0096] 42) Based on the three-dimensional model of the breach, a spiral welding planning path is constructed, wherein the spiral welding planning path is x=rcos(θ), y=rsin(θ), z=hθ, wherein r represents the spiral radius, θ represents the angle parameter, and h represents the pitch.

[0097] Among them, a spiral path planning strategy is adopted, starting from the center of the breach, and welding is performed along the edge of the breach in a spiral manner. For the spiral welding planning path, the parameters of the spiral are adjusted according to the size and shape of the breach to ensure that the welding path covers the entire breach area and the spacing of the welding path is uniform to ensure the consistency of welding quality.

[0098] 43) Calculate the welding robot joint angle corresponding to each path point in the spiral welding planning path.

[0099] In the process of path planning, the kinematic constraints and workspace limitations of the welding robot are considered, the robot kinematic model is introduced, and the robot joint angle corresponding to each path point is calculated. Let the robot joint variable be θ i , where i = 1, 2, ..., n, n is the number of robot joints. The position of the end effector in the world coordinate system is P e (x e ,y e ,z e ), through the forward kinematics equation P e (θ 1 ,θ 2 ,…,θ n ) Calculate the robot joint angles. At the same time, check whether the path point is within the robot's workspace to avoid collision or exceeding the working range of the robot.

[0100] 44) Inputting the rupture depth, rupture width and pipeline material type into the welding parameter adjustment model to obtain welding parameters, wherein the welding parameters include welding current, welding voltage and welding speed.

[0101] Among them, in the real-time adjustment of welding parameters, the welding parameters are adjusted in real time according to the geometric characteristics and material properties of the breach. The breach depth d, breach width u and pipeline material type M are introduced as input parameters to establish a welding parameter adjustment model of welding current I, welding voltage U and welding speed v. The welding parameter adjustment model is modeled using a neural network model, and the neural network structure includes an input layer (breach depth, breach width, material type), a hidden layer (multiple neurons, using a nonlinear activation function, such as RelU function f(x)=max(0,x)) and an output layer (welding current, welding voltage, welding speed).

[0102] In the present invention, the welding parameter adjustment model is trained by a large amount of experimental teaching data. The loss function in the training process adopts the mean square error function Where N is the number of training samples, I i ,U i ,v i is the actual welding parameter, The welding parameters predicted by the model. After training, the welding parameter adjustment model can accurately predict the appropriate welding parameters according to the fracture characteristics and material information obtained in real time.

[0103] 45) Fine-tune the welding parameters based on the welding temperature field distribution and the molten pool shape.

[0104] Among them, during the welding process, the temperature field distribution T (x, y, z, t) and the molten pool shape S are monitored in real time. m (x m ,y m ,z m ), where x, y, z are spatial coordinates, t is time, x m ,y m ,z m is the coordinate of the molten pool boundary point. The temperature field distribution is obtained by infrared thermal imager, and the molten pool shape is extracted by image processing algorithm. According to the temperature field distribution and molten pool shape, the welding parameters are further fine-tuned. If the pool temperature is found to be too high or the molten pool shape is unstable, the welding current is appropriately reduced or the welding speed is increased to ensure the stability of the welding process and the quality of the seam.

[0105] 46) The adjusted welding parameters are input into the welding prediction model to obtain welding state parameters, wherein the welding dynamic parameters include weld penetration, weld width and welding deformation.

[0106] Among them, in the welding process stability control, the model predictive control method is used to ensure the stability of the welding process. A welding prediction model is established, and the welding current I, welding voltage U, and welding speed v are used as input variables. The weld penetration h d 、Weld width h w and welding deformation δ as output variables. The welding prediction model is expressed as Where k is the time step, A, B, C are the model coefficient matrices, which are obtained through experimental data identification.

[0107] 47) Adjusting the welding parameters based on the welding state parameters and the welding quality target.

[0108] Among them, according to the set welding quality goals such as the expected penetration depth, welding width range and allowable welding deformation, the model predictive controller calculates the optimal control input sequence in multiple future time steps, namely welding current, welding voltage and welding speed. The optimization objective function is: Where p is the prediction time domain, m is the control time domain, is the predicted output variable, y ref (k+j) is the reference output value, and Δu(k+j) is the change in the control input). By solving the optimization problem, the optimal control input at the current moment is obtained and applied to the welding robot.

[0109] It should be noted that after each control cycle, the welding prediction model parameters and state variables are updated, and the welding prediction model is corrected according to the actual measured welding quality data (penetration depth, weld width, deformation) to improve the prediction accuracy of the welding prediction model, thereby achieving real-time optimization control of the welding process and ensuring stable and reliable welding quality. Through the above-mentioned innovative welding repair method, which is closely connected with the breach identification and positioning steps, it is possible to adaptively plan the welding path according to the specific situation of the breach, adjust the welding parameters in real time, and effectively control the stability of the welding process, thereby improving the quality and efficiency of welding repair and ensuring that the performance of the pipeline after repair meets the requirements.

[0110] Step S5: Perform welding inspection and evaluation on the repaired breach.

[0111] In one embodiment, when performing welding inspection and evaluation on the repaired breach, a welding quality evaluation model is constructed. A large number of pipeline image samples after welding are collected, including pipeline images of qualified welds and various types of defective welds (such as pores, cracks, and lack of fusion). These pipeline images are annotated, and the annotation information includes detailed information such as weld quality category (qualified or defective type) and defect location and size.

[0112] A convolutional neural network model is constructed as a welding quality assessment model. The network structure uses multiple convolutional layers, pooling layers and fully connected layers. The convolutional layer is used to extract the features of the image, the pooling layer performs downsampling to reduce the data dimension, and the fully connected layer is used for classification and regression tasks. The network structure can include 5 convolutional layers, with convolution kernel sizes of 3x3, 3x3, 5x5, 5x5, 3x3, and a step size of 1 or 2. Each convolutional layer is followed by a ReLu activation function and a 2x2 maximum pooling layer; the fully connected layer includes 3 hidden layers, with the number of nodes being 256, 128, and 64, respectively. The last output layer sets the corresponding number of nodes according to the classification task (qualified or defect type).

[0113] Define the loss function. For classification tasks (prediction of qualified or defective types), the cross entropy loss function is used. Where N is the number of samples, y i is the true label, is the probability value predicted by the model. For regression tasks (such as defect location and size prediction), the mean square error loss function is used. where ri is the true value, is the model prediction value. The welding quality assessment model is trained by the back propagation algorithm to optimize the model parameters so that the model can accurately identify the quality after welding. Finally, the pipeline image after the breach is repaired is input into the welding quality assessment model to obtain the welding quality assessment result.

[0114] In another embodiment, the welding quality is evaluated by multimodal data fusion detection. An ultrasonic sensor is used to detect defects inside the weld and obtain the internal structure information of the weld; a laser displacement sensor is used to measure the surface flatness and deformation of the pipe after welding.

[0115] For the detection data obtained by the ultrasonic sensor, the presence of defects inside the weld and the location and size of the defects are determined by analyzing the amplitude, time delay and other characteristics of the echo signal. A mathematical model between the ultrasonic echo signal and the defect characteristics is established, such as using a signal processing algorithm to extract the characteristic parameters of the echo signal, and then using a pattern recognition method to determine the defect type.

[0116] For the pipe surface height data obtained by the laser displacement sensor, the welding deformation is evaluated by calculating the change in surface height. Assume that the coordinates of the measurement point are (x i ,y i ,z i ), i = 1, 2, ..., n, n is the number of measurement points, calculate the surface flatness in is the average height value, welding deformation Δ=max(z i )-min(z i ).

[0117] The pipeline image, the detection data of the ultrasonic sensor and the detection data of the laser displacement sensor are fused, and the data layer fusion method is used to fuse the data of different sensors in the feature extraction stage. The detection data of the ultrasonic sensor and the detection data of the laser displacement sensor are converted into image form, such as mapping the ultrasonic echo signal intensity to grayscale value, and representing the laser displacement data as height image, and then input into the welding quality assessment model together with the pipeline image data for training and detection, so that the model can integrate multi-modal information for more accurate welding quality assessment, thereby obtaining welding quality assessment results.

[0118] It should be noted that for the detection data of ultrasonic sensors and laser displacement sensors, when performing defect identification, the defect identification threshold is dynamically adjusted according to the welding parameters (such as welding current, welding voltage, welding speed, etc.) and pipeline material characteristics during the welding process, and a threshold adjustment model is established, with welding parameters and material characteristics as input and defect identification threshold as output. Among them, for the defect identification threshold in image detection, by analyzing the relationship between welding current I, welding voltage U and weld defect formation, a defect identification threshold adjustment formula T = aI + bU + c is established, where a, b, c are coefficients determined based on experimental data. When the welding current is large, the defect identification threshold is appropriately increased to avoid misjudgment; when the voltage fluctuates greatly, the defect identification threshold is dynamically adjusted according to the voltage change to ensure that defects can be accurately detected.

[0119] For the defect recognition threshold adjustment of the detection data of the ultrasonic sensor and the detection data of the laser displacement sensor, the influence of welding parameters and material properties is also considered. The ultrasonic defect recognition threshold is adjusted according to the acoustic properties of the material, and the deformation value of the laser displacement detection is adjusted according to the mechanical properties such as the elastic modulus of the material. The defect recognition threshold is dynamically and adaptively adjusted to improve the adaptability of the detection system to different welding conditions and materials and reduce the false detection and missed detection rates.

[0120] Through the above innovative post-welding inspection method, which is closely integrated with the welding repair steps, the welding quality can be comprehensively and accurately evaluated, welding defects can be discovered and feedback can be given in a timely manner, and the quality and safety of pipeline welding repair can be ensured. At the same time, polymorphic data fusion and dynamic welding recognition threshold adjustment technology improve the reliability and adaptability of the inspection system to meet the welding inspection needs under different working conditions.

[0121] The protection scope of the pipeline rupture automatic welding identification and repair method described in the embodiment of the present invention is not limited to the execution order of the steps listed in this embodiment. All solutions implemented by adding, reducing or replacing steps in the prior art based on the principles of the present invention are included in the protection scope of the present invention.

[0122] An embodiment of the present invention further provides a pipeline rupture automatic welding identification and repair system, which can implement the pipeline rupture automatic welding identification and repair method described in the present invention. However, the implementation device of the pipeline rupture automatic welding identification and repair system described in the present invention includes but is not limited to the structure of the pipeline rupture automatic welding identification and repair system listed in this embodiment. All structural deformations and replacements of the prior art made according to the principles of the present invention are included in the protection scope of the present invention.

[0123] like Figure 2As shown, in one embodiment, the pipeline breach automatic welding identification and repair system of the present invention includes an acquisition module 21, a preprocessing module 22, an identification and positioning module 23, a repair module 24 and an evaluation module 25.

[0124] The acquisition module 21 is used to acquire pipeline images of different viewing angles captured by multiple cameras.

[0125] The preprocessing module 22 is connected to the acquisition module 21 and is used to preprocess the pipeline image.

[0126] The identification and positioning module 23 is connected to the preprocessing module 22 and is used to identify and locate the breach based on the preprocessed pipeline image.

[0127] The repair module 24 is connected to the identification and positioning module 23 and is used for performing welding repair on the identified and located breaches.

[0128] The evaluation module 25 is connected to the repair module 24 and is used to perform welding detection and evaluation on the repaired breach.

[0129] Among them, the structures and principles of the acquisition module 21, the preprocessing module 22, the identification and positioning module 23, the repair module 24 and the evaluation module 25 correspond one by one to the steps in the above-mentioned EEG signal adaptive identification method, so they are not repeated here.

[0130] In the several embodiments provided by the present invention, it should be understood that the disclosed system, device or method can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of modules / units is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules or units can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules or units, which can be electrical, mechanical or other forms.

[0131] The modules / units described as separate components may or may not be physically separated, and the components displayed as modules / units may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules / units may be selected according to actual needs to achieve the purpose of the embodiments of the present invention. For example, the functional modules / units in the various embodiments of the present invention may be integrated into one processing module, or each module / unit may exist physically separately, or two or more modules / units may be integrated into one module / unit.

[0132] Those skilled in the art should further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0133] The embodiment of the present invention also provides a computer-readable storage medium. A person of ordinary skill in the art can understand that all or part of the steps in the method for adaptively identifying EEG signals in the above embodiment can be completed by instructing a processor through a program, and the program can be stored in a computer-readable storage medium, and the storage medium is a non-transitory medium, such as a random access memory, a read-only memory, a flash memory, a hard disk, a solid-state hard disk, a magnetic tape, a floppy disk, an optical disc, and any combination thereof. The above storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more available media. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a digital video disc (DVD)), or a semiconductor medium (e.g., a solid-state disk (SSD)), etc.

[0134] The embodiment of the present invention further provides a terminal, which includes a processor and a memory.

[0135] The memory is used to store computer programs.

[0136] The memory includes: ROM, RAM, disk, USB flash drive, memory card or CD and other media that can store program codes.

[0137] The processor is connected to the memory and is used to execute the computer program stored in the memory so that the terminal executes the above-mentioned electroencephalogram signal adaptive recognition method.

[0138] Preferably, the processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0139] like Figure 3 As shown, the terminal of the present invention is in the form of a general computing device. The components of the terminal may include but are not limited to: one or more processors or processing units 31, a memory 32, and a bus 33 connecting different system components (including the memory 32 and the processing unit 31).

[0140] Bus 33 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor or a local bus using any of a variety of bus architectures. For example, these architectures include but are not limited to Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MAC) bus, Enhanced ISA bus, Video Electronics Standards Association (VESA) local bus and Peripheral Component Interconnect (PCI) bus.

[0141] The terminal typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the terminal, including volatile and non-volatile media, removable and non-removable media.

[0142] The memory 32 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 321 and / or cache memory 322. The terminal may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 323 may be used to read and write non-removable, non-volatile magnetic media ( Figure 3 not shown, usually called a "hard drive"). Although Figure 3Not shown in the figure, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to the bus 33 via one or more data medium interfaces. The memory 32 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of the embodiments of the present invention.

[0143] A program / utility 324 having a set (at least one) of program modules 3241 may be stored, for example, in the memory 32, such program modules 3241 including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. The program modules 3241 generally perform the functions and / or methods of the embodiments described herein.

[0144] The terminal may also communicate with one or more external devices (e.g., keyboards, pointing devices, displays, etc.), one or more devices that enable a user to interact with the terminal, and / or any device that enables the terminal to communicate with one or more other computing devices (e.g., network cards, modems, etc.). Such communication may be performed via input / output (I / O) interface 34. Furthermore, the terminal may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via network adapter 35. Figure 3 As shown, the network adapter 35 communicates with other modules of the terminal via the bus 33. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the terminal, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0145] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A pipeline rupture automatic welding identification and repair method, characterized in that: The method comprises the following steps: Obtain pipeline images from different perspectives collected by multiple cameras; Preprocessing the pipeline image; Identify and locate breaches based on preprocessed pipeline images; For the identified and located breaches, welding repair is performed; Weld inspection and evaluation of repaired breaches.

2. The pipeline rupture automatic welding identification and repair method according to claim 1 is characterized by: Acquiring pipeline images from different angles collected by multiple cameras includes the following steps: Setting the shooting angles and shooting distances of the multiple cameras; according to Adjust the camera acquisition parameters, which include exposure time T, gain G and aperture size A, L represents the ambient light intensity, C represents the complexity of the pipeline surface texture, and k1, k2, k3, k4, k5, k6 represent empirical coefficients; The pipeline image I is acquired based on the shooting angle, the shooting distance and the camera acquisition parameters. i =f(θ i ,d i ,P), where I i represents the pipeline image captured by the i-th camera, P represents the pipeline surface features, θ i represents the shooting angle of the i-th camera, d i Represents the shooting distance of the i-th camera.

3. The pipeline rupture automatic welding identification and repair method according to claim 1 is characterized in that: Preprocessing the pipeline image includes the following steps: According to LoG new =w(x,y)×▽ 2 G filters the pipeline image, where LoG new represents the filtered pipeline image, w(x,y) represents the adaptive weight factor, ▽ 2 G represents LoG filtering, (x, y) represents the pixel of the pipeline image; Perform image enhancement on the filtered pipeline image based on deep learning algorithm; Perform adaptive threshold binarization on the pipeline image after image enhancement; Perform morphological gradient edge detection and optimization on the binarized pipeline image.

4. The pipeline rupture automatic welding identification and repair method according to claim 1 is characterized in that: The breach identification and location based on the preprocessed pipeline image includes the following steps: Based on the preprocessed pipeline image, an image pyramid is constructed to obtain pipeline images of different scales; For each scale of pipeline image, extract local binary pattern features and oriented gradient histogram features; Fusing the local binary pattern feature and the oriented gradient histogram feature to obtain a fused feature; Classifying the fused features based on a support vector machine to identify breaches in the pipeline image, and using pixel points corresponding to the identified breaches as breach feature points; The identified breach feature points under different perspectives are matched and fused to obtain the fused breach position.

5. The pipeline rupture automatic welding identification and repair method according to claim 1 is characterized by: For the identified and located breach, welding repair includes the following steps: Constructing a three-dimensional model of the breach based on the shape and position information of the breach; Constructing a spiral welding planning path based on the three-dimensional model of the breach, wherein the spiral welding planning path is x=rcos(θ), y=rsin(θ), z=hθ, wherein r represents the spiral radius, θ represents the angle parameter, and h represents the pitch; Calculating the welding robot joint angle corresponding to each path point in the spiral wire welding planning path; Inputting the rupture depth, rupture width and pipeline material type into a welding parameter adjustment model to obtain welding parameters, wherein the welding parameters include welding current, welding voltage and welding speed; Fine-tuning the welding parameters based on the welding temperature field distribution and the molten pool shape; Inputting the adjusted welding parameters into the welding prediction model to obtain welding state parameters, wherein the welding dynamic parameters include weld penetration, weld width and welding deformation; The welding parameters are adjusted based on the welding state parameters and a welding quality target.

6. The pipeline rupture automatic welding identification and repair method according to claim 1 is characterized by: The following steps are involved in the weld inspection and evaluation of a repaired breach: Construct welding quality assessment model; The image of the pipeline after the breach is repaired is input into the welding quality assessment model to obtain the welding quality assessment result.

7. The pipeline rupture automatic welding identification and repair method according to claim 1 is characterized by: Weld inspection and evaluation of repaired breaches also includes the following steps: Obtain the internal structure information of the weld collected by the ultrasonic sensor; Obtain the surface flatness and deformation information of the pipe after welding collected by the laser displacement sensor; Based on the pipeline image after welding, the internal structure information of the weld, and the surface flatness and deformation information of the pipeline, a welding quality assessment result is obtained through a welding quality assessment model.

8. A pipeline rupture automatic welding identification and repair system, characterized by: The system includes an acquisition module, a preprocessing module, an identification and positioning module, a repair module and an evaluation module; The acquisition module is used to acquire pipeline images of different viewing angles collected by multiple cameras; The preprocessing module is used to preprocess the pipeline image; The identification and positioning module is used to identify and locate the breach based on the preprocessed pipeline image; The repair module is used to perform welding repair on the identified and located breach; The evaluation module is used to perform welding detection evaluation on the repaired breach.

9. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the pipeline rupture automatic welding identification and repair method described in any one of claims 1 to 7 is implemented.

10. A terminal, characterized in that: include: Processor and memory; The memory is used to store computer programs; The processor is used to execute the computer program stored in the memory, so that the terminal executes the pipeline rupture automatic welding identification and repair method according to any one of claims 1 to 7.

Images