Small line welding seam tracking correction control system and method with adaptive function

Abstract

The present application belongs to the technical field of welding, and discloses a small-line scanning welding seam tracking correction control system and method with self-adaptive function; the method comprises the following steps: collecting welding seam image data and welding seam depth; processing the welding seam image data to obtain a welding seam characteristic image; obtaining welding seam first characteristic data of a HFSL strip welding seam edge contour based on the welding seam characteristic image and the welding seam depth; obtaining welding seam second characteristic data of the HFSL strip welding seam edge contour based on the welding seam characteristic image and the welding seam first characteristic data of the HFSL strip welding seam edge contour; constructing a set based on the welding seam second characteristic data of the HFSL strip welding seam edge contour to obtain a variable-gap welding seam set and a cross-welding seam set; intelligently and adaptively adjusting a welding control process based on the welding seam first characteristic data, the variable-gap welding seam set and the cross-welding seam set; the present application intelligently and adaptively adjusts the variable-gap welding seam and the cross-welding seam, and improves the welding quality.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of welding, more particularly, the present application relates to a small-line scanning welding seam tracking correction control system and method with adaptive function. BACKGROUND

[0002] In the field of modern industrial manufacturing, welding is widely used in high-precision manufacturing. With the complication of manufacturing process, the diversity of welding seam morphology puts forward higher requirements for welding precision. Among them, variable gap welding seam and cross welding seam are relatively typical complex welding seam types. The irregularity of welding seam morphology makes it difficult for traditional fixed path welding method to meet the demand of high-precision welding.

[0003] In the prior art, although a certain degree of welding seam tracking automatic welding can be achieved, the welding seam tracking automatic welding of the prior art mainly relies on fixed trajectory planning or adjustment method based on simple sensing feedback. In complex welding seam, the adaptability is insufficient, and it is difficult to deal with variable gap welding seam and cross welding seam, so as to meet the requirement of high-precision welding. That is, it is impossible to optimize the welding gun movement for variable gap welding seam, resulting in uneven welding seam filling and affecting the welding quality. It is also impossible to dynamically optimize the welding seam characteristics for cross welding seam, resulting in welding gun speed adjustment lag or unreasonable adjustment range, affecting the stability of the welding process.

[0004] In view of the above problems, the present application provides a small-line scanning welding seam tracking correction control system and method with adaptive function, which uses a small-line scanning sensor to scan the welding seam morphology in real time, and dynamically adjusts the welding parameters in combination with the welding control system, so as to realize intelligent adaptive adjustment of variable gap welding seam and cross welding seam. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned purpose, the present application provides the following technical scheme: a small-line scanning welding seam tracking correction control method with adaptive function, comprising:

[0006] Collecting welding seam image data and welding seam depth;

[0007] Processing the welding seam image data to obtain a welding seam characteristic image;

[0008] Based on the welding seam characteristic image and the welding seam depth, welding seam first characteristic data corresponding to the HFSL strip welding seam edge contour is obtained;

[0009] Based on the welding seam characteristic image and the welding seam first characteristic data corresponding to the HFSL strip welding seam edge contour, welding seam second characteristic data corresponding to the HFSL strip welding seam edge contour is obtained;

[0010] Based on the welding seam second characteristic data corresponding to the HFSL strip welding seam edge contour, a set is constructed to obtain a variable gap welding seam set and a cross welding seam set;

[0011] Intelligently and adaptively adjusting the welding control process based on the first weld feature data, the variable-gap weld set and the cross-weld set.

[0012] Further, the method of intelligently and adaptively adjusting the welding control process comprises:

[0013] If the variable-gap weld set is not empty, intelligently and adaptively adjusting the welding control process for the variable-gap welds;

[0014] If the cross-weld set is not empty, intelligently and adaptively adjusting the welding control process for the cross-welds.

[0015] Further, the method of intelligently and adaptively adjusting the welding control process for the variable-gap welds comprises:

[0016] Extracting the variable-gap welds from the variable-gap weld set in sequence and obtaining the first weld feature data corresponding to the variable-gap welds; based on the first weld feature data, extracting the curvature change set of the variable-gap welds; setting a curvature change threshold value and marking the data sampling points with curvature values greater than the curvature change threshold value in the curvature change set as variable-gap points;

[0017] Pre-setting a welding gun moving speed threshold value; pre-setting a welding gun distance threshold value one and a welding gun distance threshold value two, the welding gun distance threshold value one being smaller than the welding gun distance threshold value two;

[0018] If the variable-gap point is not adjacent to the variable-gap point, marking the corresponding variable-gap point as an isolated variable-gap point;

[0019] If the variable-gap point is adjacent to the variable-gap point, marking the corresponding variable-gap point as a continuous variable-gap point;

[0020] When the distance between the welding gun and the isolated variable-gap point reaches the welding gun distance threshold value one, adjusting the welding gun moving speed to the welding gun moving speed threshold value;

[0021] When the distance between the welding gun and the continuous variable-gap point reaches the welding gun distance threshold value two, calculating the latest welding gun moving speed according to the current welding gun moving speed, the number of variable-gap points in the continuous variable-gap point and the curvature of the variable-gap points in the continuous variable-gap point; adjusting the welding gun moving speed to the latest welding gun moving speed.

[0022] Further, the method of intelligently and adaptively adjusting the welding control process for the cross-welds comprises:

[0023] Extract the cross-welds from the cross-weld set in turn, and obtain the number of weld edge contours of the cross-weld, denoted as JCSL, extract the weld first feature data corresponding to the JCSL weld edge contours, and input the weld first feature data into the weld weight setting model to obtain the cross-welding weight corresponding to the weld edge contour;

[0024] Sort the JCSL weld edge contours in descending order according to the cross-welding weight values to obtain the cross-weld welding sequence, and perform the welding operation according to the cross-weld welding sequence.

[0025] Further, the method for obtaining the weld first feature data corresponding to the HFSL weld edge contours comprises:

[0026] S100: Let the initial value of hfsl be 1, and the value range of hfsl be 1 to HFSL; mark the three-dimensional coordinates of the position of the small line scanning camera as the center coordinate position;

[0027] S101: Divide the first hfsl weld edge contour into M data sampling points; calculate the curvature corresponding to the M data sampling points according to the pixel point two-dimensional coordinates of the M data sampling points, and construct the curvature variation set of the first hfsl weld edge contour from the curvature corresponding to the M data sampling points;

[0028] Calculate the weld angle of the M data sampling points based on the pixel point three-dimensional coordinates of the M data sampling points and the center coordinate position, and construct the weld angle variation set from the weld angle of the M data sampling points;

[0029] S102: Construct the weld first feature data of the first hfsl weld edge contour from the curvature variation set and the weld angle variation set;

[0030] S103: Let hfsl = hfsl + 1, if hfsl is less than or equal to HFSL, continue to execute S101 to S102; if hfsl is greater than HFSL, obtain the weld first feature data corresponding to the HFSL weld edge contours, and end the current process.

[0031] Further, the method for obtaining the curvature corresponding to the M data sampling points comprises:

[0032] S400: Let the initial value of m be 1, and the value range of m be 1 to M;

[0033] S401: Select two data sampling points adjacent to the mth data sampling point, denoted as adjacent data point one and adjacent data point two; substitute the pixel point two-dimensional coordinates of the mth data sampling point, adjacent data point one and adjacent data point two into the standard equation of a circle to construct a standard equation system of a circle;

[0034] S402: Solve the standard equations of a circle to obtain the radius of the circle of curvature;

[0035] S403: Take the reciprocal of the radius of the curvature circle to obtain the curvature of the m-th data sampling point;

[0036] S404: Let m = m + 1. If m is less than or equal to M, continue executing S401 to S403; if m is greater than M, obtain the curvature of M data sampling points and stop the current process.

[0037] Furthermore, the method for obtaining the weld angle of the M data sampling points includes:

[0038] S500: Let the initial value of m be 1, and the range of m is from 1 to M;

[0039] S501: Select a data sampling point adjacent to the m-th data sampling point and denot it as the first data sampling point. Select a data sampling point adjacent to the first data sampling point and denot it as the second data sampling point. The second data sampling point is different from the m-th data sampling point.

[0040] S502: Based on the three-dimensional coordinates of the m-th data sampling point and the first data sampling point, a first direction vector is constructed; based on the three-dimensional coordinates of the first data sampling point and the second data sampling point, a second direction vector is constructed.

[0041] S503: Calculate the magnitude of the first direction vector, denoted as the first direction vector magnitude; calculate the magnitude of the second direction vector, denoted as the second direction vector magnitude.

[0042] The weld angle of the m-th data sampling point is calculated based on the first direction vector, the second direction vector, the magnitude of the first direction vector and the magnitude of the second direction vector.

[0043] S504: Let m = m + 1. If m is less than or equal to M, continue executing S501 to S503; if m is greater than M, obtain the weld angles of M data sampling points and end the current process.

[0044] Furthermore, the method for obtaining the second feature data of the weld corresponding to the edge contour of the HFSL weld includes:

[0045] S200: Set the initial value of hfsl to 1, and the value range of hfsl is from 1 to HFSL;

[0046] S201: Select the edge contour of the hfsl-th weld seam, extract the weld seam edge contours adjacent to the hfsl-th weld seam edge contour from the weld seam feature image, and record them as adjacent weld seam edge contours. Record the number of adjacent weld seam edge contours as... ; a number of adjacent weld seam edge profiles representing the first hfsl weld seam edge profile;

[0047] S202: based on the weld feature image, the first hfsl weld seam edge profile, and perform feature extraction on the weld first feature data of the i-th adjacent weld seam edge profile to obtain weld second feature data corresponding to the first hfsl weld seam edge profile;

[0048] S203: let hfsl = hfsl + 1, if hfsl is less than or equal to HFSL, continue to execute S201 to S202; if hfsl is greater than HFSL, obtain the weld second feature data corresponding to the HFSL weld seam edge profile, and end the current process.

[0049] Further, the method for obtaining the weld second feature data corresponding to the first hfsl weld seam edge profile comprises:

[0050] S300: let the initial value of i be 1, and the value range of i be 1 to ;

[0051] S301: divide the first hfsl weld seam edge profile and the i-th adjacent weld seam edge profile in the weld feature image into N data sampling points, take the distance between each data sampling point and the first hfsl weld seam edge profile and the i-th adjacent weld seam edge profile as the weld gap corresponding to the data sampling point, thereby obtaining the weld gap corresponding to the N data sampling points, and construct the weld gap corresponding to the N data sampling points into a weld gap change set;

[0052] obtain the weld first feature data of the first hfsl weld seam edge profile and the weld first feature data of the i-th adjacent weld seam edge profile;

[0053] obtain the curvature change set and the weld angle change set from the weld first feature data of the first hfsl weld seam edge profile, and obtain the curvature change set and the weld angle change set from the weld first feature data of the i-th adjacent weld seam edge profile, which are denoted as an adjacent curvature change set and an adjacent weld angle change set, respectively;

[0054] S302: perform point-by-point difference on the curvature change set and the adjacent curvature change set to obtain a curvature change trend set, and perform point-by-point difference on the weld angle change set and the adjacent weld angle change set to obtain a weld angle change trend set;

[0055] S303: construct the weld gap change set, the curvature change trend set, and the weld angle change trend set into a to-be-processed data subset corresponding to the i-th adjacent weld seam edge profile;

[0056] S304: Let i = i + 1, if i is less than or equal to... If i is greater than 1, then continue executing S301 to S303; Then we get For the subset of data to be processed corresponding to the edge contours of adjacent weld seams, execute S305;

[0057] S305: Yes The subsets of data to be processed corresponding to the edge contours of adjacent welds are processed separately to obtain the second feature data of the weld corresponding to the edge contour of the hfslth weld.

[0058] Furthermore, on The methods for processing the subsets of data to be processed corresponding to the edge contours of adjacent weld seams include:

[0059] Will The subsets of data to be processed corresponding to the edge contours of adjacent welds are respectively input into the weld variation diagnosis model to obtain... The weld change diagnosis results corresponding to the edge contours of adjacent welds, wherein the weld change diagnosis results include normal and abnormal;

[0060] If the weld change diagnosis result is normal, the corresponding subset of data to be processed does not require further processing;

[0061] If the weld change diagnosis result is abnormal, the corresponding subset of data to be processed will be added to the second feature data of the weld.

[0062] Furthermore, the method for constructing the variable gap weld set and the cross weld set includes:

[0063] The second feature data of the weld corresponding to the edge contour of the HFSL weld are input into the weld type diagnosis model to obtain the weld type corresponding to the edge contour of the HFSL weld; the weld type includes variable gap weld and cross weld;

[0064] The weld edge profile of the weld type variable gap weld is denoted as variable gap weld and added to the variable gap weld collection;

[0065] The weld edge profile of the weld type is cross weld, and it is added to the cross weld collection.

[0066] An adaptive small-line sweep weld seam tracking and correction control system, implementing the aforementioned adaptive small-line sweep weld seam tracking and correction control method, includes:

[0067] The first acquisition module is used to acquire weld image data and weld depth;

[0068] The image processing module is used to process weld image data to obtain weld feature images;

[0069] The first processing module obtains weld first feature data corresponding to the HFSL strip weld edge contour based on the weld feature image and the weld depth;

[0070] The second processing module obtains weld second feature data corresponding to the HFSL strip weld edge contour based on the weld feature image and the weld first feature data corresponding to the HFSL strip weld edge contour;

[0071] The set construction module performs set construction based on the weld second feature data corresponding to the HFSL strip weld edge contour to obtain a variable gap weld set and a cross weld set;

[0072] The intelligent optimization module intelligently and adaptively adjusts the welding control process based on the weld first feature data, the variable gap weld set and the cross weld set.

[0073] Compared with the prior art, the small-line scanning weld tracking correction control system and method with the adaptive function have the following technical effects and advantages:

[0074] For the variable gap weld, the present scheme extracts variable gap points through the curvature change set, and distinguishes between isolated variable gap points and continuous variable gap points. In combination with the relative distance between the welding torch and the variable gap points, the welding torch moving speed is dynamically adjusted. When the welding torch approaches the isolated variable gap point, the moving speed is reduced in advance to improve the fusion quality. When the welding torch enters the continuous variable gap point area, the latest welding torch speed is calculated based on the number of variable gap points and the curvature change, so as to ensure the stability of the penetration depth and the uniformity of the weld filling during the welding process, thereby effectively dealing with the welding defects caused by the sudden change of the weld gap.

[0075] For the cross weld, the present scheme extracts the weld edge contour and calculates the cross welding weight to prioritize the welds and perform welding in the order of decreasing weight, so as to ensure the rationality of the weld fusion sequence and reduce the welding residual stress and deformation. Meanwhile, the present scheme further combines the weld first feature data and the cross weld set to dynamically adjust the distance threshold of the welding torch and the cross point of the cross weld and the welding torch moving speed when the welding torch approaches the cross point, so that the welding process is more stable, the weld quality is improved, and the operation stability is improved.

[0076] Compared with the traditional weld tracking method, the present scheme has higher intelligent adaptive ability, can adjust the welding path and speed in real time according to the weld characteristics, optimizes the welding process in a complex weld environment, improves the weld quality consistency, reduces the welding defects, and improves the welding automation level and industrial production efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0077] Figure 1The figure is a small line scanning welding seam tracking correction control system with adaptive function of the embodiment 1 of the present application.

[0078] Figure 2 The figure is a small line scanning welding seam tracking correction control method flow chart with adaptive function of the embodiment 3 of the present application.

[0079] Figure 3 The figure is a small line scanning welding seam tracking correction control system with adaptive function of the embodiment 2 of the present application.

[0080] Figure 4 The figure is a method flow chart for intelligent adaptive adjustment of a welding control process.

[0081] Figure 5 The figure is a method flow chart for intelligent adaptive adjustment of a welding control process for a variable gap welding seam.

[0082] Figure 6 The figure is a method flow chart for acquiring welding seam first feature data corresponding to a HFSL strip welding seam edge profile.

[0083] Figure 7 The figure is a method flow chart for acquiring welding seam second feature data corresponding to a HFSL strip welding seam edge profile. DETAILED DESCRIPTION

[0084] The technical solutions in the embodiments of the present application will be described in detail, clearly and completely below with reference to the drawings in the embodiments of the present application. It should be particularly noted that the specific embodiments described below are only used to better illustrate and describe the technical solutions of the present application, and are intended to enable those skilled in the art to better understand and implement the present application, and should not be understood as limiting the protection scope of the present application. Those skilled in the art can modify, adjust or equivalently replace the present application according to the content disclosed in the present application without departing from the spirit and essence of the present application, and these should be regarded as the protection scope of the present application.

[0085] Embodiment 1

[0086] Please refer to Figure 1 As shown in the figure, the embodiment discloses a small line scanning welding seam tracking correction control system with adaptive function, which comprises a first acquisition module, an image processing module, a first processing module, a second processing module, a set construction module and an intelligent optimization module. Each module is connected through wired and / or wireless connection to realize data transmission.

[0087] The first acquisition module is used for acquiring welding seam image data and welding seam depth. The welding seam image data comprises a welding seam image and pixel point two-dimensional coordinates corresponding to pixel points in the welding seam image. That is, each pixel point in the welding seam image has corresponding pixel point two-dimensional coordinates in the welding seam image, and the pixel point two-dimensional coordinates are used for subsequent welding seam feature extraction and trajectory calculation.

[0088] The weld seam image is acquired by a small line scan camera, which is an image acquisition device. In the weld seam tracking scene, the small line scan camera plays a key role. It acquires image information of the weld seam by means of line-by-line scanning. Compared with a traditional area array camera, the small line scan camera has the characteristics of high resolution and high scanning speed, and can capture detailed information of the weld seam more clearly and quickly. In the welding environment, the small line scan camera can scan along the direction of the weld seam to acquire key data such as the contour and position of the weld seam. The weld seam depth is acquired by a laser sensor.

[0089] The image processing module is configured to process the weld seam image data to obtain a weld seam feature image.

[0090] The method for obtaining the weld seam feature image comprises:

[0091] acquiring the weld seam image from the weld seam image data and converting the weld seam image into a weld seam grayscale image;

[0092] processing the weld seam grayscale image using an edge detection algorithm to extract the weld seam edge contour;

[0093] counting the number of weld seam edge contours in the weld seam grayscale image, and denoting the number of weld seam edge contours as HFSL; and respectively marking each weld seam edge contour in the weld seam grayscale image to obtain HFSL weld seam edge contours marked with values.

[0094] The HFSL weld seam edge contours marked with values constitute the weld seam feature image.

[0095] The first processing module is configured to obtain weld seam first feature data corresponding to the HFSL weld seam edge contours based on the weld seam feature image and the weld seam depth.

[0096] As shown in Figure 6 the method for obtaining the weld seam first feature data corresponding to the HFSL weld seam edge contours comprises:

[0097] S100: setting the initial value of hfsl to 1, and the value range of hfsl to 1 to HFSL; and marking the three-dimensional coordinates of the position of the small line scan camera as the central coordinate position;

[0098] S101: dividing the hfsl weld seam edge contour into M data sampling points; calculating the curvatures corresponding to the M data sampling points according to the two-dimensional coordinates of the pixel points of the M data sampling points, and constructing the curvatures corresponding to the M data sampling points into a curvature variation set of the hfsl weld seam edge contour.

[0099] The weld angle of the M data sampling points is calculated based on the three-dimensional coordinates of the pixel points and the center coordinate position of the M data sampling points, and the weld angle of the M data sampling points is constructed into a weld angle change set;

[0100] S102: The curvature change set and the weld angle change set are constructed into weld first feature data of the HFSL weld edge contours;

[0101] S103: Let hfsl = hfsl + 1, if hfsl is less than or equal to HFSL, continue to execute S101 to S102; if hfsl is greater than HFSL, the weld first feature data corresponding to the HFSL weld edge contours is obtained, and the current process is ended.

[0102] The method for obtaining the three-dimensional coordinates of the pixel points comprises:

[0103] The two-dimensional coordinates of the pixel points corresponding to the HFSL weld edge contours in the weld feature image are obtained, and based on the two-dimensional coordinates of the pixel points corresponding to the HFSL weld edge contours and the weld depth, for each weld edge contour, the two-dimensional coordinates of the pixel points of the weld edge contour are extracted, and the corresponding weld depth is taken as the Z-axis coordinate of the three-dimensional coordinates, that is, the two-dimensional coordinates of the pixel points are combined with the corresponding weld depth to be converted into the corresponding three-dimensional coordinates of the pixel points;

[0104] For example, the two-dimensional coordinate information of the pixel point is (3, 4), the weld depth corresponding to the pixel point is 5, and the three-dimensional coordinates of the pixel point are (3, 4, 5).

[0105] The method for obtaining the curvatures corresponding to the M data sampling points comprises:

[0106] S400: Let the initial value of m be 1, and the value range of m be 1 to M;

[0107] S401: Select two data sampling points adjacent to the mth data sampling point, and mark them as adjacent data point one and adjacent data point two respectively; the two-dimensional coordinates of the pixel points of the mth data sampling point, the adjacent data point one and the adjacent data point two are substituted into the standard equation of a circle to construct a standard equation set of a circle;

[0108] S402: The standard equation set of the circle is solved to obtain the radius of the curvature circle;

[0109] S403: The radius of the curvature circle is taken as the reciprocal to obtain the curvature of the mth data sampling point;

[0110] S404: Let m = m + 1, if m is less than or equal to M, continue to execute S401 to S403; if m is greater than M, the curvatures of the M data sampling points are obtained, and the current process is ended.

[0111] The standard equation set of the circle is:

[0112] ;

[0113] wherein, is the two-dimensional coordinate of the mth data sampling point, is the two-dimensional coordinate of the 1st data sampling point, is the two-dimensional coordinate of the 2nd data sampling point; is the center coordinate of the curvature circle, obtained by solving the equation group, is the radius of the curvature circle, obtained by solving the equation group.

[0114] The process of solving the equation group is as follows:

[0115] Let be denoted as equation (1);

[0116] Let be denoted as equation (2);

[0117] Let be denoted as equation (3);

[0118] Subtract equation (2) from equation (1) to obtain:

[0119] ;

[0120] Expand to obtain:

[0121] ;

[0122] Eliminate and , and rearrange to obtain:

[0123] ; denoted as equation (4);

[0124] Subtract equation (3) from equation (2) to obtain:

[0125] ;

[0126] Expand to obtain:

[0127] ;

[0128] Eliminate and , and rearrange to obtain:

[0129] ; denoted as equation (5);

[0130] Equations (4) and (5) form a binary linear equation group about a and b:

[0131] ;

[0132] Let , , , , , ;

[0133] The binary linear equation group becomes:

[0134] ;

[0135] Simplify by ;

[0136] Substitute into to get ;

[0137] Simplify to get ;

[0138] Substitute into , and solve for the value of a; thus obtaining the values of a and b;

[0139] Substitute the values of a and b into the standard equation of the circle, and thus solve for the value of r.

[0140] The method for obtaining the weld angle of M data sampling points comprises the following steps:

[0141] S500: Let the initial value of m be 1, and the value range of m be 1 to M;

[0142] S501: Select one data sampling point adjacent to the mth data sampling point as a first data sampling point, and select one data sampling point adjacent to the first data sampling point as a second data sampling point, wherein the second data sampling point is different from the mth data sampling point;

[0143] S502: Based on the three-dimensional coordinates of the mth data sampling point and the first data sampling point, a first direction vector is constructed; based on the three-dimensional coordinates of the first data sampling point and the second data sampling point, a second direction vector is constructed;

[0144] S503: Calculate the module length of the first direction vector, denoted as the first direction vector module length; calculate the module length of the second direction vector, denoted as the second direction vector module length;

[0145] Based on the first direction vector, the second direction vector, the first direction vector module length and the second direction vector module length, the weld angle of the mth data sampling point is calculated;

[0146] ​S504: m = m + 1, if m is less than or equal to M, continue to execute S501 to S503; if m is greater than M, the weld angle of the M data sampling points is obtained, and the current process is ended.

[0147] The construction method of the first direction vector comprises:

[0148] ;

[0149] wherein, is the first direction vector of the mth data sampling point, is the three-dimensional coordinates of the m data sampling points; is the three-dimensional coordinates of the first data sampling point.

[0150] The construction method of the second direction vector comprises:

[0151] ;

[0152] wherein, is the second direction vector of the mth data sampling point, is the three-dimensional coordinates of the second data sampling point.

[0153] The method for obtaining the first direction vector length comprises:

[0154] ;

[0155] wherein, is the first direction vector length of the mth data sampling point.

[0156] The method for obtaining the second direction vector length comprises:

[0157] ;

[0158] wherein, is the second direction vector length of the mth data sampling point.

[0159] The method for obtaining the weld angle comprises:

[0160] ;

[0161] wherein, is the weld angle of the mth data sampling point, is the inverse cosine function.

[0162] The second processing module obtains the weld second feature data corresponding to the HFSL strip weld edge contour based on the weld feature image and the weld first feature data corresponding to the HFSL strip weld edge contour.

[0163] AsFigure 7 As shown, the method for obtaining the weld second feature data corresponding to the HFSL weld edge profile comprises:

[0164] S200: let the initial value of hfsl be 1, and the value range of hfsl be 1 to HFSL;

[0165] S201: select the first hfsl weld edge profile, extract the weld edge profiles adjacent to the first hfsl weld edge profile from the weld feature image, and mark the adjacent weld edge profiles as adjacent weld edge profiles; the number of adjacent weld edge profiles is marked as ; The number of adjacent weld edge profiles of the first hfsl weld edge profile is represented as

[0166] S202: based on the weld feature image, the first hfsl weld edge profile, and the weld first feature data of the adjacent weld edge profile, perform feature extraction to obtain the weld second feature data corresponding to the first hfsl weld edge profile;

[0167] S203: let hfsl = hfsl + 1, if hfsl is less than or equal to HFSL, continue to execute S201 to S202; if hfsl is greater than HFSL, obtain the weld second feature data corresponding to the HFSL weld edge profile, and end the current process.

[0168] The method for obtaining the weld second feature data corresponding to the first hfsl weld edge profile comprises:

[0169] S300: let the initial value of i be 1, and the value range of i be 1 to ;

[0170] S301: divide the first hfsl weld edge profile and the first i adjacent weld edge profile into N data sampling points in the weld feature image, and take the distance between each data sampling point and the first i adjacent weld edge profile as the weld gap corresponding to the data sampling point, thereby obtaining the weld gap corresponding to the N data sampling points, and constructing the weld gap variation set from the weld gap corresponding to the N data sampling points;

[0171] Obtain the weld first feature data of the first hfsl weld edge profile and the weld first feature data of the first i adjacent weld edge profile;

[0172] Obtain the curvature variation set and the weld angle variation set from the weld first feature data of the first hfsl weld edge profile; obtain the curvature variation set and the weld angle variation set from the weld first feature data of the first i adjacent weld edge profile, and mark them as adjacent curvature variation set and adjacent weld angle variation set, respectively;​

[0173] S302: Subtract the curvature change set from the adjacent curvature change set point by point to obtain the curvature change trend set; subtract the weld angle change set from the adjacent weld angle change set point by point to obtain the weld angle change trend set.

[0174] S303: Construct the set of weld gap variation, the set of curvature variation trend, and the set of weld angle variation trend into a subset of data to be processed corresponding to the edge contour of the i-th adjacent weld;

[0175] S304: Let i = i + 1, if i is less than or equal to... If i is greater than 1, then continue executing S301 to S303; Then we get For the subset of data to be processed corresponding to the edge contours of adjacent weld seams, execute S305;

[0176] S305: Yes The subsets of data to be processed corresponding to the edge contours of adjacent welds are processed separately to obtain the second feature data of the weld corresponding to the edge contour of the hfslth weld.

[0177] Methods for point-by-point subtraction between the set of curvature changes and adjacent sets of curvature changes include:

[0178] h = 1, 2, 3, ..., M;

[0179] in, This represents the h-th curvature change trend value in the set of curvature change trends. Let h be the h-th curvature in the set of curvature changes of the edge profile of the hfsl-th weld. Let h be the curvature of the set of adjacent curvature changes of the i-th adjacent weld edge profile.

[0180] It should be noted that M data sampling points correspond to M curvatures, that is, the set of curvature changes contains M curvatures, and the adjacent set of curvature changes contains M curvatures. Therefore, the value range of h is from 1 to M.

[0181] The method of subtracting the set of weld angle changes and the set of angle changes of adjacent welds point by point includes:

[0182] k = 1, 2, 3, ..., M;

[0183] in, This represents the k-th weld angle variation trend value in the set of weld angle variation trends. Let k be the weld angle in the set of weld angle variations of the edge profile of the hfsl-th weld. The angle of the kth weld is in the set of angle changes of adjacent welds on the edge profile of the i-th adjacent weld.

[0184] It should be noted that M data sampling points correspond to M weld angles, that is, the set of weld angle changes contains M weld angles, and the set of adjacent weld angle changes contains M weld angles. Therefore, the value of k ranges from 1 to M.

[0185] right The methods for processing the subsets of data to be processed corresponding to the edge contours of adjacent weld seams include:

[0186] Will The subsets of data to be processed corresponding to the edge contours of adjacent welds are respectively input into the weld variation diagnosis model to obtain... The weld change diagnosis results corresponding to the edge contours of adjacent welds, wherein the weld change diagnosis results include normal and abnormal;

[0187] If the weld change diagnosis result is normal, the corresponding subset of data to be processed does not require further processing;

[0188] If the weld change diagnosis result is abnormal, the corresponding subset of data to be processed will be added to the second feature data of the weld.

[0189] The training method for the weld change diagnostic model includes:

[0190] A pre-constructed weld change diagnosis dataset is provided, comprising Q sets of weld change diagnosis data and corresponding weld change diagnosis results, where Q is a positive integer greater than 0. The weld change diagnosis data includes a subset of data to be processed, comprising sets of weld gap changes, curvature change trends, and weld angle change trends. The weld change diagnosis dataset is divided into a weld change diagnosis data training set and a weld change diagnosis data validation set. The weld change diagnosis data training set is used for parameter learning of the weld change diagnosis model, while the weld change diagnosis data validation set is used for real-time evaluation of the generalization ability of the weld change diagnosis model.

[0191] In the training process of the weld change diagnosis model, a deep neural network structure based on a multilayer perceptron is adopted, weld change diagnosis data are converted into feature vectors as input, nonlinear features in the data are extracted through a hidden layer, and finally a probability distribution of the weld change diagnosis result is generated by using a softmax activation function in the output layer, and the weld change diagnosis result corresponding to the maximum probability is output as the final prediction result; the training process aims to minimize the cross-entropy loss function, and an early stopping strategy is introduced to monitor the performance of the weld change diagnosis data verification set; when the prediction accuracy on the weld change diagnosis data verification set reaches a preset threshold, it is considered that the weld change diagnosis model has converged, and the training is stopped.

[0192] The softmax activation function is:

[0193] ;

[0194] wherein, is the output probability corresponding to the num-th feature vector, is the num-th feature vector, is the total number of feature vectors, and e is a constant.

[0195] It should be noted that the weld change diagnosis result is used to evaluate whether the current selected weld and its adjacent welds have mutual influence in the welding process, that is, whether the welding quality or shape of the adjacent welds is disturbed when one weld is welded.

[0196] When the weld change diagnosis result is normal, the weld gap change, curvature change and angle change trend of the current selected weld and the adjacent welds remain stable, and the welding process does not interfere with each other, thereby ensuring that the weld quality is not affected.

[0197] When the weld change diagnosis result is abnormal, it indicates that the gap change, curvature change or angle change trend of the current selected weld and the adjacent welds has instability or severe fluctuation, which leads to mutual interference in the welding process, and further affects the welding quality. For example, structural deformation may occur due to stress transmission during welding, or weld forming may be affected due to uneven heat input, thereby causing quality problems such as weld deviation and poor fusion.

[0198] The set construction module constructs a set based on the second weld feature data corresponding to the edge profile of the HFSL strip weld, and obtains a variable gap weld set and a cross weld set.

[0199] The construction method of the variable gap weld set and the cross weld set includes:

[0200] input the second weld feature data corresponding to the edge profile of the HFSL strip weld seam to the weld type diagnosis model respectively, to obtain the weld type corresponding to the edge profile of the HFSL strip weld seam; the weld type includes a variable gap weld seam and a cross weld seam;

[0201] record the weld edge profile with the weld type of variable gap weld seam as variable gap weld seam, and add it to the variable gap weld seam set;

[0202] record the weld edge profile with the weld type of cross weld seam as cross weld seam, and add it to the cross weld seam set.

[0203] It should be noted that the variable gap weld seam refers to the dynamic change of the weld gap along the welding direction. If the welding parameters are fixed, it may cause the problems of incomplete fusion, burn-through or uneven weld formation. The cross weld seam is formed by the mutual intersection of multiple weld seams. The welding sequence and stress distribution of the cross weld seam are crucial. It is difficult to reasonably plan the welding priority of the cross weld seam by traditional welding method, resulting in insufficient weld joint strength and unstable fusion quality. In addition, in complex weld seam structure, the rationality of the welding gun moving speed has a direct impact on the welding quality. If the welding path and the welding gun moving speed cannot adapt to the dynamic change of the weld shape, welding defects will be caused, and the welding consistency will be reduced.

[0204] The training method of the weld type diagnosis model includes:

[0205] Pre-construct a weld type diagnosis data set, which includes G sets of weld type diagnosis data and the weld types corresponding to the G sets of weld type diagnosis data, G being a positive integer greater than 0, the weld type diagnosis data including second weld feature data; divide the weld type diagnosis data set into a weld type diagnosis data training set and a weld type diagnosis data validation set, wherein the weld type diagnosis data training set is used for parameter learning of the weld type diagnosis model, and the weld type diagnosis data validation set is used for real-time evaluation of the generalization ability of the weld type diagnosis model;

[0206] In the training process of the weld type diagnosis model, a deep neural network structure based on a multilayer perceptron is adopted, the weld type diagnosis data is converted into a feature vector as input, the nonlinear features in the data are extracted through the hidden layer, and finally the probability distribution of the weld type is generated by using the softmax activation function in the output layer. The weld type corresponding to the maximum probability is output as the final prediction result; the training process aims to minimize the cross-entropy loss function, and an early stopping strategy is introduced to monitor the performance of the weld type diagnosis data validation set. When the prediction accuracy on the weld type diagnosis data validation set reaches a preset threshold, it is considered that the weld type diagnosis model has converged, and the training is stopped.

[0207] The intelligent optimization module intelligently and adaptively adjusts the welding control process based on the first characteristic data of the weld, the variable-gap weld set and the cross-weld set.

[0208] As shown in Figure 4 The method of intelligently and adaptively adjusting the welding control process comprises:

[0209] If the variable-gap weld set is not empty, the welding control process for the variable-gap weld is intelligently and adaptively adjusted;

[0210] If the cross-weld set is not empty, the welding control process for the cross-weld is intelligently and adaptively adjusted.

[0211] As shown in Figure 5 The method of intelligently and adaptively adjusting the welding control process for the variable-gap weld comprises:

[0212] The variable-gap welds are sequentially extracted from the variable-gap weld set, and the first characteristic data of the weld corresponding to the variable-gap weld is obtained; based on the first characteristic data of the weld, the curvature change set of the variable-gap weld is extracted; the curvature change threshold is set, and the data sampling points with the curvature value greater than the curvature change threshold in the curvature change set are marked as variable-gap points;

[0213] The welding gun moving speed threshold is preset; the welding gun distance threshold one and the welding gun distance threshold two are preset, and the welding gun distance threshold one is smaller than the welding gun distance threshold two;

[0214] If the variable-gap point is not adjacent to the variable-gap point, the corresponding variable-gap point is marked as an isolated variable-gap point;

[0215] If the variable-gap point is adjacent to the variable-gap point, the corresponding variable-gap point is marked as a continuous variable-gap point;

[0216] When the distance between the welding gun and the isolated variable-gap point reaches the welding gun distance threshold one, the welding gun moving speed is adjusted to the welding gun moving speed threshold;

[0217] When the distance between the welding gun and the continuous variable-gap point reaches the welding gun distance threshold two, the latest welding gun moving speed is calculated according to the current welding gun moving speed, the number of variable-gap points in the continuous variable-gap point and the curvature of the variable-gap points in the continuous variable-gap point; the welding gun moving speed is adjusted to the latest welding gun moving speed.

[0218] The calculation method of the latest welding gun moving speed comprises:

[0219] ;

[0220] Wherein, is the latest welding gun moving speed, is the current welding gun moving speed, the number of the continuous variable gap points, the curvature of the first variable gap point in the continuous variable gap points.

[0221] It should be noted that the curvature value greater than the curvature change threshold indicates that the weld trajectory is suddenly changed, resulting in discontinuity of the weld, thereby causing unstable weld quality, and thus it is necessary to adjust the welding gun moving speed at the variable gap point and determine the trigger timing of adjusting the welding gun moving speed, to improve the fusion quality and ensure stable penetration and uniform weld formation during the welding process, thereby effectively dealing with welding defects caused by sudden change of the weld gap, such as incomplete fusion, burn-through or uneven weld, etc.

[0222] The method for intelligently and adaptively adjusting the welding control process of the cross weld includes:

[0223] extracting the cross welds from the cross weld set in sequence, and obtaining the number of weld edge contours of the cross weld, denoted as JCSL, extracting the first weld feature data corresponding to the JCSL weld edge contours, and inputting the first weld feature data into a weld weight setting model to obtain the cross welding weight value corresponding to the weld edge contour;

[0224] sequentially sorting the JCSL weld edge contours according to the cross welding weight values to obtain the cross weld welding sequence, and performing the welding operation according to the cross weld welding sequence.

[0225] The training method of the weld weight setting model includes:

[0226] pre-collecting a weld weight setting dataset, the weld weight setting dataset including D groups of weld weight setting data and the cross welding weight values corresponding to the D groups of weld weight setting data, D being a positive integer greater than 0, the weld weight setting data including the first weld feature data; dividing the weld weight setting dataset into a training set and a validation set, wherein the training set is used to train the weld weight setting model, and the validation set is used to evaluate the generalization performance of the weld weight setting model;

[0227] During the training process of the weld weight setting model, the cross-entropy loss function is minimized as the optimization objective, the early stopping strategy is used to monitor the performance of the validation set, the model performance is optimized by continuously adjusting the network parameters; when the prediction accuracy on the validation set reaches the expected accuracy, it is considered that the weld weight setting model has converged, and the training is stopped; the weld weight setting model is trained using a deep neural network based on a multilayer perceptron;

[0228] The weld weight setting data is converted into a feature vector; the input layer of the weld weight setting model receives the feature vector, extracts the nonlinear relationship in the data through the hidden layer, and finally the output layer of the weld weight setting model calculates the probability distribution of the cross-weld weight value through the softmax activation function, and outputs the cross-weld weight value corresponding to the maximum probability as the final prediction result.

[0229] Embodiment 2

[0230] Please refer to Figure 3 The embodiment provides a small-line scanning weld tracking and correction control system with an adaptive function, and further comprises:

[0231] The cross optimization module further intelligently and adaptively adjusts the cross weld based on the weld first feature data, the cross weld welding sequence and the cross weld set.

[0232] The method for further intelligently and adaptively adjusting the cross weld comprises:

[0233] S600: Let the initial value of jcs1 be 1, and the value range of jcs1 be 1 to JCSL, and JCSL be the number of weld edge contours of the cross weld;

[0234] S601: Record jcs1 as the cross weld welding sequence number; input the weld first feature data of the jcs1th cross weld in the cross weld welding sequence, the cross weld welding sequence number and the cross weld set into the welding parameter setting model to obtain the distance threshold of the welding gun and the cross weld intersection point and the welding gun moving speed threshold;

[0235] S602: When the distance between the welding gun and the cross weld intersection point reaches the distance threshold of the welding gun and the cross weld intersection point, adjust the welding gun moving speed to the welding gun moving speed threshold;

[0236] S603: Let jcs1 = jcs1 + 1, if jcs1 is less than or equal to JCSL, continue to execute S601 to S602; if jcs1 is greater than JCSL, end the current process.

[0237] It should be noted that by further intelligently and adaptively adjusting the cross weld, i.e., dynamically adjusting the welding gun moving speed and the distance threshold of the welding gun and the cross weld intersection point, the welding process is more stable, the weld quality is prevented from being reduced due to uneven heat input, and thus the welding quality and operation stability are improved.

[0238] The training method of the welding parameter setting model comprises:

[0239] Pre-collect a welding parameter setting data set, the welding parameter setting data set includes F sets of welding parameter setting data and the distance threshold value and the welding gun moving speed threshold value of the welding gun and the intersection point of the cross weld corresponding to the F sets of welding parameter setting data, F is a positive integer greater than 0, the welding parameter setting data includes weld first feature data, cross weld welding serial number and cross weld set; the welding parameter setting data set is divided into training set and validation set, wherein the training set is used for training the welding parameter setting model, and the validation set is used for evaluating the generalization performance of the welding parameter setting model;

[0240] During the welding parameter setting model training process, the cross entropy loss function is minimized as the optimization target, the performance of the validation set is monitored using the early stopping strategy, and the model performance is optimized by continuously adjusting the network parameters; when the prediction accuracy on the validation set reaches the expected accuracy, it is considered that the welding parameter setting model has converged, and the training is stopped; the welding parameter setting model is trained using a deep neural network based on a multilayer perceptron;

[0241] The welding parameter setting data is converted into a feature vector; the input layer of the welding parameter setting model receives the feature vector, extracts the nonlinear relationship in the data through the hidden layer, and finally the output layer of the welding parameter setting model calculates the probability distribution of the distance threshold value and the welding gun moving speed threshold value of the welding gun and the intersection point of the cross weld through the softmax activation function, and outputs the distance threshold value and the welding gun moving speed threshold value corresponding to the maximum probability as the final prediction result.

[0242] Embodiment 3

[0243] Please refer to Figure 2 As shown in the figure, the embodiment provides a small line scanning weld tracking correction control method with adaptive function, which comprises:

[0244] Collecting weld image data and weld depth;

[0245] Processing the weld image data to obtain a weld feature image;

[0246] Based on the weld feature image and the weld depth, the weld first feature data corresponding to the HFSL strip weld edge contour is obtained;

[0247] Based on the weld feature image and the weld first feature data corresponding to the HFSL strip weld edge contour, the weld second feature data corresponding to the HFSL strip weld edge contour is obtained;

[0248] Based on the weld second feature data corresponding to the HFSL strip weld edge contour, a set is constructed to obtain a variable gap weld set and a cross weld set;

[0249] Intelligently and self-adaptively adjust the welding control process based on the first characteristic data of the weld, the set of variable-gap welds, and the set of cross-welds.

[0250] The above merely provides the preferred embodiments of the present application, but the protection scope of the present application is not limited thereto, and any modification, equivalent replacement, improvement, etc. within the technical range disclosed by the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0251] Finally: the above merely provides the preferred embodiments of the present application, but the protection scope of the present application is not limited thereto, and any modification, equivalent replacement, improvement, etc. within the technical range disclosed by the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A small wire stitch tracking correction control method with adaptive function, characterized by, The method comprises the following steps: Collecting weld image data and weld depth; Processing the weld image data to obtain a weld feature image; Based on the weld feature image and the weld depth, obtain the first weld feature data corresponding to the HFSL strip weld edge contour; Based on the weld feature image and the first weld feature data corresponding to the HFSL strip weld edge contour, obtain the second weld feature data corresponding to the HFSL strip weld edge contour; Based on the second weld feature data corresponding to the HFSL strip weld edge contour, construct a set to obtain a variable gap weld set and a cross weld set; Intelligently and adaptively adjust the welding control process based on the first weld feature data, the variable gap weld set and the cross weld set; The method for intelligently and adaptively adjusting the welding control process comprises the following steps: If the variable gap weld set is not empty, intelligently and adaptively adjust the welding control process for the variable gap weld; the method for intelligently and adaptively adjusting the welding control process for the variable gap weld comprises the following steps: Extract the variable gap weld from the variable gap weld set in sequence, and obtain the first weld feature data corresponding to the variable gap weld; based on the first weld feature data, extract the curvature change set of the variable gap weld; set the curvature change threshold, and mark the data sampling points with curvature values greater than the curvature change threshold in the curvature change set as variable gap points; Pre-set a welding gun moving speed threshold; pre-set a welding gun distance threshold one and a welding gun distance threshold two, the welding gun distance threshold one is smaller than the welding gun distance threshold two; If the variable gap points are not adjacent, mark the corresponding variable gap points as isolated variable gap points; If the variable gap points are adjacent, mark the corresponding variable gap points as continuous variable gap points; When the distance between the welding gun and the isolated variable gap point reaches the welding gun distance threshold one, adjust the welding gun moving speed to the welding gun moving speed threshold; When the distance between the welding gun and the continuous variable gap point reaches the welding gun distance threshold two, calculate the latest welding gun moving speed according to the current welding gun moving speed, the number of variable gap points in the continuous variable gap point and the curvature of the variable gap points in the continuous variable gap point; adjust the welding gun moving speed to the latest welding gun moving speed; If the cross weld set is not empty, intelligently and adaptively adjust the welding control process for the cross weld; the method for intelligently and adaptively adjusting the welding control process for the cross weld comprises the following steps: Extract the cross weld from the cross weld set in sequence, and obtain the number of weld edge contours of the cross weld, denoted as JCSL; extract the first weld feature data corresponding to the JCSL strip weld edge contour, and input the first weld feature data into a weld weight setting model to obtain the cross welding weight corresponding to the weld edge contour; Sort the JCSL strip weld edge contour in descending order according to the cross welding weight value to obtain the cross weld welding sequence, and perform the welding operation according to the cross weld welding sequence.

2. The small line scan weld tracking correction control method with adaptive function according to claim 1, characterized in that, The method for obtaining the first weld feature data corresponding to the HFSL strip weld edge contour comprises the following steps: S100: set the initial value of hfsl as 1, and the value range of hfsl as 1 to HFSL; mark the three-dimensional coordinates of the position of the small line scanning camera as the center coordinate position; S101: divide the first hfsl weld seam edge profile into M data sampling points; calculate the curvatures corresponding to the M data sampling points according to the two-dimensional coordinates of the pixel points of the M data sampling points, and construct the curvatures corresponding to the M data sampling points into a curvature variation set of the first hfsl weld seam edge profile; calculate the weld angles of the M data sampling points based on the three-dimensional coordinates of the pixel points of the M data sampling points and the center coordinate position, and construct the weld angles of the M data sampling points into a weld angle variation set; S102: construct the curvature variation set and the weld angle variation set into the weld first feature data of the first hfsl weld seam edge profile; S103: let hfsl = hfsl + 1, if hfsl is less than or equal to HFSL, continue to execute S101 to S102; if hfsl is greater than HFSL, obtain the weld first feature data corresponding to the HFSL weld seam edge profile, and end the current process.

3. The small wire scan welding seam tracking correction control method with adaptive function according to claim 2, characterized in that, The method for obtaining the curvatures corresponding to the M data sampling points comprises: S400: let the initial value of m be 1, and the value range of m be 1 to M; S401: select two data sampling points adjacent to the mth data sampling point, and mark them as adjacent data point one and adjacent data point two respectively; substitute the two-dimensional coordinates of the pixel points of the mth data sampling point, the adjacent data point one and the adjacent data point two into the standard equation of a circle respectively, and construct a standard equation set of a circle; S402: solve the standard equation set of a circle to obtain the radius of the curvature circle; S403: take the reciprocal of the radius of the curvature circle to obtain the curvature of the mth data sampling point; S404: let m = m + 1, if m is less than or equal to M, continue to execute S401 to S403; if m is greater than M, obtain the curvatures of the M data sampling points, and stop the current process.

4. The small wire scan welding seam tracking correction control method with adaptive function according to claim 3, characterized in that, The method for obtaining the weld angles of the M data sampling points comprises: S500: let the initial value of m be 1; S501: select one data sampling point adjacent to the mth data sampling point as a first data sampling point, and select one data sampling point adjacent to the first data sampling point as a second data sampling point, which is different from the mth data sampling point; S502: based on the three-dimensional coordinates of the mth data sampling point and the first data sampling point, a first direction vector is constructed; based on the three-dimensional coordinates of the first data sampling point and the second data sampling point, a second direction vector is constructed; S503: calculate the module length of the first direction vector, and mark it as the first direction vector module length; calculate the module length of the second direction vector, and mark it as the second direction vector module length; calculate the weld angle of the mth data sampling point based on the first direction vector, the second direction vector, the first direction vector module length and the second direction vector module length; S504: let m = m + 1, if m is less than or equal to M, continue to execute S501 to S503; if m is greater than M, obtain the weld angles of the M data sampling points, and end the current process.

5. The small wire scan welding seam tracking correction control method with adaptive function according to claim 1, characterized in that, The method for obtaining the weld second feature data corresponding to the HFSL weld seam edge profile comprises: S200: let the initial value of hfsl be 1, and the value range of hfsl be 1 to HFSL; S201: select the first hfsl piece of weld edge contour, extract the weld edge contour adjacent to the first hfsl piece of weld edge contour from the weld feature image, mark as adjacent weld edge contour, mark the number of adjacent weld edge contours as ; the number of adjacent weld edge contours of the first hfsl piece of weld edge contour; S202: based on the weld feature image, the first hfsl weld edge contour and S202: based on the weld feature image, the first hfsl weld edge contour and S202: based on the weld feature image, the first hfsl weld edge contour and S203: set hfsl=hfsl+1, if hfsl is less than or equal to HFSL, continue to execute S201 to S202; if hfsl is greater than HFSL, obtain the second weld feature data corresponding to the HFSLth weld edge contour, and end the current process.

6. The small wire scan welding seam tracking correction control method with adaptive function according to claim 5, characterized in that, The method for obtaining the second weld feature data corresponding to the HFSLth weld edge contour comprises: S300: Let the initial value of i be 1, and the value range of i be 1 to ; S301: divide the HFSLth weld edge contour and the ith adjacent weld edge contour into N data sampling points in the weld feature image, take the distance between each data sampling point and the ith adjacent weld edge contour as the weld gap corresponding to the data sampling point, and thus obtain the weld gaps corresponding to the N data sampling points, and construct the weld gap variation set from the weld gaps corresponding to the N data sampling points; obtain the first weld feature data of the HFSLth weld edge contour and the first weld feature data of the ith adjacent weld edge contour; obtain the curvature variation set and the weld angle variation set from the first weld feature data of the HFSLth weld edge contour, and obtain the curvature variation set and the weld angle variation set from the first weld feature data of the ith adjacent weld edge contour, and denote the adjacent curvature variation set and the adjacent weld angle variation set respectively; S302: perform point-by-point difference between the curvature variation set and the adjacent curvature variation set to obtain the curvature trend set, and perform point-by-point difference between the weld angle variation set and the adjacent weld angle variation set to obtain the weld angle trend set; S303: construct the ith adjacent weld edge contour into a to-be-processed data subset; S304: let i = i + 1, if i is less than or equal to , then continue to execute S301 to S303; if i is greater than , then obtain the subset of data to be processed corresponding to the adjacent weld edge contour, and execute S305; S305: process the subsets of data corresponding to the first hfsl weld seam edge profile and the second hfsl weld seam edge profile, respectively, to obtain weld seam characteristic data corresponding to the first hfsl weld seam edge profile and the second hfsl weld seam edge profile, respectively. S305: process the subsets of data corresponding to the first hfsl weld seam edge profile and the second hfsl weld seam edge profile, respectively, to obtain weld seam characteristic data corresponding to the first hfsl weld seam edge profile and the second hfsl weld seam edge profile, respectively.

7. The small wire scan welding seam tracking correction control method with adaptive function according to claim 6, characterized in that, To The method for processing the subsets of data corresponding to the adjacent weld seam edge profiles comprises: The adjacent weld seam edge contour corresponding to the to-be-processed data subset is input into the weld seam change diagnosis model, and the weld seam change diagnosis result of the adjacent weld seam edge contour is obtained. adjacent weld seam edge contour corresponding to the to-be-processed data subset is input into the weld seam change diagnosis model, and the weld seam change diagnosis result of the adjacent weld seam edge contour is obtained. if the weld change diagnosis result is normal, the to-be-processed data subset does not need to be further processed; if the weld change diagnosis result is abnormal, the to-be-processed data subset is added to the second weld feature data.

8. The small line scan weld tracking correction control method with adaptive functionality of claim 1, wherein, The method for constructing the variable-gap weld set and the cross-weld set comprises: input the second weld feature data corresponding to the HFSLth weld edge contour into the weld type diagnosis model respectively to obtain the weld type corresponding to the HFSLth weld edge contour; the weld type comprises a variable-gap weld and a cross-weld; denote the weld edge contour with the variable-gap weld type as a variable-gap weld, and add it to the variable-gap weld set; denote the weld edge contour with the cross-weld type as a cross-weld, and add it to the cross-weld set.

9. A small wire scan welding seam tracking correction control system with adaptive function for implementing the small wire scan welding seam tracking correction control method with adaptive function according to any one of claims 1 to 8, characterized in that It comprises: a first acquisition module for acquiring weld image data and weld depth; an image processing module for processing the weld image data to obtain a weld feature image; a first processing module for obtaining the first weld feature data corresponding to the HFSLth weld edge contour based on the weld feature image and the weld depth; a second processing module for obtaining the second weld feature data corresponding to the HFSLth weld edge contour based on the weld feature image and the first weld feature data corresponding to the HFSLth weld edge contour; The collection building module performs collection building on the second weld feature data corresponding to the HFSL butt joint edge profile, and obtains a variable gap weld collection and a cross weld collection; The intelligent optimization module performs intelligent adaptive adjustment on the welding control process based on the first weld feature data, the variable gap weld collection and the cross weld collection.

Images