Fillet weld extraction method and system based on line laser, welding method and medium

Abstract

The invention provides a fillet weld extraction method based on line laser. The fillet weld extraction method comprises the following steps: step 1, image processing; step 2, establishing an RO I window; step 3, removing other splashes; and step 4, extracting welding seam feature points. According to the method, the characteristic that the pixel sum of a real light bar is maximum within a certain angle range is utilized, and the light bar and strong noise are segmented; and then the inflection point is found through the distance method, the situation that when the quality of the light bar is poor, the inflection point can be stably extracted with high precision is avoided, the mathematical model of the method is small, welding seam feature points can be rapidly extracted, the requirement for real-time welding of a welding robot can be met, and automatic welding is facilitated.

Description

Technical Field

[0001] The present invention relates to the technical field of fillet weld extraction, in particular to a fillet weld extraction method and system based on line laser, a welding method, and a medium. Background Art

[0002] With the development of machine vision technology, the robot welding technology based on line structured light vision has become a popular research field of automated welding. Among them, weld feature extraction is the key to ensuring high-quality welding of welding robots.

[0003] Steger C et al. published an article titled "Unbiased extraction of lines with parabolic and Gaussian profiles" in the journal "Computer Vision and Image Understanding". This literature calculates the light stripe direction using the Hessian matrix, and takes the zero-crossing point of the first-order derivative of the grayscale on the cross-section as the center. The accuracy of this method is very high, but it has poor robustness to light stripes with different brightness and widths. At the same time, if the reflection coefficient of the weld surface is relatively high, the laser stripe will have strong reflection, interfering with the correct extraction of the center line of the light stripe.

[0004] Deng Jingyu et al. published an article titled "Extraction of Weld Feature Lines Based on Dynamic Small Window Hough Transform Method" in the Transactions of the China Welding Institution. The method of this article improves the speed of the robot to extract weld features, but it depends on the similarity of the front and back series of images during the weld tracking process.

[0005] Patent No. CN108898158A "A Weld Identification Method Based on Piecewise Linear Fitting" uses a recursive subdivision method to segment the discrete points. Within each segment of discrete points, the method of random sample consensus is used to find the inliers of the segment of discrete points, perform linear fitting on the inliers, and then use the distance ratio judgment method to remove the lines with similar slopes. Finally, the weld feature points are found by line intersection. However, this invention uses the medium distance ratio judgment method, which often requires parameter adjustment and is not conducive to the requirements of automated production.

[0006] Hong Lei et al. published an article titled "Analysis of Weld Stripe Linear Feature Extraction Based on Slope Analysis Method" in the Transactions of the China Welding Institution, and proposed a method based on the slope analysis method of the laser center point. This article calculates the distribution quantity of the light points on each stripe line through the slope change feature to complete the point set division, and fits the lines according to each point set after division as the obtained weld feature lines. However, its weld extraction is easily affected by the center point of the light stripe, which is not conducive to the robot to extract weld features in real time in the industrial field.

[0007] With the rapid development of new energy vehicles (including plug-in hybrid, pure electric, and fuel cell-driven models), the welding quality of their core components (such as battery casings, motor brackets, and body structural parts) is crucial for vehicle safety and performance. In the manufacturing of new energy vehicles, lightweight materials such as aluminum alloys and carbon fibers, as well as high-precision laser welding and arc welding processes, are widely used, posing higher requirements for weld seam recognition technology.

[0008] Weld seam tracking is an important means to improve welding accuracy and achieve automated welding. Traditional weld seam tracking algorithms often struggle to solve the problem of interference from strong welding noise. Therefore, an image processing method resistant to strong noise interference is proposed, which can effectively segment out the complete light stripe. After segmenting out the complete light stripe, the next step is to extract the weld seam feature points. The main idea of existing technologies for extracting weld seam points is to find the point with the largest slope change or inflection point among the center points of the line laser light stripe after extracting the center points of the line laser light stripe, and then consider this point as the weld seam point. These methods require high requirements for the continuity and smoothness of discrete points. When the continuity and smoothness of the center points are relatively poor, incorrect extraction may occur. Summary of the Invention

[0009] The present invention provides a method and system for extracting fillet welds based on line laser, a welding method, and a medium. By utilizing the characteristic that the pixel sum of the real light stripe is the largest within a certain angle range, the light stripe and strong noise are separated; then the distance method is used to find the inflection point, avoiding the situation where when the quality of the light stripe is relatively poor, the inflection point can still be stably and accurately extracted. Moreover, the mathematical model of this method is relatively small, enabling rapid extraction of weld seam feature points, meeting the real-time welding requirements of welding robots, and being beneficial to automated welding.

[0010] To achieve the above object, the present invention adopts the following technical solutions:

[0011] This specification discloses a method for extracting fillet welds based on line laser, including:

[0012] Step 1: Image processing;

[0013] First, convert the complete light stripe image into a single-channel image, and then traverse the gray values of the entire single-channel image column by column from right to left. When there is a gray value greater than the threshold d, record this point as point A, and then stop the search to obtain the pixel-level coordinates (x1, y1) of point A;

[0014] Step 2: Establish an ROI window;

[0015] Take point A as the demarcation point to establish ROI_1 window and ROI_2 window, obtaining an image with some noise removed;

[0016] Step 3: Remove the remaining spatter;

[0017] Based on the image with some noise removed, using the principle that the cumulative sum of the gray values on the line laser light stripe is the largest, a coordinate system o-xy is established with the upper left corner of the image as the origin, and a coordinate system o'-x'y' is established with A(x1, y1) as the origin. After performing rotation and translation operations, an image with splash noise removed is obtained;

[0018] Step 4: Extract the weld feature points;

[0019] After subjecting the image with splash noise removed to threshold processing and image enhancement, the coordinates of the center point of the line laser after denoising are extracted using a traditional light stripe extraction algorithm. Based on the coordinates of the center point of the line laser after denoising, the inflection points are obtained using the distance method, which are the weld feature points.

[0020] In this specification, in Step 3, after establishing the coordinate system, assume that the gray value of the image at (x i , y j ) is f(x i , y j ), and set the rotation center as A(x1, y1):

[0021]

[0022] where (x i ', y' j ) is the new coordinate with (x1, y1) as the origin;

[0023] Establish the image pixel accumulation formula:

[0024]

[0025] where ρ k is the distance from the straight line to the origin: ρ k = x i cosθ l + y j sinθ l , θ l is the angle between the image and the x-axis, δ is the Dirac delta function, and R(ρ k , θ l ) is the cumulative sum of the gray values on the straight line when the angle is θ l , and the image size is n×m;

[0026] Combined with formula (1), formula (2) becomes:

[0027]

[0028] where f(x i '+ x1, y' j + y1) is the translated image function, and θ l' is the angle between the translated formula and the x'-axis, ρ' k is the distance from the straight line to the new origin (x1, y1), that is: ρ' k = x i 'cosθ l '+ y' j sinθ l ';

[0029] Filter the value of R'(ρ' k , θ l '), that is:

[0030]

[0031] Inverse solve Equation (3) according to Equation (5) to find the f(x k , θ l ') corresponding to the value of f(x i '+ x1, y' j + y1), and obtain the image with splash noise removed;

[0032]

[0033] Among them, Δθ is the discrete step of θ l ', Δρ is the discrete step of ρ' k , L is the discretization number of the angle θ l '(l = 1, 2,..., L), and K is the discretization number of the angle ρ' k (l = 1, 2,..., K).

[0034] In this specification, in step 4, based on the extracted coordinates of the center point of the denoised line laser, the center point set is denoted as C, and its number is denoted as m;

[0035] Select the coordinates of the first point and the d-th point in the center point set C, 1 < d < 5, and obtain the straight line equation L1 of these two point coordinates. Then calculate the distances from all coordinates between the coordinates of the first point and the d-th point to the straight line L1. When the distance is less than the threshold D m , calculate the straight line equation L id of the two coordinate points of the coordinates of the first point and the i*d-th point, i = 2, 3,..., m; then calculate the distances from all coordinates between the coordinates of the first point and the i*d-th point to the straight line L id . Until the distances from the w-th, (w + 1)-th, and (w + 2)-th coordinate points between the coordinates of the first point and the i*d-th point to the straight line L id are all greater than the threshold D m , then the w-th coordinate is defaulted as the inflection point coordinate, and the w-th coordinate point is P w (xw , y w ), with P w (x w , y w ) as the demarcation point, divide the center point set C into two sets C1 and C2;

[0036]

[0037] For sets C1 and C2 respectively, use the line fitting algorithm to find the corresponding line equations, and then find their intersection point, which is the weld feature point.

[0038] In this specification, the threshold D m = 10 mm.

[0039] This specification also discloses a fillet weld extraction system based on line laser, which is used to implement the fillet weld extraction method based on line laser described in any one of the above, and the fillet weld extraction system based on line laser includes:

[0040] An image processing module, used for:

[0041] First, convert the complete light strip image into a single-channel image, and then traverse the gray values of the entire single-channel image column by column from right to left. When there is a gray value greater than the threshold d, record this point as point A, and then stop the search to obtain the pixel-level coordinates (x1, y1) of point A;

[0042] An ROI window establishment module, used for:

[0043] Establish ROI_1 window and ROI_2 window with point A as the demarcation point to obtain an image with some noise removed;

[0044] A module for removing the remaining spatter, used for:

[0045] Based on the image with some noise removed, using the principle that the sum of gray values on the line laser light strip is the largest, establish a coordinate system o-xy with the upper left corner of the image as the origin and a coordinate system o'-x'y' with A(x1, y1) as the origin, and after performing rotation and translation operations, obtain an image with spatter noise removed;

[0046] A module for extracting weld feature points, used for:

[0047] After threshold processing and image enhancement of the image with spatter noise removed, use the traditional light strip extraction algorithm to extract the coordinates of the center points of the line laser after denoising. Based on the coordinates of the center points of the line laser after denoising, use the distance method to obtain the inflection points, which are the weld feature points.

[0048] This specification also discloses a welding method, including:

[0049] Collect weld images;

[0050] Based on the weld image, the complete light stripe is segmented to obtain a complete light stripe map;

[0051] Based on the complete light stripe map, the weld feature points are obtained by using the method for extracting fillet welds based on line laser described in any one of the above;

[0052] Welding is performed based on the weld feature points.

[0053] This specification also discloses a computer-readable storage medium storing computer instructions, and when a computer reads the computer instructions, the computer executes the method for extracting fillet welds based on line laser described in any one of the above.

[0054] In summary, the present invention has at least the following beneficial effects:

[0055] The present invention utilizes the characteristic that the pixel sum of the real light stripe is the largest within a certain angle range to separate the light stripe from strong noise; then the distance method is used to find the inflection point, avoiding the situation that when the quality of the light stripe is relatively poor, the inflection point can still be stably and accurately extracted, and the mathematical model of this method is small, enabling rapid extraction of weld feature points, which can meet the real-time welding of welding robots and is beneficial to automated welding. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0057] Figure 1 It is a schematic diagram of the method for extracting fillet welds based on line laser involved in the present invention.

[0058] Figure 2 It is a schematic diagram of the complete light stripe map involved in the present invention.

[0059] Figure 3 It is a schematic diagram of the image during the welding process involved in the present invention.

[0060] Figure 4 It is a schematic diagram of the single-channel image involved in the present invention.

[0061] Figure 5 It is a schematic diagram of the image with some noise removed involved in the present invention.

[0062] Figure 6 It is a schematic diagram of the image with splash noise removed involved in the present invention.

[0063] Figure 7 Schematic diagram of the image after threshold processing and image enhancement involved in the present invention.

[0064] Figure 8 Schematic diagram of the weld feature point P involved in the present invention. Detailed implementation manners

[0065] In the following, only some exemplary embodiments are briefly described. As those skilled in the art can recognize, the described embodiments can be modified in various different ways without departing from the spirit or scope of the embodiments of the present invention. Therefore, the drawings and the description are considered to be exemplary in nature and not restrictive.

[0066] The following disclosure provides many different implementation manners or examples for implementing different structures of the embodiments of the present invention. To simplify the disclosure of the embodiments of the present invention, the components and settings of specific examples are described below. Of course, they are only examples and are not intended to limit the embodiments of the present invention. In addition, the embodiments of the present invention may repeat reference numerals and / or reference letters in different examples. This repetition is for the purpose of simplification and clarity, and does not itself indicate the relationship between the various implementation manners and / or settings discussed.

[0067] The embodiments of the present invention will be described in detail below with reference to the drawings.

[0068] As Figure 1 shown, this embodiment provides a method for extracting fillet welds based on line lasers, including:

[0069] Step 1: Image processing;

[0070] First, convert the complete light stripe image into a single-channel image, and then traverse the gray values of the entire single-channel image column by column from right to left. When there is a gray value greater than the threshold d, record this point as point A, and then stop the search to obtain the pixel-level coordinates (x1, y1) of point A;

[0071] Step 2: Establish an ROI window;

[0072] Take point A as the demarcation point to establish ROI_1 window and ROI_2 window, and obtain an image with some noise removed;

[0073] Step 3: Remove the remaining spatter;

[0074] Based on the image with some noise removed, using the principle that the sum of the gray values on the line laser light stripe is the largest, establish a coordinate system o-xy with the upper left corner of the image as the origin and a coordinate system o'-x'y' with A ( (x1, y1) as the origin, and after performing rotation and translation operations, obtain an image with spatter noise removed;

[0075] Step 4: Extract the weld feature points;

[0076] After the image with splash noise removed is subjected to threshold processing and image enhancement, the center coordinates of the line laser after denoising are extracted using a traditional light stripe extraction algorithm. Based on the center coordinates of the line laser after denoising, the inflection points are obtained using the distance method, which are the weld feature points.

[0077] In some embodiments, in Step 3, after establishing the coordinate system, let the gray value of the image at (x i , y j ) be f(x i , y j ), and set the rotation center as A(x1, y1):

[0078]

[0079] where (x i , y' j ) is the new coordinate with (x1, y1) as the origin;

[0080] Establish the image pixel accumulation formula:

[0081]

[0082] where ρ k is the distance from the line to the origin: ρ k = x i cosθ l + y j sinθ l , θ l is the angle between the image and the x-axis, δ is the Dirac delta function, and R(ρ k , θ l ) is the sum of the gray values on the line when the angle is θ l , and the image size is n×m;

[0083] Combined with formula (1), formula (2) becomes:

[0084]

[0085] where f(x i '+x1, y' j +y1) is the translated image function, θ l ' is the angle between the translated formula and the x'-axis, ρ' k is the distance from the line to the new origin (x1, y1), that is: ρ' k = x i 'cosθ l '+ y' j sinθ l ';

[0086] Filter the value of R'(ρ' k , θ l '), which is:

[0087]

[0088] Solve equation (3) inversely according to equation (5) to obtain the value of f(x k , θ l ') corresponding to R'(ρ' i '+ x1, y' j + y1) in formula (4), and obtain the image with splash noise removed;

[0089]

[0090] Among them, Δθ is the discrete step of θ l ', Δρ is the discrete step of ρ' k , L is the discretization number of the angle θ l '(l = 1, 2,..., L), and K is the discretization number of the angle ρ' k (l = 1, 2,..., K).

[0091] In some embodiments, in step 4, based on the extracted coordinates of the center point of the denoised line laser, the set of center points is denoted as C, and the number of them is denoted as m;

[0092] Select the first point coordinate and the d-th point coordinate in the set of center points C, 1 < d < 5 (when d is 2 / 3 / 4, the following calculations can be performed without following the order until the distance is less than the threshold; d can also be defaulted to 4). Obtain the straight line equation of these two point coordinates as L1, and then calculate the distances from all coordinates between the first point coordinate and the d-th point coordinate to the straight line L1. When the distance is less than the threshold D m , calculate the straight line equation L id of the two coordinate points of the first point coordinate and the i*d-th point coordinate, i = 2, 3,..., m; then calculate the distances from all coordinates between the first point coordinate and the i*d-th point coordinate to the straight line L id . Until the distances from the w-th, (w + 1)-th, and (w + 2)-th coordinate points between the first point coordinate and the i*d-th point coordinate to the straight line L id are all greater than the threshold D m (that is, when the distances from three consecutive coordinate points to the straight line are all greater than the threshold), then default the w-th coordinate as the inflection point coordinate, and the w-th coordinate point is P w (x w , y w ), with P w (x w , yw ) Take the central point set C as a demarcation point and divide it into two sets C1 and C2;

[0093]

[0094] For the sets C1 and C2 respectively, use the line fitting algorithm to find the corresponding line equations, and then find their intersection point, which is the weld feature point.

[0095] In some embodiments, the threshold D m = 10mm; it can also be set according to the actual situation.

[0096] This embodiment provides a fillet weld extraction system based on line laser for implementing the fillet weld extraction method based on line laser described in any one of the above, and the fillet weld extraction system based on line laser includes:

[0097] An image processing module for:

[0098] First, convert the complete light stripe image into a single-channel image, and then traverse the gray values of the entire single-channel image column by column from right to left. When there is a gray value greater than the threshold d, record this point as point A, and then stop the search to obtain the pixel-level coordinates (x1, y1) of point A;

[0099] An ROI window establishment module for:

[0100] Take point A as a demarcation point to establish ROI_1 window and ROI_2 window, and obtain an image with some noise removed;

[0101] A module for removing the remaining spatter for:

[0102] Based on the image with some noise removed, using the principle that the sum of gray values on the line laser light stripe is the largest, establish a coordinate system o-xy with the upper left corner of the image as the origin and a coordinate system o'-x'y' with A(x1, y1) as the origin, and after rotation and translation operations, obtain an image with spatter noise removed;

[0103] A module for extracting weld feature points for:

[0104] After threshold processing and image enhancement of the image with spatter noise removed, use the traditional light stripe extraction algorithm to extract the coordinates of the center points of the line laser after denoising. Based on the coordinates of the center points of the line laser after denoising, use the distance method to obtain the inflection points, which are the weld feature points.

[0105] This embodiment provides a welding method, including:

[0106] Collect weld images;

[0107] Based on the weld images, segment the complete light stripe to obtain a complete light stripe image;

[0108] Based on the complete light stripe image, using the line laser-based fillet weld extraction method described in any one of the above, weld feature points are obtained;

[0109] Welding is performed based on the weld feature points.

[0110] This embodiment provides a computer-readable storage medium, and the storage medium stores computer instructions. When a computer reads the computer instructions, the computer executes the line laser-based fillet weld extraction method described in any one of the above.

[0111] The technical idea of the present invention is as follows:

[0112] As Figure 2 is a complete light stripe image. Figure 3 is an image during the welding process, and the image has a large amount of spatter noise. How to remove the noise and extract the weld feature points is the task of the present invention.

[0113] Step 1: Image processing. First, convert Figure 2 to a single-channel image, and then traverse the gray values of the entire image column by column from right to left. When there is a gray value greater than the threshold d, record the point and then stop the search. Finally, the pixel-level coordinates of point A(x1, y1) in Figure 4 can be obtained.

[0114] Step 2: Establish an ROI window. Through experiments, it can be known that the position of point A remains basically unchanged during the welding process. Then, in Figure 3 , establish ROI_1 and ROI_2 windows with A as the demarcation point, and thus Figure 5 can be obtained. It can be seen from Figure 5 that a large amount of noise has been removed.

[0115] Step 3: Remove the remaining spatter. From the ROI_1 and ROI_2 in Figure 5 , it can be observed that the spatter is basically a linear light stripe. Among all the linear light stripes, the sum of the gray values on the line laser light stripe is the largest. Establish a coordinate system o-xy with the upper left corner of the image as the origin and a coordinate system o'-x'y' with A(x1, y1) as the origin.

[0116] Let the gray value of the image at (x i , y j ) be f(x i , y j ), and set the rotation center as A(x1, y1):

[0117]

[0118] where (x i ', y' j) is the new coordinate with the origin at (x1, y1).

[0119] That is, the image pixel accumulation formula:

[0120]

[0121] Where ρ k is the distance from the straight line to the origin: ρ k = x i cosθ l + y j sinθ l , θ l is the angle between the image and the x-axis, δ is the Dirac delta function, R(ρ k , θ l ) is the accumulated sum of the gray values on the straight line when the angle is θ l , and the image size is n × m.

[0122] Combined with formula (1), formula (2) can be changed to:

[0123]

[0124] Where f(x i '+ x1, y' j + y1) is the translated image function, θ l ' is the angle between the translated formula and the x'-axis, ρ' k is the distance from the straight line to the new origin (x1, y1), that is: ρ' k = x i 'cosθ l '+ y' j sinθ l '.

[0125] Then screen the value of R'(ρ' k , θ l '), which is:

[0126]

[0127] Inverse solve formula (3) to find the value of f(x k , θ l ') corresponding to R'(ρ' i '+ x1, y' j + y1) in formula (4).

[0128]

[0129] Where Δθ is the discrete step of θ l ', Δρ is the discrete step of ρ' k ', and L is the angle θl The discretization quantity of '(l = 1, 2,..., L), and K is the angle ρ' k (l = 1, 2,..., K). Finally, the image with splash noise removed can be obtained, as shown in Figure 6 .

[0130] Step 4: Extract the weld feature points. After Figure 6 threshold processing and image enhancement, we get Figure 7 . Using the traditional light stripe extraction algorithm, the coordinates of the center points of the line laser after denoising can be extracted (visualized in red). The set of center points is denoted as C, and the number of them is denoted as m.

[0131] Select the coordinates of the first point and the d-th point (d < 5) in the set C of center points. The straight-line equation of these two points is L1. Then calculate the distances from all the coordinates between the first point and the d-th point to the straight line L1. When the distance is less than the threshold D m , calculate the straight-line equation L id of these two coordinate points of the first point and the (i * d)-th point (i = 2, 3,..., m). Then calculate the distances from all the coordinates between the first point and the (i * d)-th point (i = 2, 3,..., m) to the straight line L id . Until the distances from the w-th, (w + 1)-th, and (w + 2)-th coordinate points between the first point and the (i * d)-th point (i = 2, 3,..., m) to the straight line L id are all greater than the threshold D m , then the w-th coordinate is defaulted to be the inflection point coordinate, and the coordinate point P w (x w , y w ) corresponding to the w-th coordinate. Then, with P w (x w , y w ) as the demarcation point, the coordinates of the center points of the line laser are divided into two sets C1 and C2.

[0132]

[0133] For the sets C1 and C2 respectively, use the straight-line fitting algorithm to find the corresponding straight-line equations, and then find their intersection point, which is the weld feature point, denoted as P, as shown in Figure 8 .

[0134] The above-described embodiments are used to illustrate the present invention, not to limit the present invention. Therefore, changes in the example numerical values or replacement of equivalent elements still belong to the scope of the present invention.

[0135] From the above detailed description, those of ordinary skill in the art can clearly understand that the present invention can indeed achieve the foregoing objectives and has actually met the requirements of the patent law.

[0136] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they know the basic creative concept. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications falling within the scope of the present invention. The above description is only the preferred embodiments of the present invention and is not intended to limit the present invention. It should be noted that any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

[0137] It should be noted that the above description of the process is only for illustration and explanation and does not limit the scope of application of this specification. Those skilled in the art can make various corrections and changes to the process under the guidance of this specification. However, these corrections and changes are still within the scope of this specification.

[0138] The basic concept has been described above. Obviously, for those of ordinary skill in the art who read this application, the above invention disclosure is only for illustration and does not constitute a limitation to this application. Although not explicitly stated here, those of ordinary skill in the art may make various modifications, improvements, and corrections to this application. Such modifications, improvements, and corrections are suggested in this application, so such modifications, improvements, and corrections still belong to the spirit and scope of the exemplary embodiments of this application.

[0139] Meanwhile, this application uses specific terms to describe the embodiments of this application. For example, "one embodiment", "an embodiment", and / or "some embodiments" mean a certain feature, structure, or characteristic related to at least one embodiment of this application. Therefore, it should be emphasized and noted that the "an embodiment" or "one embodiment" or "an alternative embodiment" mentioned two or more times in different positions in this specification does not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of this application can be appropriately combined.

[0140] In addition, those of ordinary skill in the art can understand that various aspects of this application can be illustrated and described by several patentable types or situations, including any new and useful process, machine, product, or combination of substances, or any new and useful improvement thereof. Therefore, various aspects of this application can be implemented entirely by hardware, can be implemented entirely by software (including firmware, resident software, microcode, etc.), or can be implemented by a combination of hardware and software. The above hardware or software can all be referred to as "units", "modules", or "systems". In addition, various aspects of this application can take the form of a computer program product embodied in one or more computer-readable media, in which computer-readable program code is included.

[0141] The computer program code required for the operations of each part of this application can be written in any one or more programming languages, including object-oriented programming languages such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python, etc., conventional procedural programming languages such as C programming language, VisualBasic, Fortran2103, Perl, COBOL2102, PHP, ABAP, dynamic programming languages such as Python, Ruby and Groovy, or other programming languages, etc. This program code can run entirely on the user's computer, or run on the user's computer as an independent software package, or run partially on the user's computer and partially on a remote computer, or run entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer through any network form, such as a local area network (LAN) or a wide area network (WAN), or connected to an external computer (for example, through the Internet), or in a cloud computing environment, or used as a service such as software as a service (SaaS).

[0142] In addition, unless clearly stated in the claims, the order of the processing elements and sequences, the use of numbers and letters, or the use of other names in this application are not used to limit the order of the processes and methods of this application. Although some currently considered useful embodiments of the invention are discussed through various examples in the above disclosure, it should be understood that such details only serve the purpose of illustration, and the appended claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that conform to the essence and scope of the embodiments of this application. For example, although the implementation of the above various components can be embodied in a hardware device, it can also be implemented as a pure software solution. For example, it can be installed on an existing server or mobile device.

[0143] Similarly, it should be noted that, in order to simplify the description of this application disclosure and thus help the understanding of one or more embodiments of the invention, in the previous description of the embodiments of this application, sometimes multiple features are merged into one embodiment, drawing or description thereof. However, this method of this application should not be construed as reflecting the intention that the claimed subject matter requires more features than those clearly recited in each claim. On the contrary, the subject matter of the invention should have fewer features than the above single embodiment.

Claims

1. A method for extracting fillet welds based on line laser, characterized in that, Including: Step 1: Image processing; First, convert the complete light stripe image into a single-channel image, and then traverse the gray values of the entire single-channel image column by column from right to left. When there is a gray value greater than the threshold d, record this point as point A, and then stop the search to obtain the pixel-level coordinates (x1, y1) of point A; Step 2: Establish an ROI window; Take point A as the demarcation point to establish ROI_1 window and ROI_2 window, and obtain an image with some noise removed; Step 3: Remove the remaining splashes; Based on the image with some noise removed, using the principle that the sum of gray values on the line laser light strip is the largest, a coordinate system o-xy is established with the upper left corner of the image as the origin, and a coordinate system o'-x'y' is established with as the origin. After performing rotation and translation operations, an image with splash noise removed is obtained; Step 4: Extract weld feature points; After the image with splashes and noise removed undergoes threshold processing and image enhancement, use the traditional light stripe extraction algorithm to extract the coordinates of the center point of the line laser after denoising. Based on the coordinates of the center point of the line laser after denoising, use the distance method to obtain the inflection point, which is the weld feature point.

2. The method for extracting fillet welds based on line laser according to claim 1, wherein In step 3, after establishing the coordinate system, assume that the gray value of the image at (x i , y j ) is f(x i , y j ), and set the rotation center as A(x1, y1): where (x′ i , y' j ) is the new coordinate with (x1, y1) as the origin; Establish an image pixel accumulation formula: where ρ k is the distance from the straight line to the origin: ρ k = x i cosθ l + y j sinθ l , θ l is the angle between the image and the x-axis, δ is the Dirac delta function, R(ρ k , θ l ) is the cumulative sum of the gray values on the straight line at the angle θ l , and the image size is n×m; Combined with formula (1), formula (2) becomes: where f(x′ i + x1, y' j + y1) is the translated image function, and θ′ l is the angle between the translated formula and the x'-axis. ρ' k is the distance from the straight line to the new origin (x1, y1), that is: ρ' k = x′ i cosθ′ l + y' j sinθ′ l ; For the value of R'(ρ' k , θ′ l ), the screening is as follows: Inverse solution of formula (3) according to formula (5) to obtain the value of f(x′ k , θ′ l ) corresponding to (ρ′ i + x1, y′ j + y1), and obtain the image with splash noise removed; where Δθ is the discrete step size of θ′ l and Δρ is the discrete step size of ρ' k ; L is the discretization number of the angle θ′ l (l = 1, 2,..., L), and K is the discretization number of the angle ρ' k (l = 1, 2,..., K).

3. The method for extracting fillet welds based on line laser according to claim 1, characterized in that, In step 4, based on the coordinates of the center point of the line laser after extraction and denoising, the set of center points is denoted as C, and its number is denoted as m; Select the coordinates of the 1st point and the dth point in the set C of center points, where 1 < d < 5. Obtain the straight-line equation L1 of these two points' coordinates. Then calculate the distances from all coordinates between the 1st point's coordinates and the dth point's coordinates to the straight line L1. When the distance is less than the threshold D m calculate the straight-line equation L of the two coordinate points of the 1st point's coordinates and the (i * d)th point's coordinates id , where i = 2, 3,..., m; then calculate the distances from all coordinates between the 1st point's coordinates and the (i * d)th point's coordinates to the straight line L id until the distances from the wth, (w + 1)th, and (w + 2)th coordinate points between the 1st point's coordinates and the (i * d)th point's coordinates to the straight line L id are all greater than the threshold D m At this time, default the wth coordinate as the inflection point coordinate, and the wth coordinate point as P w (x w , y w ). Divide the set C of center points into two sets C1 and C2 with P w (x w , y w ) as the demarcation point Use the linear fitting algorithm for sets C1 and C2 respectively, find the corresponding linear equations, and then find their intersection point, which is the weld feature point.

4. The method for extracting fillet welds based on line laser according to claim 3, characterized in that Threshold D m = 10 mm.

5. The angular weld extraction system based on line laser is characterized in that, For implementing the method for extracting fillet welds based on line laser described in any one of claims 1 to 4, the system for extracting fillet welds based on line laser includes: An image processing module, used for: First, convert the complete light stripe image into a single-channel image, and then traverse the gray values of the entire single-channel image column by column from right to left. When there is a gray value greater than the threshold d, record this point as point A, and then stop the search to obtain the pixel-level coordinates (x1, y1) of point A; An ROI window establishment module, used for: Take point A as the demarcation point to establish ROI_1 window and ROI_2 window, and obtain an image with some noise removed; A module for removing the remaining splashes, used for: Based on the image with some noise removed, using the principle that the sum of gray values on the line laser light stripe is the largest, establish a coordinate system o-xy with the upper left corner of the image as the origin and a coordinate system o'-x'y' with A(x1, y1) as the origin, and after rotation and translation operations, obtain an image with splashes and noise removed; A module for extracting weld feature points, used for: After the image with splashes and noise removed undergoes threshold processing and image enhancement, use the traditional light stripe extraction algorithm to extract the coordinates of the center point of the line laser after denoising. Based on the coordinates of the center point of the line laser after denoising, use the distance method to obtain the inflection point, which is the weld feature point.

6. A welding method, characterized in that, Including: Collect a weld image; Based on the weld image, segment out the complete light stripe to obtain a complete light stripe image; Based on the complete light stripe image, use the method for extracting fillet welds based on line laser described in any one of claims 1 to 4 to obtain weld feature points; Perform welding based on the weld feature points.

7. A computer-readable storage medium, characterized in that, The storage medium stores computer instructions, and when the computer reads the computer instructions, the computer executes the method for extracting fillet welds based on line laser described in any one of claims 1 to 4.

Images