Multi-point reinforcing mesh welding device based on numerical control positioning

Abstract

The invention relates to the technical field of reinforcing steel bar machining, in particular to a multi-point reinforcing steel bar mesh welding device based on numerical control positioning, which comprises fixed seats, a welding seat arranged between the two fixed seats, and a welding table movably arranged below the welding seat, the welding gun is movably arranged on the welding table; the controller is arranged on the side wall of the fixing base and used for adjusting the position of the welding gun. According to the multi-point reinforcing mesh welding device based on numerical control positioning, by accurately obtaining the three-dimensional coordinates of the welding points, the precision and stability of welding operation can be greatly improved, the welding quality problem caused by position deviation is avoided, and the welding accuracy is improved based on welding movement data predicted by a machine learning model. Automatic control of the welding device can be achieved, manual intervention is reduced, the production efficiency is improved, the model is trained through historical welding data, the welding requirements of different types of reinforcing meshes can be met, and good adaptability is achieved in various production environments.

Description

Technical Field

[0001] The invention relates to the technical field of steel bar processing, in particular to a multi-point steel bar mesh welding device based on numerical control positioning. Background Art

[0002] Steel mesh is a reinforcing material commonly used in concrete structures. It is formed by arranging multiple longitudinal steel bars and multiple transverse steel bars at a certain interval and welding all the intersections together to form a mesh structure. This structural form is widely used in various engineering fields such as bridges, roads, tunnels, etc. It has significant advantages such as improving the strength of concrete structures, increasing structural rigidity, and improving seismic performance. When welding steel mesh, the existing steel mesh welding device usually needs to arrange the straightened and sheared longitudinal steel bars and transverse steel bars in a criss-cross pattern, and then move the welding device to weld and fix each intersection of the longitudinal steel bars and the transverse steel bars in turn. The welding process is complicated, the welding efficiency is low, and the device has low integration, occupies a large area, and has high cost.

[0003] For example, a Chinese patent with authorization announcement number CN107363429B discloses a continuous automatic welding device for isolation pier steel mesh, which mainly includes a semi-circular steel bar pusher, a guide rail, a welding module, a longitudinal steel bar support plate, a longitudinal steel bar support plate, a longitudinal steel bar clip support frame, a longitudinal steel bar clip and a control system; a semi-circular steel bar pusher is set at the rear, a longitudinal steel bar clip is set at the upper front, and a longitudinal steel bar support plate and a longitudinal steel bar support plate are respectively set below the longitudinal steel bar clip; an opening is set at the lower part of the longitudinal steel bar clip, and the longitudinal steel bar support plate can move upward to take out the longitudinal steel bars in the longitudinal steel bar clip; fixing positions for the longitudinal steel bars are set on both sides of the support plate; a longitudinal guide rail is installed between the semi-circular steel bar pusher and the longitudinal steel bar clip support frame, and a welding module is installed on the guide rail.

[0004] As mentioned above, in the conventional steel mesh welding process, the positioning of welding points usually relies on manual operation or simple mechanical devices, which makes it difficult to ensure the precise positioning of each welding point. The deviation of the welding area may lead to unstable welding quality, which in turn affects the overall strength and stability of the steel mesh. Especially in projects with high precision requirements, such errors can lead to serious structural problems. In addition, conventional welding equipment requires a lot of manual debugging and intervention, and operators need to adjust the welding path based on experience and intuition. This not only increases the complexity of the operation, but also easily leads to inconsistent welding accuracy due to human factors, which in turn affects product quality. Summary of the invention

[0005] In order to solve the above problems, the present invention provides a multi-point steel mesh welding device based on numerical control positioning.

[0006] The present invention adopts the following technical scheme, a multi-point steel mesh welding device based on numerical control positioning, the welding device is arranged on a steel mesh conveying frame, and is used to weld the steel mesh, the welding device comprises:

[0007] A fixing seat, of which two are provided;

[0008] A welding seat is arranged between the two fixing seats, and the welding seat can be lifted up and down between the two fixing seats;

[0009] A welding table, which is movably arranged below the welding seat;

[0010] A welding gun, which is movably arranged on the welding table, wherein a plurality of welding guns are arranged and the plurality of welding guns are arranged on the same horizontal line;

[0011] A controller, which is arranged on the side wall of the fixing seat and is used to adjust the position of the welding gun, and the controller includes:

[0012] The data acquisition module collects the image data of the steel mesh through a CCD camera installed on the lower surface of the welding table, and pre-processes the collected image data;

[0013] The data processing module extracts the pre-processed image data, analyzes and processes the image data, and obtains the three-dimensional coordinates of the welding point.

[0014] The welding control module obtains real-time welding characteristic parameters, inputs them into a pre-built machine learning model that predicts welding movement data, predicts the welding movement data, and controls the operation of the welding device based on the welding movement data. The welding movement data is the movement data of the welding table in the horizontal, longitudinal and vertical directions. The welding characteristic parameters include the diameters of the horizontal and longitudinal steel bars of the steel mesh and the three-dimensional coordinates of the welding points.

[0015] As a further description of the above technical solution: the method of analyzing and processing the image data to obtain the three-dimensional coordinates of the welding point includes:

[0016] Step 1, using image processing algorithms to extract straight lines from image data and locate the position of the steel bars;

[0017] Step 2, locate the intersection of the steel mesh by calculating the intersection of the straight lines, and mark the intersection as the welding point;

[0018] Step 3: perform coordinate transformation on the welding point coordinates to generate the three-dimensional coordinates of the welding point.

[0019] As a further description of the above technical solution: Said Methods for extracting straight lines from image data using image processing algorithms include:

[0020] By performing edge detection on the image data, all edge points (a, b) are extracted;

[0021] Convert each edge point (a, b) in the image data into coordinates (ρ, θ) in the polar coordinate system to represent a straight line, where ρ represents the shortest distance from the straight line to the coordinate (0.0), and θ represents the angle of the straight line, which is between 0° and 90°;

[0022] For each edge point (a, b) in the image, the coordinates in the polar coordinate system are calculated using the formula;

[0023] The calculation formula is: ρ = acosθ + bsinθ;

[0024] For each edge point (a, b), repeat the above calculation, traverse different θ values, and obtain a series of (ρ, θ) parameter pairs;

[0025] Create a two-dimensional accumulator, where each pair of ρ and θ corresponds to a spatial position, traverse all edge points in the image, map their corresponding ρ and θ values ​​to the accumulator, and accumulate them;

[0026] Set the accumulation threshold, remove the sum combination whose median value of the accumulator is lower than the accumulation threshold, because they correspond to noise or unclear straight lines, and obtain the ρ and θ combination whose median value of the accumulator is higher than the threshold, ρ and θ correspond to the straight lines in the steel mesh in the image;

[0027] According to the obtained combination of ρ and θ, draw the corresponding straight line on the image;

[0028] The ρ and θ parameters of each line are converted back into image data to plot the actual reinforcement positions;

[0029] To locate the welding point, after extracting the straight lines of the steel bars, calculate the coordinates of the intersection of each intersecting line. The coordinates of the intersection are the coordinates of the welding point.

[0030] As a further description of the above technical solution: the method of analyzing and processing the image data to obtain the three-dimensional coordinates of the welding point includes:

[0031] Perform grayscale processing on the collected image data, obtain the grayscale value of each pixel block in the image data, and mark it as a real-time grayscale value, recorded as HDs, compare and analyze the real-time grayscale value of each block in the image data with the preset welding position grayscale threshold to determine whether to mark the pixel block as a welding area;

[0032] Pixel blocks with welding area marks are extracted and mapped to a pre-constructed two-dimensional coordinate system. Adjacent pixel blocks are marked as the same welding area to obtain multiple welding areas. The centroid of each welding area is calculated as the coordinate, recorded as the welding point coordinate. The depth value Z of the centroid is obtained through the depth camera to calculate the three-dimensional coordinates of the welding point.

[0033] As a further description of the above technical solution: the method for determining whether to mark the welding area of ​​the pixel block includes:

[0034] Preset grayscale threshold HD 1 , HD 2 , and HD 1 <HD 2 ; Grayscale threshold HD 1 , HD 2 It is determined by those skilled in the art based on a large number of experiments.

[0035] When HDs<HD 1 , the pixel block is not marked as a welding area;

[0036] When HD 1 ≤HDs≤HD 2 , mark the pixel block as a welding area;

[0037] When HDs>HD 2 , the pixel block is not marked as a welding area.

[0038] As a further description of the above technical solution: the method of calculating the centroid of each welding area as the coordinate includes:

[0039] The area of ​​the welding area is calculated by the formula,

[0040] Calculate the weighted average of the horizontal and vertical coordinates of the pixels in the welding area;

[0041] The centroid coordinates are obtained using the centroid calculation formula.

[0042] As a further description of the above technical solution:

[0043] The method for calculating the three-dimensional coordinates of the welding point based on the centroid coordinates is:

[0044] The three-dimensional coordinates of the welding point are calculated by combining the focal length and principal point coordinates of the CCD camera with the centroid coordinates and depth information.

[0045] As a further description of the above technical solution: the training method of the machine learning model for predicting welding movement data includes:

[0046] The historical welding training parameters of the welding device are collected when the welding device is in a state of meeting welding standards. The historical welding training parameters include welding characteristic parameters and welding movement data.

[0047] The collected historical welding training parameters are converted into a corresponding set of feature vectors;

[0048] Each group of feature vectors is used as the input of the machine learning model, and the machine learning model takes the welding movement data corresponding to each group of welding feature parameters as the output, and the welding movement data actually corresponding to each group of welding feature parameters is used as the prediction target, and minimizing the loss function value of the machine learning model is used as the training target; training is stopped when the loss function value of the machine learning model is less than or equal to the preset target loss value.

[0049] As a further description of the above technical solution: in the step 1, after extracting the straight line from the image data using the image processing algorithm, it also includes:

[0050] By sequentially obtaining the spacing between adjacent horizontal lines and the spacing between vertical lines, a set is established to determine the density of the steel mesh, and the lines corresponding to the abnormal spacing values ​​are obtained and removed.

[0051] As a further description of the above technical solution: the method of performing coordinate conversion on the welding point coordinates to generate the three-dimensional coordinates of the welding point includes:

[0052] Perform camera calibration to obtain the camera's internal and external parameters, including focal length, principal point coordinates, and distortion parameters;

[0053] According to the calibrated internal and external parameters of the camera, the welding point coordinates are converted into three-dimensional coordinates of the welding point through the coordinate mapping formula.

[0054] Beneficial effects:

[0055] The multi-point steel mesh welding device based on CNC positioning provided by the present invention can greatly improve the accuracy and stability of welding operations and avoid welding quality problems caused by position deviations by accurately acquiring the three-dimensional coordinates and depth information of welding points. The welding movement data predicted by the machine learning model can realize automatic control of the welding device, reduce manual intervention, and improve production efficiency. The model is trained through historical welding data and can adapt to the welding requirements of different types of steel meshes, thereby achieving good adaptability in a variety of production environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The present invention will be further explained below in conjunction with the accompanying drawings and embodiments:

[0057] Figure 1A schematic diagram of the structure of a multi-point steel mesh welding device based on numerical control positioning provided by an embodiment of the present invention;

[0058] Figure 2 A schematic diagram of the structure of a welding base provided in an embodiment of the present invention;

[0059] Figure 3 A schematic diagram of the structure of a welding station provided in an embodiment of the present invention;

[0060] Figure 4 The embodiment of the present invention provides Figure 3 A magnified view of area A in;

[0061] Figure 5 A schematic diagram of the structure of a welding gun provided in an embodiment of the present invention;

[0062] Figure 6 A module connection diagram of a controller provided in an embodiment of the present invention.

[0063] Description of the drawings: 1. Fixed seat; 11. Fixed hole; 12. Hydraulic telescopic rod; 2. Welding seat; 21. First slide groove; 22. Sliding seat; 23. Inverted concave seat; 24. Electric push rod; 3. Welding table; 31. First slide groove; 32. Strip groove; 33. Positioning hole; 4. Welding gun; 41. Welding gun seat; 42. Crossbeam; 43. Sliding block; 44. Base; 45. Electric telescopic rod; 5. Controller. DETAILED DESCRIPTION

[0064] In order to make the technical means, creative features, objectives and effects of the present invention easy to understand, the present invention is further described below with reference to specific diagrams. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.

[0065] Example 1

[0066] See also Figure 1-Figure 5 The embodiment of the present invention provides a technical solution: a multi-point steel mesh welding device based on numerical control positioning, the welding device is arranged on a steel mesh conveying frame, and is used to weld the steel mesh. The welding device includes:

[0067] There are two fixed seats 1 in total. The fixed seats 1 are of a Z-shaped structure. A fixing hole 11 is opened on the bottom horizontal end of the fixed seat 1. A hydraulic telescopic rod 12 is installed on the upper horizontal end of the fixed seat 1. The bottom end of the hydraulic telescopic rod 12 is fixedly connected to the welding seat 2.

[0068] Specifically, the two fixed seats 1 are provided to fix the welding seat 2. When in use, the two fixed seats 1 are installed and fixed on both sides of the steel mesh conveying frame so that the welding seat 2 is located directly above the steel mesh conveying frame. When in use, by controlling the extension and retraction of the hydraulic telescopic rod 12, the welding seat 2 can be driven to move up and down, thereby realizing the vertical height adjustment of the welding gun 4.

[0069] The welding seat 2 is arranged between the two fixed seats 1. The welding seat 2 can be lifted up and down between the two fixed seats 1. Two first slide grooves 21 are provided at both ends of the lower surface of the welding seat 2 along the width direction. The two first slide grooves 21 are slidably connected with sliding seats 22. An inverted concave seat 23 is fixed between the bottom ends of the two sliding seats 22. The welding table 3 is movably arranged in the inverted concave seat 23. An electric push rod 24 is embedded and installed on one side wall of the inverted concave seat 23 along the length direction. The movable end of the electric push rod 24 extends to the inside of the inverted concave seat 23 and is fixedly connected to the welding table 3.

[0070] Specifically, an electric push rod 24 is also embedded and installed at one end of the first slide groove 21, and the movable end of the electric push rod 24 is fixed to the sliding seat 22. When in use, the electric push rod 24 in the first slide groove 21 is controlled to extend and retract, and the sliding seat 22 can be pushed to move forward and backward in the first slide groove 21, so as to adjust the position of the welding gun 4 horizontally;

[0071] When the longitudinal position of the welding gun 4 needs to be adjusted, the electric push rod 24 in the inverted concave seat 23 is controlled to extend and retract, pushing the welding table 3 to move back and forth in the inverted concave seat 23 to adjust the longitudinal position of the welding gun 4.

[0072] A welding table 3, which is movably arranged below the welding seat 2;

[0073] The welding gun 4 is movably arranged on the welding table 3. A plurality of welding guns 4 are arranged, and the plurality of welding guns 4 are on the same horizontal line.

[0074] A second slide groove 31 and a strip groove 32 are provided on the upper surface of the welding table 3 along the length direction, and a plurality of positioning holes 33 are indirectly provided on the bottom plate of the strip groove 32 along the length direction;

[0075] A slider 43 is slidably connected in the second slide groove 31, a crossbeam 42 is vertically welded on the slider 43, a welding gun seat 41 is welded to the other end of the crossbeam 42, the welding gun 4 is bolted to the welding gun seat 41, a base 44 is provided on the lower surface of the crossbeam 42 at a position opposite to the strip groove 32, the base 44 is slidably connected in the strip groove 32, and an electric telescopic rod 45 is embedded and installed at the center of the lower surface of the base 44.

[0076] Specifically, when in use, the setting of multiple welding guns 4 can simultaneously weld the intersections of multiple horizontal and vertical steel bars located on the same horizontal line, that is, multi-point welding is realized, thereby improving the welding efficiency. Secondly, when multiple welding guns 4 are in use, the spacing between the multiple welding guns 4 can be adjusted according to the lateral spacing set for the steel mesh. The welding gun 4 is adjusted in a manner that the electric telescopic rod 45 is controlled to contract, and the sliding block 43 is pushed down to slide in the second slide groove 31 to adjust the position of the welding gun 4. After the adjustment is completed, the electric telescopic rod 45 is controlled to extend so that it is inserted into the positioning hole 33 opened in the strip groove 32 for limiting the position.

[0077] Example 2

[0078] See also Figure 1-Figure 6 , this embodiment adds a controller 5 on the basis of the above embodiment, and the position of the welding gun 4 is automatically adjusted by the controller 5;

[0079] The controller 5 is arranged on the side wall of the fixing seat 1 and is used to adjust the position of the welding gun 4. The controller 5 includes:

[0080] The data acquisition module collects the image data of the steel mesh through the CCD camera installed on the lower surface of the welding table 3, and pre-processes the collected image data;

[0081] The preprocessing includes using filtering algorithms (such as Gaussian filtering, mean filtering, etc.) to remove noise in the image, reduce unnecessary interference, convert it into a grayscale image to simplify the complexity of subsequent processing, enhance contrast, adjust brightness, and histogram equalization to ensure that the structure of the steel mesh is more obvious.

[0082] The data processing module extracts the pre-processed image data, analyzes and processes the image data, and obtains the three-dimensional coordinates of the welding point.

[0083] The method of analyzing and processing the image data to obtain the three-dimensional coordinates of the welding point includes:

[0084] Perform grayscale processing on the collected image data, obtain the grayscale value of each pixel block in the image data, and mark it as a real-time grayscale value, recorded as HDs, compare and analyze the real-time grayscale value of each block in the image data with the preset welding position grayscale threshold to determine whether to mark the pixel block as a welding area;

[0085] The method for determining whether to mark the welding area of ​​the pixel block includes:

[0086] Preset grayscale threshold HD 1 , HD 2 , and HD 1 <HD 2 ; Grayscale threshold HD 1 , HD 2It is determined by those skilled in the art based on a large number of experiments.

[0087] When HDs<HD 1 , the pixel block is not marked as a welding area;

[0088] When HD 1 ≤HDs≤HD 2 , mark the pixel block as a welding area;

[0089] When HDs>HD 2 , the pixel block is not marked as a welding area.

[0090] Extract pixel blocks with welding area marks, map them to a pre-built two-dimensional coordinate system, mark adjacent pixel blocks as the same welding area, obtain multiple welding areas, and calculate the centroid of each welding area as the coordinate, recorded as the welding point coordinate, obtain the depth value Z of the centroid through the depth camera, and thus calculate the three-dimensional coordinates of the welding point;

[0091] Methods for calculating the centroid of each weld area as coordinates include:

[0092] The area of ​​the welding area is calculated by the formula:

[0093] Where M 00 is the area of ​​the welding area, and I(x,y) is the (x,y) value of the pixel point in the welding area

[0094] Calculate the weighted average of the horizontal and vertical coordinates of the pixel points in the welding area. The calculation formula for the weighted average of the horizontal and vertical coordinates is:

[0095] Where M 10 and M 01 are the weighted averages of the horizontal and vertical coordinates of the pixel points respectively;

[0096] The centroid coordinates (cX, cY) are obtained using the centroid calculation formula.

[0097] The calculation formula of the centroid coordinates is:

[0098]

[0099] Where M 00 is the area of ​​the welding area, M 10 and M 01 are the weighted averages of the horizontal and vertical coordinates of the pixel points, respectively.

[0100] The method for calculating the three-dimensional coordinates of the welding point based on the centroid coordinates is:

[0101] The three-dimensional coordinates of the welding point are calculated by combining the focal length and principal point coordinates of the CCD camera with the centroid coordinates and depth information.

[0102] Preferably, the calculation method of the three-dimensional coordinates A, B, C of the welding point includes:

[0103]

[0104] C = Z;

[0105] Among them, f x , f y is the focal length of the CCD camera, cx, cy are the principal point coordinates of the CCD camera.

[0106] The welding movement data are the movement data of the welding table 3 in the horizontal, longitudinal and vertical directions;

[0107] The welding characteristic parameters include the diameter of the horizontal and vertical steel bars of the steel mesh and the three-dimensional coordinates of the welding points.

[0108] The welding control module obtains real-time welding characteristic parameters, inputs them into a pre-built machine learning model that predicts welding movement data, predicts the welding movement data, and controls the operation of the welding device based on the welding movement data. The welding movement data is the movement data of the welding table 3 in the horizontal, longitudinal and vertical directions. The welding characteristic parameters include the diameters of the horizontal and longitudinal steel bars of the steel mesh and the three-dimensional coordinates of the welding points.

[0109] The training method of the machine learning model that predicts welding movement data includes:

[0110] Collecting historical welding training parameters of the welding device, the historical welding training parameters are collected when the welding device is in a state of meeting welding standards, and the historical welding training parameters include welding characteristic parameters and welding movement data;

[0111] Convert the collected historical welding training parameters into a corresponding set of feature vectors;

[0112] Each set of feature vectors is used as the input of the machine learning model. The machine learning model takes the welding movement data corresponding to each set of welding feature parameters as output, the welding movement data actually corresponding to each set of welding feature parameters is used as the prediction target, and minimizing the loss function value of the machine learning model is used as the training goal. Training is stopped when the loss function value of the machine learning model is less than or equal to the preset target loss value.

[0113] The machine learning model may be one of support vector machine regression, random forest regression or neural network regression models.

[0114] The loss function value of the machine learning model is the mean square error.

[0115] Mean square error is one of the commonly used loss functions. The model is trained with minimization as the goal, so that the machine learning model can better fit the data, thereby improving the performance and accuracy of the model.

[0116] In the loss function, MSE is the loss function value of the machine learning model, x is the feature vector group number; m is the number of feature vector groups; y is the number of feature vector groups; x is the welding movement data corresponding to the xth group of feature vectors, It is the welding movement data corresponding to the xth group of feature vectors in real time.

[0117] Other model parameters of the machine learning model, target loss value, optimization algorithm, training set test set validation set ratio, and loss function optimization are all achieved through actual engineering implementation and continuous experimental tuning.

[0118] In this embodiment, by collecting historical welding training parameters of the welding device, a machine learning model that predicts welding movement data is trained based on the historical welding training parameters, and then the image data of the steel mesh is collected, and then grayscale processing is performed to obtain the position of the welding area, and the center of mass of each welding area is calculated as the coordinate, recorded as the welding point coordinate, and the depth value Z of the center of mass is obtained by the depth camera, thereby calculating the three-dimensional coordinates of the welding point, and collecting the diameters of the horizontal and vertical steel bars of the steel mesh, The diameters of the horizontal and vertical steel bars of the steel mesh and the welding point coordinates are input into the machine learning model that predicts the welding movement data, and the welding movement data is predicted based on the trained machine learning model, and the welding device is controlled to work based on the welding movement data. By accurately obtaining the three-dimensional coordinates and depth information of the welding point, the accuracy and stability of the welding operation can be greatly improved, and welding quality problems caused by position deviation can be avoided. Based on the welding movement data predicted by the machine learning model, the welding device can be automatically controlled, reducing manual intervention and improving production efficiency. The model is trained by historical welding data, and can adapt to the welding requirements of different types of steel meshes, and achieve good adaptability in a variety of production environments.

[0119] Example 3

[0120] The method of analyzing and processing the image data to obtain the three-dimensional coordinates of the welding point includes:

[0121] Step 1, using image processing algorithms to extract straight lines from image data and locate the position of the steel bars;

[0122] It should be noted that in step 1, after extracting straight lines from image data using an image processing algorithm, the following steps are further included:

[0123] By sequentially obtaining the spacing between adjacent horizontal lines and the spacing between vertical lines, a set is established to determine the density of the steel mesh, and the lines corresponding to the abnormal spacing values ​​are obtained and removed.

[0124] Step 2, locate the intersection of the steel mesh by calculating the intersection of the straight lines, and mark the intersection as the welding point;

[0125] Step 3: perform coordinate transformation on the welding point coordinates to generate the three-dimensional coordinates of the welding point.

[0126] Said Methods for extracting straight lines from image data using image processing algorithms include:

[0127] By performing edge detection on the image data (such as using the Canny algorithm), all edge points are extracted; it should be noted that edge points refer to pixels in the image, which are part of the edge, that is, the pixels around these points have obvious changes in brightness, color or grayscale value. By performing edge detection on the image data (such as using the Canny algorithm), all edge points are extracted;

[0128] Convert each edge point (a, b) in the image data into coordinates (ρ, θ) in the polar coordinate system to represent a straight line, where ρ represents the shortest distance from the straight line to the coordinate (0.0), and θ represents the angle of the straight line, which is between 0° and 90°;

[0129] For each edge point (a, b) in the image, the coordinates in the polar coordinate system are calculated using the following formula:

[0130] ρ = acosθ + bsinθ;

[0131] For each edge point (a, b), repeat the above calculation, traverse different θ values, and obtain a series of (ρ, θ) parameter pairs.

[0132] It should be noted that θ is the angle range of detection, which is between 0° and 90°. In order to calculate the position of the straight line, it is necessary to traverse each edge point in the image and convert it into a point in the polar coordinate system.

[0133] Create a two-dimensional accumulator, where each pair of ρ and θ corresponds to a spatial position, traverse all edge points in the image, map their corresponding ρ and θ values ​​to the accumulator, and accumulate them;

[0134] It should be noted that the value in the accumulator represents the number of "votes" for a line corresponding to a certain pair of ρ and θ combinations. Each vote indicates that the line passes through a certain edge point in the image.

[0135] Set the accumulation threshold, remove the sum combination whose median value of the accumulator is lower than the accumulation threshold, because they correspond to noise or unclear straight lines, and obtain the ρ and θ combination whose median value of the accumulator is higher than the threshold, ρ and θ correspond to the straight lines in the steel mesh in the image;

[0136] According to the obtained combination of ρ and θ, draw the corresponding straight line on the image;

[0137] It should be noted that the drawing of the straight line is based on the following formula:

[0138] a=ρcosθ+bsinθ,

[0139] The ρ and θ parameters of each line are converted back into image data to plot the actual reinforcement positions;

[0140] To locate the welding point, after extracting the straight lines of the steel bars, calculate the coordinates of the intersection of each intersecting line. The coordinates of the intersection are the coordinates of the welding point.

[0141] It should be noted that the equation of line 1: y = m 1 x+b 1 ; Line 2 equation: y = m 2 x+b 2

[0142] The intersection point (x,y) can be found by solving the system of equations of the two lines:

[0143] m 1 x+b 1 =m 2 x+b 2 ,

[0144] The values ​​of x and y obtained are the coordinates of the intersection point.

[0145] The method for performing coordinate conversion on the welding point coordinates to generate the three-dimensional coordinates of the welding point comprises:

[0146] Perform camera calibration to obtain the camera's internal and external parameters, including focal length, principal point coordinates, and distortion parameters;

[0147] According to the calibrated internal and external parameters of the camera, the welding point coordinates are converted into three-dimensional coordinates of the welding point through the coordinate mapping formula.

[0148] The formula is as follows:

[0149]

[0150] Among them: X, Y, Z are the three-dimensional coordinates of the welding point. u, v are the coordinates of the welding point. K is the camera's intrinsic matrix, and R and T are the camera's rotation matrix and displacement vector (external parameters). Through this transformation, the three-dimensional coordinates of the welding point are obtained.

[0151] In this embodiment, the process of analyzing the position of the welding point and generating coordinates based on the steel mesh image data collected by the CCD camera can greatly improve the welding accuracy, production efficiency, and welding quality, and promote the intelligence and automation of the welding system, reduce labor costs, optimize the production process, and enhance the control ability of the production process.

[0152] The basic principles, main features and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and the description in the specification are only to illustrate the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A multi-point steel mesh welding device based on numerical control positioning, which is arranged on a steel mesh conveying frame and is used to weld steel mesh, and is characterized in that: The welding device comprises: Two fixing seats (1) are provided; A welding seat (2) is arranged between the two fixing seats (1), and the welding seat (2) can be lifted up and down between the two fixing seats (1); A welding table (3) movably arranged below the welding seat (2); A welding gun (4) is movably arranged on the welding table (3), wherein a plurality of welding guns (4) are arranged, and the plurality of welding guns (4) are located on the same horizontal line; A controller (5) is arranged on a side wall of the fixing seat (1) and is used to adjust the position of the welding gun (4). The controller (5) comprises: A data acquisition module, which acquires image data of the steel mesh through a CCD camera installed on the lower surface of the welding table (3), and pre-processes the acquired image data; The data processing module extracts the pre-processed image data, analyzes and processes the image data, and obtains the three-dimensional coordinates of the welding point; The welding control module obtains real-time welding characteristic parameters, inputs them into a pre-built machine learning model for predicting welding movement data, predicts the welding movement data, and controls the operation of the welding device based on the welding movement data, wherein the welding movement data is the movement data of the welding table (3) in the horizontal, longitudinal and vertical directions, and the welding characteristic parameters include the diameters of the horizontal and longitudinal steel bars of the steel mesh and the three-dimensional coordinates of the welding points.

2. The multi-point steel mesh welding device based on numerical control positioning according to claim 1 is characterized in that: The method of analyzing and processing the image data to obtain the three-dimensional coordinates of the welding point includes: Step 1, using image processing algorithms to extract straight lines from image data and locate the position of the steel bars; Step 2, locate the intersection of the steel mesh by calculating the intersection of the straight lines, and mark the intersection as the welding point; Step 3: perform coordinate transformation on the welding point coordinates to generate the three-dimensional coordinates of the welding point.

3. The multi-point steel mesh welding device based on numerical control positioning according to claim 2 is characterized in that: Said Methods for extracting straight lines from image data using image processing algorithms include: By performing edge detection on the image data, all edge points (a, b) are extracted; Convert each edge point (a, b) in the image data into coordinates (ρ, θ) in the polar coordinate system to represent a straight line, where ρ represents the shortest distance from the straight line to the coordinate (0.0), and θ represents the angle of the straight line, which is between 0° and 90°; For each edge point (a, b) in the image, the coordinates in the polar coordinate system are calculated using the formula: For each edge point (a, b), repeat the above calculation, traverse different θ values, and obtain a series of (ρ, θ) parameter pairs; Create a two-dimensional accumulator, where each pair of ρ and θ corresponds to a spatial position, traverse all edge points in the image, map their corresponding ρ and θ values ​​to the accumulator, and accumulate them; Set the accumulation threshold, remove the ρ and θ combinations whose accumulator median values ​​are lower than the accumulation threshold, and obtain the ρ and θ combinations whose accumulator median values ​​are higher than the threshold, where ρ and θ correspond to the straight lines in the steel mesh in the image; According to the obtained combination of ρ and θ, draw the corresponding straight line on the image; The ρ and θ parameters of each line are converted back into image data to plot the actual reinforcement positions; To locate the welding point, after extracting the straight lines of the steel bars, calculate the coordinates of the intersection of each intersecting line. The coordinates of the intersection are the coordinates of the welding point.

4. The multi-point steel mesh welding device based on numerical control positioning according to claim 1 is characterized in that: The method of analyzing and processing the image data to obtain the three-dimensional coordinates of the welding point includes: Perform grayscale processing on the collected image data, obtain the grayscale value of each pixel block in the image data, and mark it as a real-time grayscale value, recorded as HDs, compare and analyze the real-time grayscale value of each block in the image data with the preset welding position grayscale threshold to determine whether to mark the pixel block as a welding area; Pixel blocks with welding area marks are extracted and mapped to a pre-constructed two-dimensional coordinate system. Adjacent pixel blocks are marked as the same welding area to obtain multiple welding areas. The centroid of each welding area is calculated as the coordinate, recorded as the welding point coordinate. The depth value Z of the centroid is obtained through the depth camera to calculate the three-dimensional coordinates of the welding point.

5. The multi-point steel mesh welding device based on numerical control positioning according to claim 4 is characterized in that: The method for determining whether to mark a welding area on a pixel block comprises: Preset grayscale thresholds HD1 and HD2, and HD1<HD2; When HDs<HD1, the pixel block is not marked as a welding area; When HD1≤HDs≤HD2, the pixel block is marked as a welding area; When HDs>HD2, the welding area mark is not performed on the pixel block.

6. The multi-point steel mesh welding device based on numerical control positioning according to claim 5 is characterized in that: Methods for calculating the centroid of each weld area as coordinates include: Calculate the area of ​​the welding area by summing the formula; Calculate the weighted average of the horizontal and vertical coordinates of the pixels in the welding area; The centroid coordinates are obtained using the centroid calculation formula.

7. The multi-point steel mesh welding device based on numerical control positioning according to claim 6 is characterized in that: The method for calculating the three-dimensional coordinates of the welding point based on the centroid coordinates is: The three-dimensional coordinates of the welding point are calculated by combining the focal length and principal point coordinates of the CCD camera with the centroid coordinates and depth information.

8. The multi-point steel mesh welding device based on numerical control positioning according to claim 1 is characterized in that: The training method of the machine learning model for predicting welding movement data includes: Collecting historical welding training parameters of the welding device, the historical welding training parameters are collected when the welding device is in a state of meeting welding standards, and the historical welding training parameters include welding characteristic parameters and welding movement data; Converting the collected historical welding training parameters into a corresponding set of feature vectors; Each group of feature vectors is used as the input of the machine learning model, and the machine learning model takes the welding movement data corresponding to each group of welding feature parameters as the output, and the welding movement data actually corresponding to each group of welding feature parameters is used as the prediction target, and minimizing the loss function value of the machine learning model is used as the training target; training is stopped when the loss function value of the machine learning model is less than or equal to the preset target loss value.

9. The multi-point steel mesh welding device based on numerical control positioning according to claim 3 is characterized in that: In the step 1, after extracting straight lines from the image data using an image processing algorithm, the step further includes: By sequentially obtaining the spacing between adjacent horizontal lines and the spacing between vertical lines, a set is established to determine the density of the steel mesh, and the lines corresponding to the abnormal spacing values ​​are obtained and removed.

10. The multi-point steel mesh welding device based on numerical control positioning according to claim 2 is characterized in that: The method for performing coordinate conversion on the welding point coordinates to generate the three-dimensional coordinates of the welding point comprises: Perform camera calibration to obtain the camera's internal and external parameters, including focal length, principal point coordinates, and distortion parameters; According to the calibrated internal and external parameters of the camera, the welding point coordinates are converted into three-dimensional coordinates of the welding point through the coordinate mapping formula.

Images