A tracking welding control method, system and device for irregular welds

Abstract

The present invention relates to a tracking welding control method, system and equipment for irregular welds, and relates to the field of welding technology, including: in the welding process of irregular welds, laser projection is performed on the surface of the workpiece, the workpiece image is collected, and the welding power, welding position information, and welding speed are obtained, and irregularity analysis is performed to obtain the irregularity; according to the welding power sequence and welding speed sequence in the most recent preset time window, adjacent welding power analysis and adjacent welding speed analysis are performed to obtain the welding power coefficient and welding speed coefficient, and according to the irregularity, the irregularity coefficient is calculated; according to the irregularity decision, image downsampling parameters are obtained, the workpiece image is downsampled, image recognition resource decisions are made, recognition is performed, weld position information is obtained, and tracking welding control is performed. The present invention solves the technical problems of irregular weld tracking welding control being difficult, having low recognition accuracy, and having low timeliness in the prior art.

Description

Technical Field

[0001] The present invention relates to the field of welding technology, and in particular to a tracking welding control method, system and equipment for irregular welds. Background Art

[0002] Irregular welds appear as curves with large changes in curvature, which makes the welding process more difficult. It is also difficult to identify the exact weld position during automatic tracking welding control, and welding deviations are prone to occur.

[0003] During the welding process of irregular welds, laser recognition sensing is often used to identify the weld position. When the curvature of the irregular weld is large and the welding speed is fast, the recognition is difficult, and the bright light generated during welding will cause the image to be overexposed, which increases the difficulty of recognition, resulting in low accuracy and timeliness of weld position recognition. Therefore, the existing technology has technical problems such as high difficulty in controlling irregular weld tracking welding, low recognition accuracy, and low timeliness. Summary of the invention

[0004] The present invention aims to solve the technical problems in the prior art of irregular weld tracking welding control being difficult, having low recognition accuracy and low timeliness, and provides a tracking welding control method, system and equipment for irregular welds to solve the problems.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] In a first aspect, the present invention provides a tracking welding control method for an irregular weld, comprising: during the welding process of the irregular weld, performing laser projection on the surface of a workpiece, collecting a workpiece image, and obtaining welding power, welding position information, and welding speed, performing irregularity analysis according to the welding position information, and obtaining the irregularity;

[0007] According to the welding power sequence and welding speed sequence in the most recent preset time window, adjacent welding power analysis and adjacent welding speed analysis are performed to obtain welding power coefficient and welding speed coefficient, and according to the irregularity, an irregularity coefficient is calculated;

[0008] According to the irregularity decision, image downsampling parameters are obtained, and the workpiece image is downsampled. According to the welding power, welding speed and irregularity, in combination with the welding power coefficient, welding speed coefficient and irregularity coefficient, an image recognition resource decision is made, the downsampled workpiece image is recognized, the weld position information is obtained, and tracking welding control is performed.

[0009] In a second aspect, the present invention provides a tracking welding control system for irregular welds, comprising: a welding information acquisition module, which performs laser projection on the workpiece surface during the welding process of the irregular weld, collects workpiece images, and obtains welding power, welding position information, and welding speed, and performs irregularity analysis according to the welding position information to obtain irregularity;

[0010] The welding change analysis module performs adjacent welding power analysis and adjacent welding speed analysis according to the welding power sequence and welding speed sequence in the most recent preset time window to obtain the welding power coefficient and the welding speed coefficient, and calculates the irregularity coefficient according to the irregularity;

[0011] The identification welding control module is used to obtain image downsampling parameters according to the irregularity decision, downsample the workpiece image, make image recognition resource decisions according to the welding power, welding speed and irregularity, combined with the welding power coefficient, welding speed coefficient and irregularity coefficient, identify the downsampled workpiece image, obtain weld position information, and perform tracking welding control.

[0012] In a third aspect, the present invention provides a tracking welding control device for irregular welds, including the tracking welding control system for irregular welds in the second aspect.

[0013] The beneficial effects of the present invention are as follows: during the welding process, the present invention uses laser projection technology to project the weld trajectory on the workpiece surface and collect the workpiece image. Then, by obtaining the welding power, welding position information and welding speed, and performing irregularity analysis, the curvature and irregularity of the weld can be dynamically evaluated during the real-time welding process. Further, by analyzing the changes in welding power and welding speed, the welding power coefficient and welding speed coefficient are calculated, which provides an important basis for the dynamic adjustment of the welding state during the welding process. According to the irregularity decision, the image downsampling parameters are obtained and the workpiece image is downsampled. This feature can effectively reduce the computational burden of image processing when the welding speed is fast, improve the response speed, and maintain the image recognition accuracy. The data processing after image downsampling can save computing resources, and the downsampling processing according to the irregularity can ensure the recognition quality, which improves the real-time performance of the recognition processing during the welding process. By combining the welding power coefficient and the welding speed coefficient to make image recognition resource decisions and perform weld recognition, the accuracy and real-time performance of weld position recognition are further improved. The present invention provides a tracking welding control method for irregular welds, which significantly improves the accuracy and real-time performance of weld position recognition during the welding process by accurately analyzing and adjusting multiple key factors in the welding process, and effectively solves the problems existing in the prior art of irregular welds being difficult to accurately track, having low recognition accuracy and poor timeliness. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A schematic flow chart of a tracking welding control method for an irregular weld provided by the present invention;

[0015] Figure 2 A schematic structural diagram of a tracking welding control system for irregular welds provided by the present invention.

[0016] In the accompanying drawings, the components represented by the reference numerals are described as follows:

[0017] Welding information acquisition module 11, welding change analysis module 12, and welding identification control module 13. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0019] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0020] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0021] Embodiment 1, as Figure 1 As shown, an embodiment of the present invention provides a tracking welding control method for an irregular weld, which specifically includes the following steps:

[0022] S10: During the welding process of the irregular weld, laser projection is performed on the workpiece surface to collect the workpiece image, and obtain the welding power, welding position information, and welding speed, and irregularity analysis is performed according to the welding position information to obtain the irregularity.

[0023] In the embodiment of the present application, during the welding process of irregular welds, laser projection technology is first required to ensure high-precision weld trajectory acquisition. Specifically, by performing laser projection on the workpiece surface, the laser line is projected onto the welding area, and the laser pattern at the weld will be deformed or disappear, based on which the weld position is identified.

[0024] After laser projection, the image of the welding area on the workpiece surface is collected, the workpiece image is obtained, and the weld position is identified.

[0025] Furthermore, the welding power, welding position information, and welding speed in the current welding are collected, and the irregularity of the weld at the current welding position is analyzed according to the welding position information to obtain the irregularity.

[0026] Step S10 in the method provided in the embodiment of the present application includes:

[0027] During the welding process of irregular welds, laser projection is performed on the workpiece surface to collect workpiece images;

[0028] Collect the current welding power and welding moving speed;

[0029] A coordinate system is constructed on the workpiece surface to obtain the coordinates of the current welding position as welding position information.

[0030] In an embodiment of the present application, during the welding process of irregular welds, in order to accurately track the weld trajectory and precisely control the welding quality, it is first necessary to perform laser projection on the workpiece surface, and then collect the workpiece image of the welding area. For example, a rectangular laser pattern can be projected by laser projection, and the laser pattern at the weld will be deformed, and an industrial camera is used to collect the workpiece surface image as the workpiece image. The workpiece image contains the outline of the weld, as well as the details of the weld and the surrounding area. The collected image will serve as the basis for subsequent weld position identification to ensure accurate positioning and tracking of irregular welds during the welding process.

[0031] Furthermore, the current welding power and welding moving speed are collected. The welding power is the energy output generated by the arc during the welding process, which is usually controlled by the power supply of the welding equipment and can be obtained through circuit sensor detection. The size of the welding power directly affects the heat input of the welding, and then affects the intensity of the arc light generated by the welding. The arc light will cause the collected workpiece image to be overexposed, thereby affecting the recognition accuracy. Exemplarily, the welding power is 5000W.

[0032] The welding moving speed is the speed at which the welding head moves along the weld path on the workpiece surface, which can be obtained by arranging a speed sensor on the welding equipment. When the welding speed is high, the workpiece image recognition speed cannot keep up with the welding moving speed, which leads to insufficient timeliness of weld position recognition and affects the welding quality. For example, the welding moving speed is 200 mm / min.

[0033] And build a coordinate system on the surface of the workpiece. The purpose of building a coordinate system is to convert the position data in the welding process into an operational digital signal for precise control. Usually, the coordinate system adopts a three-dimensional coordinate system (X, Y, Z axis), which is built according to the length, width and height of the workpiece, and then the position coordinates of the current welding position are collected as welding position information. Or the coordinate system is defined based on the mobile device of the welding robot arm to obtain the real-time welding position coordinates as welding position information.

[0034] The embodiment of the present application collects the welding position, welding power and welding speed during the welding process as basic data for subsequent adjustment and identification of the workpiece image, and then analyzes the timeliness requirement of identifying the workpiece image at the current welding position, thereby ensuring the quality of the weld and the efficiency of the welding process.

[0035] Step S10 in the method provided in the embodiment of the present application also includes:

[0036] According to the preset length interval, the welding position interval is divided with the welding position information as the end point;

[0037] The average curvature of the welding position interval is calculated as the irregularity.

[0038] In the embodiment of the present application, an analysis area is determined on the welding trajectory according to the welding position information to analyze the irregularity of the weld at the current welding position. For example, with the welding position information as the end point, the welds completed in the past are divided according to the preset length interval to obtain the welding position interval. The preset length interval refers to a range set on the weld path completed by welding, which is usually a fixed length range and is used to analyze the changes in the welding path. For example, the preset length interval is 50 mm.

[0039] The welding position interval where welding is completed includes a section of the weld path before the current welding position. The average curvature, that is, the bending degree of the welding trajectory, is calculated according to the curve formed by the weld trajectory in the welding position interval.

[0040] Specifically, the coordinates of multiple welding points within the welding position range can be obtained, and then the welding trajectory curve can be obtained by least square fitting, and then the tangent vector of each welding point and the second-order derivative of the curve can be calculated, and then the curvature of multiple points can be calculated according to the welding trajectory curve, and then the mean can be calculated to obtain the average curvature. The average curvature is used as the irregularity of the current welding position.

[0041] Generally speaking, the greater the irregularity of the adjacent welding rule, the greater the irregularity of the subsequent weld. The greater the irregularity, the higher the accuracy requirement for weld position recognition. By analyzing the irregularity, a data basis can be provided for subsequent workpiece image recognition preprocessing, thereby improving the accuracy of weld position recognition.

[0042] S20: performing adjacent welding power analysis and adjacent welding speed analysis according to the welding power sequence and welding speed sequence in the most recent preset time window to obtain the welding power coefficient and the welding speed coefficient, and calculating the irregularity coefficient according to the irregularity.

[0043] In the embodiment of the present application, the welding power during the welding process will affect the welding arc intensity, and thus affect the image quality and the difficulty of workpiece image recognition. The welding speed will require different workpiece image recognition efficiencies. Irregularity will affect the difficulty of workpiece image recognition. Therefore, based on the welding power sequence and welding speed sequence in the recent time, adjacent welding power analysis and welding speed analysis are performed, as well as irregularity analysis, to obtain the welding power coefficient, welding speed coefficient and irregularity coefficient, which respectively reflect the degree of influence of the three types of parameters on workpiece image recognition.

[0044] Step S20 in the method provided in the embodiment of the present application includes:

[0045] Collect welding power and welding speed at multiple time stamps within the latest preset time window to obtain welding power sequence and welding speed sequence;

[0046] According to the welding power sequence and the welding speed sequence, an average welding power and an average welding speed are calculated;

[0047] The maximum welding power and the maximum welding speed are obtained, and the ratios of the average welding power and the average welding speed to the maximum welding power and the maximum welding speed are calculated as the welding power coefficient and the welding speed coefficient respectively;

[0048] The maximum irregularity is obtained, and the ratio of the irregularity to the maximum irregularity is calculated as the irregularity coefficient.

[0049] In an embodiment of the present application, during the welding process, a preset time window (e.g., 5 seconds) is set, and within this time range, welding parameter data is collected once every fixed time interval (e.g., 0.1 seconds) to form a welding power sequence and a welding speed sequence.

[0050] Exemplarily, the welding power sequence is {3500 W, 3520 W, 3495 W, 3510 W, 3550 W}, and the welding speed sequence is {250 mm / min, 248 mm / min, 253 mm / min, 249 mm / min, 260 mm / min}.

[0051] Furthermore, the average values ​​of the welding power and the welding speed in the welding power sequence and the welding speed sequence are calculated to obtain the average welding power and the average welding speed. For example, the average welding power is 3515W and the average welding speed is 252mm / min.

[0052] Further, the maximum welding power and the maximum welding speed are obtained, which are the maximum welding power value and the maximum welding speed value recorded in the historical welding data, such as the maximum welding power value and the maximum welding speed value recorded in the past month, for example, 5000W and 300mm / min respectively.

[0053] Further, the ratios of the average welding power and the average welding speed to the maximum welding power and the maximum welding speed are calculated as the welding power coefficient and the welding speed coefficient, for example, 3515W / 5000W=0.703 and 252mm / min / 300mm / min=0.84, respectively.

[0054] The welding power coefficient and welding speed coefficient reflect the size of the welding power and welding speed in the recent time. The greater the welding power, the greater the impact on the welding image quality, and the higher the image recognition quality is required, that is, the greater the image recognition resources are required. The greater the welding speed, the faster the image recognition speed is required to improve the weld recognition response speed, that is, the fewer image recognition resources are used to improve the recognition efficiency.

[0055] Furthermore, the maximum irregularity is obtained, for example, the maximum value of the irregularity of the processing records in the past month, and the processing method of the irregularity is the same as the above content.

[0056] The ratio of the current irregularity to the maximum irregularity is calculated as the irregularity coefficient. The irregularity coefficient reflects the degree of irregularity of the current welding position. The larger the irregularity, the greater the weld irregularity, the greater the recognition difficulty, and the higher the image recognition quality required, that is, the greater the image recognition resources required.

[0057] By calculating the welding power coefficient, welding speed coefficient and irregularity coefficient, the recognition computing resources can be used for subsequent decision-making to identify the current workpiece image, thereby ensuring that the efficiency, quality and timeliness of workpiece image recognition meet the requirements of current welding characteristics.

[0058] S30: Obtain image downsampling parameters based on the irregularity decision, perform downsampling processing on the workpiece image, make image recognition resource decisions based on the welding power and welding speed, combined with the welding power coefficient and the welding speed coefficient, identify the downsampled workpiece image, obtain weld position information, and perform tracking welding control.

[0059] In the embodiment of the present application, in order to improve the efficiency of irregular weld identification, the image needs to be downsampled. However, in order to ensure image quality and recognition accuracy, the downsampling processing parameters need to be configured. The image downsampling parameters are obtained based on the irregularity. The greater the irregularity, the greater the image downsampling parameters and the greater the resolution of the downsampled image, so as to ensure image quality and identify workpiece images at locations with greater irregularities.

[0060] Furthermore, after downsampling the workpiece image, the decision configuration of the image recognition resources is performed according to the welding power, welding speed and irregularity, combined with the welding power coefficient, welding speed coefficient and irregularity coefficient, so as to improve the image recognition quality while ensuring the recognition timeliness, adapt to the current welding characteristics, and avoid the situation where the welding recognition tracking is deviated due to insufficient image recognition quality or insufficient timeliness.

[0061] Step S30 in the method provided in the embodiment of the present application includes:

[0062] Obtaining a downsampling parameter interval for downsampling the workpiece image, wherein the size of the downsampling parameter is positively correlated with the size of the image resolution after downsampling;

[0063] Get the maximum irregularity;

[0064] Calculating the ratio of the irregularity to the maximum irregularity, and obtaining the image downsampling parameter by calculation within the downsampling parameter interval;

[0065] The image downsampling parameters are used to perform downsampling processing on the workpiece image.

[0066] In an embodiment of the present application, a downsampling parameter interval for downsampling the workpiece image is first obtained. The downsampling parameter is specifically the ratio of the resolution of the workpiece image after downsampling to the original resolution. The size of the downsampling parameter is positively correlated with the size of the image resolution after downsampling.

[0067] An exemplary downsampling parameter interval is (0.1, 1), where if the downsampling parameter is 0.1, the ratio of the resolution of the workpiece image after downsampling to the original resolution is 0.1. For example, the workpiece image resolution is , the resolution after downsampling is If the downsampling parameter is 1, the workpiece image will not be downsampled.

[0068] The greater the irregularity, the more difficult it is to identify the weld position, the higher the image quality required, and the larger the downsampling parameter. The smaller the irregularity, the easier it is to identify the weld position, the lower the image quality required, and the smaller the downsampling parameter.

[0069] Furthermore, the maximum irregularity is obtained, for example, the maximum irregularity recorded in the past month, for example, 0.2. Then the ratio of the irregularity to the maximum irregularity is calculated, and the image downsampling parameters are calculated within the downsampling parameter interval. Specifically, the product of the ratio and the length of the downsampling parameter interval is calculated, and then the minimum downsampling parameter plus the product is used as the image downsampling parameter. For example, the length of the downsampling parameter interval (0.1, 1) is 0.9, and the current irregularity is 0.1, then 0.1 / 0.2=50% is calculated, and the image downsampling parameter is used. , using the minimum downsampling parameter 0.1+0.45=0.55, the downsampling parameter is 0.55. For example, the resolution of the workpiece image after downsampling is .

[0070] For example, during the downsampling process, the workpiece image can be downsampled to several resolution levels, such as 960×540 and 480×270, etc., and the resolution level closest to the workpiece image after calculation or downsampling is selected to perform downsampling processing on the workpiece image, which is more efficient, for example The closest resolution level is 960×540, and the workpiece image is downsampled according to this resolution level.

[0071] Finally, the workpiece image is downsampled according to the image downsampling parameters. The downsampling method may be bilinear interpolation, etc., to downsample and reduce the workpiece image.

[0072] In the embodiment of the present application, by calculating the optimal downsampling parameters in combination with the irregularity ratio, the resolution of the workpiece image can be dynamically adjusted to ensure that the computing cost can be reduced while maintaining the integrity of the weld features, thereby ensuring the accuracy and timeliness of the workpiece image recognition of the weld position.

[0073] Step S30 in the method provided in the embodiment of the present application further includes:

[0074] According to the welding tracking data in the historical time, the weld identification path of the integrated weld identification quantity is trained to obtain the weld identification channel;

[0075] According to the welding power, welding speed and irregularity, combined with the welding power coefficient, welding speed coefficient and irregularity coefficient, the number of identification paths of the integrated weld identification quantity is calculated and determined to obtain the number of weld identification paths;

[0076] Randomly selecting the weld recognition paths of the number of weld recognition paths in the weld recognition channel, inputting the downsampled workpiece image, identifying and obtaining the weld position information of the number of weld recognition paths, and calculating the mean to obtain the weld position information;

[0077] Tracking welding control is performed according to the weld position information.

[0078] In the embodiment of the present application, firstly, the weld recognition path of the integrated weld recognition number is trained according to the welding tracking data in the historical time to obtain the weld recognition channel. By training the weld recognition channel including multiple weld recognition paths, the weld position recognition of the workpiece image under the welding characteristics of different welding positions can be handled, thereby ensuring the recognition accuracy and timeliness.

[0079] The step of "training a weld identification path integrating a number of weld identifications based on welding tracking data in historical time to obtain a weld identification channel" in the method provided in the embodiment of the present application includes:

[0080] According to the welding tracking data in the historical time, a set of sample workpiece images is collected, and the weld position information in each sample workpiece image is identified and marked to obtain a set of sample weld position information;

[0081] The sample workpiece image set and the sample weld position information set are randomly divided multiple times to obtain weld position recognition training data with an integrated weld recognition quantity;

[0082] The weld position recognition training data of the integrated weld recognition quantity is used to train the weld recognition path of the integrated weld recognition quantity to obtain the weld recognition channel.

[0083] In the embodiment of the present application, in the historical welding process, the tracking data of the weld seam, including the image of the welded workpiece, is collected in real time to obtain a set of sample workpiece images. Preferably, the workpiece images recorded during the welding process of similar workpieces can be collected as sample workpiece images.

[0084] Furthermore, the position of the weld in each sample workpiece image is marked, and computer recognition or manual auxiliary marking can be used to obtain a sample weld position information set. The sample weld position information includes the specific pixel coordinates of the weld in the image, such as (x, y) = (120, 300).

[0085] Furthermore, in order to train the weld recognition path of the integrated weld recognition number, the sample workpiece image set and the sample weld position information set are randomly divided multiple times. For example, 60% of the data is randomly divided each time, and the sample workpiece image set and the sample weld position information set are put back after the division. The division is performed 10 times to obtain 10 groups of weld position recognition training data, and then 10 weld recognition paths are trained. The integrated weld recognition number is 10.

[0086] Furthermore, the weld position recognition training data of the integrated weld recognition quantity is used to train the weld recognition paths of the integrated weld recognition quantity respectively, and preferably a convolutional neural network is used to train the weld recognition paths.

[0087] Taking the training process of one of the weld recognition paths as an example, the training steps of the weld recognition paths with integrated weld recognition numbers are the same, but the training data are different. Therefore, weld recognition paths with different integrated weld recognition numbers can be trained to perform weld position recognition of different workpiece images. Integrating the recognition results of multiple paths can improve accuracy.

[0088] For example, based on the convolutional neural network, a weld recognition path is constructed, including an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer. The convolution kernel size of the convolutional layer is , including 3 convolutional layers and pooling layers, to perform multi-layer extraction of workpiece image features. The fully connected layer uses the mean square error loss function to calculate the error between the weld position information output by the weld recognition path and the sample weld position information as the loss. Adam is used to optimize the network parameters with a learning rate of 0.001. The weld position recognition training data is then used to train the weld recognition path. During the training, the network parameters are optimized according to the loss adjustment, so that the loss value is gradually reduced until the requirements are met, for example, until the error loss is less than 5%. The training is completed and further testing can be performed. If the test meets the requirements, the training is completed. If it does not meet the requirements, the training continues until the requirements are met.

[0089] In this way, the weld position recognition training data of the integrated weld recognition quantity is used to respectively train and obtain the weld recognition paths of the integrated weld recognition quantity, and the multiple weld recognition paths that have been trained are combined to obtain the weld recognition channel.

[0090] After the weld recognition channel training is completed, it is necessary to make a calculation decision on the number of recognition paths based on the welding power, welding speed and irregularity, combined with the welding power coefficient, welding speed coefficient and irregularity coefficient, that is, the calculation decision of the image recognition resources, to adapt to the position characteristics of the current welding position and ensure the accuracy and timeliness of weld position recognition.

[0091] The step of "calculating and deciding the number of identification paths for the integrated weld identification number based on the welding power, welding speed and irregularity, combined with the welding power coefficient, welding speed coefficient and irregularity coefficient, to obtain the number of weld identification paths" in the method provided in the embodiment of the present application includes:

[0092] Get the maximum welding power, maximum welding speed, and maximum irregularity;

[0093] Calculate the ratio of the welding power and irregularity to the maximum welding power and the maximum irregularity, and subtract the ratio of the welding speed to the maximum welding speed from 1 to obtain a power decision coefficient, a speed decision coefficient, and an irregularity decision coefficient, and multiply them by the integrated weld identification number and retain integers to obtain a power identification number, a speed identification number, and an irregularity identification number;

[0094] According to the welding power coefficient, welding speed coefficient and irregularity coefficient, the power identification quantity, speed identification quantity and irregularity identification quantity are weightedly calculated to obtain the weld identification path quantity.

[0095] In the embodiment of the present application, the maximum welding power, the maximum welding speed, and the maximum irregularity are first obtained, for example, the maximum welding power, welding speed, and irregularity recorded in welding within the past month.

[0096] Then, the ratios of the welding power and the irregularity to the maximum welding power and the maximum irregularity are calculated, for example, 0.703, 0.84 and 0.5 respectively, and the welding speed and the maximum welding speed are subtracted from 1, for example, 1-0.84=0.16.

[0097] Furthermore, the power decision coefficient, the speed decision coefficient and the irregularity decision coefficient are obtained by subtracting the difference between 1 and the ratio of the welding power and the irregularity to the maximum welding power and the maximum irregularity and the ratio of the welding speed to the maximum welding speed.

[0098] Further, the power decision coefficient, speed decision coefficient and irregularity decision coefficient are respectively multiplied by the integrated weld identification number and integers are retained to obtain the power identification number, speed identification number and irregularity identification number. For example, if the integrated weld identification number is 10, the power identification number, speed identification number and irregularity identification number are respectively: And round up to 8, And round up to 2, And round up to 5.

[0099] Among them, in the weld seam recognition of the welding workpiece image, the greater the welding power, the greater the impact of the welding arc on the image quality, the more accurate the image recognition is required, and the more weld seam recognition paths are required for recognition to improve the recognition accuracy. The higher the welding speed, the more efficient the image recognition is required, and the fewer weld seam recognition paths are required for recognition to improve the recognition timeliness. Therefore, the ratio of 1 minus the welding speed and the maximum welding speed is used as the speed decision coefficient. The greater the irregularity, the more difficult it is to recognize the weld position in the workpiece image, the more accurate the image recognition is required, and the more weld seam recognition paths are required for recognition to improve the recognition accuracy.

[0100] Furthermore, according to the welding power coefficient, welding speed coefficient and irregularity coefficient in the aforementioned content, the power identification quantity, speed identification quantity and irregularity identification quantity are weightedly calculated and rounded to obtain the number of weld identification paths.

[0101] For example, the number of weld identification paths is And round it up to 5.

[0102] Among them, the parameters with greater influence in the current welding have a larger proportion of corresponding identification quantities, which improves the accuracy of the weld identification path quantity decision calculation and its adaptability to the current welding position characteristics.

[0103] In an embodiment of the present application, a number of weld recognition paths are randomly selected in a trained weld recognition channel, for example, 5 weld recognition paths are randomly selected, the currently downsampled workpiece image is input, and the recognition output obtains the recognition weld position information of the number of weld recognition paths, for example, the position coordinates of 5 welds, and the mean of the 5 weld position coordinates is further calculated, for example, the mean of 5 X-axis coordinates and 5 Y-axis coordinates is calculated, as the weld coordinates after mean processing, as the weld position information.

[0104] Tracking welding control is performed based on the weld position information.

[0105] The embodiment of the present application considers the influence of irregularity, welding power and welding speed in irregular weld welding on the weld coordinates of workpiece image recognition, and makes decision configuration of image recognition resources, specifically configuring the number of weld recognition paths, thereby improving the accuracy of weld position recognition and ensuring the timeliness of recognition, and then performing weld recognition that is most suitable for the current welding position, improving the quality of tracking welding control, avoiding deviations, and improving efficiency.

[0106] The tracking welding control method of an irregular weld provided by an embodiment of the present invention has at least the following technical effects:

[0107] In the welding process, the embodiment of the present invention adopts laser projection technology to project the weld trajectory on the workpiece surface and collect the workpiece image. Then, by obtaining the welding power, welding position information and welding speed, and performing irregularity analysis, the curvature and irregularity of the weld can be dynamically evaluated in the real-time welding process. Further, by analyzing the changes in welding power and welding speed, the welding power coefficient and welding speed coefficient are calculated, which provides an important basis for the dynamic adjustment of the welding state during the welding process. According to the irregularity decision, the image downsampling parameters are obtained and the workpiece image is downsampled. This feature can effectively reduce the computational burden of image processing when the welding speed is fast, improve the response speed, and maintain the image recognition accuracy. The data processing after image downsampling can save computing resources, and the downsampling processing according to the irregularity can ensure the recognition quality, which improves the real-time performance of the recognition processing during the welding process. By combining the welding power coefficient and the welding speed coefficient to make image recognition resource decisions and perform weld recognition, the accuracy and real-time performance of weld position recognition are further improved. The present invention provides a tracking welding control method for irregular welds, which significantly improves the accuracy and real-time performance of weld position recognition during the welding process by accurately analyzing and adjusting multiple key factors in the welding process, and effectively solves the problems existing in the prior art of irregular welds being difficult to accurately track, having low recognition accuracy and poor timeliness.

[0108] Embodiment 2, as Figure 2 As shown, based on the same inventive concept as the tracking welding control method for an irregular weld provided in the first embodiment, the embodiment of the present invention further provides a tracking welding control system for an irregular weld, including: a welding information acquisition module 11, during the welding process of the irregular weld, laser projection is performed on the surface of the workpiece, the workpiece image is collected, and the welding power, welding position information, and welding speed are obtained, and irregularity analysis is performed according to the welding position information to obtain the irregularity;

[0109] The welding change analysis module 12 performs adjacent welding power analysis and adjacent welding speed analysis according to the welding power sequence and welding speed sequence in the most recent preset time window to obtain the welding power coefficient and the welding speed coefficient, and calculates the irregularity coefficient according to the irregularity;

[0110] The identification welding control module 13 is used to obtain image downsampling parameters according to the irregularity decision, downsample the workpiece image, make image recognition resource decisions according to the welding power, welding speed and irregularity, combined with the welding power coefficient, welding speed coefficient and irregularity coefficient, identify the downsampled workpiece image, obtain weld position information, and perform tracking welding control.

[0111] Furthermore, the welding information acquisition module 11 is also used for:

[0112] During the welding process of irregular welds, laser projection is performed on the workpiece surface to collect workpiece images;

[0113] Collect the current welding power and welding moving speed;

[0114] A coordinate system is constructed on the workpiece surface to obtain the coordinates of the current welding position as welding position information.

[0115] Furthermore, the welding information acquisition module 11 is also used for:

[0116] According to the preset length interval, the welding position interval is divided with the welding position information as the end point;

[0117] The average curvature of the welding position interval is calculated as the irregularity.

[0118] Furthermore, the welding variation analysis module 12 is also used for:

[0119] Collect welding power and welding speed at multiple time stamps within the latest preset time window to obtain welding power sequence and welding speed sequence;

[0120] According to the welding power sequence and the welding speed sequence, an average welding power and an average welding speed are calculated;

[0121] The maximum welding power and the maximum welding speed are obtained, and the ratios of the average welding power and the average welding speed to the maximum welding power and the maximum welding speed are calculated as the welding power coefficient and the welding speed coefficient respectively;

[0122] The maximum irregularity is obtained, and the ratio of the irregularity to the maximum irregularity is calculated as the irregularity coefficient.

[0123] Furthermore, the identification welding control module 13 is also used for:

[0124] Obtaining a downsampling parameter interval for downsampling the workpiece image, wherein the size of the downsampling parameter is positively correlated with the size of the image resolution after downsampling;

[0125] Get the maximum irregularity;

[0126] Calculating the ratio of the irregularity to the maximum irregularity, and obtaining the image downsampling parameter by calculation within the downsampling parameter interval;

[0127] The image downsampling parameters are used to perform downsampling processing on the workpiece image.

[0128] Furthermore, the identification welding control module 13 is also used for:

[0129] According to the welding tracking data in the historical time, the weld identification path of the integrated weld identification quantity is trained to obtain the weld identification channel;

[0130] According to the welding power, welding speed and irregularity, combined with the welding power coefficient, welding speed coefficient and irregularity coefficient, the number of identification paths of the integrated weld identification quantity is calculated and determined to obtain the number of weld identification paths;

[0131] Randomly selecting the weld recognition paths of the number of weld recognition paths in the weld recognition channel, inputting the downsampled workpiece image, identifying and obtaining the weld position information of the number of weld recognition paths, and calculating the mean to obtain the weld position information;

[0132] Tracking welding control is performed according to the weld position information.

[0133] Among them, according to the welding tracking data in the historical time, the training of the weld recognition path integrating the number of weld recognition is carried out, including:

[0134] According to the welding tracking data in the historical time, a set of sample workpiece images is collected, and the weld position information in each sample workpiece image is identified and marked to obtain a set of sample weld position information;

[0135] The sample workpiece image set and the sample weld position information set are randomly divided multiple times to obtain weld position recognition training data with an integrated weld recognition quantity;

[0136] The weld position recognition training data of the integrated weld recognition quantity is used to train the weld recognition path of the integrated weld recognition quantity to obtain the weld recognition channel.

[0137] According to the welding power, welding speed and irregularity, combined with the welding power coefficient, welding speed coefficient and irregularity coefficient, the number of identification paths of the integrated weld identification number is calculated and determined to obtain the number of weld identification paths, including:

[0138] Get the maximum welding power, maximum welding speed, and maximum irregularity;

[0139] Calculate the ratio of the welding power and irregularity to the maximum welding power and the maximum irregularity, and subtract the ratio of the welding speed to the maximum welding speed from 1 to obtain a power decision coefficient, a speed decision coefficient, and an irregularity decision coefficient, and multiply them by the integrated weld identification number and retain integers to obtain a power identification number, a speed identification number, and an irregularity identification number;

[0140] According to the welding power coefficient, welding speed coefficient and irregularity coefficient, the power identification quantity, speed identification quantity and irregularity identification quantity are weightedly calculated to obtain the weld identification path quantity.

[0141] Embodiment 3: This embodiment provides a tracking welding control device for irregular welds, which includes a tracking welding control system for irregular welds in Embodiment 2. The device may also include any automatic welding parts and devices in the prior art.

[0142] Although preferred embodiments of the present invention have been described, additional changes and modifications may occur to these embodiments once those skilled in the art understand the basic inventive concepts.

[0143] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention belong to the scope of the present invention and its equivalent technologies, the present invention is also intended to include these changes and variations.

Claims

1. A tracking welding control method for irregular welds, characterized in that: The method comprises: During the welding process of irregular welds, laser projection is performed on the workpiece surface to collect workpiece images, and obtain welding power, welding position information, and welding speed. Irregularity analysis is performed based on the welding position information to obtain irregularity, including: According to the preset length interval, the welding position interval is divided with the welding position information as the end point; Calculating the average curvature of the welding position interval as the irregularity; According to the welding power sequence and welding speed sequence in the most recent preset time window, adjacent welding power analysis and adjacent welding speed analysis are performed to obtain welding power coefficient and welding speed coefficient, and according to the irregularity, the irregularity coefficient is calculated, including: Collect welding power and welding speed at multiple time stamps within the latest preset time window to obtain welding power sequence and welding speed sequence; According to the welding power sequence and the welding speed sequence, an average welding power and an average welding speed are calculated; The maximum welding power and the maximum welding speed are obtained, and the ratios of the average welding power and the average welding speed to the maximum welding power and the maximum welding speed are calculated as the welding power coefficient and the welding speed coefficient respectively; Obtaining a maximum irregularity, and calculating a ratio of the irregularity to the maximum irregularity as an irregularity coefficient; The image downsampling parameters are obtained according to the irregularity decision, the workpiece image is downsampled, and the image recognition resource decision is made according to the welding power, welding speed and irregularity, in combination with the welding power coefficient, welding speed coefficient and irregularity coefficient, the downsampled workpiece image is recognized, the weld position information is obtained, and tracking welding control is performed, wherein the image downsampling parameters are obtained according to the irregularity decision, including: Obtaining a downsampling parameter interval for downsampling the workpiece image, wherein the size of the downsampling parameter is positively correlated with the size of the image resolution after downsampling; Get the maximum irregularity; Calculating the ratio of the irregularity to the maximum irregularity, and obtaining the image downsampling parameter by calculation within the downsampling parameter interval; The image downsampling parameters are used to perform downsampling processing on the workpiece image.

2. The tracking welding control method of an irregular weld according to claim 1, characterized in that: During the welding process of irregular welds, laser projection is performed on the workpiece surface to collect workpiece images and obtain welding power, welding position information, and welding speed, including: During the welding process of irregular welds, laser projection is performed on the workpiece surface to collect workpiece images; Collect the current welding power and welding moving speed; A coordinate system is constructed on the workpiece surface to obtain the coordinates of the current welding position as welding position information.

3. The tracking welding control method of an irregular weld according to claim 1, characterized in that: According to the welding power, welding speed and irregularity, combined with the welding power coefficient, welding speed coefficient and irregularity coefficient, image recognition resource decision is made, the downsampled workpiece image is recognized, and weld position information is obtained, including: According to the welding tracking data in the historical time, the weld identification path of the integrated weld identification quantity is trained to obtain the weld identification channel; According to the welding power, welding speed and irregularity, combined with the welding power coefficient, welding speed coefficient and irregularity coefficient, the number of identification paths of the integrated weld identification quantity is calculated and determined to obtain the number of weld identification paths; Randomly select the weld recognition path of the number of weld recognition paths in the weld recognition channel, input the downsampled workpiece image, identify and obtain the recognition weld position information of the number of weld recognition paths, calculate the mean of the recognition weld position information of the number of weld recognition paths to obtain the weld position information, wherein the recognition weld position information includes the weld position coordinates; Tracking welding control is performed according to the weld position information.

4. The tracking welding control method for irregular welds according to claim 3 is characterized in that: Based on the welding tracking data in the historical time, the weld recognition path training of the integrated weld recognition quantity is carried out, including: According to the welding tracking data in the historical time, a set of sample workpiece images is collected, and the weld position information in each sample workpiece image is identified and marked to obtain a set of sample weld position information; The sample workpiece image set and the sample weld position information set are randomly divided multiple times to obtain weld position recognition training data with an integrated weld recognition quantity; The weld position recognition training data of the integrated weld recognition quantity is used to train the weld recognition path of the integrated weld recognition quantity to obtain the weld recognition channel.

5. The tracking welding control method for irregular welds according to claim 3 is characterized in that: According to the welding power, welding speed and irregularity, combined with the welding power coefficient, welding speed coefficient and irregularity coefficient, the number of identification paths of the integrated weld identification quantity is calculated and determined to obtain the number of weld identification paths, including: Get the maximum welding power, maximum welding speed, and maximum irregularity; Calculate the ratio of the welding power and irregularity to the maximum welding power and the maximum irregularity, and subtract the ratio of the welding speed to the maximum welding speed from 1 to obtain a power decision coefficient, a speed decision coefficient, and an irregularity decision coefficient, and multiply them by the integrated weld identification number and retain integers to obtain a power identification number, a speed identification number, and an irregularity identification number; According to the welding power coefficient, welding speed coefficient and irregularity coefficient, the power identification quantity, speed identification quantity and irregularity identification quantity are weightedly calculated to obtain the weld identification path quantity.

6. A tracking welding control system for irregular welds, characterized in that: The steps for implementing the tracking welding control method of an irregular weld as described in any one of claims 1 to 5 include: The welding information acquisition module performs laser projection on the workpiece surface during the welding process of the irregular weld, collects the workpiece image, and obtains the welding power, welding position information, and welding speed, and performs irregularity analysis based on the welding position information to obtain the irregularity; The welding change analysis module performs adjacent welding power analysis and adjacent welding speed analysis according to the welding power sequence and welding speed sequence in the most recent preset time window to obtain the welding power coefficient and the welding speed coefficient, and calculates the irregularity coefficient according to the irregularity; The identification welding control module is used to obtain image downsampling parameters according to the irregularity decision, downsample the workpiece image, make image recognition resource decisions according to the welding power, welding speed and irregularity, combined with the welding power coefficient, welding speed coefficient and irregularity coefficient, identify the downsampled workpiece image, obtain weld position information, and perform tracking welding control.

7. A tracking welding control device for irregular welds, characterized in that: The device includes a tracking welding control system for irregular welds as described in claim 6.

Images