Welding Method for Hot End Connecting Pipe Assembly Based on Machine Vision

Abstract

The present invention discloses a welding method for a hot end connecting pipe assembly based on machine vision, which relates to the technical field of pipeline welding. It includes step S1 of docking and assembling a plurality of unit pipes into a hot end connecting pipe assembly; step S2 of acquiring a first picture and adjusting the docking deviation of the unit pipes according to the first picture; step S3 of acquiring a second picture and determining the welding position of the sensor nuts on the unit pipes according to the second picture; step S4 of welding the unit pipes and welding and fixing the sensor nuts on the corresponding unit pipes according to the nut welding positions. The welding method for the hot end connecting pipe assembly based on machine vision of the present invention is used to guide the welding and forming of the hot end connecting pipe assembly, so as to improve the automation degree of the welding and forming process of the hot end connecting pipe assembly and improve the forming quality of the hot end connecting pipe assembly.

Description

Technical Field

[0001] The present invention relates to the technical field of pipeline welding, and particularly relates to a welding method for a hot-end connecting pipe assembly based on machine vision. Background Art

[0002] The exhaust system is one of the key systems of the vehicle powertrain. It is responsible for discharging the combustion exhaust gas during the engine operation into the atmosphere, and plays a very crucial role in purifying the engine exhaust gas, absorbing the engine vibration, reducing noise, etc. A variety of sensors are welded and installed on the hot-end connecting pipe of the exhaust system. The various sensors are used to detect the exhaust system and the exhaust gas condition. Therefore, the installation quality of the sensors is of great significance for improving the vehicle power performance, economy, reducing air pollution, ride comfort, etc.

[0003] The main body of the hot-end connecting pipe is a special-shaped metal pipe, which includes several short pipes welded together continuously. Most of these short pipes are short straight pipes, and a small part are arc-shaped short bent pipes, and the orientations of adjacent welded short pipes are also different. In the prior art, the method of welding these short pipes together to form the complete main body of the hot-end connecting pipe is mostly manual welding. However, due to the requirement of orientation when the short pipes are welded together, it is difficult to perform the butt joint positioning of adjacent short pipes under manual control before welding. The accuracy of the butt joint positioning of adjacent short pipes before welding is poor, which seriously reduces the forming quality of the hot-end connecting pipe.

[0004] In addition, after the main body part of the hot-end connecting pipe is welded, technicians also need to weld sensor nuts on the hot-end connecting pipe to facilitate the installation of oxygen sensors, so as to detect the exhaust gas emission quality through the sensors. In the prior art, the process of welding the sensor nuts is implemented separately from the welding process of the hot-end connecting pipe, and the former is also a manual welding and installation method, which seriously reduces the forming speed of the hot-end connecting pipe and also reduces the forming quality of the hot-end connecting pipe. Summary of the Invention

[0005] The technical solution of the present invention to solve the above problems is: to provide a welding method for a hot-end connecting pipe assembly based on machine vision, which is used to weld and form the hot-end connecting pipe assembly, replace the traditional manual welding and assembly method, improve the forming quality of the hot-end connecting pipe assembly, and greatly shorten the construction period to solve the above technical problems.

[0006] The present invention provides a welding method for a hot-end connecting pipe assembly based on machine vision. The hot-end connecting pipe assembly includes a plurality of unit pipes welded together in sequence. The method is characterized by including the following steps:

[0007] Step S1, docking and assembling a plurality of unit pipes into a hot-end connecting pipe assembly;

[0008] Step S2, obtaining a first picture and adjusting the docking deviation of the unit pipes according to the first picture;

[0009] Step S3, obtain the second picture, and determine the welding position of the sensor nut on the unit pipeline according to the second picture;

[0010] Step S4, weld the unit pipeline, and weld and fix the sensor nut on the corresponding unit pipeline according to the nut welding position.

[0011] Furthermore, before step S1, number the multiple unit pipelines according to the unit pipeline assembly and welding sequence, so as to facilitate the robotic arm to automatically grasp the corresponding unit pipelines.

[0012] Furthermore, in step S1, the robotic arm grasps the corresponding unit pipeline according to the number, and the robotic arm docks the unit pipelines it grasps with each other according to the number.

[0013] Furthermore, in the said step S2, the first picture is a visible picture at the docking position of any adjacent and docked two unit pipelines.

[0014] Furthermore, the method for determining the docking deviation of the unit pipeline according to the first picture is: use MATLAB software to extract the contour map of the first picture, calculate the deviation size at the docking position of the unit pipeline according to the contour map, and compare the calculated deviation size with the allowable maximum pipeline docking deviation size to determine whether the unit pipeline docking meets the standard; if the calculated deviation size does not meet the standard, control the robotic arm to correspondingly adjust the unit pipeline docking position until the calculated deviation size meets the standard, and if the calculated deviation size meets the standard, then execute step S3.

[0015] Furthermore, the device corresponding to obtaining the first picture includes multiple movable first CCD cameras, and the shooting direction of each said first CCD camera forms an acute angle less than 45° with the pipeline axis of one of the corresponding docked unit pipelines.

[0016] Furthermore, in step S3, the second picture is an overall structure view when multiple unit pipelines are docked and assembled.

[0017] Furthermore, the method for determining the welding position of the sensor nut according to the second picture is: use MATLAB software to extract the contour map of the second picture, compare it with the overall design drawing of the hot end connection pipe assembly, calculate and determine the unit pipeline corresponding to the welded sensor nut, and further determine the specific welding and installation position of the sensor nut on the corresponding unit pipeline.

[0018] Further, the device for obtaining the second picture includes two groups of second CCD cameras respectively arranged on both sides of the unit pipeline in the docking state. The distance between each group of the second CCD cameras and the manipulator gripping the unit pipeline in the docking state is fixed, and the height position and shooting direction of the second CCD cameras are fixed.

[0019] Further, the welding of the unit pipeline includes successively welding the unit pipelines in butt joint, or simultaneously welding multiple unit pipelines in butt joint; the welding of the sensor nuts includes successively welding the sensor nuts on each unit pipeline, or simultaneously welding the sensor nuts on multiple unit pipelines.

[0020] A welding method for a hot-end connecting pipe assembly based on machine vision according to the present invention has the following specific beneficial effects:

[0021] First, in step S2, when the unit pipelines are welded and connected to assemble the hot-end connecting pipe assembly, the present invention can calculate and judge the docking deviation (which can also be understood as the pipe docking coaxiality) at the docking position of adjacent unit pipelines through machine vision. The docking deviation size calculated by machine vision undoubtedly has higher accuracy compared with visual observation by the naked eye, and there will be no problems such as visual fatigue. Therefore, the docking accuracy between unit pipelines is greatly improved, and the pipe body quality of the hot-end connecting pipe assembly is further enhanced.

[0022] Second, in step S3, the welding position of the sensor nuts can be quickly determined on the unit pipeline or multiple unit pipelines that have been docked through machine vision. The algorithm is simple and the accuracy is also high, without the need for simulation.

[0023] While the position of the sensor nuts is determined by machine vision, the UG software in the computer system can automatically generate the corresponding numerical control program to control the manipulator to grab the sensor nuts to the corresponding pipe welding positions for welding. That is to say, both the method of determining the welding position of the sensor nuts based on machine vision and the method of determining the welding position of the sensor nuts through modeling can ensure a high degree of automation of the entire welding process. The method of determining the welding position of the sensor nuts by machine vision has low difficulty in data acquisition, and the accuracy of the data information obtained through pictures is relatively high.

[0024] Generally speaking, a welding method for a hot-end connecting pipe assembly based on machine vision according to the present invention is used to guide the welding forming of the hot-end connecting pipe assembly, so as to improve the automation degree of the welding forming process of the hot-end connecting pipe assembly and enhance the forming quality of the hot-end connecting pipe assembly. Description of the Drawings

[0025] The accompanying drawings incorporated herein and constituting a part of this specification illustrate embodiments of the present invention and, together with the description, serve to explain the principles of the present invention. In these drawings, like reference numerals are used to denote like elements. The drawings in the following description are some embodiments of the present invention, not all embodiments. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0026] Figure 1 It is the overall process schematic diagram of a specific embodiment of the present invention;

[0027] Figure 2 It is the schematic diagram of the pipeline docking state contour obtained after image processing of the first picture in a specific embodiment of the present invention. Specific Embodiments

[0028] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, rather than all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention. It should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other arbitrarily.

[0029] It should be noted that, in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or sequence between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the element.

[0030] Please refer to Figure 1 , a welding method for a hot-end connecting pipe assembly based on machine vision in a specific embodiment of the present invention. The hot-end connecting pipe assembly includes a plurality of unit pipes welded together in sequence. The method is characterized by including the following steps:

[0031] Step S1, docking and assembling a plurality of unit pipes into a hot-end connecting pipe assembly;

[0032] Step S2, obtain the first picture and adjust the docking deviation of the unit pipelines according to the first picture;

[0033] Step S3, obtain the second picture and determine the welding position of the sensor nuts on the unit pipelines according to the second picture;

[0034] Step S4, weld the unit pipelines and weld and fix the sensor nuts on the corresponding unit pipelines according to the nut welding positions.

[0035] Specifically, in this embodiment, before implementing step S1, technicians can number the multiple unit pipelines that make up the hot end connecting pipe assembly; the technicians can number them according to the sequential docking and assembly order of the multiple unit pipelines, or can also number them according to the position order of each unit pipeline in the hot end connecting pipe assembly. For example, number the multiple unit pipelines that make up the hot end connecting pipe assembly sequentially from left to right. The specific numbers can be numbers such as 1, 2, 3, 4... or other symbols; in addition, if the specific method of welding and assembling the hot end connecting pipe assembly is: weld a unit pipeline as soon as it is docked, the multiple unit pipelines can also be numbered according to the specific welding or docking sequence. The specific numbers can be numbers such as 1, 2, 3, 4... or other symbols.

[0036] The above numbering method can also be executed by a computer.

[0037] Furthermore, in this embodiment, the actuators for grasping the unit pipelines and docking and assembling the unit pipelines preferably use intelligent robots or numerically controlled manipulators.

[0038] In step S1, the computer control center controls the manipulator to grasp the unit pipeline with the corresponding number and dock the corresponding unit pipelines with each other; moreover, the computer control center can also control the manipulator to grasp the welding equipment to weld the docked unit pipelines, and finally weld to form the main pipeline part of the hot end connecting pipe assembly.

[0039] Furthermore, in step S2, the first picture is a visible picture at the docking position of any adjacent and docked two unit pipelines. In this embodiment, the devices for obtaining the first picture are multiple movable first CCD cameras, and the first CCD cameras are grasped by some manipulators. During the process of shooting and obtaining the first picture, the shooting direction of the first CCD camera forms an angle less than 45° with the pipeline axis of one of the corresponding docked unit pipelines.

[0040] The numerical value of the included angle between the shooting direction of the first CCD camera and the axis of the corresponding unit pipeline is obtained by comprehensively calculating based on the included angle between the axis of the corresponding unit pipeline and the manipulator holding the corresponding unit pipeline, the included angle between the shooting direction of the first CCD camera and the manipulator holding the first CCD camera, the included angle between the manipulator holding the corresponding unit pipeline and the manipulator holding the first CCD camera, and their positional relationship. For example, first, ensure that the manipulator holding the first CCD camera and the manipulator holding the corresponding unit pipeline are parallel to each other. Then, control the manipulator holding the first CCD camera to rotate around the unit pipeline by any angle. Next, determine the included angle between the first CCD camera and the manipulator holding the first CCD camera. The difference between the included angle between the axis of the corresponding unit pipeline and the manipulator holding the corresponding unit pipeline and the included angle between the first CCD camera and the manipulator holding the first CCD camera is the included angle between the first CCD camera and the corresponding unit pipeline.

[0041] Please refer to Figure 2 , in this embodiment, after the first CCD camera captures the first picture, it sends the first picture to the computer control center. The computer control center receives the corresponding first picture and extracts the pipeline contour in the first picture through MATLAB software and the corresponding image processing program, and forms a new contour map that can display the pipeline docking status.

[0042] For the convenience of those skilled in the art to understand this solution, in this embodiment, Figure 2 is a simulation schematic diagram of the contour map of the docking position of the unit pipeline, mainly used to describe the calculation principle of the pipeline docking deviation size.

[0043] Please refer to Figure 2 , when using MATLAB software to process the first picture, the wavelet denoising algorithm or the canny edge detection algorithm can be selected to extract the contour line map at the docking position of the unit pipeline in the first picture; when there is a deviation in the docking of the two unit pipelines, the Figure 2 shown shaded area will appear. By calculating the maximum width of this shaded area and converting it through the scale, the actual deviation size of the unit pipeline docking can be calculated, thereby providing a basis for the manipulator to adjust the accuracy of the unit pipeline docking. The computer control center compares the calculated actual deviation size of the unit pipeline docking with the allowable maximum deviation size of the unit pipeline docking to determine whether the unit pipeline docking meets the standard; if the calculated deviation size does not meet the standard, control the manipulator to correspondingly adjust the docking position of the unit pipeline until the calculated deviation size meets the standard. If the calculated deviation size meets the standard, then execute step S3.

[0044] In this embodiment, the first CCD camera can be moved under the grasping of the manipulator to facilitate angle adjustment. However, after the shooting angle of the first CCD camera is adjusted, the manipulator still needs to move the first CCD camera along the established shooting angle to a position at a determined distance from the unit pipeline docking position. This determined distance is the distance from the first CCD camera to the unit pipeline docking position when determining the scale.

[0045] Further, in step S3, the second CCD camera takes pictures of the overall structure of the hot-end connecting pipe assembly in the assembled state at fixed positions on both sides of the multiple unit pipelines in the docking state. The second CCD camera is fixed on both sides of the pipeline docking area, and the distance between the docked unit pipelines remains fixed. The pictures taken by the second CCD camera are the second pictures.

[0046] Specifically, a neural convolutional network model is adopted. A batch of training data is prepared in advance. The training data includes the contour maps of the second pictures of the hot-end connecting pipe assembly at different angles, as well as the target pictures manually marked with the welding positions and specific angle values of the sensor nuts on the contour maps of the second pictures in the early stage. This batch of training data is input into the neural convolutional network model for training to obtain a trained neural convolutional network model.

[0047] The second CCD camera sends the taken second pictures to the computer control center. The computer control center can extract the line contour map of the second pictures through MATLAB software; the processed line contour map of the second pictures is input into the above neural convolutional network model to obtain the target pictures marked with the welding positions of the sensor nuts. The calculation control center controls the manipulator to grasp the corresponding sensor nuts to the corresponding sensor welding positions according to the coordinates of the sensor nut welding positions. [[ID=,10]]

[0048] In this embodiment, when using MATLAB software to extract the contour of the second pictures, the wavelet denoising algorithm or the canny edge detection algorithm can be selected to extract the outer pipe contour lines of each unit pipeline in the second pictures. The device for obtaining the second pictures is specifically two groups of second CCD cameras respectively arranged on both sides of the unit pipelines in the docking state. The distance between each group of the second CCD cameras and the manipulator grasping the unit pipelines in the docking state is fixed, and the height position and shooting direction of the second CCD cameras are fixed.

[0049] Further, in step S4, welding the unit pipelines includes successively welding the docked unit pipelines or simultaneously welding multiple docked unit pipelines; welding the sensor nuts includes successively welding the sensor nuts on each unit pipeline or simultaneously welding the sensor nuts on multiple unit pipelines.

[0050] Specifically, the operation of welding the unit pipelines and the operation of welding and installing the sensor nuts can be carried out simultaneously; the operation of welding the unit pipelines and the operation of welding and installing the sensor nuts can also be carried out alternately; it is also possible to first weld and install the corresponding sensor nuts on the corresponding unit pipelines, then butt the unit pipelines with the welded sensor nuts, and then weld the unit pipelines to form the hot end connection pipe assembly.

[0051] Wherever not mentioned above, the prior art applies.

[0052] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than limiting it. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A welding method for a hot end connecting pipe assembly based on machine vision, wherein the hot end connecting pipe assembly comprises a plurality of unit pipes connected by welding in sequence, characterized in that: The steps include: Step S1, assembling a plurality of unit pipes into a hot end connecting pipe assembly; Step S2: obtaining a first image, and adjusting the unit pipeline docking deviation according to the first image; Step S3, obtaining a second image, and determining the sensor nut welding position on the unit pipe according to the second image; Step S4, welding the unit pipes, and welding and fixing the sensor nuts on the corresponding unit pipes according to the nut welding positions; In step S2, the first picture is a visual picture of the docking position of any group of two adjacent and docked unit pipes; The method for determining the unit pipe docking deviation based on the first image is as follows: using MATLAB software to extract a contour image of the first image, calculating the deviation size at the unit pipe docking position based on the contour image, and comparing the calculated deviation size with the maximum allowable pipe docking deviation size to determine whether the unit pipe docking meets the standard; if the calculated deviation size does not meet the standard, controlling the manipulator to adjust the unit pipe docking position accordingly until the calculated deviation size meets the standard; if the calculated deviation size meets the standard, executing step S3; In step S3, the second picture is an overall structural view of a plurality of unit pipes when they are butt-jointed and assembled; The method for determining the welding position of the sensor nut based on the second image is as follows: using MATLAB software to extract the contour image of the second image, comparing it with the overall design drawing of the hot end connecting pipe assembly, calculating and determining the corresponding welded unit pipe of the sensor nut, and further determining the specific welding installation position of the sensor nut on the corresponding unit pipe; The equipment for obtaining the second picture includes two groups of second CCD cameras respectively arranged on both sides of the unit pipe in the docking state. The distance between each group of second CCD cameras and the corresponding manipulator grasping the unit pipe in the docking state is fixed, and the height position and shooting angle of the second CCD camera are fixed.

2. The welding method of the hot end connecting pipe assembly based on machine vision according to claim 1, characterized in that: Before step S1, a plurality of unit pipes are numbered according to the assembly and welding sequence of the unit pipes, so that a robot arm can automatically grab the corresponding unit pipes.

3. The welding method of the hot end connecting pipe assembly based on machine vision according to claim 2, characterized in that: In step S1 , a numerically controlled robot grasps corresponding unit pipes according to the numbers, and the numerically controlled robot docks the grasped unit pipes with each other according to the numbers.

4. The welding method of a hot end connecting pipe assembly based on machine vision according to claim 1, characterized in that: The device for acquiring the first image includes a plurality of movable first CCD cameras, and the angle between the shooting direction of each first CCD camera and the pipe axis of one of the docking unit pipes being photographed is an acute angle less than 45°.

5. The welding method of a hot end connecting pipe assembly based on machine vision according to claim 1, characterized in that: Welding the unit pipes includes welding the butted unit pipes one by one, or welding a plurality of butted unit pipes at the same time; welding the sensor nuts includes welding the sensor nuts on each unit pipe one by one, or welding the sensor nuts on a plurality of unit pipes at the same time.

Images