Robot teaching-free welding system and welding method based on visual positioning

Abstract

The invention relates to the technical field of robot teaching-free welding, and particularly discloses a robot teaching-free welding system based on visual localization and a welding method.The robot teaching-free welding system comprises a PLC, a power module, a wireless communication module and a visual localization coordinate conversion module. The path planning optimization module directly generates a welding path according to a visual positioning result, tedious operation is reduced, the problem of poor dependence and adaptability on pre-stored data is solved, the visual acquisition module and the positioning module work cooperatively, workpiece position and posture information is updated in real time, and the positioning precision is improved. The teaching-free intelligent decision-making and optimization module is combined with real-time data to adjust a path and parameters, the image processing and path planning optimization module generates a proper path according to image information for a complex workpiece, the sensor fusion module fuses vision and other sensor data, the adaptability of the system to the complex workpiece is enhanced, and the welding quality is guaranteed.

Description

Technical Field

[0001] The invention relates to the technical field of robot non-test teaching welding, in particular to a robot non-test teaching welding system and a welding method based on visual positioning. Background Art

[0002] Robotic teaching-free welding is an advanced welding technology. In the traditional robot welding process, a complex teaching process is usually required, that is, the operator manually guides the robot's end effector (welding tool) to move along the welding path, records the path, speed, welding parameters and other information, and allows the robot to learn the welding trajectory and operation methods. Robotic teaching-free welding is different. It uses advanced sensor technology (such as visual sensors, laser sensors, etc.) and intelligent control algorithms. These sensors can scan the welding workpiece to obtain key information such as the shape, size, and weld position of the workpiece, and then transmit the data to the control system. The control system automatically generates welding paths and parameters based on this information, and directs the robot to directly perform high-quality welding operations, which greatly improves welding efficiency and reduces the time and workload of manual teaching. It can also better adapt to the welding needs of workpieces of various shapes and specifications, showing huge advantages in the field of welding in modern manufacturing.

[0003] In the related technology, the teaching-free welding method uses the indirect electrical positioning technology of the welding equipment itself to determine the actual position of the workpiece to be welded on the workstation, and uses the position of the workpiece to be welded pre-stored in the welding equipment to determine the deviation between the two, and adjusts the welding path in the welding equipment to determine the actual welding path corresponding to the actual position. It does not require expensive and complex path correction equipment, nor does it require teaching of the welding equipment before welding, which greatly improves welding adaptability and work efficiency.

[0004] First, its positioning accuracy is susceptible to interference. Indirect electrical positioning relies on contact signals to determine path data, but this process is easily interfered by many factors. For example, the robot's own mechanical motion accuracy will affect positioning. If the mechanical motion is not accurate enough, it may cause deviations in the position recording of the contact signal. At the same time, the sensitivity of the contact signal trigger is also an important factor. Unstable sensitivity may cause errors in the recorded contact signal position. Moreover, the errors generated by each contact will accumulate. As the positioning process proceeds, the errors continue to accumulate, eventually leading to a decrease in the accuracy of the positioning path data, which in turn has an adverse effect on the welding quality. Secondly, the indirect electrical positioning efficiency is low. This positioning method has cumbersome steps and requires linear movement and signal recording from multiple position points. This process takes a long time, especially on complex-shaped workpieces. Since more position points are required to determine the positioning path, the efficiency is lower. For example, for irregularly shaped workpieces with multiple curved surfaces or corners, contact signals need to be recorded at more positions, which greatly increases the time cost of positioning.

[0005] Furthermore, indirect electrical positioning is highly dependent on pre-stored data and has poor actual adaptability. Determining the actual welding path data depends to a large extent on pre-stored total path data, and the pre-stored data is obtained by scanning and marking the welding path through a laser radar. However, in actual production, the workpiece position and posture may deviate from the reference position due to errors such as clamping and processing. If the pre-stored data cannot accurately reflect the actual workpiece situation, the welding path data will have large errors. In addition, for workpieces with complex shapes, multi-layer structures or complex cavities, this method has poor adaptability. For example, when welding multi-layer workpieces, the indirect electrical positioning to determine the welding information of the lower structure may be interfered by the upper structure, making the positioning inaccurate, and the welding edges of complex shapes are also difficult to accurately position, resulting in difficulty in ensuring the welding quality. For example, for workpieces with complex curved surfaces and internal cavities, indirect electrical positioning may not be able to accurately find a suitable welding path, thereby affecting the welding effect and quality. Summary of the invention

[0006] The purpose of the present invention is to provide a robot teaching-free welding system and welding method based on visual positioning, so as to solve the problems in the prior art raised in the above background technology that the positioning accuracy is easily disturbed and the efficiency is low, and it is highly dependent on pre-stored data and has poor practical adaptability.

[0007] To achieve the above-mentioned purpose, the present invention provides a robot teaching-free welding system and welding method based on visual positioning, comprising a PLC controller, a power module, a wireless communication module, a visual positioning coordinate conversion module, an image processing module, a motion control module, a visual acquisition module, a path planning optimization module, an image processing module and a comprehensive management module, wherein the power module is electrically connected to the PLC controller, and the PLC controller is connected to the visual positioning coordinate conversion module, the image processing module, the motion control module, the visual acquisition module, the path planning optimization module, the image processing module and the comprehensive management module through the wireless communication module;

[0008] The visual positioning coordinate conversion module is composed of an image feature extraction module, a feature matching module, a three-dimensional coordinate calculation module and a coordinate conversion matrix calculation module. The image feature extraction module, the feature matching module, the three-dimensional coordinate calculation module and the coordinate conversion matrix calculation module are all connected through a wireless communication module and a PLC controller signal.

[0009] Furthermore, the visual positioning coordinate conversion module also includes a real-time positioning update module and a positioning accuracy assessment module, and the real-time positioning update module and the positioning accuracy assessment module are connected via a wireless communication module and a PLC controller signal.

[0010] Furthermore, the motion control module is composed of a motion instruction receiving module, a joint motion control module, a motion trajectory tracking module and a motion speed adjustment module, and the motion instruction receiving module, the joint motion control module, the motion trajectory tracking module and the motion speed adjustment module are all connected through the wireless communication module and the PLC controller signal.

[0011] Furthermore, the motion control module also includes a motion stability control module and a motion safety protection module, and the motion stability control module and the motion safety protection module are both connected via a wireless communication module and a PLC controller signal.

[0012] Furthermore, the path planning optimization module is composed of a path planning optimization module, a welding path generation module, a path parameter setting module, a path smoothing processing module and a path optimization algorithm module. The path planning optimization module, the welding path generation module, the path parameter setting module, the path smoothing processing module and the path optimization algorithm module are all connected through the wireless communication module and the PLC controller signal.

[0013] Furthermore, the path planning optimization module also includes a path adaptability adjustment module and a path storage calling module, and the path adaptability adjustment module and the path storage calling module are connected via a wireless communication module and a PLC controller signal.

[0014] Furthermore, the image processing module is composed of an image acquisition preprocessing module, an image enhancement module, a feature extraction algorithm module, an image segmentation module, an image analysis and recognition module, and an image post-processing module. The image acquisition preprocessing module, the image enhancement module, the feature extraction algorithm module, the image segmentation module, the image analysis and recognition module, and the image post-processing module are all connected through a wireless communication module and a PLC controller signal.

[0015] Furthermore, the visual acquisition module is composed of a camera selection and configuration module, a lighting system design module, an image acquisition control module, an image transmission and storage module, an image quality monitoring module and a system calibration and maintenance module. The camera selection and configuration module, the lighting system design module, the image acquisition control module, the image transmission and storage module, the image quality monitoring module and the system calibration and maintenance module are connected through a wireless communication module and a PLC controller signal.

[0016] Furthermore, the comprehensive management module is composed of a welding execution module, a system control management module, a sensor fusion module, a calibration and calibration module, a teaching-free intelligent decision-making and optimization module and a safety protection module. The welding execution module, the system control management module, the sensor fusion module, the calibration and calibration module, the teaching-free intelligent decision-making and optimization module and the safety protection module are connected through the wireless communication module and the PLC controller signal.

[0017] A robot teaching-free welding method based on visual positioning comprises the following steps:

[0018] Step 1: Preparation, system initialization and parameter setting, select an industrial camera with a resolution of 25 million pixels for visual acquisition, configure a suitable lens focal length of 70mm-90mm according to the size of the welding workpiece and the working distance, design the lighting system so that its light intensity reaches 1200Lux-1800Lux, and use a multi-angle LED light source with an angle set at 40°-50° to ensure uniform lighting on the workpiece surface and reduce shadows and reflections;

[0019] Step 2: Image acquisition and processing. Start the image acquisition control module and set the acquisition frequency to 70-90 frames per second. The image acquisition preprocessing module grayscales the acquired image and uses the median filter algorithm with a filter window size of 3x3-5x5 ​​to remove noise. The image enhancement module adjusts the image contrast to 1.6-2.0 times to highlight the features of the welding area. The feature extraction algorithm module uses the Canny edge detection algorithm and sets the low threshold to 70-90 and the high threshold to 160-200 to extract the edge features of the welding area. The image segmentation module accurately segments the welding area from the background based on color and texture features.

[0020] Step 3: Visual positioning and coordinate conversion. The image feature extraction module extracts key feature points, such as the start and end points and corner points of the weld. The feature matching module matches the extracted features with the pre-stored standard features, requiring the matching accuracy to reach 92%-98% or more. The three-dimensional coordinate calculation module uses binocular vision technology to calculate the coordinates of the workpiece in three-dimensional space. The coordinate accuracy is controlled within ±0.06mm-±0.1mm. The coordinate transformation matrix calculation module accurately calculates the transformation matrix between the image coordinate system and the robot coordinate system, and the conversion accuracy is within ±0.03mm-±0.05mm. The real-time positioning update module continuously tracks the position changes of the workpiece at a frequency of 20Hz-30Hz. The positioning accuracy evaluation module regularly evaluates the positioning accuracy, and the error range is kept within ±0.12mm-±0.18mm.

[0021] Step 4: Path planning optimization. The welding path generation module automatically generates the welding path according to the visual positioning results and welding process requirements. The path curvature radius is not less than 70mm-90mm. The path parameter setting module sets the welding speed at 7mm / s-9mm / s, and accurately sets the arc starting and arc ending positions. The path optimization algorithm module uses a genetic algorithm, with the goal of reducing the path length by 12%-18% and the welding time by 18%-22%. The path smoothing module controls the path curvature change rate within 0.7 degrees / mm-0.9 degrees / mm. The path adaptability adjustment module can adjust the path in real time within 0.7s-1.0s according to the situation during the actual welding process. The path storage and call module stores the optimized path for next use.

[0022] Step 5: Motion control and welding execution, the motion command receiving module quickly receives the command from the path planning module, and the command transmission delay is controlled within 70ms-90ms. The joint motion control module accurately controls the robot joint motion, and the joint motion accuracy is within ±0.05mm-±0.08mm. The motion trajectory tracking module tracks the robot motion trajectory in real time, and the deviation from the planned trajectory is controlled within ±0.10mm-±0.15mm. The motion speed adjustment module adjusts the robot motion speed within the range of 4mm / s-12mm / s according to the actual situation. The motion stability control module takes measures to ensure the stability of the robot motion, and the vibration amplitude is controlled within ±0.05mm-±0.08mm. The motion safety protection module sets limit switches and emergency stop buttons, etc. When an abnormality occurs, the robot motion is stopped within 70ms-90ms. The welding execution module controls the welding power output, the welding current is 100A-150A, and the voltage is 22V-28V;

[0023] Step 6: Comprehensive management and communication. The system control management module in the comprehensive management module manages the entire system in a coordinated manner. The task scheduling response time is within 1.2s-1.8s. The sensor fusion module integrates the data of visual sensors, force sensors, laser sensors, etc. The measurement accuracy of force sensors is within ±0.6N-±1.0N, and the measurement accuracy of laser sensors is within ±0.05mm-±0.08mm. The calibration and calibration module regularly calibrates and calibrates the system. The calibration cycle is 1.5 months-2.5 months. The teaching-free intelligent decision-making and optimization module automatically adjusts the welding process parameters within 1.2s-1.8s according to the real-time situation. The safety protection module ensures the safe operation of the system, and the safety light curtain detection accuracy is set within ±3mm-±5mm. The wireless communication module realizes wireless communication between the system and external devices. The communication distance is 70m-90m and the data transmission rate is 25Mbps-35Mbps, which is convenient for operators to remotely monitor and adjust system parameters.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] First, in the present invention, the robot teaching-free welding system based on visual positioning adopts a visual positioning coordinate conversion module to accurately determine the three-dimensional position and posture of the workpiece through image feature extraction, matching and three-dimensional coordinate calculation, which is not affected by mechanical motion accuracy and contact signal error. In the welding of complex-shaped workpieces, the weld position and direction can be accurately identified, and the real-time positioning update module can track the change of the workpiece position to ensure welding accuracy. The path planning optimization module automatically generates and optimizes the welding path according to the visual positioning result. The path smoothing processing module reduces the vibration impact of the robot movement and prolongs the service life. The adaptive adjustment module can adjust the path in real time to enhance adaptability and flexibility. The image processing module provides accurate image information for visual positioning and path planning. The image analysis and recognition module detects welding defects and adjusts parameters. The post-processing module provides a basis for quality assessment and tracing.

[0026] Secondly, in the present invention, the welding system is highly intelligent and adaptable. The teaching-free intelligent decision-making and optimization module automatically adjusts the process parameters according to the workpiece material, shape and welding requirements to realize intelligent welding, improve production efficiency and quality stability, and combines the sensor fusion module data to make more accurate decisions and optimizations, such as fusing vision and force sensor data to adjust the path and parameters to adapt to different working conditions. The comprehensive management module coordinates the management system and coordinates the work of each module. The operation efficiency and stability are improved through intelligent task scheduling and parameter configuration. The sensor fusion module fuses multiple sensor data to realize all-round monitoring and control, improve the system adaptability and intelligence level, and the system can automatically adjust the welding path and process parameters to adapt to different workpieces and environments, and can also operate stably in harsh environments.

[0027] Third, in the present invention, the system provides multiple safety guarantees through the motion safety protection module of the motion control module, sets limit switches and emergency stop buttons, monitors and controls the robot's motion speed and acceleration in real time, ensures safe operation, reduces vibration and shake, improves welding quality and operational safety, and the system calibration and maintenance module of the visual acquisition module makes calibration and maintenance more convenient. Regular maintenance of the camera and system ensures long-term stable operation and accuracy. The calibration and calibration module accurately calibrates and calibrates the system to ensure coordinated work of each module and accurate data, improves reliability and reduces maintenance costs. The wireless communication module enhances convenience, and operators can remotely monitor and adjust parameters, improve work efficiency and operational flexibility, and facilitate communication and integration with other equipment to realize intelligent production.

[0028] Fourthly, in the present invention, the system uses a visual acquisition and processing module to replace the indirect electrical positioning that is susceptible to interference, thereby improving the positioning efficiency and accuracy. The path planning optimization module directly generates a welding path based on the visual positioning results, reducing tedious operations. For the problems of dependence on pre-stored data and poor adaptability, the visual acquisition and positioning modules work together to update the workpiece position and posture information in real time. The teaching-free intelligent decision-making and optimization module adjusts the path and parameters based on real-time data. For complex workpieces, the image processing and path planning optimization module generates a suitable path based on image information. The sensor fusion module fuses vision and other sensor data to enhance the system's adaptability to complex workpieces and ensure welding quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is the overall system module diagram of the present invention;

[0030] Figure 2 This is a schematic diagram of the motion control module in the present invention;

[0031] Figure 3 This is a schematic diagram of the path planning optimization module in the present invention;

[0032] Figure 4 This is a schematic diagram of the visual positioning coordinate conversion module in the present invention;

[0033] Figure 5 This is a schematic diagram of the image processing module in the present invention;

[0034] Figure 6 This is a schematic diagram of the visual acquisition module in the present invention;

[0035] Figure 7 It is a schematic diagram of the comprehensive management module in the present invention. DETAILED DESCRIPTION

[0036] Example

[0037] See also Figure 1-Figure 7In an embodiment of the present invention, a robot teaching-free welding system based on visual positioning includes a PLC controller, a power module, a wireless communication module, a visual positioning coordinate conversion module, an image processing module, a motion control module, a visual acquisition module, a path planning optimization module, an image processing module and a comprehensive management module. The power module is electrically connected to the PLC controller, and the PLC controller is connected to the visual positioning coordinate conversion module, the image processing module, the motion control module, the visual acquisition module, the path planning optimization module, the image processing module and the comprehensive management module through the wireless communication module.

[0038] The visual positioning coordinate conversion module is composed of an image feature extraction module, a feature matching module, a three-dimensional coordinate calculation module and a coordinate conversion matrix calculation module. The image feature extraction module, the feature matching module, the three-dimensional coordinate calculation module and the coordinate conversion matrix calculation module are all connected through the wireless communication module and the PLC controller signal. The system is composed of a PLC controller, a power supply module, a wireless communication module, a visual positioning coordinate conversion module, an image processing module, a motion control module, a visual acquisition module, a path planning optimization module, an image processing module and a comprehensive management module. The power supply module provides stable power for the system. The PLC controller realizes signal connection with each module through the wireless communication module to ensure efficient collaborative work. The visual positioning coordinate conversion module is composed of an image feature extraction module, a feature matching module, a three-dimensional coordinate calculation module and a coordinate conversion matrix calculation module. The image feature extraction module can accurately extract key features from the visually acquired image, and the feature matching module performs Efficient feature matching, the three-dimensional coordinate calculation module can accurately calculate the three-dimensional coordinates of the object, and the coordinate conversion matrix calculation module further ensures the accurate conversion of coordinates. The collaborative work of these modules enables the system to achieve high-precision positioning during the welding process. The image processing module of the system can perform fine processing on the collected images to improve the image quality and the accuracy of feature recognition. The motion control module can accurately control the movement of the robot according to instructions to ensure the accuracy and stability of welding. The path planning optimization module can optimize the welding path according to actual conditions to improve welding efficiency and quality. The comprehensive management module comprehensively manages the entire system to ensure the coordinated operation of each module. In short, this robot teaching-free welding system based on visual positioning has the advantages of high-precision positioning, efficient image processing, precise motion control, optimized path planning and comprehensive integrated management. It can improve the efficiency and quality of welding and bring great convenience to the welding work in the modern manufacturing industry.

[0039] See also Figure 1 and Figure 4The visual positioning coordinate conversion module also includes a real-time positioning update module and a positioning accuracy evaluation module. The real-time positioning update module and the positioning accuracy evaluation module are connected to the PLC controller signal through the wireless communication module. The motion control module is composed of a motion instruction receiving module, a joint motion control module, a motion trajectory tracking module and a motion speed adjustment module. The motion instruction receiving module, the joint motion control module, the motion trajectory tracking module and the motion speed adjustment module are all connected to the PLC controller signal through the wireless communication module. The motion control module also includes a motion stability control module and a motion safety protection module. The motion stability control module and the motion safety protection module are all connected to the PLC controller signal through the wireless communication module. The visual positioning coordinate conversion module also includes a real-time positioning update module and a positioning accuracy evaluation module. The two are connected to the PLC controller signal through the wireless communication module. The real-time positioning update module can ensure that the system can update the positioning information according to the actual situation at any time during the welding process to ensure the accuracy and timeliness of positioning. The positioning accuracy evaluation module can evaluate the positioning accuracy and provide a basis for system adjustment. The motion control module is composed of a motion command receiving module, a joint motion control module, a motion trajectory tracking module, a motion speed adjustment module, a motion stability control module and a motion safety protection module. These modules are connected through the wireless communication module and the PLC controller signal. The motion command receiving module receives instructions, the joint motion control module controls the robot joint movement, the motion trajectory tracking module ensures that the robot moves according to the planned path, the motion speed adjustment module can adjust the motion speed according to actual needs, the motion stability control module ensures the stability of the robot movement, and the motion safety protection module provides safety protection for the system to prevent accidents.

[0040] See also Figure 1 and Figure 3The path planning optimization module is composed of a path planning optimization module, a welding path generation module, a path parameter setting module, a path smoothing processing module and a path optimization algorithm module. The path planning optimization module, the welding path generation module, the path parameter setting module, the path smoothing processing module and the path optimization algorithm module are all connected through the wireless communication module and the PLC controller signal. The path planning optimization module also includes a path adaptability adjustment module and a path storage calling module. The path adaptability adjustment module and the path storage calling module are connected through the wireless communication module and the PLC controller signal. The path planning optimization module is composed of a path planning optimization module, a welding path generation module, a path parameter setting module, a path smoothing processing module, The system is composed of a path optimization algorithm module, a path adaptability adjustment module and a path storage and calling module. These modules are connected through wireless communication modules and PLC controller signals. The welding path generation module can generate a welding path according to the input information of the system. The path parameter setting module can set various parameters of the path. The path smoothing processing module makes the generated path smoother and reduces vibration during robot movement. The path optimization algorithm module optimizes the path through a specific algorithm to improve welding efficiency. The path adaptability adjustment module can adaptively adjust the path according to actual conditions to meet different welding requirements. The path storage and calling module can store the generated path to facilitate the next call and improve work efficiency.

[0041] See also Figure 1-Figure 5The image processing module is composed of an image acquisition preprocessing module, an image enhancement module, a feature extraction algorithm module, an image segmentation module, an image analysis and recognition module, and an image post-processing module. The image acquisition preprocessing module, the image enhancement module, the feature extraction algorithm module, the image segmentation module, the image analysis and recognition module, and the image post-processing module are all connected to the PLC controller signal through the wireless communication module. The visual acquisition module is composed of a camera selection and configuration module, a lighting system design module, an image acquisition control module, an image transmission and storage module, an image quality monitoring module, and a system calibration and maintenance module. The camera selection and configuration module, the lighting system design module, the image acquisition control module, the image transmission and storage module, the image quality monitoring module, and the system calibration and maintenance module are connected to the PLC controller signal through the wireless communication module. The image processing module is composed of an image acquisition preprocessing module, an image enhancement module, a feature extraction algorithm module, an image segmentation module, an image analysis and recognition module, and an image post-processing module. These modules are all connected to the PLC controller signal through the wireless communication module. The line communication module is connected to the PLC controller signal, the image acquisition preprocessing module performs preliminary processing on the acquired images to prepare for subsequent processing, the image enhancement module improves the image quality, the feature extraction algorithm module extracts image features, the image segmentation module segments the image, the image analysis and recognition module analyzes and recognizes the image, and the image post-processing module further optimizes the processed image. The visual acquisition module is composed of a camera selection and configuration module, a lighting system design module, an image acquisition control module, an image transmission and storage module, an image quality monitoring module and a system calibration and maintenance module. These modules are connected to the PLC controller signal through the wireless communication module. The camera selection and configuration module selects a suitable camera according to the requirements, the lighting system design module designs a suitable lighting system to ensure image quality, the image acquisition control module controls the image acquisition process, the image transmission and storage module is responsible for the transmission and storage of the image, the image quality monitoring module monitors the image quality, and the system calibration and maintenance module calibrates and maintains the system.

[0042] See also Figure 7The comprehensive management module is composed of a welding execution module, a system control management module, a sensor fusion module, a calibration and calibration module, a teaching-free intelligent decision-making and optimization module, and a safety protection module. The welding execution module, the system control management module, the sensor fusion module, the calibration and calibration module, the teaching-free intelligent decision-making and optimization module, and the safety protection module are connected through the wireless communication module and the PLC controller signal. The welding execution module ensures the accurate performance of the welding operation and can efficiently complete the welding task. The system control management module comprehensively controls the entire system to ensure the coordinated operation of each module and improve the overall efficiency of the system. The sensor fusion module integrates different sensor data so that the system can obtain more comprehensive and accurate information and provide a more reliable basis for the welding process. The calibration and calibration module ensures the accuracy and stability of the system, making the welding results more reliable. The teaching-free intelligent decision-making and optimization module can make intelligent decisions and optimizations according to actual conditions to improve the quality and efficiency of welding. The safety protection module provides the system with all-round security protection, reduces the risks during system operation, and ensures the safety of personnel and equipment.

[0043] A robot teaching-free welding method based on visual positioning comprises the following steps:

[0044] Step 1: Preparation, system initialization and parameter setting, select an industrial camera with a resolution of 25 million pixels for visual acquisition, configure a suitable lens focal length of 70mm-90mm according to the size of the welding workpiece and the working distance, design the lighting system so that its light intensity reaches 1200Lux-1800Lux, and use a multi-angle LED light source with an angle set at 40°-50° to ensure uniform lighting on the workpiece surface and reduce shadows and reflections;

[0045] Step 2: Image acquisition and processing. Start the image acquisition control module and set the acquisition frequency to 70-90 frames per second. The image acquisition preprocessing module grayscales the acquired image and uses the median filter algorithm with a filter window size of 3x3-5x5 ​​to remove noise. The image enhancement module adjusts the image contrast to 1.6-2.0 times to highlight the features of the welding area. The feature extraction algorithm module uses the Canny edge detection algorithm and sets the low threshold to 70-90 and the high threshold to 160-200 to extract the edge features of the welding area. The image segmentation module accurately segments the welding area from the background based on color and texture features.

[0046] Step 3: Visual positioning and coordinate conversion. The image feature extraction module extracts key feature points, such as the start and end points and corner points of the weld. The feature matching module matches the extracted features with the pre-stored standard features, requiring the matching accuracy to reach 92%-98% or more. The three-dimensional coordinate calculation module uses binocular vision technology to calculate the coordinates of the workpiece in three-dimensional space. The coordinate accuracy is controlled within ±0.06mm-±0.1mm. The coordinate transformation matrix calculation module accurately calculates the transformation matrix between the image coordinate system and the robot coordinate system, and the conversion accuracy is within ±0.03mm-±0.05mm. The real-time positioning update module continuously tracks the position changes of the workpiece at a frequency of 20Hz-30Hz. The positioning accuracy evaluation module regularly evaluates the positioning accuracy, and the error range is kept within ±0.12mm-±0.18mm.

[0047] Step 4: Path planning optimization. The welding path generation module automatically generates the welding path according to the visual positioning results and welding process requirements. The path curvature radius is not less than 70mm-90mm. The path parameter setting module sets the welding speed at 7mm / s-9mm / s, and accurately sets the arc starting and arc ending positions. The path optimization algorithm module uses a genetic algorithm, with the goal of reducing the path length by 12%-18% and the welding time by 18%-22%. The path smoothing module controls the path curvature change rate within 0.7 degrees / mm-0.9 degrees / mm. The path adaptability adjustment module can adjust the path in real time within 0.7s-1.0s according to the situation during the actual welding process. The path storage and call module stores the optimized path for next use.

[0048] Step 5: Motion control and welding execution, the motion command receiving module quickly receives the command from the path planning module, and the command transmission delay is controlled within 70ms-90ms. The joint motion control module accurately controls the robot joint motion, and the joint motion accuracy is within ±0.05mm-±0.08mm. The motion trajectory tracking module tracks the robot motion trajectory in real time, and the deviation from the planned trajectory is controlled within ±0.10mm-±0.15mm. The motion speed adjustment module adjusts the robot motion speed within the range of 4mm / s-12mm / s according to the actual situation. The motion stability control module takes measures to ensure the stability of the robot motion, and the vibration amplitude is controlled within ±0.05mm-±0.08mm. The motion safety protection module sets limit switches and emergency stop buttons, etc. When an abnormality occurs, the robot motion is stopped within 70ms-90ms. The welding execution module controls the welding power output, the welding current is 100A-150A, and the voltage is 22V-28V;

[0049] Step 6: Comprehensive management and communication. The system control management module in the comprehensive management module manages the entire system in a coordinated manner. The task scheduling response time is within 1.2s-1.8s. The sensor fusion module integrates the data of visual sensors, force sensors, laser sensors, etc. The measurement accuracy of force sensors is within ±0.6N-±1.0N, and the measurement accuracy of laser sensors is within ±0.05mm-±0.08mm. The calibration and calibration module regularly calibrates and calibrates the system. The calibration cycle is 1.5 months-2.5 months. The teaching-free intelligent decision-making and optimization module automatically adjusts the welding process parameters within 1.2s-1.8s according to the real-time situation. The safety protection module ensures the safe operation of the system, and the safety light curtain detection accuracy is set within ±3mm-±5mm. The wireless communication module realizes wireless communication between the system and external devices. The communication distance is 70m-90m and the data transmission rate is 25Mbps-35Mbps, which is convenient for operators to remotely monitor and adjust system parameters.

[0050] Image feature extraction of the visual positioning coordinate transformation module: The SIFT (Scale-Invariant Feature Transform) algorithm can be used. Its main steps include constructing a scale space, detecting key points, and determining the direction of key points.

[0051] The scale space function L(x, y, q) of the key point is obtained by convolving the image I(x, y) with the Gaussian kernel function G(x, y, q): L(x, y, σ) = G(x, y, σ) * I(x, y)

[0052] Where * represents the convolution operation;

[0053]

[0054] Three-dimensional coordinate calculation: Triangulation can be used. Assume that the intrinsic parameter matrix of the camera is K, and the coordinates of the spatial point P in the two camera coordinate systems are: P1 = (X1, Y1, Z1) and P2 = (X2, Y2, Z2);

[0055] The corresponding pixel coordinates are:

[0056] p1 = (u1, v1) and p2 = (u2, v2);

[0057] Visual positioning coordinate conversion module - proportional scaling positioning formula:

[0058] L2 = L1 * (PL2 / PL1);

[0059] W2 = W1*(PW2 / PW1);

[0060] Motion control module-speed adjustment time calculation:

[0061] t=(V2-V1) / a;

[0062] Path planning optimization module-path length estimation formula:

[0063] For each path, the length l = sqrt((x2-x1)^2+(y2-y1)^2), and the total path length L = suml;

[0064] Image processing module - average gray value calculation:

[0065] Gavg=sum(Gi) / n(i=1,2,...,n);

[0066] Visual acquisition module-exposure time and brightness relationship:

[0067] T2 = T1*(B2 / B1).

[0068] Motion command receiving module: Siemens S7-1200 series PLC (for receiving commands), joint motion control module: Beckhoff AX5000 series servo drive, motion trajectory tracking module: Renishaw OMP400 laser interferometer (for detecting tracking accuracy), motion speed adjustment module: Schneider ATV320 inverter, motion stability control module: Rexroth MSK series motor (can assist stability control), motion safety protection module: Omron G9SX-NS202-DC24 safety relay.

[0069] Welding path generation module: ABB RobotStudio software (for path generation part functions), path parameter setting module: Advantech IPC-610 industrial computer can be used for parameter setting operations, path optimization algorithm module: MATLAB software (for running optimization algorithm), path smoothing processing module: SolidWorks software can be used (for simulation and processing of partial path smoothing), path adaptability adjustment module: FANUC robot controller (for path adaptability adjustment), path storage call module: Western Digital WDPurple series hard disk is used to store path data.

[0070] Image feature extraction module: OpenCV software library (for feature extraction code running), feature matching module: Halcon machine vision software (for feature matching function), three-dimensional coordinate calculation module: Hexagon MetrologX4 software (for coordinate calculation related), coordinate transformation matrix calculation module: NVIDIA JetsonXavierNX developer kit can be used to assist in calculation, real-time positioning update module: Keyence CV-X400 series vision system, positioning accuracy evaluation module: Mitutoyo SJ-410 surface roughness tester (can be indirectly used for accuracy evaluation and comparison).

[0071] Image acquisition and preprocessing module: BasleracA1300-200um industrial camera (for acquisition), and NIVisionAssistant software (for preprocessing), image enhancement module: AdobePhotoshop (used to simulate some image enhancement algorithm ideas), feature extraction algorithm module: TensorFlow deep learning framework (used to run feature extraction algorithm), image segmentation module: ImageJ software (for image segmentation operations), image analysis and recognition module: PyTorch deep learning framework (for analysis and recognition), image post-processing module: MATLAB software can be used (for post-processing related operations).

[0072] Camera selection and configuration module: Hikvision MV-CA013-21UC industrial camera, lighting system design module: Osram LRTBF30066LED ring light source, image acquisition control module: ADLINK PCI-9820 image acquisition card, image transmission and storage module: SanDisk ExtremePROSD card (for storing transmitted images), image quality monitoring module: X-Rite i1Pro3Plus spectrophotometer (for image quality related monitoring), system calibration and maintenance module: Cognex DM1000-14M40-P vision system calibration tool.

[0073] System control and management module: Siemens S7-1500 series PLC (for system control and management), sensor fusion module: Texas Instruments TMS320F28379D digital signal processor (for sensor data fusion), calibration and calibration module: Miiris optoNCDTILR1181-10 laser displacement sensor (for calibration and calibration), teaching-free intelligent decision-making and optimization module: Intel Core i7-12700H processor (for running decision optimization algorithm), safety protection module: Honeywell SafetyManager safety management system.

[0074] The working principle of the present invention is that the robot teaching-free welding system based on visual positioning performs well in terms of accuracy and efficiency. First, the visual positioning coordinate conversion module uses modules such as image feature extraction, feature matching and three-dimensional coordinate calculation to accurately determine the position and posture of the workpiece in three-dimensional space. Compared with traditional contact positioning, visual positioning is not affected by mechanical motion accuracy and contact signal error. It can accurately identify the position and direction of the weld in the welding of complex-shaped workpieces, greatly improving the positioning accuracy. At the same time, the real-time positioning update module can continuously track the position changes of the workpiece during the welding process to ensure welding accuracy. Secondly, the path planning optimization module operates efficiently. It can automatically generate welding paths based on the visual positioning results and optimize them through path optimization algorithms. The path smoothing processing module makes the welding path smoother, reduces robot movement vibration and impact, and extends the service life. The path adaptability adjustment module can adjust the path in real time according to actual conditions to enhance the system adaptability and flexibility. Furthermore, the image processing module performs preprocessing, enhancement, feature extraction and other operations on the image to provide accurate image information for visual positioning and path planning, further improving the system accuracy. The image analysis and recognition module can analyze the welding process image in real time, detect welding defects and adjust parameters to improve welding quality. The image post-processing module provides a basis for quality assessment and traceability.

[0075] This welding system is highly intelligent and adaptable. The teaching-free intelligent decision-making and optimization module can automatically adjust the welding process parameters according to the workpiece material, shape and welding requirements to achieve intelligent welding without manual teaching, improve production efficiency and welding quality stability. Combined with the data of the sensor fusion module, it can make more accurate decisions and optimizations, such as integrating vision, force and other sensor data, and adjusting the welding path and parameters in real time to adapt to different working conditions. The system control management module in the comprehensive management module manages the entire system and coordinates the work of each module. Through intelligent task scheduling and parameter configuration, the system operation efficiency and stability are improved. The sensor fusion module integrates multiple sensor data to provide the system with comprehensive and accurate information, realize all-round monitoring and control of the welding process, and improve the system adaptability and intelligence level. For workpieces of different shapes, sizes and materials, the system can automatically adjust the welding path and process parameters through visual positioning and intelligent decision optimization modules, and easily deal with simple straight welds and complex curved welds. Under different welding environments, the system operates stably with high reliability design and anti-interference ability, and adapts to harsh industrial environments such as high temperature, high humidity, and strong electromagnetic interference.

[0076] This system provides multiple safety guarantees through the motion safety protection module in the motion control module, sets limit switches, emergency stop buttons, etc. When the robot moves abnormally, it stops in time to avoid safety accidents. The robot's motion speed and acceleration are monitored and controlled in real time to ensure safe operation. At the same time, the motion stability control module reduces vibration and shake, improves welding quality and operational safety. The system calibration and maintenance module of the visual acquisition module makes system calibration and maintenance more convenient. Regular camera calibration and system maintenance ensure long-term stable operation and accuracy. The calibration and calibration module accurately calibrates and calibrates the system to ensure that each module works in coordination and data is accurate, improves reliability and reduces maintenance costs and time. The wireless communication module enhances the convenience of the system. Operators can remotely monitor the welding process and adjust parameters through wireless devices, improve work efficiency and operational flexibility, and facilitate communication and integration of the system with other equipment to achieve intelligent production, such as seamless connection with factory automation production lines, improving production automation levels and overall efficiency.

[0077] This system uses the camera selection configuration module and lighting system design module in the visual acquisition module to ensure that the high-resolution camera captures clear images and provides stable lighting. The image acquisition preprocessing module of the image processing module removes interference, the feature extraction algorithm module extracts key features, and the visual positioning coordinate conversion module determines the three-dimensional position and posture of the workpiece based on these features, replacing the indirect electrical positioning that is susceptible to interference. At the same time, the path planning optimization module directly generates the welding path based on the visual positioning results, without tedious operations, and improves the positioning efficiency. For the problem of strong dependence on pre-stored data and poor actual adaptability, the visual acquisition module and the visual positioning coordinate conversion module work together to continuously collect images and update the workpiece position and posture information in real time during the welding process. The teaching-free intelligent decision-making and optimization module combines real-time data to intelligently adjust the welding path and process parameters to avoid errors caused by relying on inaccurate pre-stored data. For complex workpieces, the image segmentation and analysis recognition module of the image processing module and the path planning optimization module can generate a suitable welding path based on detailed image information. The sensor fusion module fuses vision and other sensor data, such as assisting in determining the underlying structure information in the welding of multi-layer structure workpieces, enhancing the system's adaptability to complex workpieces and ensuring welding quality.

Claims

1. A robot teaching-free welding system based on visual positioning, characterized in that: It includes a PLC controller, a power module, a wireless communication module, a visual positioning coordinate conversion module, an image processing module, a motion control module, a visual acquisition module, a path planning optimization module, an image processing module and an integrated management module, wherein the power module is electrically connected to the PLC controller, and the PLC controller is connected to the visual positioning coordinate conversion module, the image processing module, the motion control module, the visual acquisition module, the path planning optimization module, the image processing module and the integrated management module through the wireless communication module; The visual positioning coordinate conversion module is composed of an image feature extraction module, a feature matching module, a three-dimensional coordinate calculation module and a coordinate conversion matrix calculation module. The image feature extraction module, the feature matching module, the three-dimensional coordinate calculation module and the coordinate conversion matrix calculation module are all connected through a wireless communication module and a PLC controller signal.

2. According to the vision-based robot teaching-free welding system of claim 1, it is characterized in that: The visual positioning coordinate conversion module also includes a real-time positioning update module and a positioning accuracy evaluation module, and the real-time positioning update module and the positioning accuracy evaluation module are connected via a wireless communication module and a PLC controller signal.

3. The robot teaching-free welding system based on visual positioning according to claim 1 is characterized in that: The motion control module is composed of a motion instruction receiving module, a joint motion control module, a motion trajectory tracking module and a motion speed adjustment module. The motion instruction receiving module, the joint motion control module, the motion trajectory tracking module and the motion speed adjustment module are all connected through a wireless communication module and a PLC controller signal.

4. The robot teaching-free welding system based on visual positioning according to claim 1 is characterized in that: The motion control module also includes a motion stability control module and a motion safety protection module, and the motion stability control module and the motion safety protection module are both connected via a wireless communication module and a PLC controller signal.

5. The robot teaching-free welding system based on visual positioning according to claim 1 is characterized in that: The path planning optimization module is composed of a path planning optimization module, a welding path generation module, a path parameter setting module, a path smoothing processing module and a path optimization algorithm module. The path planning optimization module, the welding path generation module, the path parameter setting module, the path smoothing processing module and the path optimization algorithm module are all connected through a wireless communication module and a PLC controller signal.

6. The robot teaching-free welding system based on visual positioning according to claim 1 is characterized in that: The path planning optimization module also includes a path adaptability adjustment module and a path storage calling module, and the path adaptability adjustment module and the path storage calling module are connected via a wireless communication module and a PLC controller signal.

7. The robot teaching-free welding system based on visual positioning according to claim 1 is characterized in that: The image processing module is composed of an image acquisition preprocessing module, an image enhancement module, a feature extraction algorithm module, an image segmentation module, an image analysis and recognition module, and an image post-processing module. The image acquisition preprocessing module, the image enhancement module, the feature extraction algorithm module, the image segmentation module, the image analysis and recognition module, and the image post-processing module are all connected through a wireless communication module and a PLC controller signal.

8. The robot teaching-free welding system based on visual positioning according to claim 1 is characterized in that: The visual acquisition module is composed of a camera selection and configuration module, a lighting system design module, an image acquisition control module, an image transmission and storage module, an image quality monitoring module and a system calibration and maintenance module. The camera selection and configuration module, the lighting system design module, the image acquisition control module, the image transmission and storage module, the image quality monitoring module and the system calibration and maintenance module are connected through a wireless communication module and a PLC controller signal.

9. The robot teaching-free welding system based on visual positioning according to claim 1 is characterized in that: The comprehensive management module is composed of a welding execution module, a system control management module, a sensor fusion module, a calibration and calibration module, a teaching-free intelligent decision-making and optimization module and a safety protection module. The welding execution module, the system control management module, the sensor fusion module, the calibration and calibration module, the teaching-free intelligent decision-making and optimization module and the safety protection module are connected through a wireless communication module and a PLC controller signal.

10. A robot teaching-free welding method based on visual positioning, using a robot teaching-free welding system based on visual positioning according to claims 1-9, characterized in that: The steps include: Step 1: Preparation, system initialization and parameter setting, select an industrial camera with a resolution of 25 million pixels for visual acquisition, configure a suitable lens focal length of 70mm-90mm according to the size of the welding workpiece and the working distance, design the lighting system so that its light intensity reaches 1200Lux-1800Lux, and use a multi-angle LED light source with an angle set at 40°-50° to ensure uniform lighting on the workpiece surface and reduce shadows and reflections; Step 2: Image acquisition and processing. Start the image acquisition control module and set the acquisition frequency to 70-90 frames per second. The image acquisition preprocessing module grayscales the acquired image and uses the median filter algorithm with a filter window size of 3x3-5x5 ​​to remove noise. The image enhancement module adjusts the image contrast to 1.6-2.0 times to highlight the features of the welding area. The feature extraction algorithm module uses the Canny edge detection algorithm and sets the low threshold to 70-90 and the high threshold to 160-200 to extract the edge features of the welding area. The image segmentation module accurately segments the welding area from the background based on color and texture features. Step 3: Visual positioning and coordinate conversion. The image feature extraction module extracts key feature points, such as the start and end points and corner points of the weld. The feature matching module matches the extracted features with the pre-stored standard features, requiring the matching accuracy to reach 92%-98% or more. The three-dimensional coordinate calculation module uses binocular vision technology to calculate the coordinates of the workpiece in three-dimensional space. The coordinate accuracy is controlled within ±0.06mm-±0.1mm. The coordinate transformation matrix calculation module accurately calculates the transformation matrix between the image coordinate system and the robot coordinate system, and the conversion accuracy is within ±0.03mm-±0.05mm. The real-time positioning update module continuously tracks the position changes of the workpiece at a frequency of 20Hz-30Hz. The positioning accuracy evaluation module regularly evaluates the positioning accuracy, and the error range is kept within ±0.12mm-±0.18mm. Step 4: Path planning optimization. The welding path generation module automatically generates the welding path according to the visual positioning results and welding process requirements. The path curvature radius is not less than 70mm-90mm. The path parameter setting module sets the welding speed at 7mm / s-9mm / s, and accurately sets the arc starting and arc ending positions. The path optimization algorithm module uses a genetic algorithm, with the goal of reducing the path length by 12%-18% and the welding time by 18%-22%. The path smoothing module controls the path curvature change rate within 0.7 degrees / mm-0.9 degrees / mm. The path adaptability adjustment module can adjust the path in real time within 0.7s-1.0s according to the situation during the actual welding process. The path storage and call module stores the optimized path for next use. Step 5: Motion control and welding execution, the motion command receiving module quickly receives the command from the path planning module, and the command transmission delay is controlled within 70ms-90ms. The joint motion control module accurately controls the robot joint motion, and the joint motion accuracy is within ±0.05mm-±0.08mm. The motion trajectory tracking module tracks the robot motion trajectory in real time, and the deviation from the planned trajectory is controlled within ±0.10mm-±0.15mm. The motion speed adjustment module adjusts the robot motion speed within the range of 4mm / s-12mm / s according to the actual situation. The motion stability control module takes measures to ensure the stability of the robot motion, and the vibration amplitude is controlled within ±0.05mm-±0.08mm. The motion safety protection module sets limit switches and emergency stop buttons, etc. When an abnormality occurs, the robot motion is stopped within 70ms-90ms. The welding execution module controls the welding power output, the welding current is 100A-150A, and the voltage is 22V-28V; Step 6: Comprehensive management and communication. The system control management module in the comprehensive management module manages the entire system in a coordinated manner. The task scheduling response time is within 1.2s-1.8s. The sensor fusion module integrates the data of the visual sensor, force sensor, laser sensor, etc. The measurement accuracy of the force sensor is within ±0.6N-±1.0N, and the measurement accuracy of the laser sensor is within ±0.05mm-±0.08mm. The calibration and calibration module regularly calibrates and calibrates the system. The calibration cycle is 1.5 months-2.5 months. The teaching-free intelligent decision-making and optimization module automatically adjusts the welding process parameters within 1.2s-1.8s according to the real-time situation. The safety protection module ensures the safe operation of the system, and the safety light curtain detection accuracy is set within ±3mm-±5mm. The wireless communication module realizes wireless communication between the system and external devices. The communication distance is 70m-90m and the data transmission rate is 25Mbps-35Mbps.

Images