Adaptive welding method for stainless steel furniture modules in continuous feed

Abstract

The application relates to a self-adaptive welding method for a stainless steel furniture module in continuous conveying. The self-adaptive welding method for the stainless steel furniture module in continuous conveying comprises the following steps: S2, continuous conveying and profile acquisition; continuous follow-up conveying is performed, and profile acquisition is performed through laser to obtain a laser signal of a keel in a continuous conveying state; S3, relative position determination; the laser signal of the keel obtained in step S2 is solved through an algorithm to obtain the relative position of the keel and a panel; S6, welding fixation; the relative position of the keel and the panel obtained in step S3 is used for welding the keel and the panel. The self-adaptive welding method for the stainless steel furniture module in continuous conveying has the advantages of automation and high efficiency.

Description

Technical Field

[0001] This application relates to the field of stainless steel furniture technology, and in particular to an adaptive welding method for stainless steel furniture modules under continuous conveying. Background Technology

[0002] Stainless steel furniture, with its advantages of corrosion resistance, high strength, easy cleaning, and long service life, is widely used in various settings such as homes, offices, hotels, and medical facilities, and market demand is increasing year by year. In the manufacturing process of stainless steel furniture, panel splicing and assembly is one of the core processes. The panel mainly consists of the panel body and the keel. As the supporting structure of the panel, the quality of the welding between the keel and the panel body directly determines the structural stability, load-bearing capacity, and service life of the stainless steel furniture.

[0003] Currently, there are two main welding methods for the keel and the main body of stainless steel furniture panels: one is manual welding, where a pre-fabricated keel is placed in a predetermined position on the main body of the panel, and the keel's posture is manually adjusted to align with the welding position before welding. This method has obvious drawbacks: it is labor-intensive, has extremely low welding efficiency, and the precision of manually adjusting the keel's posture is limited, easily leading to problems such as weld misalignment, incomplete welds, and uneven welds, resulting in unstable welding quality. Furthermore, manual welding requires a high level of skill from the operators, resulting in high training costs and certain safety hazards.

[0004] Another approach is semi-automatic or automated welding. To improve production efficiency and welding quality, the industry is gradually adopting automated conveyor lines in conjunction with welding equipment. This involves prefabricating the keel and placing it on the main body of the panel, then transporting the panel and keel together to the welding station via a conveyor line, where the welding equipment completes the welding. However, this automated welding method still faces many technical bottlenecks and struggles to meet the demands of large-scale, high-precision production of stainless steel furniture panels.

[0005] Specifically, in automated welding processes, due to various factors such as the prefabrication precision of the keel, errors in material feeding operations, and vibrations of the conveyor line, the keel often becomes misaligned or tilted when placed on the main body of the plate. If the tilt of the keel cannot be accurately identified and addressed, it can easily lead to weld misalignment, affecting welding quality and even causing product scrap. To solve this problem, existing technologies typically employ a "stop-and-correct" method. When keel misalignment is detected, the conveyor line operation is paused, and the position and orientation of the keel are adjusted using mechanical clamps or other mechanisms to align it with the preset welding position before conveying and welding resume.

[0006] However, this method of stopping the line for correction has the following drawbacks: First, stopping the line for correction severely affects production efficiency. Each correction requires pausing the conveyor line, slowing down the entire production line's cycle time and making continuous production impossible, thus failing to meet the demands of large-scale production. Second, the mechanical correction mechanism has a complex structure and high cost. Furthermore, the correction process can easily cause scratches, extrusion deformation, and other damage to the stainless steel plate and keel, affecting the product's appearance quality and structural strength. Third, the correction process requires additional control logic and actuators, increasing the complexity and failure rate of the automation system and resulting in high maintenance costs. Fourth, for slightly misaligned keels, stopping the line for correction is extremely inefficient and wastes production time.

[0007] Furthermore, stainless steel has strong specular reflective properties, which can easily lead to problems such as overexposure, blurred edges, and difficulty in feature extraction during visual recognition. This results in low accuracy in recognizing the position and tilt angle of the keel, which in turn affects the correction effect of the welding trajectory and makes it impossible to guarantee welding accuracy. At the same time, slight vibrations are unavoidable during continuous conveying. These vibrations can cause momentary shifts in the position of the plate body and the keel, further increasing the risk of welding misalignment. Existing technologies lack effective countermeasures.

[0008] In summary, existing stainless steel furniture panel keel welding technologies, whether manual or traditional automated, suffer from problems such as low welding efficiency, unstable welding quality, high production costs, and the inability to achieve continuous high-precision production. In particular, when the keel is misaligned, the method of stopping the line to correct the deviation is difficult to balance production efficiency and welding quality, and it cannot effectively solve the problem of weld deviation caused by the high reflectivity of stainless steel and conveyor vibration. Summary of the Invention

[0009] Based on this, the purpose of this application is to provide an adaptive welding method for stainless steel furniture modules under continuous conveying, so as to achieve precise welding of the keel and the main body of the panel through dual-dimensional adaptive compensation of the welding gun posture and position without continuous conveying or mechanical adjustment of the keel position. This method balances production efficiency and welding quality, reduces production costs, and meets the needs of large-scale and high-precision production of stainless steel furniture panels.

[0010] Step S1: Pre-process the panel and keel.

[0011] S1.1 Workpiece Pretreatment and Screening

[0012] Pre-treatment is performed on the stainless steel panels and prefabricated stainless steel keels to remove impurities such as oil, oxide scale, and rust from the surface of the panels and keels, ensuring that the surface is flat, without protrusions or depressions, and avoiding impurities or surface defects from affecting the height difference recognition accuracy. At the same time, the dimensions of the prefabricated keels are inspected to select keels that meet the preset size requirements, and the cross-sectional height of each type of keel is determined (i.e., the height difference benchmark value H0 after the keel and the panel are stacked). The dimensions of the panels are inspected to avoid height benchmark deviations caused by unevenness of the panels themselves.

[0013] S1.2, Standard for the placement of panels and keel.

[0014] S1.2.1 Panel Placement: Place the pre-treated stainless steel panel stably on the conveyor table of the continuous conveyor line; adjust the placement of the panel so that the length direction of the panel is as parallel as possible to the conveying direction of the continuous conveyor line, so as to provide a basis for establishing a global coordinate system based on the panel.

[0015] S1.2.2, Keel Placement: The keels shall be placed on the panel with overlapping joints. A certain degree of skewness, local offset, and unilateral misalignment are allowed. Specific placement requirements are as follows:

[0016] (1) Placement method: The keel is overlapped on the panel perpendicular to the length of the panel, and the bottom surface of the keel is in contact with the surface of the panel;

[0017] (2) Skewing range: The overall skew angle α∈[0°, 15°] of the keel can be along the width direction (horizontal) or the length direction (longitudinal) of the panel to adapt to the feeding error in actual production;

[0018] (3) Offset range: The lateral offset of the keel relative to the preset welding area of ​​the panel is Δx∈[-50mm, 50mm], and the longitudinal offset is Δy∈[-30mm, 30mm];

[0019] (4) Multiple keel placement: If multiple keels need to be placed on the panel, the multiple keels are evenly distributed along the length of the panel. Each keel is placed independently, and each keel is allowed to have different tilt and offset states, which do not affect the recognition.

[0020] S1.3 Equipment Installation and Commissioning

[0021] S1.3.1 Installation of Laser Contour Recognition System: The laser contour recognition system is aligned with the center line of the continuous conveyor line, and the scanning range covers the width direction of the panel and the length direction of the keel to ensure that the complete contour of the keel can be captured.

[0022] The laser contour recognition system includes a line laser emitter, a laser receiver, a signal processing module, and a data transmission module. The line laser emitter emits a linear laser beam that illuminates the panel and keel surface. The laser receiver receives the laser echo signal. The signal processing module preprocesses the echo signal, and the data transmission module transmits the processed signal to the central control system.

[0023] S1.3.2 System Calibration: Perform collaborative calibration of the laser contour recognition system and the central control system to ensure recognition accuracy and speed. The specific calibration steps are as follows:

[0024] (1) Altitude reference calibration;

[0025] (2) Coordinate system calibration:

[0026] (3) Recognition accuracy calibration:

[0027] (4) Recognition speed calibration:

[0028] S1.4 Database Establishment: Input the dimensional parameters of panels of different specifications, the dimensional parameters of keels of different specifications (section height H0, length, width), height difference reference value H0, calibration data and other information into the database of the central control system to establish the panel-keel parameter matching relationship, so as to facilitate the quick call of corresponding parameters during subsequent identification and improve identification efficiency; at the same time, input the identification error threshold, the allowable range of skew angle, the allowable range of offset and other parameters for subsequent identification result judgment and screening.

[0029] Step S2: Continuous transport and contour acquisition. Narrow-band multi-point cross-sectional acquisition is performed using laser.

[0030] S2.1 Continuous Conveying Start-up: Start the continuous conveyor line, set the conveying speed v, and adjust the conveying speed according to production needs; during the conveying process, the speed control system maintains a stable conveying speed, reduces laser scanning deviation caused by conveying speed fluctuations, and ensures recognition accuracy; the positioning and limiting mechanism continuously limits the panel to prevent lateral displacement of the panel and ensures the stability of the panel's reference position.

[0031] S2.2 Laser Scan Start-up: When the panel and keel are transported to the scanning area of ​​the laser contour recognition system, the control system issues a command to start the laser contour recognition system. The line laser emitter continuously emits a linear laser to scan the upper surface of the panel and keel along the transport direction (X-axis direction). The laser scanning width covers the width direction of the panel (Y-axis direction). The line laser emitter starts single-line narrow-band scanning along the width direction of the panel, thereby reducing the amount of scanning calculation and recognition processing.

[0032] S2.3 Laser Echo Signal Acquisition: After the laser shines on the surface of the panel and the keel, it generates a reflected echo. The laser receiver acquires the echo signal in real time. The acquisition frequency is consistent with the laser scanning frequency to ensure that the height information of each scanning point can be captured quickly. Since the panel is a flat thin plate and the keel is a raised profile, there is a fixed height difference H0 between the two in the Z-axis direction (height direction). Therefore, the height data of the laser echo signal will show a significant abrupt change at the boundary between the keel and the panel, that is, from the height value Z1 of the panel to the height value Z2 of the keel (Z2=Z1+H0). This height abrupt change point is the boundary point between the panel and the keel, which is also the core basis for subsequent contour segmentation and pose calculation.

[0033] S2.4 Signal Preprocessing: The echo signal acquired by the laser receiver may contain a small amount of noise (such as signal interference caused by welding fumes or workshop dust). The signal processing module performs real-time preprocessing on the echo signal to remove noise interference and extract the effective height data. The preprocessing steps are as follows:

[0034] (1) Signal filtering: An adaptive Gaussian filtering algorithm is used to filter the echo signal, remove random noise and interference signals, and ensure the stability of the altitude data;

[0035] (2) Signal amplification: The filtered echo signal is amplified appropriately by a factor of 10 to 20 to enhance the signal strength at the height change point, which is convenient for subsequent boundary point extraction;

[0036] (3) Baseline calibration: Using the height data of the upper surface of the panel as the baseline, the height measurement reference is calibrated in real time to avoid height measurement deviation caused by slight fluctuations of the panel;

[0037] (4) Valid signal screening: Based on the height difference benchmark value H0 stored in the database, the signals with height differences in the range of [H0-0.1mm, H0+0.1mm] are selected as valid signals corresponding to the keel, and abnormal signals are eliminated to ensure the accuracy of subsequent identification.

[0038] S2.5 Dynamic Synchronous Acquisition: During continuous conveying, the laser contour recognition system remains synchronized with the continuous conveying line, and the laser scanning speed matches the conveying speed to ensure that each scanning point can accurately correspond to the fixed position on the panel and keel, avoiding contour deformation and recognition deviation caused by asynchronous scanning and conveying; at the same time, the control system receives the effective height data transmitted by the signal processing module in real time, continuously accumulates scanning data, and provides sufficient data support for subsequent keel contour reconstruction and pose calculation.

[0039] In this step, the scanning speed of the laser contour recognition system can be automatically adjusted according to the conveying speed. That is, when the conveying speed increases, the scanning frequency increases synchronously, and when the conveying speed decreases, the scanning frequency decreases synchronously. This ensures recognition accuracy while avoiding resource waste, achieving high efficiency and energy saving. At the same time, the scanning process is not affected by stainless steel mirror reflection, strong workshop light, welding fumes, etc., because laser recognition relies on height information, rather than surface reflection or texture. Even if there are overexposed spots or oil stains on the panel and keel surface, it will not affect the recognition of height differences, ensuring the stability of recognition.

[0040] Step S3: Determine the relative position.

[0041] S3.1 Height Abrupt Change Point Extraction and Boundary Segmentation: The central control system analyzes the pre-processed effective height data and extracts height abrupt change points, namely the points where the height value changes abruptly from Z1 (panel height) to Z2 (keel height) and from Z2 back to Z1. These abrupt change points are the boundary points between the panel and the keel. Based on the coordinates (X, Y, Z) of the boundary points, the scanning area is divided into two parts: the panel area with a height value of Z1 and the keel area with a height value of Z2, achieving precise segmentation between the panel and the keel, completely unaffected by surface reflection and texture.

[0042] Specific extraction method: The gradient threshold method is adopted, and the height gradient threshold is set to 0.5mm / pixel. When the height difference between two adjacent scan points is greater than the gradient threshold, the point is determined to be a height change point, that is, the boundary point between the keel and the panel. At the same time, combined with the coordinate continuity of adjacent scan points, isolated change points (caused by noise interference) are eliminated to ensure the accuracy of the boundary points. For cases where the keel is partially occluded or the edge is missing, the coordinates of adjacent boundary points are fitted to supplement the complete boundary contour, avoid recognition failure, and improve the fault tolerance of recognition.

[0043] S3.2 Keel Contour Reconstruction (Core Step for Accurate Identification): Based on the extracted boundary point coordinates, the central control system uses a B-spline curve fitting algorithm to fit the boundary points of the keel and reconstruct the complete contour of the keel, including the contours of the two side edges, the contours of the two end faces, and the contour of the upper surface. At the same time, based on the cross-sectional height H0 of the keel and the fitted contour, the model of the keel is reconstructed to clarify the spatial posture of the keel.

[0044] The specific steps of contour reconstruction are as follows:

[0045] (1) Fitting the contours of both sides of the keel: Extract all boundary points along the length direction (X-axis direction) of the keel and fit them with B-spline curves to obtain the edge lines of both sides of the keel (boundaries along the Y-axis direction).

[0046] (2) Fitting the contours of the two end faces: Extract all boundary points in the width direction (Y-axis direction) of the keel, and fit the two end face lines of the keel (boundaries in the X-axis direction) to determine the length range of the keel;

[0047] (3) Upper surface contour fitting: Extract all height points on the upper surface of the keel, fit the flat contour of the upper surface of the keel, and verify whether there are defects such as local protrusions or depressions in the keel. If there are defects, mark them as abnormal and prompt the staff to handle them.

[0048] (4) Three-dimensional contour reconstruction: Combine the two side edge lines, the two end face lines and the upper surface contour to reconstruct the three-dimensional contour model of the keel and clarify the spatial position of the keel in the global coordinate system (based on the panel).

[0049] In this step, the contour reconstruction algorithm adopts a parallel computing approach, utilizing the multi-core processor of the central control system to split the contour fitting task into multiple cores for parallel processing; at the same time, the fitting algorithm is adaptive, automatically adjusting the fitting parameters according to the contour characteristics of different specifications of keel, adapting to stainless steel keel of any specification without the need for recalibration.

[0050] S3.3 Precise Position Identification of Keel Relative to Panel: Using the panel as a reference (global coordinate system), the specific positional parameters of the keel relative to the panel are calculated based on the reconstructed 3D contour of the keel, including the following core parameters:

[0051] (1) Lateral offset Δx: The offset of the keel centerline relative to the panel centerline in the Y-axis direction, that is, the offset distance of the keel in the width direction of the panel. If Δx is positive, it means that the keel is offset to the right side of the panel; if Δx is negative, it means that the keel is offset to the left side of the panel.

[0052] (2) Longitudinal offset Δy: The offset of the keel centerline relative to the centerline of the preset welding area of ​​the panel in the X-axis direction, that is, the offset distance of the keel in the length direction of the panel. If Δy is positive, it means that the keel is offset forward in the conveying direction; if Δy is negative, it means that the keel is offset backward in the conveying direction.

[0053] (3) Distance from edge to panel boundary: The distance from the two sides of the keel to the two sides of the panel boundary is denoted as L1 (left side distance) and L2 (right side distance), respectively, and is used to determine whether the keel is within the preset welding area of ​​the panel;

[0054] (4) Single-sided misalignment: The misalignment distance between the two ends of the keel relative to the preset welding area of ​​the panel, denoted as Δs;

[0055] (5) Keel center coordinates (X0, Y0): The coordinates of the midpoint of the keel center line in the global coordinate system, which clarifies the specific position of the keel on the panel.

[0056] Specific calculation method: Taking the reference boundary of the panel as a reference, calculate the coordinate difference between the center line of the keel and the center line of the panel according to the two side edge lines and the center line of the keel, and obtain the lateral offset Δx and the longitudinal offset Δy; calculate the distance L1 and L2 from the edge to the panel boundary according to the Y-axis coordinates of the two side edge lines of the keel and the Y-axis coordinates of the two side boundaries of the panel; calculate the single-sided misalignment Δs according to the X-axis coordinates of the end face lines of the two ends of the keel and the X-axis coordinates of the two ends of the preset welding area of ​​the panel; determine the center coordinates (X0, Y0) of the keel according to the coordinates of the center line of the keel.

[0057] S3.4 Accurate Identification of Keel Placement Status: Based on the reconstructed 3D contour and centerline of the keel, the placement status parameters of the keel are calculated, with a focus on identifying the keel's tilt angle. Simultaneously, it is determined whether the keel exhibits local offset, twisting, or other abnormalities. The specific identification method is as follows:

[0058] (1) Overall tilt angle α: The angle between the center line of the keel and the center line of the panel (Y-axis direction) is the overall tilt angle of the keel, α∈[0°, 15°]; if α=0°, it means that the keel is placed upright and there is no tilt; if α>0°, it means that the keel is tilted. Based on the direction of the angle, the tilt direction (left tilt or right tilt) can be determined.

[0059] Specific calculation method: The angle between the fitted straight line of the keel centerline and the panel centerline (Y-axis) is calculated by using the vector dot product formula, that is, α=arccos((vector a·vector b) / (|vector a|×|vector b|)), where vector a is the direction vector of the keel centerline and vector b is the direction vector of the panel centerline (Y-axis direction).

[0060] (2) Local skew angle α1: If the keel has local twisting or local offset, calculate the skew angle of different sections of the keel. That is, along the length direction of the keel (X-axis direction), every 50mm is a section. Calculate the angle between the center line of each section and the center line of the panel to obtain the local skew angle α1, which is used to determine whether the keel has local deformation. If the difference between the local skew angle and the overall skew angle is ≤0.1°, it is judged as qualified; if the difference is >0.1°, it is judged as local twisting and marked as abnormal.

[0061] (3) Judgment of placement status: The placement status of the keel is judged based on parameters such as the overall tilt angle α, the local tilt angle α1, and the single-sided misalignment Δs.

[0062] S3.5 Real-time output and storage of recognition results: The central control system will output the calculated position parameters (Δx, Δy, L1, L2, Δs, (X0, Y0)) and placement parameters (α, α1, placement result) of the keel relative to the panel to the welding torch execution system in real time, providing accurate data support for subsequent welding torch adaptive trajectory compensation; at the same time, all recognition data will be stored in the database, including recognition time, keel specifications, panel specifications, conveying speed, recognition parameters, and placement status, for subsequent data analysis, production optimization, and quality traceability.

[0063] Step S4: Continuously optimize recognition accuracy and speed.

[0064] S4.1 Recognition Accuracy Optimization Measures: To further improve recognition accuracy and avoid recognition deviations caused by various errors, the following optimization measures are adopted:

[0065] (1) Real-time calibration of height reference: During continuous transmission, the central control system automatically extracts the height data of the panel area and recalibrates the height reference every 100 frames of data scanned, so as to avoid height measurement deviation caused by slight panel fluctuations and laser emitter drift.

[0066] (2) Boundary point optimization extraction: The dual extraction method of "gradient threshold + neighboring point verification" is adopted. Not only are height change points extracted through gradient threshold, but the effectiveness of change points is also verified through the coordinate continuity of neighboring points to eliminate isolated noise points and ensure the accuracy of boundary points.

[0067] (3) Fitting algorithm optimization: The adaptive B-spline curve fitting algorithm is adopted. The number of fitting nodes is automatically adjusted according to the complexity of the keel contour. When the contour is complex, the number of nodes is increased, and when the contour is simple, the number of nodes is reduced, which ensures the fitting accuracy and reduces the fitting time.

[0068] (4) Error compensation: Establish an identification error model, predict possible identification errors in advance based on parameters such as conveying speed, scanning frequency, and keel specifications, and perform real-time compensation to ensure the final identification accuracy and the accuracy of skew angle identification;

[0069] (5) Dual laser redundancy recognition: Add a spare line laser emitter to the laser contour recognition system. The two laser emitters scan synchronously and compare the recognition results of the two sets of scan data. If the difference is ≤ ±0.03mm, the average value is used as the final recognition result; if the difference is > ±0.03mm, an alarm signal is issued to prompt the staff to check the equipment to avoid excessive recognition error.

[0070] S4.2 Optimization Measures for Recognition Speed: To achieve efficient and rapid recognition and adapt to the high-speed production rhythm of continuous follow-up conveying, the following optimization measures are adopted:

[0071] (1) Parallel computing optimization: The central control system adopts a multi-core processor to split the tasks such as signal preprocessing, boundary point extraction, contour fitting, and pose calculation into different cores for parallel processing, which greatly shortens the total recognition time. When multiple keels are recognized at the same time, the recognition time is shortened synchronously without affecting the conveying speed.

[0072] (2) Adaptive adjustment of scanning parameters: The laser scanning frequency and scanning spacing are automatically adjusted according to the keel specifications and conveying speed. For example, when the keel size is large and the conveying speed is slow, the scanning frequency is reduced and the scanning spacing is increased to improve the recognition efficiency; when the keel size is small and the conveying speed is fast, the scanning frequency is increased and the scanning spacing is reduced to ensure the recognition accuracy.

[0073] (3) Data compression and transmission: The laser echo signal and identification data are compressed using a lossless compression algorithm to reduce the data volume to less than 50% of the original, thereby reducing the data transmission time;

[0074] (4) Preloading mechanism: Commonly used keel parameters, panel parameters, fitting parameters, etc. are preloaded into the memory of the central control system and can be directly called during recognition without reading from the database, thus shortening the recognition time;

[0075] (5) Invalid data filtering: In the signal preprocessing stage, invalid scanning data in the panel area (such as blank areas at the edge of the panel) are filtered out in advance, and only valid data in the keel area is processed to reduce the amount of data processing and improve the recognition speed.

[0076] S4.3 Recognition Stability Optimization: To ensure the recognition system can operate stably in harsh industrial environments (welding fumes, strong light, vibration), the following optimization measures are taken:

[0077] (1) Dustproof and splashproof design: A dustproof and splashproof protective cover is set on the outside of the laser contour recognition system. The protective cover is made of transparent high temperature resistant material, which does not affect laser scanning and can prevent welding fumes and sparks from contaminating the laser transmitter and receiver, thus extending the service life of the equipment.

[0078] (2) Vibration-resistant design: The laser contour recognition system is installed on a vibration damping bracket to reduce the impact of continuous conveyor line vibration on the equipment and ensure the stability of laser scanning;

[0079] (3) Temperature compensation: The laser contour recognition system has a built-in temperature sensor to monitor the operating temperature of the equipment in real time. When the temperature exceeds the preset range (0~40℃), the laser power and signal processing parameters are automatically adjusted to avoid the decrease in recognition accuracy caused by excessive temperature.

[0080] (4) Fault self-diagnosis: The central control system monitors the working status of the laser contour recognition system in real time. If there is a fault in the laser transmitter, receiver, or data transmission, an alarm signal will be issued immediately, and the fault location and handling method will be indicated, so that staff can quickly troubleshoot and reduce downtime.

[0081] Step S5: Simultaneous identification and batch processing of multiple keels.

[0082] S5.1 Multi-keel synchronous scanning: When multiple keels are placed on the panel, the scanning range of the laser contour recognition system covers all keels. The line laser scans all keels and the panel surface synchronously along the conveying direction, and the laser receiver collects the echo signals of all keels synchronously. There is no need to scan in batches, ensuring the synchronicity of multi-keel recognition.

[0083] S5.2 Multi-keel pose calculation: The central control system adopts a multi-threaded processing method to process the echo signal of each keel independently, extracting boundary points, reconstructing contours, calculating relative pose and placement status, and ensuring that the recognition process of each keel does not interfere with each other, thus ensuring the efficiency and accuracy of multi-keel recognition; at the same time, the placement status of each keel is identified based on the relative position of multiple keels.

[0084] S5.3 Batch Identification Result Output and Feedback: After the identification of multiple keels is completed, the central control system outputs the identification result (position parameters, placement status) of each keel to the welding torch execution system to ensure that each welding torch can perform adaptive compensation for the corresponding keel. At the same time, the batch identification data is summarized and stored in the database for batch analysis to explore the feeding error pattern, and fed back to the feeding mechanism to optimize the feeding operation, reduce keel skew and offset, and improve the pass rate of batch production.

[0085] S6. Welding and fixing: Based on the relative positions of the keel and the panel obtained in S3, weld the keel and the panel.

[0086] Step S6 includes: setting the starting position, lateral movement speed and direction of the welding torch according to the position parameters of the keel relative to the panel and the placement parameters of the keel, so that the weld seam is consistent with the contour of the keel.

[0087] Multiple welding torches are mounted on the truss, each equipped with an independent lateral movement mechanism to drive the welding torch to move laterally; the speed of the lateral movement of the welding torch is adapted to the tilt angle of the keel, so that the weld seam is adapted to the placement of the keel.

[0088] When the panel and keel pass through the welding gun, they are first pressed together by the pressure roller, which stabilizes the relative position of the keel and the panel to ensure the quality of the welding.

[0089] Beneficial effects

[0090] Compared with the prior art, the present invention has the following advantages through the above technical solution: it fits the production scenario of continuous follow-up welding of stainless steel furniture panels without trajectory compensation.

[0091] 1. High recognition accuracy, solving the problem of reflection interference; This invention is based on laser recognition of stacked height difference. Without relying on surface reflection, texture, or color, it avoids interference from stainless steel mirror reflection, overexposure, smoke and dust. The position recognition accuracy and tilt angle recognition accuracy exceed those of existing traditional visual recognition methods. It can accurately identify the relative position and placement of the keel and the panel, providing accurate data support for subsequent welding gun adaptive compensation, and greatly reducing welding defects such as welding deviation and incomplete welding.

[0092] 2. High production efficiency; fast recognition speed, adaptable to continuous follow-up conveying.

[0093] 3. Highly adaptable and versatile; requires no workpiece calibration, adaptable to stainless steel keels and panels of any specifications, capable of recognizing various placement states such as overall skewness, local offset, and unilateral misalignment of the keel, highly fault-tolerant, requires no re-adjustment of equipment for different specifications of workpieces, adaptable to the production needs of different types of stainless steel furniture panels, and highly versatile.

[0094] To better understand and implement this application, the following detailed description is provided in conjunction with the accompanying drawings. Attached Figure Description

[0095] Figure 1 A flowchart illustrating an adaptive welding method for stainless steel furniture modules under continuous conveying, as exemplified in this application;

[0096] Figure 2 Another flowchart illustrating an adaptive welding method for stainless steel furniture modules under continuous conveying, as exemplified in this application;

[0097] Figure 3 This is a working principle diagram from one perspective of the adaptive welding method for stainless steel furniture modules under continuous conveying, which is an exemplary embodiment of this application.

[0098] Figure 4 This is a working principle diagram from another perspective of the adaptive welding method for stainless steel furniture modules under continuous conveying, which is an example of this application. Detailed Implementation

[0099] In the description of this application, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on this application. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0100] like Figures 1-4 As shown, the present invention provides an adaptive welding method for stainless steel furniture modules under continuous conveying, comprising the following steps:

[0101] S2. Continuous conveying and contour acquisition; continuous follow-up conveying, and contour acquisition by laser to obtain the laser signal of the keel in the continuous conveying state;

[0102] S3. Determine the relative position; Based on the laser signal of the keel obtained in step S2, calculate the relative position between the keel and the panel using an algorithm.

[0103] S6. Welding and fixing: Based on the relative positions of the keel and the panel obtained in S3, weld the keel and the panel.

[0104] In a preferred embodiment, step S2 includes the following steps:

[0105] S2.1 The panel and the keel placed on the panel are continuously conveyed at a constant speed.

[0106] S2.2 When the panel and keel pass through the scanning area, the panel and keel are continuously laser scanned;

[0107] S2.3 Acquire laser echo signals and determine the boundary points of the keel relative to the panel based on the obvious abrupt change in the laser echo signal data in the height direction;

[0108] S2.4 Preprocess the laser signal to remove noise interference;

[0109] S2.5 Dynamic synchronous acquisition to ensure that the laser scanning speed is consistent with the delivery speed.

[0110] In a preferred embodiment, step S3 includes the following steps:

[0111] S3.1 Extract and segment height abrupt change points to determine the boundary between the keel and the panel;

[0112] S3.2. Based on the mutation points, fit and reconstruct the keel outline;

[0113] S3.3. Based on the reconstructed keel outline, calculate the position parameters of the keel relative to the panel;

[0114] S3.4. Based on the reconstructed keel outline and centerline, calculate the keel placement parameters;

[0115] S3.5. Real-time output of recognition results.

[0116] In a preferred embodiment, the position parameters in step S3.3 include lateral offset, longitudinal offset, distance between the edge of the keel and the edge of the panel, and the coordinates of the keel center.

[0117] The calculation method in step S3.3 is as follows: taking the boundary of the panel as a reference, the coordinate difference between the center line of the keel and the center line of the panel is calculated based on the edge lines on both sides of the keel and the center line of the keel, and the horizontal offset Δx and the vertical offset Δy are obtained respectively; the distances L1 and L2 from the edge to the panel boundary are calculated based on the Y-axis coordinates of the edge lines on both sides of the keel and the Y-axis coordinates of the boundary on both sides of the panel respectively; and the center coordinates of the keel are determined based on the coordinates of the center line of the keel.

[0118] In a preferred embodiment, the placement parameters in step S3.4 include the overall tilt angle of the keel and the local tilt angle of the keel.

[0119] The method for calculating the overall tilt angle of the keel is to calculate the angle between the center line of the keel and the center line of the panel, and this angle is the overall tilt angle of the keel.

[0120] The method for calculating the local skew angle of the keel is as follows: along the length of the keel, every 500mm is a segment, and the angle between the center line of each segment and the center line of the panel is calculated. This angle is the local skew angle of that segment. If the difference between the local skew angle and the overall skew angle is ≤0.1°, it is judged as qualified; if the difference between the local skew angle and the overall skew angle is 0.1°, it is judged as local distortion.

[0121] In a preferred embodiment, step S6 includes: setting the starting position, lateral movement speed, and direction of the welding torch according to the position parameters of the keel relative to the panel and the placement parameters of the keel, so that the weld seam is consistent with the contour of the keel.

[0122] In a preferred embodiment, before step S2, the method further includes the following step:

[0123] S1. Perform pre-processing of the panel and keel;

[0124] In step S1, multiple keels are placed on the panel, with one surface of the panel facing down and the other surface facing up, and multiple keels are placed on it.

[0125] The panel is placed on a conveyor belt and transported by the conveyor belt.

[0126] In a preferred embodiment, after step S3, the method further includes the following step:

[0127] S4. Continuously optimize recognition accuracy and speed;

[0128] S5. Perform simultaneous identification and batch processing of multiple keels.

[0129] In a preferred embodiment, step S4 includes the following steps:

[0130] S4.1 Continuous optimization of recognition accuracy, specifically including real-time calibration of height benchmark, optimized extraction of boundary points, and dynamic adjustment of the number of fitting nodes to optimize the fitting algorithm;

[0131] S4.2 Continuous optimization of recognition speed, including parallel computing optimization, adaptive adjustment of scanning parameters, data compression and transmission, establishment of a preloading mechanism, and filtering of invalid data;

[0132] S4.3, Optimization of recognition stability.

[0133] In a preferred embodiment, step S5 includes the following steps:

[0134] S5.1, Multi-keel synchronous scanning;

[0135] S5.2 Calculate the position and attitude of each multi-keel structure separately;

[0136] S5.3 Batch processing of recognition results and matching with the corresponding welding torch welding path.

[0137] The following is a specific implementation method combined with parameters to facilitate understanding of the technical content of the present invention.

[0138] Example 1: Basic Implementation Method (Single Keel Identification)

[0139] This embodiment provides a laser recognition method for the relative pose of stainless steel keel and panel based on the stacking height difference, which is applied to the production of stainless steel dining table panels. The panel of the stainless steel dining table is made of 304 stainless steel; the prefabricated keel is made of 304 stainless steel, and each panel is equipped with one keel, which is placed in the center of the panel.

[0140] The equipment used in this embodiment includes: a continuous conveyor line, a laser contour recognition system, a control system, a positioning and limiting mechanism, and a quality inspection system.

[0141] The specific implementation steps are as follows:

[0142] Step S1: Preliminary preparation and equipment debugging.

[0143] S1.1 Workpiece pretreatment and screening: Detect the keel dimensions; detect the panel dimensions; enter the keel cross-section height H0=10mm into the central control system database.

[0144] S1.2, Layout guidelines for panels and keel:

[0145] (1) Panel placement: Place the panel stably on the conveyor table of the continuous conveyor line, adjust the panel position so that the length direction of the panel is parallel to the conveying direction and the center line of the panel coincides with the center line of the conveyor line;

[0146] (2) Keel placement: The keel is placed on the panel by manual or robotic arms.

[0147] S1.3 Equipment Installation and Commissioning:

[0148] (1) Installation of the laser contour recognition system;

[0149] (2) Debugging of the positioning and limiting mechanism;

[0150] (3) System calibration:

[0151] ① Height reference calibration: Place a standard calibration block with a height of 10mm on the panel, scan it and adjust the laser parameters to make the height detection accuracy ±0.008mm;

[0152] ② Coordinate system calibration: Using the top surface of the panel as a reference, establish a global coordinate system (X-axis: conveying direction, Y-axis: panel width direction, Z-axis: height direction), and enter the coordinates of the four corners of the panel;

[0153] ③ Recognition accuracy calibration: Place the keel on the panel, compare the recognition result with the standard value, and adjust the algorithm parameters;

[0154] ④ Recognition speed calibration: Set the conveying speed to 2m / min and the detection and recognition speed to 7m / min.

[0155] S1.4: Database establishment: Enter panel dimensions, keel dimensions, height difference reference value H0=10mm, calibration data, identification error threshold, etc. into the database.

[0156] Step S2: Continuous follow-up transport and laser contour acquisition.

[0157] S2.1 Continuous Conveying Start: Start the continuous conveying line and set the conveying speed v.

[0158] S2.2 Laser Scan Start-up: When the panel and keel are transported to the scanning area, the laser contour recognition system is activated. The line laser emitter emits a linear laser to scan the upper surface of the panel and keel along the transport direction, covering the entire area of ​​the panel and keel.

[0159] S2.3 Laser Echo Signal Acquisition: The laser receiver acquires the laser echo signal in real time, captures height change points, and transmits the echo signal to the signal processing module.

[0160] S2.4 Signal Preprocessing: The signal processing module uses Gaussian filtering to remove noise signals; amplifies the signal by 15 times to enhance the signal at height change points; uses the panel height Z1 as the baseline to calibrate the height reference in real time; filters valid signals with height differences in the range of [9.9mm, 10.1mm] and removes abnormal signals.

[0161] S2.5 Dynamic Synchronous Acquisition: The laser scanning speed is synchronized with the conveying speed. Each scanning point precisely corresponds to a fixed position on the panel and keel. The central control system receives effective height data in real time and continuously accumulates scanning data.

[0162] Step S3: Accurately identify the relative pose of the keel and the panel.

[0163] S3.1, Height abrupt change point extraction and boundary segmentation: The gradient thresholding method (gradient threshold = 0.5mm / pixel) is used to extract height abrupt change points. Combined with the verification of adjacent points, isolated noise points are removed, and the panel area (Z1 = 1.5mm) and the keel area (Z2 = 11.5mm) are segmented.

[0164] S3.2 Keel contour reconstruction: Using the B-spline curve fitting algorithm, the boundary points are fitted to reconstruct the contours of the two side edges, the two end faces, and the upper surface of the keel; the three-dimensional contour model of the keel is reconstructed to clarify the spatial posture of the keel.

[0165] The specific steps of contour reconstruction are as follows:

[0166] (1) Fitting the contours of both sides of the keel: Extract all boundary points along the length direction (X-axis direction) of the keel. A total of N1 boundary points were collected. B-spline curve fitting was used to obtain the boundary lines of both sides of the keel (Y-axis direction).

[0167] (2) Fitting the contours of the two ends: Extract all boundary points in the width direction (Y-axis direction) of the keel. A total of N2 boundary points were collected. Fitting the two end face lines of the keel (boundaries in the X-axis direction) clarifies the actual length of the keel.

[0168] (3) Upper surface contour fitting: Extract all height points on the upper surface of the keel. A total of N3 height points were collected and fitted to obtain the flat contour of the upper surface of the keel.

[0169] (4) Contour reconstruction: Combine the edge lines on both sides, the end face lines at both ends and the contour of the upper surface to reconstruct the contour model of the keel and clarify the specific position of the keel in the global coordinate system (based on the panel).

[0170] S3.3 Precise Position Identification of the Keel Relative to the Panel: Using the panel as a reference (global coordinate system), the specific position parameters of the keel relative to the panel are calculated based on the reconstructed keel outline. The specific calculation results are as follows:

[0171] (1) Lateral offset Δx: By calculating the difference in Y-axis coordinates between the center line of the keel and the center line of the panel, we get Δx = 2.03mm, which is a positive value, indicating that the keel is offset to the right side of the panel, and the offset is within the allowable range;

[0172] (2) Longitudinal offset Δy: By calculating the difference in X-axis coordinates between the center line of the keel and the center line of the preset welding area of ​​the panel, we get Δy=1.02mm, which is a positive value, indicating that the keel is offset forward in the conveying direction, and the offset is within the allowable range.

[0173] (3) Distance from edge to panel boundary: The panel width is 600mm and the keel width is 20mm. By calculating the difference between the Y-axis coordinates of the two sides of the keel edge line and the Y-axis coordinates of the two sides of the panel boundary, the distance on the left side L1=290.01mm and the distance on the right side L2=289.96mm are obtained, both of which are within the preset welding area.

[0174] (4) Single-sided misalignment Δs: The difference between the X-axis coordinates of the end face lines at both ends of the keel and the X-axis coordinates of the two ends of the preset welding area of ​​the panel is calculated.

[0175] (5) Keel center coordinates (X0, Y0): Based on the coordinates of the keel center line, determine the keel center coordinates as (600.05mm, 302.03mm) to clarify the specific position of the keel on the panel.

[0176] S3.4 Accurate Identification of Keel Placement Status: Based on the reconstructed 3D contour and centerline of the keel, the placement status parameters of the keel are calculated. The specific identification results are as follows:

[0177] (1) Overall tilt angle α: The angle between the direction vector of the keel centerline and the direction vector of the panel centerline (Y-axis direction) is calculated using the vector dot product formula. The result is α=5.02°, ∈[0°, 15°], indicating that the keel has a slight right-side tilt, and the tilt angle is within the allowable range.

[0178] (2) Local skew angle α1: Along the length of the keel (X-axis direction), a section is taken every 50mm, for a total of 12 sections. The angle between the center line of each section and the center line of the panel is calculated. The local skew angle α1 of each section is between 4.98° and 5.05°. The difference between the local skew angle α = 5.02° and the overall skew angle α = 5.02° is ≤0.04° and ≤0.1°, respectively. Therefore, it is determined that there is no local distortion.

[0179] (3) Judgment of placement status: Based on the overall tilt angle α=5.02° (≤15°), the difference between the local tilt angle α1 and α ≤0.04° (≤0.1°), and the single-sided misalignment Δs=1.48mm (≤2mm), the placement status of the keel is determined to be qualified and no further processing is required. It can proceed directly to the subsequent welding process.

[0180] S3.5 Real-time output and storage of identification results: The central control system will output the calculated position parameters and placement status parameters of the keel relative to the panel to the welding gun execution system in real time; at the same time, all identification data (including identification time, keel specifications, panel specifications, conveying speed v, identification parameters, placement status, etc.) will be encrypted and stored in the database of the central control system for subsequent big data analysis, production optimization and quality traceability.

[0181] Step S4: Optimize recognition accuracy and speed.

[0182] S4.1 Recognition Accuracy Optimization Measures: In this embodiment, the following optimization measures are adopted to ensure recognition accuracy:

[0183] (1) Real-time calibration of height reference: Every 100 frames of data are scanned, the central control system automatically extracts the height data of the panel area and recalibrates the height reference to avoid height measurement deviation caused by slight panel fluctuations and laser emitter drift.

[0184] (2) Boundary point optimization extraction: A dual extraction method of "gradient threshold + neighboring point verification" is adopted to remove isolated noise points;

[0185] (3) Fitting algorithm optimization: The adaptive B-spline curve fitting algorithm is adopted. Based on the simple keel contour in this embodiment, the number of fitting nodes is automatically reduced, which ensures the fitting accuracy and shortens the fitting time.

[0186] (4) Error compensation: Based on parameters such as conveying speed v, scanning frequency, and keel specifications in this embodiment, the possible recognition error is predicted by the recognition error model and real-time compensation is performed;

[0187] (5) Dual-laser redundancy recognition: Two sets of line laser emitters scan synchronously, and the recognition results of the two sets of scan data are compared. The average value of the two sets of data is used as the final recognition result to further improve the recognition accuracy.

[0188] S4.2, Optimization measures for recognition speed: In this embodiment, the following optimization measures are adopted to ensure recognition speed and adapt to the requirements of continuous follow-up conveying:

[0189] (1) Parallel computing optimization: The control system splits the tasks of signal preprocessing, boundary point extraction, contour fitting, pose calculation, etc. into 4 cores for parallel processing;

[0190] (2) Adaptive adjustment of scanning parameters: In this embodiment, the keel size is large and the conveying speed is slow. The system automatically adjusts the laser scanning frequency and scanning spacing to improve the recognition efficiency while ensuring the recognition accuracy.

[0191] (3) Data compression and transmission: The lossless compression algorithm is used to compress the laser echo signal and identification data, resulting in short data transmission time and ensuring that the transmission delay meets the requirements;

[0192] (4) Preloading mechanism: The panel size, keel size, height difference reference value H0=10mm, fitting parameters, etc. in this embodiment are preloaded into the memory of the central control system and directly called during recognition, shortening the recognition time;

[0193] (5) Invalid data filtering: In the signal preprocessing stage, invalid scanning data in the blank area of ​​the panel edge is filtered out in advance, and only valid data in the keel area is processed to reduce the amount of data processing and improve the recognition speed.

[0194] S4.3, Recognition Stability Optimization: In this embodiment, the following optimization measures are adopted to ensure that the recognition system operates stably in a workshop environment:

[0195] (1) Dustproof and splashproof design: The laser contour recognition system is equipped with a transparent, high-temperature resistant dustproof and splashproof protective cover to effectively prevent welding fumes and sparks from contaminating the laser transmitter and receiver lens, ensuring the stability of laser emission and reception;

[0196] (2) Vibration-resistant design: The laser contour recognition system is installed on the vibration damping bracket to reduce the impact of continuous conveyor line vibration on the equipment;

[0197] (3) Temperature compensation: The laser contour recognition system has a built-in temperature sensor to monitor the working temperature of the equipment in real time. In this embodiment, the ambient temperature of the workshop is 25℃, which is within the preset range (0~40℃). There is no need to adjust the laser power and signal processing parameters, and the recognition accuracy is stable.

[0198] (4) Fault self-diagnosis: The central control system monitors the working status of the laser contour recognition system in real time. In this embodiment, the equipment is operating normally and there is no fault alarm, ensuring production continuity.

[0199] Step S5: Simultaneous identification and batch processing of multiple keels.

[0200] In this embodiment, only one keel is set for each panel, so the steps of multi-keel synchronous scanning, pose calculation and batch output are not performed; however, through simulation test, when three keels are placed on the panel, the laser contour recognition system can scan all keels synchronously. The control system adopts a multi-threaded processing mode to process the echo signal of each keel independently, which verifies the adaptability and efficiency of multi-keel synchronous recognition of the present invention.

[0201] Step S6: Based on the identification results, weld the keel separately.

[0202] Each welding torch performs welding operations independently on each keel.

[0203] The embodiments described above are merely examples of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these modifications and improvements all fall within the protection scope of this application.

Claims

1. A self-adapting welding method for continuous feeding of stainless steel furniture modules, characterized by, Including the following steps: S2. Continuous conveying and contour acquisition; continuous follow-up conveying, using laser to perform narrow-band multi-point cross-section acquisition to obtain the laser signal of the keel in the continuous conveying state; S3. Determine the relative position; Based on the laser signal of the keel obtained in step S2, calculate the relative position between the keel and the panel using an algorithm. S6. Welding and fixing: Based on the relative positions of the keel and the panel obtained in S3, weld the keel and the panel.

2. The adaptive welding method for stainless steel furniture modules under continuous conveying as described in claim 1, characterized in that, Step S2 includes the following steps: S2.1 The panel and the keel placed on the panel are continuously conveyed at a constant speed. S2.2 When the panel and keel pass through the scanning area, the panel and keel are continuously laser scanned using a single-line narrow-band scanning method. S2.3 Acquire laser echo signals and determine the boundary points of the keel relative to the panel based on the obvious abrupt change in the laser echo signal data in the height direction; S2.4 Preprocess the laser signal to remove noise interference; S2.5 Dynamic synchronous acquisition to ensure that the laser scanning speed is consistent with the delivery speed.

3. The adaptive welding method for stainless steel furniture modules under continuous conveying as described in claim 2, characterized in that, Step S3 includes the following steps: S3.1 Extract and segment height abrupt change points to determine the boundary between the keel and the panel; S3.

2. Based on the mutation points, fit and reconstruct the keel outline; S3.

3. Based on the reconstructed keel outline, calculate the position parameters of the keel relative to the panel; S3.

4. Based on the reconstructed keel outline and centerline, calculate the keel placement parameters; S3.

5. Real-time output of recognition results.

4. The adaptive welding method for stainless steel furniture modules under continuous conveying as described in claim 3, characterized in that, The positional parameters in step S3.3 include lateral offset, longitudinal offset, distance between the edge of the keel and the edge of the panel, and the coordinates of the keel center. The calculation method in step S3.3 is as follows: taking the boundary of the panel as a reference, the coordinate difference between the center line of the keel and the center line of the panel is calculated based on the edge lines on both sides of the keel and the center line of the keel, and the horizontal offset Δx and the vertical offset Δy are obtained respectively; the distances L1 and L2 from the edge to the panel boundary are calculated based on the Y-axis coordinates of the edge lines on both sides of the keel and the Y-axis coordinates of the boundary on both sides of the panel respectively; and the center coordinates of the keel are determined based on the coordinates of the center line of the keel.

5. The adaptive welding method for stainless steel furniture modules under continuous conveying according to claim 3, characterized in that, The placement parameters in step S3.4 include the overall tilt angle of the keel and the local tilt angle of the keel. The method for calculating the overall tilt angle of the keel is to calculate the angle between the center line of the keel and the center line of the panel, and this angle is the overall tilt angle of the keel. The method for calculating the local skew angle of the keel is as follows: along the length of the keel, every 500mm is a segment, and the angle between the center line of each segment and the center line of the panel is calculated. This angle is the local skew angle of that segment. If the difference between the local skew angle and the overall skew angle is ≤0.1°, it is judged as qualified; if the difference between the local skew angle and the overall skew angle is 0.1°, it is judged as local distortion.

6. The adaptive welding method for stainless steel furniture modules under continuous conveying according to any one of claims 3-5, characterized in that, Step S6 includes: setting the starting position, lateral movement speed and direction of the welding torch according to the position parameters of the keel relative to the panel and the placement parameters of the keel, so that the weld seam is consistent with the contour of the keel.

7. The adaptive welding method for stainless steel furniture modules under continuous conveying according to claim 6, characterized in that, Before step S2, the following steps are also included: S1. Perform pre-processing of the panel and keel; Following step S3, the following steps are also included: S4. Continuously optimize recognition accuracy and speed; S5. Perform simultaneous identification and batch processing of multiple keels.

8. The adaptive welding method for stainless steel furniture modules under continuous conveying according to claim 7, characterized in that, Step S4 includes the following steps: S4.1 Continuous optimization of recognition accuracy, specifically including real-time calibration of height benchmark, optimized extraction of boundary points, and dynamic adjustment of the number of fitting nodes to optimize the fitting algorithm; S4.2 Continuous optimization of recognition speed, including parallel computing optimization, adaptive adjustment of scanning parameters, data compression and transmission, establishment of a preloading mechanism, and filtering of invalid data; S4.3, Optimization of recognition stability.

9. The adaptive welding method for stainless steel furniture modules under continuous conveying according to claim 7, characterized in that, Step S5 includes the following steps: S5.1, Multi-keel synchronous scanning; S5.2 Calculate the position and attitude of each multi-keel structure separately; S5.3 Batch processing of recognition results and matching with the corresponding welding torch welding path.

10. The adaptive welding method for stainless steel furniture modules under continuous conveying according to claim 7, characterized in that, In step S1, multiple keels are placed on the panel, with one surface of the panel facing down and the other surface facing up, and multiple keels are placed on it. The panel is placed on a conveyor belt and transported by the conveyor belt.

Images