Search Results (66)

A novel closed loop visual positioning system, termed YOLO–LaserGalvo (YLGS), is proposed for precise localization of welding torch tips in industrial welding automation. The proposed system integrates a monocular camera, an infrared laser distance sensor with a galvanometer scanner, and a customized deep learning detector based on an improved YOLOv11 model. In operation, the vision subsystem first detects the approximate image location of the torch tip using the YOLOv11-based model. Guided by this detection, the galvanometer steers the IR laser beam to that point and measures the distance to the torch tip. The distance feedback is then fused with the vision coordinates to compute the precise 3D position of the torch tip in real-time. Under complex illumination, the proposed YLGS system exhibits superior robustness compared with color-marker and ArUco baselines. Experimental evaluation shows that the system outperforms traditional color-marker and ArUco-based methods in terms of accuracy, robustness, and processing speed. This marker-free method provides high-precision torch positioning without requiring structured lighting or artificial markers. Its pedagogical implications in engineering education are also discussed. Potential future work includes extending the method to full 6-DOF pose estimation and integrating additional sensors for enhanced performance. Full article

The concept of symmetry is a fundamental principle in various scientific and engineering fields, including welding technology. In the context of this paper, symmetry could play a role in optimizing the welding trajectory. Welding trajectory point detection relies on machine vision perception and intelligent algorithms to extract welding trajectory, which is crucial for the automatic welding of steel parts. However, in practice, sieve-net welding still relies on manual or semi-automatic operations, which have limitations, such as fixed positions and sizes, making it unsafe and inefficient. This paper proposes a 2D skeleton extraction algorithm for detecting weld joints in a sieve-net point cloud. First, the algorithm applies principal component analysis (PCA) to transform the point cloud and projects it into a 2D image with minimal information loss. Second, the expansion corrosion method is then employed to enhance the connectivity and refinement of the sieve-net mesh to serve the extraction of 2D skeleton. Third, the algorithm extracts the skeleton of the sieve-net grid and detects solder points. The average detection accuracy of the proposed algorithm is over 95%, which confirms its feasibility and practical application value in sieve-net welding. Full article

Accurate weld defect detection is critical for ensuring structural safety and evaluating welding quality in industrial applications. Manual inspection methods have inherent limitations, including inefficiency and inadequate sensitivity to subtle defects. Existing detection models, primarily designed for natural images, struggle to adapt to the characteristic challenges of weld X-ray images, such as high noise, low contrast, and inter-defect similarity, particularly leading to missed detections and false positives for small defects. To address these challenges, a multi-dimensional feature fusion model (MADet), which is a multi-branch deep fusion network for weld defect detection, was proposed. The framework incorporates two key innovations: (1) A multi-scale feature fusion network integrated with lightweight attention residual modules to enhance the perception of fine-grained defect features by leveraging low-level texture information. (2) An anchor-based feature-selective detection head was used to improve the discrimination and localization accuracy for five typical defect categories. Extensive experiments on both public and proprietary weld defect datasets demonstrated that MADet achieved significant improvements over the state-of-the-art YOLO variants. Specifically, it surpassed the suboptimal model by 7.41% in mAP@0.5, indicating strong industrial applicability. Full article

This study aims to address the challenges posed by uneven energy amplitude and a low signal-to-noise ratio (SNR) in the total focus imaging of coarse-crystalline elastic anisotropic materials. A novel method for acoustic field correction vector-coherent total focus imaging, based on the materials’ properties, is proposed. To demonstrate the effectiveness of this method, a test specimen, an austenitic stainless steel nozzle weld, was employed. Seven side-drilled hole defects located at varying positions and depths, each with a diameter of 2 mm, were examined. An ultrasound simulation model was developed based on material backscatter diffraction results, and the scattering attenuation compensation factor was optimized. The acoustic field correction function was derived by combining acoustic field directivity with diffusion attenuation compensation. The phase coherence weighting coefficients were calculated, followed by image reconstruction. The results show that the proposed method significantly improves imaging amplitude uniformity and reduces the structural noise caused by the coarse crystal structure of austenitic stainless steel. Compared to conventional total focus imaging, the detection SNR of the seven defects increased by 2.34 dB to 10.95 dB. Additionally, the defect localization error was reduced from 0.1 mm to 0.05 mm, with a range of 0.70 mm to 0.88 mm. Full article

In response to the challenges associated with weld treatment during the on-site corrosion protection of hydraulic steel gates, this paper proposes a method utilizing a magnetic adsorption climbing robot to perform corrosion protection operations. Firstly, a magnetic adsorption climbing robot with a multi-wheel independent drive configuration is proposed as a mobile platform. The robot body consists of six joint modules, with the two middle joints featuring adjustable suspension. The joints are connected in series via an EtherCAT bus communication system. Secondly, the kinematic model of the climbing robot is analyzed and a PID trajectory tracking control method is designed, based on the kinematic model and trajectory deviation information collected by the vision system. Subsequently, the proposed kinematic model and trajectory tracking control method are validated through Python3 simulation and actual operation tests on a curved trajectory, demonstrating the rationality of the designed PID controller and control parameters. Finally, an intelligent software system for weld defect detection based on computer vision is developed. This system is demonstrated to conduct defect detection on images of the current weld position using a trained model. Full article

Background/Objectives: Additive manufacturing (AM) enables the production of cementless acetabular cups with porous surfaces that facilitate early osseointegration. Directed energy deposition (DED), a form of AM, allows the direct welding of porous structures onto metal substrates without requiring a vacuum environment, offering advantages over conventional powder bed fusion methods. Despite growing interest in DED, no prospective clinical studies evaluating DED-based acetabular components have been published to date. This study assessed short-term outcomes of a DED-based 3D-printed acetabular cup in total hip arthroplasty (THA). Methods: A total of 120 patients who underwent primary cementless THA using the Corentec Mirabo Z^® acetabular cup were prospectively enrolled. Among them, 124 hips from 100 patients who had completed a minimum of 24 months of follow-up were included in the analysis. Clinical outcomes were assessed using the Harris hip score (HHS), WOMAC, EQ-5D-5L, and pain NRS. Radiographic evaluation included measurements of cup position, osseointegration, and detection of interfacial or polar gaps on CT and plain radiographs. Implant-related complications were also recorded. Results: At a mean follow-up of 34.6 months, the implant survival rate was 99.3%, with one revision due to suspected osseointegration failure. The HHS improved from 56.6 to 91.4 at 24 months, and the NRS decreased from 6.2 to 1.1 (both p < 0.001). Interfacial gaps were observed in 58.1% of cases on CT, though most were <1 mm and not clinically significant. Common postoperative issues included greater trochanteric pain syndrome, squeaking, and iliotibial band tightness, all of which were resolved with conservative treatment. Conclusions: DED-based 3D-printed acetabular cups demonstrated favorable short-term clinical and radiographic outcomes, with high survivorship and reliable early osseointegration in cementless THA. Full article

Traditional offline programming has limitations for large parts with significant machining or assembly deviations. This study proposes a 3D scanning-assisted method that generates accurate STereoLithography (STL) models and enables multi-layer multi-bead welding trajectory planning for large intersecting line workpieces. The proposed framework implements a robust STL model processing pipeline incorporating Random Sample Consensus (RANSAC)-based cylindrical approximation, cross-sectional slicing, and automated feature detection to achieve high-precision groove feature recognition. For asymmetric variable-section grooves, a multi-layer and multi-pass path-planning algorithm based on template affine projection transformation is developed to ensure accurate deposition of welds along complex geometric contours. Experimental validation demonstrates sub-millimeter trajectory accuracy (positional errors < 1.0 mm), meeting stringent arc welding specifications and substantially expanding the applicability of offline programming systems. Full article

Probabilistic Boolean Networks (PBN) can model the dynamics of complex biological systems, as well as other non-biological systems like manufacturing systems and smart grids. In this proof-of-concept paper, we propose a PBN architecture with a learning process that significantly enhances fault and failure prediction in manufacturing systems. This concept was tested using a PBN model of an ultrasound welding process and its machines. Through various experiments, the model successfully learned to maintain a normal operating state. Leveraging the complex properties of PBNs, we utilize them as an adaptive learning tool with positive feedback, demonstrating that these networks may have broader applications than previously recognized. This multi-layered PBN architecture offers substantial improvements in fault detection performance within a positive feedback network structure that shows greater noise tolerance than other methods. Full article

In pipeline all-position welding processes, laser scanning provides critical geometric data of width-changing bevel morphology for welding torch swing control, yet conventional second-order derivative zero methods often yield pseudo-inflection points in practical applications. To address this, a third-order derivative weighted average threshold algorithm was developed, integrating image denoising, enhancement, and segmentation pre-processing with cubic spline fitting for precise bevel contour reconstruction. Bevel pixel points were captured by the laser sensor as inputs through the extracted second-order derivative eigenvalues to derive third-order derivative features, applying weighted threshold discrimination to accurately identify inflection points. Dual-angle sensors were implemented to synchronize laser-detected bevel geometry with real-time torch swing adjustments. Experimental results demonstrate that the system achieves a steady-state error of only 1.645% at the maximum swing width, a dynamic response time below 50 ms, and torch center trajectory tracking errors strictly constrained within ±0.1 mm. Compared to conventional methods, the proposed algorithm improves dynamic performance by 20.6% and exhibits unique adaptability to narrow-gap V-grooves. The results of these studies confirmed the ability of the method to provide real-time, accurate control for variable-width weld tracking, forming a swing-width adaptive control system. Full article

The precision and stability of seam position detection are critical for single-square-groove weld seams formed using two thin metal plates. However, traditional methods, such as structured laser light imaging, struggle with narrow seams that lack misalignment and have high reflectivity, while non-structured light approaches are prone to welding light interference and speckle noise. To overcome these challenges, we propose a versatile optical design that leverages differential illumination to generate differential phase contrast (DPC) images. By processing images captured under differential illumination, the DPC method notably enhances seam edge contrast and suppresses welding light noise, improving the detection robustness and reliability. This approach provides a promising solution for high-precision weld seam detection in challenging environments. Full article

The deep learning architecture YOLO (You Only Look Once) has demonstrated its superior visual detection performance in various computer vision tasks and has been widely applied in the field of automatic surface defect detection. In this paper, we propose a lightweight YOLOv8-based method for the quality inspection of car body welding spots. We developed a TA-YOLOv8 network structure which has an improved Task-Aligned (TA) head detection, designed to handle a small sample size, imbalanced positive and negative samples, and high-noise characteristics of Body-in-White welding spot data. By learning with fewer parameters, the model achieves more efficient and accurate classification. Additionally, our algorithm framework can perform anomaly segmentation and classification on our open-world raw datasets obtained from actual production environments. The experimental results show that the lightweight module improves the processing speed by an average of 2.8%, with increases in detection the mAP@50-95 and recall rate of 1.35% and 0.1226, respectively. Full article

As power system equipment gradually ages, the automated disassembly of transformers has become a critical area of research to enhance both efficiency and safety. This paper presents a transformer disassembly system designed for power systems, leveraging multimodal perception and collaborative processing. By integrating 2D images and 3D point cloud data captured by RGB-D cameras, the system enables the precise recognition and efficient disassembly of transformer covers and internal components through multimodal data fusion, deep learning models, and control technologies. The system employs an enhanced YOLOv8 model for positioning and identifying screw-fastened covers while also utilizing the STDC network for segmentation and cutting path planning of welded covers. In addition, the system captures 3D point cloud data of the transformer’s interior using multi-view RGB-D cameras and performs multimodal semantic segmentation and object detection via the ODIN model, facilitating the high-precision identification and cutting of complex components such as windings, studs, and silicon steel sheets. Experimental results show that the system achieves a recognition accuracy of 99% for both cover and internal component disassembly, with a disassembly success rate of 98%, demonstrating its high adaptability and safety in complex industrial environments. Full article

The anti-corrosion layer of the pipe provides corrosion resistance and extends the lifespan of the whole pipeline. Heat-shrinkable tape is primarily used as the pipeline joint coating material bonded to the pipeline weld connection position after heating. Delineating the bonding strength and assessing the quality of the bonded structure is crucial for pipeline safety. A detection technology based on nonlinear ultrasound is presented to quantitatively evaluate the bonding strength of a steel-EVA-polyethylene three-layer annulus bonding structure. Using the Floquet boundary condition, the dispersion curves of phase velocity and group velocity for a three-layer annulus bonding structure are obtained. Additionally, wave structure analysis is employed in theoretical study to choose guided wave modes that are appropriate for detection. In this paper, guided wave amplitude, frequency attenuation, and nonlinear harmonics are used to evaluate the structural bonding strength. The results reveal that the detection method based on amplitude and frequency attenuation can be used to preliminarily screen the poor bonding, while the acoustic nonlinear coefficient is sensitive to bonding strength changes. This study introduces a comprehensive and precise pipeline joint bonding strength detection system leveraging ultrasonic-guided wave technology for pipeline coating applications. The detection system determines the bonding strength of bonded structures with greater precision than conventional ultrasonic inspection methods. Full article

As a technological advantage of ultrasonic non-destructive testing, fully focused imaging can accurately feedback the defective characteristics of the inspected object, greatly improving the detection efficiency. This article aims to address the challenges of outdated and low detection rates in the detection technology of branch pipe fillet welds. The full matrix acquisition (FMC) and total focus method (TFM) ultrasonic detection technology are used for detection and defect image feature analysis. Firstly, a multi-mode, fully focused real-time imaging software system was developed to address the specificity of the detection object; secondly, a phased array detection system based on 64 elements was constructed; finally, a region wavelet denoising method based on TFM images was proposed to solve the problem of artifacts caused by poor coupling; and based on the feature extraction method for a minimum rectangle, we analyzed the size, position, angle, and other information regarding defects. Through experiments, it has been found that this technology can effectively improve the detection efficiency of branch pipe weld defects, with a detection rate of 100%. Based on the partition fusion denoising method, the defect imaging quality can be further improved; at the same time, based on the feature extraction method, the error is 0.1 mm, the length range of various defects is 2.3 mm–6.3 mm, the width range is 0.6 mm–0.8 mm, and the angle range is 52°–75°, which can provide an application basis for the localization, classification, and risk assessment of corner weld defects in branch pipes. Full article

Laser vision seam tracking enhances robotic welding by enabling external information acquisition, thus improving the overall intelligence of the welding process. However, camera images captured during welding often suffer from distortion due to strong noises, including arcs, splashes, and smoke, which adversely affect the accuracy and robustness of feature point detection. To mitigate these issues, we propose a feature point extraction algorithm tailored for weld images, utilizing an improved Deeplabv3+ semantic segmentation network combined with EfficientDet. By replacing Deeplabv3+’s backbone with MobileNetV2, we enhance prediction efficiency. The DenseASPP structure and attention mechanism are implemented to focus on laser stripe edge extraction, resulting in cleaner laser stripe images and minimizing noise interference. Subsequently, EfficientDet extracts feature point positions from these cleaned images. Experimental results demonstrate that, across four typical weld types, the average feature point extraction error is maintained below 1 pixel, with over 99% of errors falling below 3 pixels, indicating both high detection accuracy and reliability. Full article