Using the traditional magneto-optical detection methods, micro-weld defects parallel with the magnetic field direction may be overlooked. In order to overcome this, a non-destructive testing method based on magneto-optical imaging under a vertical combined magnetic field (VCMF) is proposed. To demonstrate this, the experimental results of the magneto-optical imaging of weld defects excited by a vertical combined magnetic field (VCMF) or parallel combined magnetic field (PCMF) are compared with those of traditional magnetic fields (constant magnetic field (CMF), alternating magnetic field (AMF), and rotating magnetic field (RMF)). It is found that the magneto-optical imaging under a VCMF can accurately detect weld defects of any shape and distribution. In addition, the center difference method is used to eliminate the influence of noise on the defect contour extraction of magneto-optical images, and the active contour of weld defects in the magneto-optical images is extracted. The results show that many noises can be identifiedby the robustness of the level set method, operating in low-pass filtering, so that much information that is usually lost can be retained.

The existing magneto-optical imaging of weld defects is carried out under a constant magnetic field or alternating magnetic field. The magneto-optical image (MOI) is easy to saturate under a constant magnetic field, resulting in the loss of defect information. It is very difficult to realize magneto-optical imaging of micro-defects under alternating magnetic fields. Therefore, magneto-optical imaging under a combined magnetic field is proposed in this article. Firstly, the effect of dynamic permeability on the law of magnetization of defects is analyzed. Then, through simulation experiments, the magnetic flux leakage law of defect samples under constant magnetic field, alternating magnetic field, and combined magnetic field are analyzed. Finally, the magneto-optical imaging experiment is used to study the imaging law of wire cutting grooves under different magnetic field excitation. The experimental results show that compared with constant magnetic field and alternating magnetic field, the difference of leakage magnetic field under combined magnetic field excitation is more obvious. This shows that using a combined magnetic field for imaging detection can effectively improve the disadvantage that it is difficult to realize an MOI of micro-defects under an alternating magnetic field.

Detecting weld surface defects under low-light conditions presents significant challenges due to limited illumination and subtle defect features. In this study, we propose a Visual Inspection System for Weld Surface Defects (VISFWSD). To support training and evaluation, we construct a dedicated Weld Surface Defect Data Set (WSDDS). Building upon the YOLOv8n architecture, we introduce DCE-YOLOv8n, a lightweight and high-performance model enhanced with Deformable Convolutional Networks (DCNv3) and the Convolutional Block Attention Module (CBAM) to improve multi-scale and fine-grained defect feature extraction. Furthermore, we adopt the Enhanced Intersection over Union (EIoU) loss function to address training sample imbalance and enhance localization accuracy. The proposed DCE-YOLOv8n model achieves superior performance across multiple evaluation metrics compared to baseline models. Specifically, it attains the highest mAP@0.5 of 92.9% and mAP@0.5:0.95 of 64.5%, outperforming YOLOv8n by 2.7 and 3.1 percentage points, respectively. Despite these accuracy gains, DCE-YOLOv8n maintains competitive computational efficiency, with 7.6 GFLOPS and 2.4 million parameters, both slightly lower than those of YOLOv8n (8.0 GFLOPS and 3.1 million parameters). Furthermore, it achieves an inference speed of 200 Frames Per Second (FPS), satisfying real-time processing requirements. These results demonstrate that DCE-YOLOv8n effectively balances detection accuracy, model size, and computational cost, making it well-suited for practical deployment in low-light weld defect inspection tasks. The proposed VISFWSD and DCE-YOLOv8n together provide an effective and efficient solution for weld surface defect detection in low-light environments.

Previous research on magneto-optical imaging of welding defects primarily focused on detecting artificially simulated known defects, and its industrial applicability remains uncertain. To optimize existing magneto-optical imaging technology, this study investigates non-destructive testing using magneto-optical imaging under vertical combined magnetic field(VCMF) excitation. First, A simulation model was established to analyze how alternating magnetic field(AMF) variations within a VCMF affect leakage magnetic flux from cross-shaped grooves, employing the controlled variable method. Subsequently, magneto-optical imaging experiments were conducted to capture images of radial cracks in spot welds under constant, alternating, and VCMFs. The study focuses on radial cracks from spot welding, which better represent industrial product requirements. By comparing VCMF magneto-optical imaging characteristics across different AMF strengths, analysis showed that alternating field intensity affects imaging in vertical composite fields. Results demonstrate that contrast variations in magneto-optical images are more pronounced under VCMFs than under single alternating fields. Furthermore, the relationship between AMF intensity and image contrast enhancement is nonlinear.

In the field of magneto-optical imaging nondestructive testing for welding defects, multi-angle detection of welding defects has already been achieved. However, research on automatic defect recognition and contour extraction remains insufficient. Therefore, to enable automatic detection of welding defects using magneto-optical imaging technology, it is essential to address the key issues of defect recognition and contour extraction in magneto-optical images. The dataset in this article includes five types of images: defect-free, lack-of-fusion, cracks, pits, and Weld reinforcement. Firstly, the Mask R-CNN detection method is used to perform defect recognition and contour segmentation on the original magneto-optical image dataset. The detection results indicate that the recognition rate of lack-of-fusion and Weld reinforcement in the original magneto-optical image is not high, and the recognition accuracy of pits and cracks is extremely low. Subsequently, the magneto-optical image dataset was preprocessed using the differential level set method, and the mask R-CNN algorithm was used to identify defect types and segment defect contours. Comparing the results of two experiments, it was found that the detection accuracy of the preprocessed dataset was higher, and the overall recognition accuracy increased by $30\%$.