Real three-dimensional volumetric ultrasonic imaging of steel weld defects: exceeding conventional inspection capabilities with a row–column addressed array

Three-dimensional (3D) characterization of weld defects is crucial for reliable integrity assessment and safe operation of critical engineering structures because two-dimensional (2D) ultrasonic imaging often provides incomplete defect information and is prone to missed detection. However, the application of 3D ultrasonic imaging to weld inspection is limited by high structural noise and the reliance of existing approaches on stacked 2D line scans, which fundamentally result in pseudo-3D representations rather than true volumetric imaging. To address these challenges, this paper proposes a 3D volumetric ultrasonic imaging method for steel weld inspection based on a Row–Column Addressing (RCA) array. A volumetric imaging framework is established by modeling ultrasonic propagation paths in a 3D coordinate system. To improve defect visibility in coarse-grained welds, nonconvex overlapping group sparse variational (NOGSV) denoising is applied to Full Matrix Capture (FMC) data, and vector coherence factor (VCF) weighting is further employed to enhance defect contrast in Total Focusing Method (TFM) imaging. Experimental validation on two steel butt-weld specimens containing representative defects demonstrates that the proposed method enables reliable 3D visualization of different defect types and reveals defect characteristics that are difficult or impossible to identify using conventional 2D ultrasonic imaging. High signal-to-noise ratio (SNR) volumetric images are achieved in weld structures, enabling accurate reconstruction of defect 3D geometry and spatial distribution. This study presents the first high-quality real 3D ultrasonic visualization of weld defects, offering an effective new approach for nondestructive testing and 3D characterization of welded structures.