Journal of Nondestructive Evaluation最新文献

{"title":"Frequency-Domain Optimized Laplacian Pyramid Network for Weld DR Image Enhancement","authors":"Lingling Liu,&nbsp;Zhenjiang Dong,&nbsp;Menghui Hu,&nbsp;Bo Ao,&nbsp;Chenxuan Hu,&nbsp;Zhipeng Liu,&nbsp;Lingfeng Wu,&nbsp;Yan Xiong","doi":"10.1007/s10921-026-01385-z","DOIUrl":"10.1007/s10921-026-01385-z","url":null,"abstract":"<div><p>Welding quality is critical to the safe operation of key structures in aerospace, nuclear power, and related fields. Digital radiography (DR), widely used as a non-destructive testing method, plays a crucial role in weld defect detection. However, DR images often suffer from blurred details and low contrast due to various factors such as imaging conditions, equipment noise, and material properties, leading to insufficient defect visibility. To address this, this paper proposes a frequency-domain optimized Laplacian pyramid network (FDO-LPN). The proposed method effectively leverages the multi-scale decomposition capability of the Laplacian pyramid to finely enhance each frequency layer. Specifically, low-frequency components are enhanced using a nonlinear encoder-decoder module to highlight weld structures and morphological features. High-frequency components are processed by a local–global transformation module, which simultaneously captures local pixel correlations and long-range dependencies, thereby generating adaptive masks to optimize high-frequency details across various frequency bands. Subsequently, a multi-scale reconstruction module progressively refines high-frequency information. Finally, a trilinear regulation module globally adjusts brightness, contrast, and grayscale distribution, further enhancing overall image quality. Experimental results show that FDO-LPN achieves the best performance on objective evaluation metrics, including peak signal-to-noise ratio, structural similarity, and brightness difference, reaching 44.33, 0.99, and 0.46, respectively, with PSNR improved by 12.10 over the second-best method. In practical applications to high-resolution DR images of small-diameter pipe and flat-plate welds from different real DR systems, FDO-LPN notably expands the grayscale dynamic range, effectively restores structural information, and improves defect visibility across various weld types, thereby demonstrating strong practical applicability and generalization ability for engineering deployment.</p></div>","PeriodicalId":655,"journal":{"name":"Journal of Nondestructive Evaluation","volume":"45 2","pages":""},"PeriodicalIF":2.4,"publicationDate":"2026-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148236616","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Thermal Load Feature Inverse Method for Honeycomb Sandwich Panel Based on Domain Adaptation Transfer Learning and ConvLSTM Neural Network","authors":"Yunhao Liu,&nbsp;Wenqi Du,&nbsp;Lingling Lu,&nbsp;Hongwei Song,&nbsp;Jixiang Song,&nbsp;Jie Le,&nbsp;Weijia Yan,&nbsp;Zhenwei Miao,&nbsp;Yacong Guo,&nbsp;Huan Tu,&nbsp;Lei Ren","doi":"10.1007/s10921-026-01381-3","DOIUrl":"10.1007/s10921-026-01381-3","url":null,"abstract":"<div><p>The identification of thermal load features is a prominent research area that aims to infer potential external thermal loads on the structure in a non-destructive way. Currently, data-driven methods are effective tools in this area. However, these methods often exhibit limited generalization because their accuracy depends heavily on the quality and coverage of the training data. To enhance the generalization of data-driven methods, this paper integrates domain adaptation transfer learning with a ConvLSTM neural network, thereby presenting a robust transfer learning framework. First, key parameters sensitive to thermal load features were extracted to build the dataset based on simulation and experimental results. Then, transfer learning was performed by embedding the ConvLSTM neural network within a domain adaptation method. The results demonstrate that, compared with a solely ConvLSTM model without transfer learning, the proposed method achieves lower prediction errors on the experiment dataset (diameter relative error decreased from 28.34% to 9.72%, power relative error decreased from 21.09% to 3.5%). Compared with fine-tuning transfer learning, the proposed method has better generalization capabilities. These improvement proves that the proposed method enhances generalization performance on datasets with significant differences in data distribution.</p></div>","PeriodicalId":655,"journal":{"name":"Journal of Nondestructive Evaluation","volume":"45 2","pages":""},"PeriodicalIF":2.4,"publicationDate":"2026-06-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148172444","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Sweeping Frequency Balanced Electromagnetic Field Distortion Effects under Buried Cracks","authors":"Jingkang Xiao,&nbsp;Fengli Zhang,&nbsp;Feng Qiu,&nbsp;Tiantian Liu,&nbsp;Jinjiang Wang","doi":"10.1007/s10921-026-01383-1","DOIUrl":"10.1007/s10921-026-01383-1","url":null,"abstract":"<div><p>Electromagnetic testing (ET) possesses the advantages of strong adaptability and low sensitivity to environmental conditions, enabling effective identification of surface cracks. However, its accuracy in detecting buried cracks is often constrained by the limitation of skin depth. To address this issue, this paper proposes sweeping frequency balanced electromagnetic technology (SFBET) and systematically investigates the disturbance characteristics of buried cracks on the sweeping balanced electromagnetic field, leading to the discovery and definition of a new characteristic parameter independent of skin depth-the amplitude zero-point frequency (AZF). First, a finite element model is established to simulate the disturbance behavior of buried cracks on the balanced electromagnetic field, and the formation mechanism of the amplitude zero-point (AZ) is revealed. Subsequently, an experimental platform for buried crack detection is constructed to validate the simulation results. Finally, the quantitative relationship between the crack buried depth and the AZ is analyzed. The results demonstrate that, under sweeping frequency excitation, the skin depth varies dynamically; buried cracks significantly disrupt the balanced state of the electromagnetic field, causing distortions in both its propagation direction and amplitude; during the sweeping process, a zero-amplitude point emerges near the crack edge, and the corresponding excitation frequency exhibits an exponential relationship with the crack buried depth, with a fitting correlation coefficient exceeding 0.99.</p></div>","PeriodicalId":655,"journal":{"name":"Journal of Nondestructive Evaluation","volume":"45 2","pages":""},"PeriodicalIF":2.4,"publicationDate":"2026-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148173597","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluation of Defect Depth Using Magnetic Memory Signals for Wire Arc Additive Manufactured Steel","authors":"Sheng Bao,&nbsp;Yan Li,&nbsp;Jingxuan Hong","doi":"10.1007/s10921-026-01386-y","DOIUrl":"10.1007/s10921-026-01386-y","url":null,"abstract":"<div><p>Wire arc additive manufacturing (WAAM) is widely used for fabricating large-scale metal structural components due to its high material utilization and short production cycle, yet its layer-by-layer deposition process easily induces defects such as cracks and circular holes. The depth of these defects directly determines the load-bearing capacity and service safety of WAAM components. This study aims to realize quantitative evaluation of defect depth in WAAM-fabricated low-carbon steel using metal magnetic memory (MMM) technology. A total of 16 rectangular WAAM specimens with artificial defects were prepared, covering two orientations (longitudinal and transverse) and four defect depths. Room-temperature incremental uniaxial tensile tests were performed, and a TSC-PC-16 magnetometer was used to measure the tangential and normal components of the residual magnetic field (RMF) at each loading step. The distribution and evolution of RMF signals were analyzed, and characteristic parameters were defined to establish the correlations between these parameters and defect depth. Results indicate that RMF signals exhibit obvious anomalies at defect locations. All characteristic parameters exhibit a significant positive linear correlation with defect depth. Longitudinal specimens exhibit stronger magnetic signal responses than transverse ones, which is attributed to higher stress concentration. Incremental loading amplifies the distortion of RMF signals by enhancing the coupling effect between external forces and defects. This study verifies the feasibility of MMM technology for quantitative evaluation of defect depth in WAAM components. It provides a non-contact and high-sensitivity non-destructive testing (NDT) method for the quality control of WAAM-fabricated structures, which is of great significance for safeguarding the safety of critical components in the aerospace, marine, and energy sectors.</p></div>","PeriodicalId":655,"journal":{"name":"Journal of Nondestructive Evaluation","volume":"45 2","pages":""},"PeriodicalIF":2.4,"publicationDate":"2026-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148173595","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An Ultrasonic Image Segmentation Detection Method Based on Automatic High-Resolution Imaging and Modified k-means Clustering Algorithm for CFRP Laminates with Delamination Defects","authors":"Xu Zhang,&nbsp;Yiting Shi,&nbsp;Congxiao Xu,&nbsp;Bo Li,&nbsp;Wei Li","doi":"10.1007/s10921-026-01382-2","DOIUrl":"10.1007/s10921-026-01382-2","url":null,"abstract":"<div><p>Ultrasonic C-scan imaging visually presents the shape and size information of defects in specimens, but its gate parameters selection and the structural and environmental noise in ultrasonic echo signals seriously affect the clarity of defect images. The <i>k</i>-means clustering algorithm can effectively distinguish between defect and background areas in the original image, but still suffers from problems such as the inability of the distance metric to differentiate sample attribute differences, the sensitivity to the choice of initial cluster centers, and the need for manual predefinition of the <i>k</i> value. Hence, a new method of ultrasonic image segmentation is proposed through improving the variational mode decomposition (VMD) method, introducing the autocorrelation algorithm, and modifying the <i>k</i>-means clustering algorithm, aiming to improve the segmentation accuracy of defect and background areas in Carbon Fiber Reinforced Polymer (CFRP) laminates. The VMD method is improved by optimizing the modal decomposition number and penalty factor, and then is used to decompose the ultrasonic echo signal. The Euclidean distance is adopted to distinguish between denoised defect and non-defect signals, achieving automatic high-resolution imaging of defects. The <i>k</i>-means clustering algorithm is modified by using the Mahalanobis distance, <i>k</i>-means +  + algorithm, and elbow method, effectively segmenting defect and background areas. The phased array equipment is employed to conduct ultrasonic testing on CFRP laminates with delamination defects of various shapes, depths, and sizes, and the experimental results are compared with those of the proposed ultrasonic image segmentation detection method. The results show that the improved VMD and autocorrelation algorithm significantly reduces the interference of background noise and improves the resolution in automatic imaging. And, the improved <i>k</i>-means clustering algorithm effectively distinguishes between defect damage and target background areas, and reduces manual parameters dependence during clustering.</p></div>","PeriodicalId":655,"journal":{"name":"Journal of Nondestructive Evaluation","volume":"45 2","pages":""},"PeriodicalIF":2.4,"publicationDate":"2026-05-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148173596","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-View Total Focusing Method with Dempster-Shafer Fusion for Nondestructive Testing of Socket Weld","authors":"Shitong Liang,&nbsp;Jingpin Jiao,&nbsp;Changhua Chen,&nbsp;Jianwei Zhang,&nbsp;Tao Liu","doi":"10.1007/s10921-026-01380-4","DOIUrl":"10.1007/s10921-026-01380-4","url":null,"abstract":"<div><p>Accurate detection of internal defects in geometrically complex socket welds is critical for structural integrity but challenging for conventional ultrasonic methods. This study proposes a novel multi-view total focusing method (TFM) approach enhanced by Dempster-Shafer (DS) theory for improved socket weld inspection. An ultrasonic measurement model incorporating sensitivity mapping is first established to guide mode selection and optimize phased array parameters for socket weld configurations. Multi-view TFM images, capturing diverse wave propagation paths (direct, half-skip, full-skip), are then generated. To address conflicts and uncertainties inherent in fusing these complementary views, an improved DS theory framework is introduced for robust data fusion. Experimental results demonstrate that the proposed DS-fused multi-view TFM significantly suppresses artifacts arising from complex geometry and mode conversions compared to single-mode views, while enhancing defect detectability and imaging clarity for reliable socket weld assessment. This integrated approach provides a more effective and reliable solution for non-destructive testing of socket weld joints.</p></div>","PeriodicalId":655,"journal":{"name":"Journal of Nondestructive Evaluation","volume":"45 2","pages":""},"PeriodicalIF":2.4,"publicationDate":"2026-05-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148172930","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Study of Accurate Measurement of Defect Depth Based on Digital Radiography","authors":"Jingdong Hu,&nbsp;Yuze Guo,&nbsp;Yuntao Li,&nbsp;Huaibei Dai,&nbsp;Xuemei Liu,&nbsp;ShuangSheng Yan","doi":"10.1007/s10921-026-01338-6","DOIUrl":"10.1007/s10921-026-01338-6","url":null,"abstract":"<div><p>Radiography plays a crucial role in nondestructive testing, particularly for detecting volumetric defects. While the characterization of these defects is generally accurate, measuring their size along the direction of the X-ray remains challenging. This paper explores the relationship between image gray value, penetration thickness, process parameters, and the linear attenuation coefficient based on radiographic theory. We utilize numerical integration and curve fitting to express the linear attenuation coefficient as a function of tube voltage and penetration thickness, establishing a mathematical model that links penetration thickness, tube voltage, and image gray levels. We then design experiments using a steel plate to determine the unknown coefficients in the model, leading to an accurate calculation model for defect size in the direction of X-ray penetration. The model's accuracy is validated through digital radiography and CT experiments conducted on small-diameter pipe welds, with measurement errors ranging from 8.16% to 8.73%</p></div>","PeriodicalId":655,"journal":{"name":"Journal of Nondestructive Evaluation","volume":"45 2","pages":""},"PeriodicalIF":2.4,"publicationDate":"2026-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148009354","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Robust Circle Detection and Dimensional Measurement of Industrial Parts Using the RCIMI Framework","authors":"Dhiviya Bharathi Duraimanickam,&nbsp;Joseph Abraham Sundar Koilraj,&nbsp;Muthaiah Rajappa","doi":"10.1007/s10921-026-01379-x","DOIUrl":"10.1007/s10921-026-01379-x","url":null,"abstract":"<div><p>Ensuring dimensional accuracy and structural correctness of metallic spare parts is essential for maintaining product reliability in industrial environments. However, inspection of reflective, non-planar components is challenging due to specular highlights, broken edges, inconsistent gradients, and perspective-induced scale variations across hole positions at different depths, which degrade the performance of conventional circle-detection methods. To address these limitations, this research proposes RCIMI (Robust Circular Identification and Metrological Inspection), a lightweight and interpretable computer-vision framework for classification and dimensional measurement of industrial parts containing circular features. RCIMI integrates a reflection-aware preprocessing strategy to stabilize edges on shiny metallic surfaces, followed by a physically guided Hough parameter stabilization mechanism that constrains radius ranges and accumulator thresholds based on tolerance specifications, improving detection repeatability under varying orientations and illumination conditions. Detected circles are refined through geometric filtering and used for part identification via hole count, followed by a pixel-to-millimetre calibration model that compensates for depth-dependent scale variation arising from non-planar tilted geometry, enabling precise dimensional verification against tolerance limits. Experimental evaluation on real-time industrial data with a mean absolute measurement error of 0.03 mm and external images demonstrates a highly stable circle localization, enabling reliable classification of visually similar components. The results confirm that RCIMI provides a practical, training-free solution for automated quality inspection of reflective non-planar components in manufacturing environments where interpretability, metrological traceability, and computational efficiency are critical.</p></div>","PeriodicalId":655,"journal":{"name":"Journal of Nondestructive Evaluation","volume":"45 2","pages":""},"PeriodicalIF":2.4,"publicationDate":"2026-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148009353","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Nondestructive Evaluation of Recrystallization in Ni-Based Single-Crystal Superalloys Using Synchrotron Radiation Diffraction Extinction Imaging","authors":"Zhimao Wang,&nbsp;Gang Li,&nbsp;Jie Zhang,&nbsp;Yanping Wang,&nbsp;Rui Sun,&nbsp;Nan Li,&nbsp;Xi Wang,&nbsp;Zengyao Xing","doi":"10.1007/s10921-026-01374-2","DOIUrl":"10.1007/s10921-026-01374-2","url":null,"abstract":"<div><p>The present study proposes a methodology for the non-destructive testing of recrystallization phenomena in nickel-based single-crystal superalloys, utilizing synchrotron radiation diffraction extinction imaging technology. Recrystallization, a critical defect in turbine blades, has been shown to degrade the material's mechanical properties. However, traditional techniques such as electron backscatter diffraction still face challenges when inspecting the internal structure of the material. The use of high-energy synchrotron X-rays (50–300 keV) allows for the simultaneous capture of both transmission extinction spots and diffraction spots during rotational scanning. This method has been demonstrated to generate recrystallization grain images with a resolution of 0.5 μm. Through meticulous analysis of spot spacing and angles, the crystallographic families [111] can be identified. The validity of these results was subsequently verified by comparing them with EBSD results. This method provides a rapid, non-destructive testing approach for aerospace quality control, addressing a critical technological gap in turbine blade manufacturing.</p></div>","PeriodicalId":655,"journal":{"name":"Journal of Nondestructive Evaluation","volume":"45 2","pages":""},"PeriodicalIF":2.4,"publicationDate":"2026-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148008915","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"SE-Attention-1D-CNN: A Deep Learning Framework for Nondestructive Mooney Viscosity Prediction in Natural Rubber Using Vis–NIR Hyperspectral Imaging","authors":"Xu Zhang,&nbsp;Ming Liu,&nbsp;Yun Li,&nbsp;Lijun Yun,&nbsp;Zaiqing Chen","doi":"10.1007/s10921-026-01362-6","DOIUrl":"10.1007/s10921-026-01362-6","url":null,"abstract":"<div><p>Conventional “sample-destructive—offline” methods for determining the Mooney viscosity (ML) of natural rubber (NR) are time-consuming, costly, and unsuitable for on-site process control. In this study, we propose an end-to-end, non-destructive framework that directly predicts ML from raw NR sheets using visible–near-infrared hyperspectral imaging (Vis–NIR HSI, 420–870 nm, 150 sampled bands at 2 nm intervals) coupled with a deep regression network. A total of 540 NR specimens were imaged under strict darkroom conditions. We developed a one-dimensional convolutional neural network augmented with a Squeeze-and-Excitation (SE) spectral attention module—termed SE-Attention-1D-CNN—to extract multi-scale spectro-chemical features while adaptively reweighting wavelength channels associated with viscosity-related moieties. The model was trained using an L2-regularized mean squared error (MSE) loss, the Adam optimizer, four-fold cross-validation, and early stopping to ensure generalization.The proposed network achieved an average correlation coefficient(R) of 0.873, coefficient of determination (R2) of 0.762, root mean square error (RMSE) of 4.037, mean absolute error (MAE) of 3.364, and residual predictive deviation (RPD) of 2.003. Ablation experiments further demonstrated the effectiveness of the SE module. Specifically, after removing the SE module, the CNN model showed inferior performance, with R and R² decreasing to 0.827 and 0.684, respectively, RMSE and MAE increasing to 4.645 and 3.775, respectively, and RPD dropping to 1.761. These results highlight the crucial role of attention-driven spectral weighting in improving predictive accuracy and model robustness. By embedding the SE mechanism within a 1D CNN, this study enables high-precision, non-destructive ML prediction and provides a novel approach for the intelligent grading of natural rubber in industrial settings.</p></div>","PeriodicalId":655,"journal":{"name":"Journal of Nondestructive Evaluation","volume":"45 2","pages":""},"PeriodicalIF":2.4,"publicationDate":"2026-05-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148009002","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}