ndt_1924-2025

Non Destructive Testing (NDT) 2024 - 2025

Click on the octopus to return to the top of the page

01 - Investigation of offshore landslides impact on bucket foundations using a coupled SPH–FEM method

Authors: Zehao Wang, Defeng Zheng, Xingsen Guo, Zhongde Gu, Yueqiang Shen, & Tingkai Nian This study explains the impact of submarine landslides on underwater structures, focusing on bucket foundations. It introduces a fluid-structure coupling system using the coupled Smoothed Particle Hydrodynamics (SPH)-Finite Element Method (FEM) to analyze the displacement response of these foundations during landslide impacts. The research aims to provide a new numerical simulation approach and offers insights into underwater structures' dynamic response and failure mechanisms under extreme conditions.

13 - Evaluation of damage performance in offshore floating photovoltaics-based hydrogen production system due to potential hydrogen release

Authors: Hao Wang, Kan Wang, Xiaolei Liu, Yang Liu, Zhijia Qian, and Sheng Ding This study explains the development and analysis of an offshore floating photovoltaics (FPV)-based hydrogen production system. It addresses the challenges, particularly safety concerns, associated with hydrogen storage in such systems. It involves creating a numerical 3D model to investigate the characteristics of accidental damage and hydrogen release under different offshore wind conditions. It also explores the dynamic development of hydrogen dispersion and evaluates the evolution of thermal hazards and damage in the marine environment.

Authors: Oleg Gaidai, Yu Cao, Yan Zhu, Fuxi Zhang, and Hongchen Li A new reliability method suitable for multi-dimensional structural dynamics is introduced in this study, using the Gaidai multivariate method on an offshore Jacket platform in Bohai Bay. It estimates collapse risks under environmental stress, addressing challenges of high dimensionality and nonlinear correlations. This method is essential for offshore engineers during design to assess operational failure, damage, or hazards.

05 - Multivariate Risk Assessment for Offshore Jacket Platformsby Gaidai Reliability Method.

Authors: Jingzhou Xin, Guangjiong Tao, Qizhi Tang, Fei Zou, Chenglong Xiang This study proposes a new damage identification method using Swin Transformer and continuous wavelet transform (CWT). It converts structural vibration data into a time-frequency diagram to capture damage characteristics. The Swin Transformer extracts damage information from this diagram. The method's accuracy is analyzed with various sample lengths and noise levels, confirming its robustness. Laboratory tests show it recognizes damage with 99. 6% accuracy for single damage and 99. 0% for multiple damages.

08 - Structural damage identification method based on Swin Transformer and continuous wavelet transform.

Authors: Andrei Nazarov, Tatiana Yurasova, and Andrey Marshakov This review examines how different aqueous electrolytes impact hydrogen absorption and self-corrosion in magnesium anodes. It discusses historical and recent studies on mechanisms of self-corrosion and hydrogen evolution under specific effects. The focus is on magnesium hydride formation and oxidation during active dissolution and how anodic dissolution occurs through protective films in other conditions. It also looks at factors influencing hydride formation and self- corrosion.

17 - Hydrogen Absorption and Self-Corrosion of Mg Anode: Influence of Aqueous Electrolyte Species

Authors: Yuheng Chen, Haicheng Zhang, Weisheng Zou, Haihua Zhang, Bin Zhou, Daolin Xu This study proposes a new ROV-based deep-sea mining system to address slippage and sinking issues of traditional vehicles. A learning-based control strategy is introduced, utilizing a nonparametric learning method to improve path-tracking performance.

23 - Dynamic modeling and learning based path tracking control for ROV-based deep-sea mining vehicle.

Authors: Beate Oswald-Tranta Inductive thermography is a strong inspection method for finding defects in metals. The technique has improved over decades from lab tests to industry use. This paper reviews its theory and key technical aspects. While it can find various defects, the focus is on detecting surface cracks in metals, with examples of current industrial applications.

26 - Inductive thermography – review of a non-destructive inspection technique for surface crack detection

Authors: Mohammadebrahim Bajgholi, Gilles Rousseau, Denis Thibault, Simon Francoeur. This study evaluates advanced Non-Destructive Testing (NDT) techniques for detecting flaws in welded repairs of martensitic stainless steel plates, focusing on Francis turbine runners. Phased Array Ultrasonic Testing (PAUT), Total Focusing Method (TFM), and Eddy Current Array (ECA) technologies were examined to assess repair integrity and identify flaws such as lack of fusion (LOF). The findings highlight the necessity of integrating advanced NDT techniques to ensure the structural integrity of critical components, particularly in high-stress applications such as turbine runners.

30 - Advanced Non-Destructive Testing Techniques for Welded Joint Repairs: A Case Study on Improving Inspection Reliability and Structural Integrity Assessment

Authors: Luca Paterlini, Andrea Marinelli, Andrea Brenna and Marco Ormellese This document proposes a risk map to manage phenomena linked to the fact that carbon steel structures that transport hydrocarbons have systems to prevent corrosion, including cathodic protection and insulating coatings, and that electrical interference, both AC and DC, can cause corrosion even with protection applied.

31 - Protection Criteria of Cathodically Protected Pipelines Under AC Interference.

32 - Natural van der Waals canalization lens for non-destructive nanoelectronic circuit imaging and inspection

Authors: Qingdong Ou, Shuwen Xue, Weiliang Ma, Jiong Yang, Guangyuan Si , Lu Liu, Gang Zhong, Jingying Liu, Zongyuan Xie, Ying Xiao, Kourosh Kalantar-Zadeh, Xiang Qi, Peining Li, Zhigao Dai, Huanyang Chen, Qiaoliang Bao Optical inspection is important in semiconductor wafer manufacturing for analyzing surfaces and defects. Traditional methods struggle with resolution and detecting buried structures. The authors suggest a new approach using a van der Waals canalization lens from a- MoO 3 crystals, achieving 15 nm resolution for both surface and buried imaging. This method offers better imaging without the issues faced by other lenses, enabling high-resolution inspection of buried nanoscale circuits, crucial for future semiconductor manufacturing.

Authors: Andrzej Katunin, Krzysztof Dragan, Marko Nagode, Krzysztof Lis, Kamil Joszko, Adam Cholewa, Jernej Klemenc, Simon Oman, Paweł Zak, Piotr Synaszko Hidden corrosion, if not found quickly, can greatly impact the strength and safety of aircraft. Enhancing non-destructive testing (NDT) methods is crucial for safe aircraft operation. The D-Sight technique is commonly used for detecting this corrosion, valued in the aerospace industry for its fast and low-cost inspections. One of the drawbacks

33 - Towards quantification of hidden corrosion using D-Sight non destructive testing technique

Authors: Emmanuella Onyinye Nwulu, Tari Yvonne Elete, Friday Emmanuel Adikwu, and Fidelis Othuke Onyeke The durability of coatings is crucial for protecting infrastructure from damage. This study suggests a framework using advanced non-destructive testing (NDT) technologies for inspecting coating durability. It incorporates methods like ultrasonic testing and data analytics to find defects and predict lifespan. Case studies show this approach improves decision-making and saves costs. The framework promotes better inspection practices across different sectors for enhanced performance and sustainability.

27 - Advancing inspection techniques for coating durability: A framework for integrating non-destructive testing technologies.

Authors: Jun Ito, Yudai Igarashi, Ryota Odagiri, Shigetaka Suzuki, Hiroshi Wagatsuma, Kazuhiro Sugiyama, and Mikihiko Oogane The magnetic hammer test (MHT) is a nondestructive inspection method that uses a tunnel magnetoresistance (TMR) sensor to detect weak magnetic fields generated by natural pipe vibration. The method can accurately evaluate wall thickness with tens of microns and underwater with less than 100 µm accuracy. This noncontact, fast, and accurate method contributes to unmanned, manpower-saving nondestructive testing (NDT) in the future.

03 - Evaluation of Pipe Thickness by Magnetic Hammer Test with a Tunnel Magnetoresistive Sensor

Authors: Jianghai He This essay discusses the practical application of remote- operated underwater vehicles for inspecting dams and bridges, addressing safety issues like leakage, erosion, and cracks. It highlights the advantages of these vehicles compared to traditional divers, their operation in low- visibility water, stabilizing positions in turbulence, and overcoming cable entanglement issues. The paper also provides an outlook for their development in underwater detection and reinforcement to address future maintenance challenges.

04 - Introduction Remote-operated Underwater Vehicle for Inspection of Underwater Structure

Authors: Manuela Ammann Switzerland's hydropower plants are the primary source of renewable energy, accounting for 57% of electricity production. However, high maintenance costs and manual inspections are costly. A workflow for underwater photogrammetry using Remotely Operated Vehicles (ROV) has been developed, enabling accurate 3D reconstruction of underwater objects. The system provides sub-centimere accuracy at object distances.

06 - Robotic photogrammetric underwater inspection of hydropower plants

Author: Ayush Kumar Ojha This study explores the use of deep learning methods for processing and interpreting underwater images and videos. The suggested methods, utilizing convolutional neural networks and generative adversarial networks (GANs), improve image quality, identify marine species, and detect anomalies for infrastructure inspection. The findings highlight the transformative potential of artificial intelligence in overcoming the limitations of traditional underwater image processing techniques, paving the way for more effective and efficient underwater exploration and conservation efforts.

07 - Deep Learning Techniques for Enhanced Underwater Remote Sensing: Applications in Marine Biodiversity and Infrastructure Inspection

Authors: Yifan Tian, Alexander Grigorievich Palaev, Ildar Ayratovich Shammazov and Yiqiang Ren This study proposes a non-destructive testing technology for pipeline corrosion based on electromagnetic ultrasound technology to improve stability and safety in energy pipeline transportation systems. This technology uses empirical mode decomposition and singular spectrum analysis to denoise electromagnetic ultrasound signals, removing mild noise pollution and accurately detecting small area defects, thereby aiding in the regular maintenance of pipeline energy transmission systems.

11 - Non-destructive testing technology for corrosion wall thickness reduction defects in pipelines based on electromagnetic ultrasound

Authors: M. Mwelango, X. Yin, M. Zhao, Z. Zhang, Z. Han, R. Fan, P. Ma, X. Yuan, and W. Li This paper explores the use of coplanar capacitive sensors for underwater structural integrity assessment. It compares the performance of bare-electrode and insulated- electrode sensors, revealing that electrical conductivity significantly influences response. The study also highlights the feasibility of defect detection, characterisation, and imaging under water, highlighting the potential of underwater NDE techniques for inspecting structures and equipment in underwater environments. This research broadens underwater NDE knowledge.

12 - Feasibility Study on the Use of the Coplanar Capacitive Sensing Technique for Underwater Non-Destructive Evaluation

Authors: Siqiang Cheng, Yi Liu, Aibin Tang, Libin Yang This document introduces an underwater robot inspection anomaly localization feedback system, which uses real-time water surface tracking, detection, and positioning. The system is housed within the underwater robot, enabling the system to issue mechanical responses based on the water surface's tracking and positioning. The system uses sonar technology to monitor the underwater robot in real-time, triggering responsive actions upon anomalies. This system identifies abnormal conditions and informs personnel for intervention.

14 - An Underwater Robot Inspection Anomaly Localization Feedback System Based on Sonar Technology

Authors: Lutz Bretschneider, Sven Bollmann, Deborah Houssin- Agbomson, Jacob Shaw, Neil Howes, Linh Nguyen, Rod Robinson, Jon Helmore, Michael Lichtenstern, Javis Nwaboh, Andrea Pogany, Volker Ebert, and Astrid Lampert The application of drone technology to pipeline leak detection is examined in this article. It covers environmental conditions, appropriate commercial sensors, routine pipeline monitoring, and contemporary uses of natural gas pipeline surveys. Additionally, the study presents a feasibility assessment conducted during a release experiment, emphasizing the minute-by-minute change of wind direction. Requirements for future applications are also produced.

15 - Concepts for drone based pipeline leak detection

Authors: Ruben Rodríguez Elizalde This study presents a novel underwater drone model for examining submerged components of historic bridges. Three bridges spanning various rivers were inspected underwater as part of the study, which showed that underwater drones are safer than existing techniques and useful for this kind of assessment. Additionally, the drones enable the investigation of characteristics like pier or butment rotation or settlement that can forecast future foundation issues that will impact the entire bridge.

16 - Underwater Inspection of Submerged Elements in Masonry Bridges and Other Old Civil Structures Using Drones

Authors: Valery Bobkov and Alexey Kudryashov. The development of unmanned underwater vehicles (AUVs) is crucial for inspecting industrial subsea structures, but GPS is impractical due to underwater environment interferences. Optical image processing offers higher accuracy for AUV navigation. This study proposes a technology to recognize underwater objects and provide coordinate references to AUVs using stereo-image processing. It uses a non-standard technique to generate a geometric model of an object, uses characteristic geometric elements, and compares visual data with an a priori model.

18 - A Method for Recognition and Coordinate Reference of Autonomous Underwater Vehicles to Inspected Objects of Industrial Subsea Structures Using Stereo Images.

Authors: Seda Karadeniz Kartal, and Recep Fatih Cantekin This study emphasizes the importance of using unmanned underwater vehicles (UUVs) for pipeline monitoring based on an experiment with a remote- controlled UUV to detect pipeline damage. It mentions that the UUV autonomously tracked the pipeline and identified damages through camera images processed by a convolutional neural network (CNN), achieving high detection accuracy. It also mentions that the UUV followed the pipeline precisely, with minimal navigation errors.

24 - Autonomous Underwater Pipe Damage Detection Positioning and Pipe Line Tracking Experiment with Unmanned Underwater Vehicle

25 - Experimental Activity with a Rover for Underwater Inspection

Authors: Erika Ottaviano, Agnese Testa, Pierluigi Rea, Marco Saccucci, Assunta Pelliccio, and Maurizio Ruggiu This paper discusses the use of an underwater rover to implement automated imaging techniques for underwater structure inspections. The research aims to improve monitoring, reduce operational complexity, and offset the financial burden of periodic inspections. The research uses the FIFISH V6 rover with a Brave 7 camera and experimental photogrammetry tests of an anchor. The results confirm the validity of the proposed approach and encourage future development for underwater inspections.

Authors: Yomi Femi, Ademola Romke Smart pipeline monitoring systems, integrating advanced technologies like sensors and AI, offer continuous, proactive monitoring of pipeline systems. These systems enable early detection of potential issues, enhance operational efficiency, and minimize environmental risks. They reduce downtime and repair costs, improve safety, and offer predictive maintenance, transforming pipeline infrastructure management in a sustainable and cost- effective manner.

28 - Smart Pipeline Monitoring System: Revolutionizing Pipeline Management

Authors: Maxime Fauconnier, Bhuvaneshwari Karunakaran, Alex Drago-González, William S. Y. Wong, Robin H. A. Ras, & Heikki J. Nieminen This study introduces waves propagating on an underwater superhydrophobic surface, allowing for the stabilization of a microscale gas layer called a plastron. The acoustic radiation force triggers kHz "plastronic waves" with a high propagation speed of up to 45 times faster than conventional deep water capillary waves. These waves are useful for non-destructive monitoring of plastron stability and spontaneous air diffusion in gas- undersaturated or supersaturated water conditions.

29 - Fast capillary waves on an underwater superhydrophobic surface

Authors: Qian Liu, Bin Xu, Xinhai Zhu, Ronglin Chen, and Hanbin Ge This study validated an interface debonding defect detection and localization approach for scaled underwater grouting jacket connections (GJCs) using surface wave measurements with piezoelectric lead zirconate titanate (PZT) actuation and sensing technology. The approach was tested and showed that the mimicked interface debonding defect was successfully detected, and the region of debonding was determined by intersecting abnormal wave propagation paths, providing a novel non-destructive debonding defect detection method for offshore wind turbine structures.

37 - Experimental Study on Interface Debonding Defect Detection and Localization in Underwater Grouting Jacket Connections with Surface Wave Measurements

Authors: Xianfeng Zeng, Wenji Ai, Zongchao Liu, and Xianling Wang This paper introduces an unsupervised enhancement framework for restoring underwater images with structural cracks. It combines physical modeling of underwater light transmission with a deep image translation architecture. The approach includes a multi- scale feature integration module and region-focused discriminator. Validated on a real-world dataset, it shows substantial improvements in visual quality and crack boundaries preservation.

38 - Unsupervised Restoration of Underwater Structural Crack Images via Physics-Constrained Image Translation and Multi- Scale Feature Retention

Authors: Renan Favarão da Silva, Edilson Gabriel Veruz, Alécio Julio Silva, Miguel Angelo de Carvalho Michalski, Gilberto Francisco Martha de Souza, Anderson Takehiro Oshiro This paper proposes a method for planning underwater inspections based on reliability and decision-making techniques. It identifies inspection needs, determines when to inspect, and selects the appropriate method. The method is demonstrated in a case study on a Brazilian FPSO platform, supporting maintenance planning and reliability.

39 - Underwater Inspection Planning Based on Reliability and Decision-Making techniques: An FPSO Platform Case Study.

Author: Alexandre Cardaillac This thesis introduces a novel and comprehensive solution for inspecting underwater ship hulls using a cost-effective Remotely Operated Vehicle (ROV). It addresses safety concerns linked to ship hull integrity and environmental impacts due to inefficient fuel consumption resulting from hull negligence. Traditional dry dock inspection methods are often slow, prompting a shift to more efficient remote inspections. The ROV is equipped with advanced navigation and perception sensors, allowing for automated inspections, a combination of acoustic-visual data, and deep learning for enhanced analysis. The solution has been tested in various harbors and on multiple ships, demonstrating adaptability for different structures and opening avenues for future research.

02 - Towards autonomous underwater navigation and perception for end-to-end ship hull inspection

Authors: Paul Dario Toasa Caiza, Daiki Shiozawa, Thomas Ummenhofer, Takahide Sakagami This article discusses a novel, portable system that utilizes the Eddy currents principle in combination with inductive thermography for the real-time detection of cracks in metallic components. Traditional systems face challenges such as high power consumption and lengthy image processing times, making them unsuitable for large structures like bridges and oil platforms. This newly developed system operates efficiently with a square-wave AC voltage, significantly reducing energy requirements while improving detection capabilities, as demonstrated by simulations and experiments conducted on structural steel specimens.

09 - Real time detection of fatigue cracks on steel structures by applying square wave induction

Authors: Jiangpei Zhu, Mei-Ling Zhuang, Yuting Qi, Bin Chen, Xiaojian Cao This study investigates the impact of ultrasonic impact treatment (UIT) on the microstructural characteristics and mechanical properties of 316L stainless steel. The findings reveal that UIT leads to a significant grain size reduction, increased hardness, and improved tensile and fatigue strengths. The analysis also uncovers the mechanisms of fatigue fracture, indicating that UIT alters the microstructure, resulting in a gradient nanostructure and martensitic transformation, which contributes to enhanced mechanical performance.

10 - Effect of ultrasonic surface impact on the microstructural characterization and mechanical properties of 316L austenitic stainless steel

Authors: Chenxi Liang, Yang Zhao, and Fei Kang This study introduces the YOLOv8s-UEC algorithm designed for the intelligent identification of underwater defects in concrete dams. Addressing the challenges posed by the limited availability of underwater defect images, the researchers constructed a dataset by creating defective concrete walls. Enhancements in the algorithm’s feature extraction and detection capabilities are achieved through the introduction of the ConvNeXt Block and Efficient-RepGFPN structure, along with the fusion of the P2 detection layer for improved recognition of small defects. The results demonstrate significant improvements in mean average precision, indicating the algorithm’s robustness in detecting underwater defects effectively.

19 - Intelligent Detection of Underwater Defects in Concrete Dams ased on YOLOv8s-UEC

Authors: Michael D. Dukes, James M Karalekas, Dustin W. Noel, Heath K. Pope, David R. Reser, Jeffrey B. Rowe, Andrew Young This article outlines the essential practices and techniques for conducting underwater inspections of bridges to ensure public safety and protect investment. It emphasizes the need to assess structural conditions of underwater components, including streambeds, utilizing various inspection methods, ranging from visual inspections in shallow waters to diving techniques in deeper waters. The manual aims to equip bridge safety personnel, including engineers and inspectors, with knowledge about inspection methods, equipment, and common defects found during assessments.

20 - Underwater Bridge Inspection Reference Manual

Authors: Yazhou Zhang, Huansheng Wu, Linpeng Liu, Yang Yang, Changchao Zhang, and Jian Duan The article discusses the development of a flexible, superhydrophobic vibration sensor designed to address the challenges faced by commercial sensors in detecting faults in engineering machinery, particularly on irregular surfaces and in complex environments. Through the use of femtosecond laser ablation to create microcracks on a PDMS film, and subsequent coating with conductive materials, the sensor demonstrates impressive performance metrics including a high-frequency response and improvements in functionality for underwater and human-machine interaction applications.

21 - Crack-Based Composite Flexible Sensor with Superhydrophobicity to Detect Strain and Vibration

Authors: Jie Liu, Deyuan Li, & Xin Xu This article introduces the Mamba-Enhanced HRNet, a novel model designed for detecting damage in bridges, essential infrastructure that supports public safety amid rapid urbanization. The model combines the strengths of HRNet’s multi-resolution parallel design and a visual state space model (VSS) to enhance its capability for global contextual understanding, while also improving computational efficiency. An extensive dataset encompassing various types of bridge damage was created for evaluation, demonstrating that this enhanced model significantly outperforms other semantic segmentation models with a Mean Intersection over Union (Mean IoU) score of 0.963.

22 - Enhancing bridge damage detection with Mamba-Enhanced HRNet for semantic segmentation

34 - Helium Leak Testing Requirements in Fabrication of Cryogenic Storage Tanks

Authors: Kamal Dhandha, Bharat Joshi, Kamlesh Mistry, Manish Mishra, and Bhavesh Limbada The article discusses the significance and technological considerations surrounding the use of liquefied gases in various industrial applications. It emphasizes the importance of thermal insulation and specialized equipment for transporting and storing cryogenic liquids at extremely low temperatures. The article outlines the design, operation, and maintenance of cryogenic storage tanks, highlighting non-destructive testing methods, particularly helium leak detection, as crucial for ensuring the integrity and safety of cryogenic systems.

Authors: Li Jiaqing, Song Fei, Xiao Zidong, Zhu Longji, Chen Lan, Wei Zheliang This article discusses the growing concern of fatigue damage in steel bridge decks due to increased traffic loads and extended service life. Traditional detection methods are often ineffective, leading to the development of a novel approach that combines K-singular value decomposition (K-SVD) dictionary learning with convolutional neural networks (CNN) for better acoustic emission (AE) signal processing. The K-SVD algorithm filters out noise and improves signal clarity, allowing CNNs to classify AE signals with high accuracy (up to 95.32%). The integration of these methods provides a scalable and real-time solution for accurate crack detection.

35 - Investigation into fatigue micro-crack identification of steel bridge decks based on acoustic emission detection technology

36 - LBA-YOLO: A novel lightweight approach for detecting micro- cracks in building structures

Authors: Wenhao Ren, & Zuowei Zhong This article presents a new algorithm designed for the detection of building cracks, specifically targeting micro- cracks, which are vital for maintaining structural integrity and safety. The model utilizes YOLOv8n and introduces two key components, AC-LayeringNetV2 and RAK-Conv, aimed at enhancing feature extraction and improving performance in complex backgrounds. Experimental results show notable improvements in precision, recall, and mean average precision while also reducing the model’s size and computational complexity, underscoring the algorithm’s effectiveness in real-time crack detection.

40 - A Review on the Advances in Underwater Inspection of Subsea Infrastructure: Tools, Technologies, and Applications

Authors: Dulo Chukwemeka Wegner The article discusses the growing importance of subsea infrastructure for various sectors, including energy, telecommunications, and environmental monitoring. It highlights the increasing number of offshore installations and the need for effective inspection methods. The limitations of traditional diver-based inspection methods are contrasted with modern approaches such as Remotely Operated Vehicles (ROVs), Autonomous Underwater Vehicles (AUVs), and semi-autonomous ROVs. The paper further explores advanced inspection tools and technologies, analyzing their principles, applications, effectiveness, and the challenges posed by underwater environments.

41 - Design of Control System for Underwater Inspection Robot in Hydropower Dam Structures

Authors: Bing Zhao, Shuo Li, Xiangbin Wang, Mingyu Yang, Xin Yu, Zhaoxu Meng, and Gang Wan This article presents a comprehensive robotic control system designed for the inspection of underwater structural surfaces of hydropower dams. By addressing the challenges posed by harsh hydraulic conditions, the system aims to enhance operational efficiency and safety through innovative hardware and software components. Key findings from field tests validate the system’s precision and stability in real-world scenarios, significantly improving inspection capabilities for critical infrastructure

42 - Edge-Enhanced CrackNet for Underwater Crack Detection in Concrete Dams

Authors: Xiaobian Wu, Weibo Zhang, Guangze Shen, and Jinbao Sheng This article presents a novel dual-stage underwater crack detection method, termed Edge-Enhanced Underwater CrackNet (E²UCN), designed to address the challenges of detecting cracks in dam structures submerged underwater. Traditional crack detection approaches struggle due to environmental factors like water disturbances and poor visibility. The proposed method employs an advanced Cycle-GAN for image enhancement, focusing on retaining edge clarity and high-frequency textures specific to cracks, followed by a YOLOv11 model for precise object detection. The results indicate a significant improvement in detection accuracy compared to conventional methods.

43 - Deep Learning for Underwater Crack Detection: Integrating Physical Models and Uncertainty-Aware Semantic Segmentation

Authors: Wenji Ai, Zongchao Liu, Shuai Teng, Shaodi Wang, and Yinghou He This article discusses the significance of underwater crack detection for the safety of submerged infrastructures and introduces a novel deep learning framework designed to overcome existing challenges such as image degradation, limited data, and model generalization. The proposed framework enhances detection accuracy by integrating physics-based image enhancement, a dual-branch segmentation network for precise crack outline delineation, and an uncertainty-aware Transformer module to assess prediction confidence. Experimental results indicate that this approach outperforms current methods in turbid water and uneven lighting conditions, achieving notable performance metrics.

44 - Deep Learning Approaches for Fault Detection in Subsea Oil and Gas Pipelines: A Focus on Leak Detection Using Visual Data

Authors: Viviane F. da Silva, Theodoro A. Netto, and Bessie A. Ribeiro The article presents a study focused on enhancing leak detection methods in subsea oil and gas pipelines through a deep learning-based framework. By addressing the shortcomings of traditional leak detection methods, the study exploits a dataset generated from controlled experiments that simulate underwater leak conditions. Different Convolutional Neural Network architectures were analyzed for their efficiency in detecting gas and liquid leaks, achieving impressive performance metrics. The work notably contributes a novel dataset for the research community, promoting reproducibility and further advancements in automated subsea inspection technologies.

45 - Next-Gen Nondestructive Testing for Marine Concrete: AI-Enabled Inspection, Prognostics, and Digital Twins

Authors: Taehwi Lee, and Min Ook Kim The article reviews recent developments in Artificial Intelligence-integrated Non-Destructive Testing (AI-NDT) technologies for marine concrete structures, emphasizing their enhanced performance and practical applicability. It highlights the limitations of traditional NDT methods in detecting early damage due to challenging marine environments and noise interference. A systematic comparison between vision-based and signal-based AI- NDT techniques reveals significant improvements in detection accuracy, showcasing the effectiveness of AI in mitigating environmental distortions. The study concludes that AI-NDT not only automates data analysis but also supports the transition from traditional maintenance strategies to more proactive, data-driven approaches.

46 - Dynamic Illumination and Visual Enhancement of Surface Inspection Images of Turbid Underwater Concrete Structures

Authors: Xiaoyan Xu, Jie Yang, Lin Cheng, Chunhui Ma, Fei Tong, Mingzhe Gao, and Xiangyu Cao This article introduces a novel underwater image enhancement algorithm called DIVE (Dynamic Illumination and Vision Enhancement), which utilizes a framework for decoupling illumination and scattering effects. The DIVE algorithm effectively corrects luminance unevenness and color deviations, allowing for real-time processing of high-resolution images. Performance tests demonstrate significant improvements in image quality, particularly in environments with elevated sediment concentrations.

47 - Visual Crack Detection in Steel Structures with Convolutional Neural Networks

Authors: Andrii Kompanets, Davide Leonetti, (Bert) Snijder As many bridges near the end of their design lifespan, accurate and efficient inspections are vital. This article explores the integration of drones, crawling robots, and image processing techniques to automate bridge inspections. The focus is on automated crack segmentation using Convolutional Neural Networks (CNN), specifically combining the ConvNext network with an encoder-decoder network for improved segmentation. The study tests the methodology on a dataset of steel bridge images with pixel-wise crack annotations, revealing that false positives are prevalent due to complex backgrounds in the images.

48 - Enabling scalable inspection of offshore mooring systems using cost-effective autonomous underwater drones

Authors: Dong Trong Nguyen, Christian Lindahl Elseth, Jakob Rude Øvstaas, Nikolai Arntzen, Geir Hamre, and Dag-Børre Lillestøl The article discusses the pressing need for improved and cost-effective inspection methods in aquaculture as it expands to meet global food demands. It proposes an autonomous inspection framework utilizing a modular, vision-based pipeline built on the open-source Robot Operating System 2 (ROS 2) and implemented on a Blueye X3 underwater drone. This framework enhances safety, scalability, and regulatory compliance by integrating real-time image enhancement, object detection, and visual servoing for mooring inspection, while significantly reducing inspection costs compared to traditional methods.

49 - Optimization of the Number of Accelerometer Placements for Dynamic Identifi cation of a Historical Masonry Bridge

Authors: Cristiano Giuseppe Coviello, Fabio Rizzo, and Maria Francesca Sabbà This article discusses the optimization of accelerometer placement for dynamic identification of a historical three- arch masonry bridge in Bitonto, Italy. Through operational modal analysis (OMA), the study demonstrates that fewer accelerometers can effectively capture the bridge’s dynamic characteristics. This approach not only streamlines monitoring processes but also helps reduce costs and logistical challenges.