Abstract The spatial characteristics of cracks are significant indicators to assess and evaluate the health existing buildings infrastructures. However, current manual crack description method is time consuming labor consuming. To improve efficiency inspection, advanced computer vision‐based techniques have been utilized detect automatically at image level grid‐cell level. But detections (high specificity) low generality inefficient, in terms that conventional approaches unable identify...

Structural health monitoring (SHM) is a multi-discipline field that involves the automatic sensing of structural loads and response by means large number sensors instruments, followed diagnosis based on collected data. Because an SHM system implemented into structure automatically senses, evaluates, warns about conditions in real time, massive data are significant feature SHM. The techniques related to referred as science engineering, include acquisition techniques, transition management...

This study conducts crack identification from real-world images containing complicated disturbance information (cracks, handwriting scripts, and background) inside steel box girders of bridges. Considering the multilevel multi-scale features input images, a modified fusion convolutional neural network architecture is proposed. As input, 350 raw are taken with consumer-grade camera divided into sub-images resolution 64 × pixels (67,200 in total). A regular structure employed as baseline to...

This paper proposed a modified faster region-based convolutional neural network (faster R-CNN) for the multitype seismic damage identification and localization (i.e., concrete cracking, spalling, rebar exposure, buckling) of damaged reinforced columns from images. Four hundred raw images containing different damages complicated background information are taken by consumer-grade camera in various locations arbitrary perspectives to simulate diverse situations where real-world postearthquake...

Medical image segmentation has important auxiliary significance for clinical diagnosis and treatment. Most of existing medical solutions adopt convolutional neural networks (CNNs). Althought these can achieve good performance, CNNs focus on local information ignore global information. Since Transformer encode the whole image, it modeling ability is effective extraction Therefore, this paper proposes a hybrid feature network, into which are integrated to utilize their advantages in...

Although structural damage recognition has been extensively investigated using deep learning and computer vision (CV) techniques, the following limitations exist for real-world applications: (1) accuracy heavily relies on a large volume of network parameters; (2) sensitivity to tiny cracks is limited due low contrast between microcrack background pixels; (3) robustness complex with various morphological features surface disturbances inadequate. To address these issues, this study proposes...

Timely acquiring the earthquake-induced damage of buildings is crucial for emergency assessment and post-disaster rescue. Optical remote sensing a typical method obtaining seismic data due to its wide coverage fast response speed. Convolutional neural networks (CNNs) are widely applied image recognition. However, insufficient extraction expression ability global correlations between local patches limit performance dense building segmentation. This paper proposes an improved Swin Transformer...

Artificial intelligence (AI) provides advanced mathematical frameworks and algorithms for further innovation vitality of classical civil engineering (CE). Plenty complex, time-consuming, laborious workloads design, construction, inspection can be enhanced upgraded by emerging AI techniques. In addition, many unsolved issues unknown laws in the field CE addressed discovered physical machine learning via merging data paradigm with laws. Intelligent science technology profoundly promote current...

Large-scale optical sensing and precise, rapid assessment of seismic building damage in urban communities are increasingly demanded disaster prevention reduction. The common method is to train a convolutional neural network (CNN) pixel-level semantic segmentation approach does not fully consider the characteristics objectives. This study developed machine-learning-derived two-stage for post-earthquake location considering data satellite remote (SRS) images with dense distribution, small...

Seismic damage assessment of reinforced concrete (RC) structures is a vital issue for post-earthquake evaluation. Conventional onsite inspection depends greatly on subjective judgments and engineering experiences human inspectors, the efficiency limited to large-scale urban areas. This study proposes computer-vision machine-learning-based seismic framework RC structures. A refined Park-Ang model built express coupled effects structural ductility energy dissipation, which reflects nonlinear...

Cable inspections revealed that severe corrosion of steel wires is one the main failure mechanisms cables. Accelerated experiments were conducted to evaluate variation in uniform and pitting depths high-strength over time. The measured depth followed a lognormal distribution with time-dependent variables at both zinc coating stage stage. block’s maximum factors from different exposure time proven be drawn same underlying continuous population Gumbel distribution. regression models scale...

Image archives of multi-class structural damages can be collected by manual inspection and then used for damage identification. On one hand, conventional image-processing-based approaches rely on optimal designs hand-crafted feature detectors lack universal adaptability various application cases; the other regular supervised learning techniques require complete types sufficient training examples to establish a robust recognition model, which brings up time-labor-consuming image collection...

To promote the development of structural health monitoring around world, 1st International Project Competition for Structural Health Monitoring (IPC-SHM, 2020) was initiated and organized in 2020 by Asia-Pacific Network Centers Research Smart Structures Technology, Harbin Institute University Illinois at Urbana-Champaign, four leading companies application technology. The goal this competition to attract more young scholars engage study monitoring, encouraging them provide creative effective...

Abstract This study develops an autonomous design method for architectural shape sketches by a novel self‐sparse generative adversarial network (self‐sparse GAN), thereby overcoming the problems regarding excessive reliance on sufficient aesthetic knowledge and time consumption in traditional human design. First, new dataset denoted “Sketch” is built using eXtended difference‐of‐Gaussians operator. Second, self‐adaptive sparse transform module (SASTM) designed following each deconvolution...

Abstract Marine microplastics are emerging as a growing environmental concern due to their potential harm marine biota. The substantial variations in physical and chemical properties pose significant challenge when it comes sampling characterizing small-sized microplastics. In this study, we introduce novel microfluidic approach that simplifies the trapping identification process of surface seawater, eliminating need for labeling. We examine various models, including support vector machine,...

Abstract This study proposesStructure‐PoseNet to recognize three‐dimensional (3D) poses and dense dynamic displacement of a structure with known geometry using monocular camera. The 3D are employed calculate the structural both visible invisible elements in videos. proposed framework consists two consecutive deep learning modules, CompNet ParaNet, provide semantic image segmentation extract pose parameters, respectively. converts natural video images into classification masks. ParaNet uses...

Recognizing and quantifying microcracks on surfaces are crucial for early detection of structural damage, as they can lead to more complex issues in engineering structures. In this study, a dataset reflecting varying surface cracks various materials from the 2018 Ecuador earthquake was constructed. Furthermore, we proposed deep learning-based methods recognizing cracks. The utilized an enhanced U-Net semantic segmentation model, along with noise reduction topological parameter extraction...