Image restoration under multiple adverse weather conditions aims to remove weather-related artifacts by using a single set of network parameters. In this paper, we find that image degradations different contain general characteristics as well their specific characteristics. Inspired observation, design an efficient unified framework with two-stage training strategy explore the weather-general and weather-specific features. The first stage learn features taking images various inputs...

Over the past two decades, an increasing number of large-scale structures have been built around world. Constructing these has a time consuming and highly expensive process. Thus, providing structural health monitoring system to guarantee their proper functionality is important. In recent years, advancement technology artificial intelligence methods based on signal processing machine learning attracted attention researchers. The challenges currently exist in field identify classify damages...

This study investigated the physicochemical and flavor quality changes in fresh-cut papaya that was stored at 4 °C. Multivariate statistical analysis used to evaluate freshness of papaya. Aerobic plate counts were selected as a predictor papaya, prediction model for established using partial least squares regression (PLSR), support vector machine (SVMR) algorithms. Freshness could be well distinguished based on analyses. The aerobic counts, significantly correlated with storage time. SVMR...

Recent advancements in sensor technology have resulted the collection of massive amounts measured data from structures that are being monitored. However, these include inherent measurement errors often cause assessment quantitative damage to be ill-conditioned. Attempts incorporate a probabilistic method into model provided promising solutions this problem by considering uncertainties as random variables, mostly modeled with Gaussian probability distribution. success methods is limited due...

Seismic response prediction is a challenging problem and significant in every stage during structure's life cycle. Deep neural network has proven to be an efficient tool the of structures. However, conventional with deterministic parameters unable predict random dynamic In this paper, deep Bayesian convolutional proposed seismic response. The Bayes-backpropagation algorithm applied train learning model. A numerical example three-dimensional building structure utilized validate performance...

Large-span precast prestressed concrete box girders have been widely used in bridge construction near or across the sea. However, this would easily lead to a hydration heat problem, including large initial tensile stress and cracks during stage of pouring. A 5 m long segment girder for Hangzhou Bay Cross-Sea Railway Bridge was continuously monitored investigate effect on long-span pouring construction. The temperature variation distribution were analyzed through finite element analysis based...

Automatic crack identification for pipeline analysis utilizes three-dimensional (3D) image technology to improve the accuracy and reliability of identification. A new technique that integrates a deep learning algorithm 3D shadow modeling (3D-SM) is proposed automatic corrosion cracks in pipelines. Since depth below surrounding area crack, projected when exposed under light sources. In this study, we analyze areas through identify evolving shape shadows. To denoise images, connected domain...

Due to unavoidable uncertainties, it is critical perform probabilistic analysis when examining the safety of coupled train-bridge system based on dynamic interaction analysis. This paper proposes a novel approach for using deep learning surrogate model. Deep neural network embedded with convolutional designed and developed construct model substituting 3D reducing computational efforts efficiently predicting series indices system. lack explicit expression performance function, automatic...

Time-history responses of the bridge induced by moving vehicle provide crucial information for design, operation, maintenance, etc. As inspired this, this work attempts to a new paradigm vehicle–bridge interaction (VBI) highlighting comparison different deep learning algorithms applied prediction time-history under vehicular loads. Particularly, three architectures with few and measurable input features developed using fully-connected feedforward neural network, long short-term memory (LSTM)...

Damage detection of civil and mechanical structures based on measured modal parameters using model updating schemes has received increasing attention in recent years. In this study, for uncertainty-oriented damage identification, a non-probabilistic structural identification (NSDI) technique an optimization algorithm interval mathematics is proposed. order to take into account the uncertainty quantification, elastic modulus described as unknown-but-bounded values proposed new scheme...

In this study, the methodology and results of ambient vibration-based investigations historical Tash Mosque in Kosovo a 3-story building Bulgaria are presented. The include full-scale situ testing both structures due to vibrations induced by micro-seismic, wind, traffic, other human activities. To aim, Ranger seismometers Kinemetric products were used. Measurements performed horizontal directions several points along structures' height utilizing high-speed data acquisition device. All...

Instance shadow detection aims to find instances paired with the objects that cast shadows. The previous work adopts a two-stage framework first predict instances, object and shadow-object associations from region proposals, then leverage post-processing match predictions form final pairs. In this paper, we present new single-stage fully-convolutional network architecture bidirectional relation learning module directly learn relations of in an end-to-end manner. Compared prior work, our...