- Landslides and related hazards
- Geotechnical Engineering and Analysis
- Geotechnical Engineering and Soil Stabilization
- Speech Recognition and Synthesis
- Speech and Audio Processing
- Soil and Unsaturated Flow
- Music and Audio Processing
- Granular flow and fluidized beds
- Flood Risk Assessment and Management
- Dam Engineering and Safety
- E-commerce and Technology Innovations
- Tree Root and Stability Studies
- Meteorological Phenomena and Simulations
- Geotechnical Engineering and Soil Mechanics
- Rock Mechanics and Modeling
- Nicotinic Acetylcholine Receptors Study
- Plant Molecular Biology Research
- Landfill Environmental Impact Studies
- 3D Surveying and Cultural Heritage
- Natural Language Processing Techniques
- Robotics and Sensor-Based Localization
- Hydrology and Sediment Transport Processes
- Tribology and Lubrication Engineering
- Remote-Sensing Image Classification
- Remote Sensing and LiDAR Applications
Qilu University of Technology
2023
Institute of Atmospheric Physics
2018-2020
Chinese Academy of Sciences
2018-2020
Hong Kong Polytechnic University
2010-2019
China Earthquake Administration
2019
Northeast Normal University
2018
Kunming University of Science and Technology
2018
Landslide displacement prediction can enhance the efficacy of landslide monitoring system, and periodic is particularly challenging. In previous studies, static regression models (e.g., support vector machine (SVM)) were mostly used for predicting displacement. These may have bad performances, when dynamic features triggers are incorporated. This paper proposes a method in manner, based on gated recurrent unit (GRU) neural network complete ensemble empirical decomposition with adaptive noise...
What the role of each phase medium plays and how their interactions do work should be essential problems to understand dynamic behaviours soils. In order disclose between solid, water, air phases soils, we applied loess samples analyse controlling effects residual deformation on pore pressure based three kinds laboratory tests. We obtained similarity difference mechanical behaviors soil under different water contents loading. Both process cause air/water pressures are independent initial...
A method capable of automatically reconstructing 3D building models with semantic information from the unstructured point cloud indoor scenes is presented in this paper. This has three main steps: segmentation using a new hybrid algorithm, room layout reconstruction, and wall-surface object reconstruction by an enriched approach. Unlike existing methods, aims to detect, cluster, model complex structures without having prior scanner or trajectory information. In addition, enables accurate...
While i-vector/PLDA framework has achieved great success, its performance still degrades dramatically under noisy conditions. To compensate for the variability of i-vectors caused by different levels background noise, this paper proposes an SNR-invariant PLDA robust speaker verification. First, nonparametric feature analysis (NFA) is employed to suppress intra-speaker variation and emphasize discriminative information inherited in boundaries between speakers i-vector space. Then,...
Small but complicated labyrinth channel emitters are easily clogged. In this study, computational fluid dynamics–discrete element method coupling approach was employed to investigate the mechanism of emitter clogging caused by particles in size 65, 100, and 150 µm. Computational dynamics used Navier–Stokes equation analyze flow characteristics continuous phase. Discrete Newton’s laws motion measure single particle group distribution disperse Particle tracking velocimetry also utilized follow...
Recent studies have examined lipophilic marine toxins (LMTs) in shellfish and toxic algae worldwide, but the occurrence seasonal variations of LMTs commercial clams (including Mactra veneriformis, Ruditapes philippinarum, Meretrix meretrix, Cyclina sinensis) at their major culturing area Jiangsu, China, remain largely unexplored. In this study, a new solid phase extraction (SPE) combination with an ultra-fast liquid chromatography triple-quadrupole linear ion trap mass spectrometry...
The mismatch between enrollment and test utterances due to different types of variabilities is a great challenge in speaker verification. Based on the observation that SNR-level variability or channel-type causes heterogeneous clusters i-vector space, this paper proposes apply supervised learning drive guide probabilistic linear discriminant analysis (PLDA) mixture models. Specifically, deep neural network (DNN) trained produce posterior probabilities SNR levels channel given i-vectors as...
Landslide identification is an increasingly important research topic in remote sensing and the study of natural hazards. It essential for hazard prevention, mitigation, vulnerability assessments. Despite great efforts over past few years, its accuracy efficiency can be further improved. Thus, this combines two most popular approaches: susceptibility analysis change detection thresholding, to derive a landslide method employing novel criteria. Through quantitative evaluation proposed masked...
Abstract. Natural as well fill slopes are commonly found in Hong Kong, China, and many other countries, slope failures with the subsequent debris flows have caused a serious loss of life property past until now. There various processes features associated flow which engineers need to know so design for precautionary measures. In this study, experiments on flume tests, friction deposition rebound tests were carried out different sizes balls determine parameters required modelling dry granular...
Abstract. Landslide is a major disaster resulting in considerable loss of human lives and property damages hilly terrain Hong Kong, China many other countries. The factor safety the critical slip surface for slope stabilization are main considerations stability analysis past, while detailed post-failure conditions slopes have not been considered sufficient detail. There however increasing interest consequences after initiation failure that includes development propagation surfaces, amount...
In speaker recognition, the mismatch between enrollment and test utterances due to noise with different signal-to-noise ratios (SNRs) is a great challenge. Based on observation that noise-level variability causes i-vectors form heterogeneous clusters, this paper proposes using an SNR-aware deep neural network (DNN) guide training of PLDA mixture models. Specifically, given i-vector, SNR posterior probabilities produced by DNN are used as posteriors indicator variables model. As result,...
For the silo flow of two materials with contrasting strengths, an automatic separation without mixing is observed during drawing. In present paper, writers observe that a columniform funnel will appear in single opening while uniform occur for double opening. The classical and popular Janssen equation hence less suitable than development arch action arising from contact force investigated through distinct element analysis different types silo. It also found there are significant differences...
To deal with the mismatch between enrollment and test utterances caused by noise different signal-to-noise ratios (SNR), we have recently proposed an SNR-invariant PLDA model for robust speaker verification. In model, SNR-specific information were separated from speaker-specific through marginalizing out SNR factors during scoring process. However, this modeling approach assumes that variabilities can be captured a single subspace regardless of level utterances. We will show in paper...
Abstract Many engineering problems are governed by partial differential equations which can be solved analytical as well numerical methods, and examples include the plasticity problem of a geotechnical system, seepage elasticity problem. Although governing either iterative finite difference method or element, there however limitations to these methods in some special cases will discussed present paper. The solutions all viewed stationary value functional. Using an approximate solution...
Classically, lateral earth-pressure, bearing-capacity, and slope-stability problems are considered separately, although these all related to the ultimate limit state of a system. The authors see that governed by similar requirements should be equivalent in terms failure mechanism, can as different view It is interesting note that, there many applications method characteristics various geotechnical problems, none devoted unification problems. Bearing-capacity, slip-line, have been treated...
Abstract. A semi-analytical solution for the three-dimensional stability analysis of ultimate uniform patched load on top a slope is developed by limit using kinematically admissible failure mechanisms. The mechanism which assumed in analytical verified strength reduction analyses and laboratory model test. Furthermore, proposed method results are further compared with some published illustrating applicability mechanism.
This paper proposes the architecture of a real-time slope monitoring system based on Internet Things (IoT), which consists sensing layer, data acquisition cloud computing and client layer. The was deployed in large-scale Shenzhen, China, estimated to be stable using stability analysis. However, locally shallow landslide had occurred slope, meaning that analysis method failed indicate failure. Conversely, through results, it is found body still under active movement. Besides, advantages risk...
A heavy-rainfall event that occurred in North China during 19–20 July 2016, resulting severe flooding, was investigated this study. In event, high-value total deformation overlapped the precipitation region, implying a close relationship between them. By deriving non-geostrophic ω equation non-uniformly saturated moist atmosphere, relation vertical velocity and diagnosed. The Q-vector divergence on right-hand side of new divided into three compositions, associated with horizontal divergence,...