A5038440653- Target Tracking and Data Fusion in Sensor Networks
- NMR spectroscopy and applications
- Image and Signal Denoising Methods
- Video Surveillance and Tracking Methods
- Remote Sensing and Land Use
- Geochemistry and Geologic Mapping
- Solar and Space Plasma Dynamics
- Radiation Effects in Electronics
- Advanced Algorithms and Applications
- Real-Time Systems Scheduling
- Time Series Analysis and Forecasting
- Human Mobility and Location-Based Analysis
- Smart Parking Systems Research
- Human Pose and Action Recognition
- Autonomous Vehicle Technology and Safety
- Hydrology and Watershed Management Studies
- Embedded Systems Design Techniques
- Adaptive optics and wavefront sensing
- Regional Development and Environment
- Hydrological Forecasting Using AI
- Advanced Graph Neural Networks
- Advanced Image Fusion Techniques
- Hydrocarbon exploration and reservoir analysis
- Petroleum Processing and Analysis
- Pulsars and Gravitational Waves Research
Beijing Meteorological Bureau
2022
China University of Petroleum, Beijing
2017-2022
Lenovo (China)
2019
Beijing Institute of Technology
2016
Zhejiang University
2011
Abstract Groundwater is unique resource for agriculture, domestic use, industry and environment in the Heihe River Basin, northwestern China. Numerical models are effective approaches to simulate analyze groundwater dynamics under changeable conditions have been widely used all over world. In this paper, of middle reaches Basin was simulated using one numerical model three machine learning algorithms (multi-layer perceptron (MLP); radial basis function network (RBF); support vector (SVM))....
Abstract In this study, the Empirical Mode Decomposition algorithm (EMD) and Long Short Term Memory neural network (LSTM) are combined into an EMD‐LSTM model, to predict variation of >2 MeV electron fluxes 1 day ahead. Input parameters include Pc5 power, AP, AE, Kp, >0.6 MeV, historical flux values, used for predictions. All time resolution daily integral values. As compared prediction results model with other classical models, show that ahead efficiency possesses a 0.80, highest can...
Low-field NMR technology has been widely used in many fields, among which the most representative method is use of CPMG sequence for T2 measurement. Important parameters such as permeability, saturation and fluid type porous media can be obtained by spectrum inversion calculation echo signal. However, easily affected noise signals. Improving signal-to-noise ratio (SNR) signal one key problems with application low-field technology. Therefore, order to obtain more accurate analysis results, it...
The laser-induced voltage (LIV) method has been used to characterize the oil shale directly. Explaining generation mechanism and response process of is expected provide a basis for real-time monitoring its pyrolysis process. Here, we focused on LIV in under irradiation 532 nm laser. influence bias laser power studied discussed. strongly depends upon exhibits an increasing trend as increases from 0.157 0.237 W. With temperature result illumination, spontaneous electric polarization results...
Gaussian particle filter algorithm provides a framework to estimate the state of moving object. However, it is known fact that parameters like noise variance and number affect effectiveness greatly. To improve performance in contour tracking, authors propose parameter adapting mechanism. simplify filter's implementation, variant sampling method also proposed. This combines step with prediction by taking advantage assumption exploring linear structure system dynamic model. Finally,...
This paper constructs a modified adaptive pulse coupled neural network (PCNN) model applied to infrared and visible image fusion. The PCNN is improved by the quantum-behaved particle swarm optimization (QPSO) set parameters automatically. Two source images are processed QPSO-PCNN respectively. hybrid fitness functions proposed for QPSO algorithm. first function made up of three evaluation criteria, spatial frequency (SF), average gradient (AG), entropy (EN). other one uses mutual information...
Parking data is vulnerably affected by spatiotemporal characteristics and surrounding societal events, causing the latent features of parking are hard to learned GANs if solely given time-series data. Hence it impossible generate desired with high quality. In this paper, we propose a multi-conditional GAN, named MCGAN refine generating process optimize quality via introducing external customized extendable conditions related samples. These conditions, in forms condition tensors MCGAN, can...
With the significant advancements in Information and Communications Technology (ICT), computer simulations have been widely used natural sciences engineering analysis.Algorithms for parameters optimization models necessity modeling due to accuracy requirements.As a principal branch of models, hydrological fundamental method researching processes.In this paper, model groundwater system middle reaches Heihe River Basin was established.Geological features different regions were characterized by...
Abstract. In recent years, great progress has been made in the study of semantic segmentation field computer vision. The accuracy constantly improved, and it widely applied fields automatic driving, medical treatment remote sensing image classification.Semantic all kinds neural network structure optimized, according to different task put forward loss function optimization algorithm improve classification, such as used classification more softmax cross entropy sigmoid is task, two functions...
Extracting entities and relations from unstructured text is the key to building a large-scale knowledge graph. In recent years, relation extraction approaches based on neural network has achieved good results. However, existing methods cannot accurately extract overlapping (i.e., one entity shared by multiple relations). this paper, we propose simple Electra-based joint model for extraction. We use subject extractor identify first, then predicate-object predict its corresponding predicates...
In the control system of a quantum gravimeter, optical signal containing phase information is weak and usually at nV level. This situation puts forward high requirements for measurement performance system. As coupling input component system, noise level preamplifier has dramatic influence on performance. this paper, current provided by photodiode. According to photodiode output characteristics, we design three preamplifiers with different gain bandwidth, which verified Under condition...
By inversion calculation of T2 spectrum echo string, the permeability, saturation, fluid type and other important parameters porous media can be obtained. However, since algorithm itself is susceptible to noise signal [1], improving SNR level has been one key problems in application low-field NMR technology. In order obtain more accurate analysis results, corresponding reduction method needs developed. past, research on mainly focused string. This paper constructs a complete model discusses...