A5100686179- Optical Polarization and Ellipsometry
- Remote-Sensing Image Classification
- Topic Modeling
- Natural Language Processing Techniques
- Calibration and Measurement Techniques
- Text and Document Classification Technologies
- Speech Recognition and Synthesis
- Geochemistry and Geologic Mapping
- Image Processing and 3D Reconstruction
- Industrial Technology and Control Systems
- Optical and Acousto-Optic Technologies
- Sparse and Compressive Sensing Techniques
- Structural Health Monitoring Techniques
- Oil and Gas Production Techniques
- Remote Sensing and Land Use
- Advanced Text Analysis Techniques
- Photonic and Optical Devices
- Infrared Target Detection Methodologies
- Reservoir Engineering and Simulation Methods
- Multimodal Machine Learning Applications
- Vehicle License Plate Recognition
- Geological Modeling and Analysis
- Stock Market Forecasting Methods
- Advanced Fiber Optic Sensors
- Advanced Image Fusion Techniques
Beihang University
2012-2020
ConocoPhillips (United States)
2019
Shanghai International Studies University
2012
Chongqing University
2010
Northeast Forestry University
2009
Xi'an Jiaotong University
2009
Northwest Institute of Nuclear Technology
2009
Southeast University
2009
University of Windsor
1995-2002
Tsinghua University
1995
Temporally-spatially modulated Fourier transform imaging spectrometers (TSMFTISs) have no internal moving parts and high throughput advantages. In recent years, TSMFTISs become a research development focus for spaceborne missions. However, the applications of airborne missions are limited because conventional spectrum reconstruction algorithm requires stability platform during pushbroom process. To overcome this limitation, we propose target-tracking-based method TSMFTISs. The direct...
Text clustering, an important part of the machine learning and pattern recognition, has extensive applications in field natural language processing. In this paper, a method is given to improve classic TFIDF algorithm on its shortcomings. This paper classifies text through Naive Bayesian classifier. And uses iterative optimize selection feature words, then classification ceaselessly. Experimental results show that preferable efficiency feature-select can increase accuracy.
To overcome the downhole space limitation, eccentric tubing was designed. Two fiber Bragg grating (FBG) sensors written in one optical were integrated into specially designed to form an on-line monitoring system, and a protective casing isolate FBG from harsh environment. The wavelengths of FBGs recorded at certain temperatures pressures order calibrate temperature pressure sensitivity, respectively. Both exhibit good linear relationships with FBGs, which suggest that they can be represented...
Multiple expert system is shown to be a promising strategy for handwritten numeral recognition. This paper presents multiple using neural networks. In the proposed system, authors have developed (1) an incremental clustering network algorithm with merging and canceling process, (2) modified directional histogram feature extraction method (3) subclass learning rejection neuron strategy. Our experimental results on large set of data show efficiency robustness system.< <ETX...
Runxin Xu, Jun Cao, Mingxuan Wang, Jiaze Chen, Hao Zhou, Ying Zeng, Yuping Li Xiang Yin, Xijin Zhang, Songcheng Jiang, Yuxuan Lei Li. Proceedings of the 58th Annual Meeting Association for Computational Linguistics: System Demonstrations. 2020.
Sensor is an important part of the camera module, dead pixels which affect image quality to a very great extent, so production process requires effective method detect sensor. This paper analyzed characteristics sensor imaging and researched point detection methods, from manual interpretation then given algorithm program interpretation. Integrated propose scientific program, has positive meaning for improving efficiency optimize product quality, points
In this paper, an approach for interpolating the missing data by support vector regression (SVR) machine is proposed. First, samples where some features are separated from original samples. Then remaining trained SVR, feature values corresponding to treated as labels. Finally, obtained hyper-surface used predict features. Experimental results show considerable effectiveness of proposed method.
The automatic bit-feeding process is the key technique of drilling technology. performance system affects directly efficiency, even success entire shaft project. In order to satisfy requirements bit-feeding, a control model combining fuzzy logic with proportional integral algorithm was proposed by analyzing operational modes. Applied winch machine, solved problem time delay and nonlinearity system. hybrid fuzzy-PI borrows ideas from both PID control, solves that precise mathematical...
A light field imaging spectrometer can obtain spatial-spectral datacube in one snapshot. We present an image formation model of the based on wave optics to demonstrate spectral process system.
Temporally-spatially modulated Fourier transform imaging spectrometers (TSMFTISs) provide high-throughout-type push-broom spectrometry with both temporal and spatial modulation features. The system requires strict registration between the detector interferogram. However, errors are unavoidable directly change corresponding optical path difference values of As a result, interferogram should be corrected before restoring spectrum. In order to obtain correct (OPD) values, an online error...
This paper proposes the building of Xiaomingbot, an intelligent, multilingual and multimodal software robot equipped with four integral capabilities: news generation, translation, reading avatar animation. Its system summarizes Chinese that it automatically generates from data tables. Next, translates summary or full article into multiple languages, reads rendition through synthesized speech. Notably, Xiaomingbot utilizes a voice cloning technology to synthesize speech trained real person's...
During the temporally and spatially modulated Fourier transform imaging spectrometer push-broom scanning process, motion state of platform can vary. Thus, target interferogram obtained from image sequence deviates ideal using high stability. The recovered spectrum will not reflect true characteristics. We adopted tracking to acquire position in via a proposed kernel regression, with relative deviation method for determining intensities, recovery spectrogram nonuniform fast algorithm. tested...
Time frequency transformations have gained increasing attention for the characterization of non-stationary signals in a broad spectrum science and engineering applications. Signals encountered rotary machine systems can be broadly classified as being either stationary or nonstationary. This study evaluates performance traditional method-polynomial fit filtering with Fourier analysis new developed method-empirical mode decomposition Hilbert transform (EMD+HT), mechanical signal decomposition....
Temporally and spatially modulated Fourier transform imaging spectrometers (TSMFTISs) can obtain images interference information of targets during the data acquisition process for remote sensing. play an important role in target classification identification, as spectrum be reconstructed with theory spectroscopy. However, defect pixels absent planar array charge-coupled device used have a significant impact on accuracy spectral recovery information, so preprocessing bad sensing is...
PreviousNext No AccessInterpretationVolume 7, Issue 3Introduction to special section: Machine learning in seismic data analysisAuthors: Haibin DiTao ZhaoVikram JayaramXinming WuLei HuangGhassan AlRegibJun CaoMauricio Araya-PoloSatinder ChopraSaleh Al-DossaryFangyu LiErwan GloaguenYouzuo LinAnne SolbergHongliu ZengHaibin DiSchlumberger, Exploration and Field Development, Data Analytics Program, Houston, Texas, USA. E-mail: .Search for more papers by this authorEmail the author at [email...
A correction method based on data reconstruction and restoration applied to interference-modulated images acquired with airborne Fourier Transform Imaging Spectrometers are introduced. Simulations done discussed demonstrated the proposed method.
Aiming at less information available in off-line signature, the accuracy of using a single character to verify is not high enough, an handwritten signature authentication method based on multi-feature fusion presented. At first, ET1DT12 feature and moment are extracted from same combined form new high-dimensional feature, then RBF neural network used for training verification. Experimental results show that can effectively improve