- Optical measurement and interference techniques
- Advanced Data Storage Technologies
- Advanced Measurement and Metrology Techniques
- Advanced Vision and Imaging
- 3D Surveying and Cultural Heritage
- Image Processing Techniques and Applications
- Cloud Computing and Resource Management
- Robotics and Sensor-Based Localization
- Industrial Vision Systems and Defect Detection
- Additive Manufacturing Materials and Processes
- Image and Object Detection Techniques
- Caching and Content Delivery
- Cloud Data Security Solutions
- Advanced Image and Video Retrieval Techniques
- Manufacturing Process and Optimization
- Distributed and Parallel Computing Systems
- Image Retrieval and Classification Techniques
- Advanced Measurement and Detection Methods
- Distributed systems and fault tolerance
- Advanced Computational Techniques and Applications
- Optical Systems and Laser Technology
- Face and Expression Recognition
- Video Surveillance and Tracking Methods
- Additive Manufacturing and 3D Printing Technologies
- Structural Health Monitoring Techniques
China University of Petroleum, East China
2015-2025
Huazhong University of Science and Technology
2014-2024
Dalian Maritime University
2024
North China Electric Power University
2024
Shenyang Institute of Computing Technology (China)
2024
Electric Power Research Institute
2024
China Electric Power Research Institute
2024
Chinese Academy of Sciences
2010-2024
Southeast University
2023
Huawei Technologies (China)
2021-2023
System calibration is crucial for any 3-D shape measurement system. An accurate method proposed to calibrate a system based on structured light technique. The projector treated as camera unify the procedures of and well-established stereo vision key realizing this establish highly correspondence between pixels generate digital micromirror device (DMD) image sets calibration. A phase-shifting used accomplish task. precalibrated lookup table linear interpolation algorithm are improve accuracy...
This Letter presents a multiview phase shifting (MPS) framework for full-resolution and high-speed reconstruction of arbitrary shape dynamic objects. Unlike conventional methods, this can directly find the corresponding points from wrapped phase-maps. Therefore, only minimum number images are required to measure objects, including discontinuous surfaces. Benefit MPS achieve full spatial resolution high, accurate 3D reconstruction. constraint is also robust discontinuities. Experimental...
Some statistical and machine learning methods have been proposed to build hard drive prediction models based on the SMART attributes, achieved good performance. However, these were not evaluated in way as they are used real-world data centers. Moreover, drives deteriorate gradually, but can describe this gradual change precisely. This paper proposes new failure Classification Regression Trees, which perform better performance well stability interpretability compared with state-of the-art...
In fast phase-measuring profilometry, phase error caused by gamma distortion is the dominant source. Previous phase-error compensation or correction methods require projector to be focused for best performance. However, in practice, as digital projectors are built with large apertures, they cannot project ideal fringe images. this Letter, a thorough theoretical model of gamma-distorted image derived from an optical perspective, and highly accurate easy implement method presented reduce...
Lack of monitoring the in situ process signatures is one challenges that has been restricting improvement Powder-Bed-Fusion Additive Manufacturing (PBF AM). Among various signatures.
Industrial robots are characterized by good flexibility and a large working space, offer new approach for the machining of complex parts with small allowances (extra material allowed subsequent machining). Parts this type (such as aircraft skin parts, wind turbine blades, etc.) easily deformed due to their scale low stiffness. Therefore, these cannot be directly machined according designed model. A feasible method is plan robotic path using point clouds after clamping from onsite measurement...
Summary Accurate resource requests prediction is essential to achieve optimal job scheduling and load balancing for cloud Computing. Existing approaches fall short in providing satisfactory accuracy because of high variances metrics. We propose a deep belief network (DBN)‐based approach predict requests. design set experiments find the most influential factors best DBN parameter performance. The innovative points proposed that it introduces analysis variance orthogonal experimental...
Traditionally, disk failure prediction accuracy is used to evaluate model. However, may not reflect their practical usage (protecting against failures, rather than only predicting failures) in cloud storage systems. In this paper, we propose two new metrics for models: migration rate, which measures how much at-risk data protected as a result of correct predictions, and mismigration migrated needlessly false predictions. To demonstrate effectiveness, compare methods: (a) classification tree...
Video data has become the largest source of big data. Owing to video data's complexities, velocity, and volume, public security other surveillance applications require efficient, intelligent runtime processing. To address these challenges, a proposed framework combines two cloud-computing technologies: Storm stream processing Hadoop batch It uses deep learning realize intelligence that can help reveal knowledge hidden in An implementation this five architecture styles: service-oriented...
Using structured light to measure the 3D shape of a high dynamic range (HDR) surface has been always challenging problem, and fusion multi-group images with different exposures is recognized as an effective solution. It tends select phase unsaturated maximum gray intensity final phase, which two problems: 1) selection criteria are too simple fully evaluate quality, 2) it affected by image noise, camera's nonlinear response, local reflection other factors best quality among initial phases may...
Optimizing ship energy efficiency is a crucial measure for reducing fuel use and emissions in the shipping industry. Accurate prediction models of consumption are essential achieving this optimization. However, external factors affecting have not been comprehensively investigated, many existing studies still face accuracy challenges. In study, we propose neural network model called TCN-GRU-MHSA (TGMA), which incorporates temporal convolutional (TCN), gated recurrent unit (GRU), multi-head...
Support vector machine (SVM) has become a popular classification tool but the main disadvantages of SVM algorithms are their large memory requirement and computation time to deal with very datasets. To speed up process training SVM, parallel methods have been proposed by splitting problem into smaller subsets network assign samples different subsets. A algorithm on large-scale problems is proposed, in which multiple classifiers applied may be trained distributed computer system. As an...
This paper proposes a high-speed FPGA architecture for the phase measuring profilometry (PMP) algorithm. The whole PMP algorithm is designed and implemented based on principle of full-pipeline parallelism. results show that accuracy system comparable with those current top-performing software implementations. achieves 3D sharp reconstruction using 12 phase-shifting images completes in 21 ms 1024 × 768 pixel resolution. To best our knowledge, this first fully pipelined systems, makes very...