A5101452989- Industrial Vision Systems and Defect Detection
- Advanced Vision and Imaging
- Advanced X-ray and CT Imaging
- Radiative Heat Transfer Studies
- Combustion and flame dynamics
- Advanced Battery Technologies Research
- Electric Vehicles and Infrastructure
- Advanced Measurement and Detection Methods
- Advanced Surface Polishing Techniques
- Remote Sensing and LiDAR Applications
- Photoacoustic and Ultrasonic Imaging
- Welding Techniques and Residual Stresses
- Opportunistic and Delay-Tolerant Networks
- Medical Imaging Techniques and Applications
- Image and Signal Denoising Methods
- MRI in cancer diagnosis
- Image and Object Detection Techniques
- Cognitive Computing and Networks
- Aquatic Ecosystems and Phytoplankton Dynamics
- Advanced Neural Network Applications
- Advanced Combustion Engine Technologies
- Satellite Communication Systems
- Machine Learning in Materials Science
- 3D Surveying and Cultural Heritage
- Metamaterials and Metasurfaces Applications
Fudan University
2025
Xi’an University of Posts and Telecommunications
2024
Affiliated Hospital of Hangzhou Normal University
2024
Hangzhou Normal University
2024
Queensland University of Technology
2024
Shenyang Research Institute of Foundry
2020-2024
Huazhong Agricultural University
2024
Hunan University
2020-2024
Beijing Institute of Technology
2010-2023
China Academy of Space Technology
2014-2023
Metal surface defect segmentation can play an important role in dealing with the issue of quality control during production and manufacturing stages. There are still two major challenges industrial applications. One is case that number metal samples severely insufficient, other most existing algorithms only be used for specific defects it difficult to generalize surfaces. In this work, a theory few-shot generic introduced solve these challenges. Simultaneously, Triplet-Graph Reasoning...
Unmanned aerial vehicles (UAVs) can obtain high-resolution topography data flexibly and efficiently at low cost. However, the georeferencing process involves use of ground control points (GCPs), which limits time cost effectiveness. Direct georeferencing, using onboard positioning sensors, significantly improve work efficiency. The purpose this study was to evaluate accuracy Global Navigation Satellite System (GNSS)-assisted UAV direct method influence number distribution GCPs. A FEIMA D2000...
Hydrogels with adhesive properties have the potential for rapid haemostasis and wound healing in uncontrolled non-pressurized surface bleeding. Herein, a typical hydrogen bond-crosslinked hydrogel above functions was constructed by directly mixing solutions of humic acid (HA) polyvinylpyrrolidone (PVP), which HA worked as crosslinking agent to form bonds PVP. By altering concentration HA, cluster stable uniform hydrogels were prepared within 10 s. The dynamic reversible nature gave HA/PVP...
Nondestructive testing of welding images is still a significant challenge due to the imaging characteristics radiographic and extremely random distribution defect types. The practical application supervised methods for nondestructive encounters challenges limited availability densely annotated samples absence prior information regarding unknown defects. Hence, this article first proposes rapid screening method using only defect-free image training screen suspected normal welds in real time....
In recent years, generating 3D human models from images has gained significant attention in reconstruction. However, deploying large neural network practical applications remains challenging, particularly on resource-constrained edge devices. This problem is primarily because require significantly higher computational power, which imposes greater demands hardware capabilities and inference time. To address this issue, we can optimize the architecture to reduce number of model parameters,...
Sufficient historical flood inventory data (FID) are crucial for accurately predicting susceptibility using supervised machine learning models. However, FID insufficient in many regions. Remote sensing provides a promising opportunity to expand the FID. whether expanded by remote can improve accuracy of modeling needs further study. In this study, framework was proposed improving prediction (FSP) combining models and Sentinel-1A radar images. Five widely used were employed verify method...
Biomass is important in monitoring global carbon storage and the cycle, which quickly accurately estimates forest biomass. Precision forestry modeling place high requirements on obtaining individual parameters of various tree species complex stands, studies have included both overall stand trees. Most existing literature focuses calculating species’ biomass a single stand, there little research stands. This paper calculates stands by combining multispectral light detection ranging (LIDAR)...
For electric vehicle DC charging station (EVCS) supplied by energy storage units (ESUs) with virtual inertia and damping control (VIDC), the dynamic interaction oscillation (DIO) might exist due to inconsistent among VIDC-controlled ESUs. this issue, a stabilization method is proposed as voltage stabilizer (DVSer), including hybrid system (HESS) architecture unified power sharing control. A super-capacitor (SC) bridge arm reserved improve high-frequency secondary regulation link design...
The mass access of a power electronics-based source/load resource to distribution network brings about complex power-quality issues, which aggravate loss problems. This paper proposes hierarchical loss-optimization design method for the multi-winding inductive filtering transformer applied in system with function harmonic elimination. parameters are classified hierarchically by means Sobol algorithm-based sensitivity analysis, contributes reducing optimization dimension and improving...
The input data of photometric stereo have three dimensions: one dimension different lights and two spatial dimensions, i.e., rows columns in the image coordinate. Recent deep-learning-based algorithms usually use 2-D/3-D convolutions to process with dimensions. Assumptions that violate natural characters problem, e.g., pixel interdependence light permutation invariance, be made due mismatch. In this article, we propose a self-attention network (SPS-Net), which can exploit information all...
Digital Image Correlation (DIC) has been widely used as a non-contact deformation measurement technique. Nevertheless, its accuracy is greatly affected by the speckle pattern on specimen. To systematically evaluate how deformability affects precision of DIC algorithms. In this study, test dataset 2D patterns with various prescribed fields was numerically generated, containing two categories speckles, i.e., deformable and non-deformable (rigid) ones. This to performance inverse compositional...
Hydrogels with complex internal structures are required for advanced drug delivery systems and tissue engineering or used as inks 3D printing. However, hydrogels lack the tunability diversity of polymeric shells require complicated postsynthesis steps to alter its structure properties. We report on first integrated approach assemble design take various functions such multilayer nanofilms, multidensity immobilization matrix, multiadhesive chromatography resins via tuning four assembly...
A controlled-source audio-frequency magnetotelluric (CSAMT) survey was used to detect geological structures beneath the thick quaternary formation in Taiyuan, Shanxi Province, northern China. Two main CSAMT lines with 182 sites were recorded. two-dimensional (2D) inversion technique interpret data. The results suggested that: 1) there are four buried faults named F1, F2, F3, and F4 dip angles about 65° across line from west east, fault displacements of these 230 m, 180 220 m 200...
Studs are small screws welded on a car's body. Measuring the key point positions of studs within production cycle time, also known as online measurement, is crucial for quality management cars. However, existing stud measuring systems can only measure one in camera position, which not efficient enough measurement. Accurately multiple position still challenge. In this article, we propose system (SOMS), with high accuracy based photometric stereo and deep learning techniques. To achieve...
The non-local means method (NLM) is widely used in image denoising. However, the performance of this heavily depends on choice smoothness parameters. In paper, we present a novel multi-scale (MNLM) for By introducing decomposition images, our can avoid difficulty choosing Compared with classical NLM method, MNLM not only improves accuracy measurement similarity, but also generates better denoising results.
In order to reduce the labor and material of library management contribute development unmanned library, an intelligent mobile robot for is proposed. The consists a platform, three-DOF rise-and-fall robotic arm multisource image recognition information fusion system. Besides recognizing data position book, also able conduct book grasp shelving in task. To adapt existing environment enlarge its move area, efficient autonomous elevator button system based on neural networks designed, which...
For the problem of rough irrigation habits and low efficiency in Hetao area China, which only pays attention to increasing yield, this paper proposes a method accurate extraction canal information based on fusion rules image enhancement order improve level management area.Accurate calibration is premise precision management.For satellite remote sensing, it difficult collect with high accuracy sub lateral ditches field.For ground stations, work move flexibly target.Therefore, UAV sensing...
The new space race is intensifying as thousands of satellites are poised for deployment into outer each year. Companies deploying low-earth orbit at an unprecedented rate to create mega-constellations. Because becoming increasingly crowded and hazardous, establishing safe, efficient, rapid inter-satellite communications now a major challenge, essential managing, maintaining, inspecting vast communication networks. Joint sensing (JSC) presents remarkable opportunity enhance performance...
Reconfigurable three-dimensional (3D) nanostructures possess additional spatial freedom and rich physical characteristics compared with their two-dimensional (2D) counterparts. Here we demonstrate a facile automated nano-kirigami method to create the 3D deformed spiral metasurfaces, which can flexibly manipulate optical waves by simply applying external voltages. Through etching Archimedean spirals into Au/SiO 2 /Si layer, multi-order localized surface plasmon resonances are excited within...