A5101943419- Energy Efficient Wireless Sensor Networks
- Advanced Database Systems and Queries
- Energy Harvesting in Wireless Networks
- Graph theory and applications
- UAV Applications and Optimization
- Advanced Graph Theory Research
- Remote Sensing and LiDAR Applications
- Date Palm Research Studies
- graph theory and CDMA systems
- Data Management and Algorithms
- Remote Sensing in Agriculture
- Remote Sensing and Land Use
- Graph Labeling and Dimension Problems
- Advanced Computational Techniques and Applications
- Forest Insect Ecology and Management
- Web Data Mining and Analysis
- Smart Agriculture and AI
- Data Mining Algorithms and Applications
- Mobile Ad Hoc Networks
- Insect and Arachnid Ecology and Behavior
- Interconnection Networks and Systems
- Age of Information Optimization
- IoT and Edge/Fog Computing
- Distributed Control Multi-Agent Systems
- Land Use and Ecosystem Services
State Forestry and Grassland Administration
2020-2025
Beijing Forestry University
2013-2025
East China University of Science and Technology
2024
Anhui University of Science and Technology
2021-2024
Massachusetts Institute of Technology
2023
American Institute of Aeronautics and Astronautics
2023
University of Electronic Science and Technology of China
2023
Fuzhou University
2022
Anhui University of Finance and Economics
2022
Shanghai Jiao Tong University
2021
Plant leaf recognition is a computer vision task used to automatically recognize plant species. It very challenging since rich morphological variations, such as sizes, textures, shapes, venation, and so on. Most existing methods typically normalize all images the same size them at one scale, resulting in unsatisfactory performances. In this letter, multiscale fusion convolutional neural network (MSF-CNN) proposed for multiple scales. First, an input image down-sampled into multiples low...
Current deep learning-based change detection approaches mostly produce convincing results by introducing attention mechanisms to traditional convolutional networks. However, given the limitation of receptive field, convolution-based methods fall short fully modelling global context and capturing long-range dependencies, thus insufficient in discriminating pseudo changes. Transformers have an efficient spatio-temporal capability, which is beneficial for feature representation changes...
Due to their distinct economic efficiency and adaptability advantages, Unmanned Aerial Vehicles (UAVs) can serve as mobile data collectors, collecting from Internet of Things Devices (IoTDs). As a promising emerging technology, the Intelligent Reflecting Surface (IRS) holds potential overcome architectural barriers improve communication quality in urban environments. This study investigates development an IoT collection system tailored for environments, leveraging synergistic operation...
Preparing a data set for analysis is generally the most time consuming task in mining project, requiring many complex SQL queries, joining tables, and aggregating columns. Existing aggregations have limitations to prepare sets because they return one column per aggregated group. In general, significant manual effort required build sets, where horizontal layout required. We propose simple, yet powerful, methods generate code columns tabular layout, returning of numbers instead number row....
This paper presents a novel physical modelling method to study projectile penetration problems. The employs recently developed transparent synthetic soil made of oil-saturated fused quartz, which represents the macroscopic behaviour silica sand. Digital image correlation (DIC, also known as particle velocimetry (PIV)) techniques were employed quantify response granular soils high-speed penetration, non-intrusively. A conical nose steel was accelerated into target at an impact velocity 13.6...
Classifying birds accurately is essential for ecological monitoring. In recent years, bird image classification has become an emerging method recognition. However, the task needs to face challenges of high intraclass variance and low inter-class among birds, as well model efficiency. this paper, we propose a fine-grained based on attention decoupled knowledge distillation. First all, attention-guided data augmentation method. Specifically, obtains images object's key part regions through...
In this paper, algorithms to describe mesoscale and microscale kinematics of granular flow using image analysis are described. At the mesoscale, digital correlation (DIC) is employed derive displacement fields, from which rigid body rotation strains calculated continuum mechanics descriptions kinematics. Moreover, Lagrangian Eulerian trajectories obtained DIC analysis. microscale, individual particle resolved identification tracking algorithms. Microscale then performed definitions affine...
Recently, the deep neural network (DNN) has become one of most advanced and powerful methods used in classification tasks. However, cost DNN models is sometimes considerable due to huge sets parameters. Therefore, it necessary compress these order reduce parameters weight matrices decrease computational consumption, while maintaining same level accuracy. In this paper, deal with compression problem, we first combine loss function into a joint function, optimize as an optimization framework....
The deployment of unmanned aerial vehicles (UAVs) has significantly improved the efficiency data collection for wireless sensor networks (WSNs). freshness collected information from sensors can be measured by age (AoI), which is an important factor to consider in collection. For during long-term mission, energy limitation UAVs may cause mission interruption, makes supplementation UAVs’ more necessary. To this end, we introduce mobile vehicle (MUV) guarantee supplementation. In paper,...
Recording vibration signals induced by larvae activity in the trunk has proven to be an efficient method for detecting trunk-boring insects. However, accuracy of detection is often limited because collected real-world environments are heavily disrupted environmental noises. To deal with this problem, we propose a deep-learning-based model that enhances signals, incorporating attention mechanism optimize its performance. The training data utilized research consist boring vibrations Agrilus...
Statistical tests represent an important technique used to formulate and validate hypotheses on a dataset. They are particularly useful in the medical domain, where link disease with measurements, risk factors, treatment. In this paper, we propose compute parametric statistical treating patient records as elements multidimensional cube. We introduce that combines dimension lattice traversal discover significant differences degree of within pairs groups. order understand cause-effect...
Wood borers, such as the emerald ash borer and holcocerus insularis staudinger, pose a significant threat to forest ecosystems, causing damage trees impacting biodiversity. This paper proposes neural network for detecting classifying wood borers based on their feeding vibration signals. We utilize piezoelectric ceramic sensors collect drilling signals introduce novel convolutional (CNN) architecture named Residual Mixed Domain Attention Module Network (RMAMNet).The RMAMNet employs both...