A5102005834- Synthetic Aperture Radar (SAR) Applications and Techniques
- Advanced SAR Imaging Techniques
- Underwater Acoustics Research
- Advanced Neural Network Applications
- Inertial Sensor and Navigation
- Robotics and Sensor-Based Localization
- Mechanical stress and fatigue analysis
- Infrared Target Detection Methodologies
- Optical measurement and interference techniques
- Advanced Sensor and Control Systems
- Geophysical Methods and Applications
- Advanced Measurement and Detection Methods
- Simulation and Modeling Applications
- GNSS positioning and interference
- Potato Plant Research
- Image Enhancement Techniques
- Vibration and Dynamic Analysis
- Irrigation Practices and Water Management
- Advanced Battery Technologies Research
- Image and Object Detection Techniques
- Modular Robots and Swarm Intelligence
- Satellite Communication Systems
- Remote Sensing and LiDAR Applications
- Advanced Numerical Analysis Techniques
- Soil Carbon and Nitrogen Dynamics
National University of Defense Technology
2016-2025
Northwest A&F University
2023-2024
Harbin Institute of Technology
2013-2024
Southwest University of Science and Technology
2024
PLA Army Engineering University
2019-2024
Ministry of Agriculture and Rural Affairs
2022-2024
First Research Institute of the Ministry of Public Security
2023
Naval University of Engineering
2017-2023
Xi'an University of Technology
2023
Agriculture and Forestry University
2023
Recent deep learning (DL) detectors adopted by radar or sonar (RS) are normally trained with transfer learning, where the typical workflow is to pretrain a convolutional neural network (CNN) on external large-scale classification datasets (e.g., ImageNet) as backbone and then finetune entire detector detection datasets. Though could effectively avoid overfitting, transferred models usually redundant might not generalize well RS To achieve high generalization eliminate dependence self-trained...
The utilization of polarimetric synthetic aperture radar (PolSAR) enables the preservation more comprehensive target scattering characteristics compared to conventional SAR. Despite significant advancements in deep learning (DL) technology for single SAR ship detection, there remains a scarcity DL research specifically focused on PolSAR detection. This paper proposed Lightweight Theory-Driven Network (LT-Net) combining domain knowledge and techniques, which is suitable detect small targets...
Underwater target detection exhibits extensive applications in marine exploration and environmental monitoring. However, conventional images of underwater targets present challenges including blurred contour information, complex conditions, pronounced scattering effects. In this work, an method based on YOLOv10 is designed, the performance compared with YOLOv5 model. Experimental results demonstrate that model has a mAP50 85.6% URPC 2020 dataset, improving by 1.2% than YOLOv5. This high...
Abstract The arm structure of the industrial robot is main bearing mechanical arm, and mass relatively large. In this paper, selected for lightweight research. Firstly, original steel analyzed by using finite element simulation software to obtain its stiffness displacement. Then, carbon fiber-reinforced thermoplastic polymer composite boom designed optimized according principle equal stiffness. model material replaced fiber material, it found that weight greatly reduced not reduced. On...
To evaluate longitudinal vibration of flexible hoisting systems with time-varying length, the governing equations are developed employing extended Hamilton's principle considering mutual influence rigid motion and deformation systems. The Galerkin's method is used to discretize equations. In mean time, A Lyapunov controller which can dissipate vibratory energy derived. motions elevator were illustrated proposed mathematical models controller. results simulation show that modeling methods...
The eutectic mixture of lauric acid (LA) and stearic (SA) is a desirable phase change material (PCM) due to the constant melting temperature large latent heat. However, its poor thermal conductivity has hampered broad utilization. In present study, pure LA, SA mixtures with various mass fractions LA-SA were used as basic PCMs, 10 wt% expanded graphite (EG) was added enhance conductivities. behaviors, microstructural analysis, conductivities stabilities PCMs investigated by differential...
Infrared target intrusion detection has significant applications in the fields of military defence and intelligent warning. In view characteristics targets as well inspection difficulties, an infrared algorithm based on feature fusion enhancement was proposed. This combines static mode analysis dynamic multi-frame correlation to extract features at different levels. Among them, LBP texture can be used effectively identify posterior patterns which have been contained library, while motion...
This paper proposes a agricultural robot's complete coverage path-planning method based on sub-region to avoid the overabundance of turns in narrow areas. First, given environment is divided into several sub-regions without preserves. Second, robot finds out covering order subregions by transforming it problem Depth-First Search (DFS), and then can cover all DFS order. When every sub-region, travels along longer side sub-region. The simulation experiment verifies that effectively reduce number turns.
This paper presents a millimeter-scale crystal-less wireless transceiver for volume-constrained insertable pills. Operating in the 402-405 MHz medical implant communication service (MICS) band, phase-tracking receiver-based over-the-air carrier recovery has ±160 ppm coverage. A fully integrated adaptive antenna impedance matching solution is proposed to calibrate variation inside body. tunable network (TMN) with single inductor performs both transmitter (TX) and receiver (RX) TX/RX mode...
When a synthetic aperture sonar (SAS) system operates under low-frequency broadband conditions, the azimuth range coupling of point target reference spectrum (PTRS) is severe, and high-resolution imaging limited. To solve above issue, we first convert multi-receivers’ signal into equivalent monostatic then divide subblocks frequency subbands within each subblock in order. The terms are converted linear based on piece-wise approximation (PLA), phase error PTRS subband less than π/4. Then, use...
In multireceiver synthetic aperture sonars (SASs), the traditional non-stop-hop-stop model normally ignores intrapulse motion of sonar platform. However, such brings Doppler scale effect (DSE) to echo signal, which will lead change in carrier frequency (CF) and modulation (FM) rate signal may eventually result image defocus. this article, considering motion, time-varying time delay range history SAS were derived by introducing into history, then was established. After approximating 2-D...
This paper presents a method for airport detection from optical satellite images using deep convolutional neural networks (CNN). To achieve fast with high accuracy, region proposal by searching adjacent parallel line segments has been applied to select candidate fields potential runways. These proposals were further classified CNN model transfer learned AlexNet identify the final regions other confusing classes. The proposed tested on remote sensing dataset consisting of 120 airports....
At present, there are no harvesters specifically adapted to process peanuts after cutting stalks. In particular, methods for harvesting stalks have not been reported thus far. Therefore, utilize peanut as feed when plastic-film-grown peanuts, and improve industry benefits, a three-stage method is proposed herein. view of the method, cut before digging, with remaining plants being shorter, thereby increasing pod-stalk ratio. To investigate adaptabilities existing in stalks, three types...
On 20 April 2013, a moment magnitude (Mw) 6.6 earthquake occurred in the Lushan region of southwestern China and caused more than 190 fatalities. In this study, we use geodetic data from nearly 30 continuously operating global positioning system (GPS) stations, two periods leveling data, interferometric synthetic aperture radar (InSAR) observations to image coseismic deformation earthquake. By using Helmert variance component estimation method, joint inversion is performed estimate source...
Precise orientation information facilitates target recognition of sonar images. Traditional estimation methods, including the Hough transform method, have poor antinoise abilities, so they cannot achieve high precision for images, which usually low signal-to-noise ratios. The convolutional neural network (CNN) regression method is sensitive to image affine transformations and requires extensive computation, it strong robustness or speed. To high-precision, high-speed, robust targets, we...