A5091073210- Robotics and Sensor-Based Localization
- Advanced Vision and Imaging
- Oceanographic and Atmospheric Processes
- 3D Surveying and Cultural Heritage
- Marine and coastal ecosystems
- Advanced Image and Video Retrieval Techniques
- Optical measurement and interference techniques
- Arctic and Antarctic ice dynamics
- Advanced Image Fusion Techniques
- Advanced Neural Network Applications
- Intracerebral and Subarachnoid Hemorrhage Research
- Mechanical stress and fatigue analysis
- Climate variability and models
- Advanced biosensing and bioanalysis techniques
- Remote Sensing in Agriculture
- Electrical Contact Performance and Analysis
- Hydrological Forecasting Using AI
- QR Code Applications and Technologies
- 3D Shape Modeling and Analysis
- Automated Road and Building Extraction
- Climate Change and Health Impacts
- Metal and Thin Film Mechanics
- Planetary Science and Exploration
- Biosensors and Analytical Detection
- Geological and Geophysical Studies
Ningbo University
2024
Huawei Technologies (China)
2018-2024
State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing
2023
Wuhan University
2023
Ministry of Natural Resources
2020-2023
Ocean University of China
2023
Nanjing University
2023
Xi'an University of Technology
2023
Tianjin University of Science and Technology
2022
Ningbo Entry-Exit Inspection And Quarantine Bureau
2020-2022
The global increase in coastal hypoxia over the past decades has resulted from a considerable rise anthropogenically-derived nutrient loading. spatial relationship between surface phytoplankton production and subsurface hypoxic zones is often explained by considering oceanographic conditions associated with basin size, shape, or bathymetry, but this not case where nutrient-enriched estuarine waters merge into complex circulation systems. We investigated physical biogeochemical processes that...
The recent fast development in computer vision and mobile sensor technology such as LiDAR RGB-D cameras is pushing the boundary of to suit need real-life applications fields Augmented Reality (AR), robotics, indoor GIS self-driving. Camera localization often a key enabling among these applications. In this paper, we developed novel camera workflow based on highly accurate 3D prior map optimized by our SLAM method conjunction with deep learning routine trained using consecutive video frames...
In the study of SLAM problem using an RGB-D camera, depth information and visual as two types primary measurement data are rarely tightly coupled during refinement camera pose estimation. this paper, a new method is proposed based on extended bundle adjustment with integrated 2D 3D basis projection model. First, geometric relationship between image plane coordinates values constructed through calibration. Then, feature points automatically extracted matched consecutive frames to build...
Temporal variation of summer hypoxia, along a repeated hydrographic section from the Changjiang River mouth to Cheju Island, is investigated using cruise data during 1997–2014. The climatological mean dissolved oxygen (DO) presents “dual-core structure” below oxycline, associated with two hypoxic centers near river and offshore Bank, respectively. DO value highly variable year year, dramatic interannual but an insignificant linear trend Composite analysis indicates that hypoxia events are...
With extensive application of RGB-D cameras in robotics, computer vision, and many other fields, accurate calibration becomes more critical to the sensors. However, most existing models for calibrating depth relative pose between a camera an RGB are not universally applicable different kinds cameras. In this paper, by using collinear equation space resection photogrammetry, we present new model correct calibrate based on 3D control field. We establish rigorous relationship two cameras; then,...
Abstract This study aims to investigate the impact of submesoscale vertical advection (SVA) on primary productivity in southern East China Sea. The analysis is based a comparison between two numerical simulations by using three‐dimensional coupled physical‐biogeochemical model. One simulation directly resolves SVA high‐resolution mesh, and other leaves unresolved low‐resolution mesh. outperforms reproducing observed chlorophyll distribution, particularly summer. Resolving results an...
In a point cloud semantic segmentation task, misclassification usually appears on the boundary. A few studies have taken boundary into consideration, but they relied complex modules for explicit prediction, which greatly increased model complexity. It is challenging to improve accuracy of points without dependence additional modules. For every point, this paper divides its neighboring different collections, and then measures entanglement with each collection. comparison measurement results...
Remote sensing images often need to be merged into a larger mosaic image support analysis on large areas in many applications. However, the performance of imagery may severely restricted if there are with cloud coverage or these used for merging have long-time span. Therefore, this paper proposes method selection full (i.e. no cloud-contaminated pixels) generation. Specifically, novel High-Frequency-Aware (HFA)-Net based Swin-Transformer region quality grading is presented provide data basis...
It is a key technology that extract and segment book call number from books’ backbone images for retrieval return robot to take an operation of sending withdrawing books or the shelves library. This paper presents image extraction segmentation method using color feature edge information adjacent with similar background. Take thick, thicker, medium thin, little thin hybrid as experiment samples, results show this effectively extracts segments number. The approach provides technical support...
The correlation between meteorological parameters and intracerebral hemorrhage (ICH) occurrence is controversial. Our research explored the effect of daily on ICH risk in a subtropical monsoon basin climate.We retrospectively analyzed patients with teaching hospital. Daily including temperature (TEM), atmospheric pressure (PRE), relative humidity (RHU), sunshine duration (SSD) were collected, diurnal variation (daily maximum minus minimum) day-to-day (average day previous day) calculated to...
Abstract. RGB-D camera allows the capture of depth and color information at high data rates, this makes it possible beneficial integrate image sequences for planetary rover mapping. The proposed mapping method consists three steps. First, strict projection relationship among 3D space, visual texture is established based on imaging principle camera, then, an extended bundle adjustment (BA) SLAM with integrated 2D measurements applied to network high-precision pose estimation. Next, as...
High precision topographic mapping of Chang'e-3 (CE-3) landing site can provide detailed terrain information to ensure the safety rover as well support scientific investigations. In this paper, high resolution sequence images acquired by lander's descent camera and stereo captured rover's Navigation cameras (Navcam) are used generate high-precision products respectively. Furthermore, taking DOM Chang'e-2(CE-2) spatial reference, CE-2 CCD images, DEM, generated from Navcam co-registered into...
This paper proposes a novel approach to determine three-dimensional (3-D) model from minimum number of viewpoint an object.Calculating the position relation location two coordinate system in order splice multiview point cloud.The Next Best Views (NBV) is calculated by visibility classification and Mean-shift algorithm then overall morphology measured object can be acquired.The view planning method not only reaches fast measurement speed, but also possesses high fidelity complex regions experiments.
Ocean eddies are typical oceanic mesoscale phenomena that numerous, widely distributed and have high energy. Traditional eddy detection methods mainly based on physical mechanisms with accuracy. However, the large number of steps complex parameter settings limit their applicability for most users. With rapid development deep learning techniques, object models been broadly used in field ocean remote sensing. This paper proposes a lightweight model, ghost YOLO (GED-YOLO), sea level anomaly...