A5002619272- Robotics and Sensor-Based Localization
- Remote Sensing and LiDAR Applications
- 3D Surveying and Cultural Heritage
- Advanced Vision and Imaging
- GNSS positioning and interference
- Visual Attention and Saliency Detection
- Target Tracking and Data Fusion in Sensor Networks
- Advanced Neural Network Applications
- Inertial Sensor and Navigation
- Ultra-Wideband Communications Technology
- Forest Ecology and Biodiversity Studies
- Automated Road and Building Extraction
- Invertebrate Immune Response Mechanisms
- Mosquito-borne diseases and control
- Advanced Adaptive Filtering Techniques
- Advanced Optical Sensing Technologies
- Venomous Animal Envenomation and Studies
Shandong University
2024-2025
State Key Laboratory of Information Engineering in Surveying Mapping and Remote Sensing
2019-2024
Wuhan University
2019-2024
Ministry of Education of the People's Republic of China
2022
Automatic 3D forest mapping and individual tree characteristics estimation are essential for management ecosystem maintenance. The low-cost unmanned aerial vehicle (UAV) laser scanning (ULS) is a newly developed tool cost-effectively collecting information attempts to use it have been made, due its capability provide with lower cost higher flexibility than the standard ULS airborne (ALS). As direct georeferenced point clouds may suffer from distortion caused by poor performance of inertial...
Interpretation of airborne laser scanning (ALS) point clouds plays a notable role in geoinformation production. As critical step for interpretation, accurate semantic segmentation can considerably broaden various applications ALS data. However, most existing methods cannot provide precise annotations and high robustness due to occlusions, varied densities, complex incomplete object structures. Therefore, we developed framework focusing on clouds. The comprises contextual feature extraction...
The safe operation of unmanned ground vehicles (UGVs) demands fundamental and essential requirements for continuous reliable positioning performance. Traditional coupled navigation systems, combining the global satellite system (GNSS) with an inertial (INS), provide continuous, drift-free position estimation. However, challenges like GNSS signal interference blockage in complex scenarios can significantly degrade Moreover, ultra-wideband (UWB) technology, known its high precision, is...
Abstract Accurate and reliable positioning is a crucial foundation for achieving sufficient autonomy intelligence in systems. The ultra-wideband (UWB) technology has significant potential applications the challenging complex field of indoor positioning, often integrated with inertial navigation system (INS) to enhance overall performance system. This paper proposes novel variational Bayesian-based cubature information filter that enhances accuracy tightly coupled UWB/INS approach aims...
Abstract. Detection And Tracking of Moving Objects (DATMO) is essential and necessary for mobile mapping system to generate clean accurate point clouds maps since dynamic targets in real-world scenarios will deteriorate the performance whole system. In this research, a robust LiDAR-SLAM presented incorporated with real-time objects removal module improve accuracy 6 DOF pose estimation precision maps. The key idea proposed method efficiently cluster sparse moving then track them independently...
Abstract. This paper proposes a real-time 3D image point clouds mapping algorithm for UAVs that is capable of effectively in weak GNSS environments. And UAV system integrated with RGB camera, an inertial measurement unit (IMU), receiver, data transmission devices, and DJI M300 flight platform. To achieve robust mapping, the utilizes visual-inertial odometry (VIO) tightly couples GNSS, image, IMU, which provides stable state estimation information mapping. Subsequently, dense matching based...
Abstract. Aiming to accomplish automatic and real-time three-dimensional mapping in both indoor outdoor scenes, a low-cost wheeled robot-borne laser scanning system is proposed this paper. The includes scanner, an inertial measurement unit, modified turtlebot3 two-wheel differential chassis etc. To achieve globally consistent map, the performs global trajectory optimization after detecting loop closure. Experiments are undertaken two typical indoor/outdoor scenes that underground car park...
Abstract. Light Detection And Ranging (LiDAR) technology has provided an impactful way to capture 3D environmental map. However, consistent mapping in sensing-degenerated and perceptually-limited scenes (e.g. multi-story buildings) or under high dynamic sensor motion rotating platform) remains a significant challenge. To this end, efficient multi-floor indoor system is proposed paper utilizing versatile LiDAR platform. In the front measurements from motor, IMU are tightly integrated track...
Abstract. To integrate the multi-spectral imagery and laser scanning data for forest management, a low-cost multisensory UAV system, named Kylin Cloud, is introduced in this paper. The Cloud composed several sensors (i.e., GNSS receiver, IMU, global shutter camera, multispectral scanner), providing fusion of reliable inventory. Experiments were undertaken park Wuhan. Results showed that registration error Digital Orthophoto Map (DOM) about one pixel, demonstrating high potential proposed system.
Abstract. Accurate registration of sparse sequential point clouds data frames acquired by a 3D light detection and ranging (LiDAR) sensor like VLP-16 is prerequisite for the back-end optimization general LiDAR SLAM algorithms to achieve globally consistent map. This process also called odometry. Aiming lower drift robust odometry in less structured outdoor scene using low-cost wheeled robot-borne laser scanning system, segment-based sampling strategy proposed this paper. Proposed method was...
At present, the anchor-based or anchor-free models that use LiDAR point clouds for 3D object detection center assigner strategy to infer bounding boxes. However, in a real world scene, can only acquire limited surface clouds, but of does not exist. Obtaining by aggregating incomplete will bring loss accuracy direction and dimension estimation. To address this problem, we propose corner-guided single-stage model (CG-SSD ).Firstly, sparse convolution backbone network composed residual layers...