Multi-sensor Arial Robots SLAM Dataset
Mr. Haotian Li (M.Phil. candidate)
Department of Mechanical Engineering
The University of Hong Kong
Date & Time
Friday, 21 April 2023
Room 7-34, Haking Wong Building, HKU
To further promote relative research, public datasets are crucial as they act key roles in saving resources from expensive instruments and cumber-some data collection processes. This seminar introduces an aerial downward shooting dataset from MaRS Lab collected in multiple scenes including airports, islands, towns, and valleys, which contains data from a unique lightweight hardware synchronized sensor package including Light Detection and Ranging (LiDAR), Inertial Measurement Units (IMUs) and a global-shutter RGB camera. Compared with existing datasets, the downward shooting scanning method is the most suitable way to collect aerial data. This method can provide the most geographic information, but it also poses great challenges for sensor fusion algorithms. In order to facilitate the evaluation of the dataset, the mapping result from DJI L1, a commercial aerial mapping solution, which is carrying a similar sensor set and collecting data together with our device is provided as the mapping ground truth, and uses the on-board Real-time kinematic (RTK) position as the position ground truth.