Searching for Features in Laser Rangefinder Scan via Combination of Local Curvature Scale and Human Obstacles Detection

0421686 - ÚT 2014 RIV CH eng C - Conference Paper (international conference)
Krejsa, Jiří - Věchet, Stanislav
Searching for Features in Laser Rangefinder Scan via Combination of Local Curvature Scale and Human Obstacles Detection.
Mechatronics 2013. Cham: Springer, 2013, s. 599-606. ISBN 978-3-319-02293-2.
[Mechatronics 2013 /10./. Brno (CZ), 07.10.2013-09.10.2013]
Institutional support: RVO:61388998
Keywords : feature extraction * obstacle detection
Subject RIV: JD - Computer Applications, Robotics

Detection of reliable features in sensor data is the base for most localization and SLAM algorithms. The paper presents a combination of local curvature scale that enables to easily find environment features such as corners and edges in laser rangefinder scan data with human obstacles detection, used to extract such dynamic obstacles reading from the scan thus improving the robustness of feature detection.
Permanent Link: http://hdl.handle.net/11104/0229622