Title :
3D mapping based VSLAM for UAVs
Author :
Li, Xiaodong ; Aouf, Nabil ; Nemra, Abdelkrim
Author_Institution :
Dept. of Inf. & Syst. Eng., Cranfield Univ., Shrivenham, UK
Abstract :
This paper addresses 3D texture mapping in Visual Simultaneous Localization And Mapping (VSLAM) for Unmanned Aerial Vehicle (UAV) applications. Landmark selection strategy based on feature detection methods such as Scale Invariant Feature Transform (SIFT) and Speed Up Robust Features (SURF) is adopted. The selected features are combined with additionally chosen features that are well distributed across the stereo views and refined by RANSAC in order to provide well visualized views for navigation. Experimental results are provided to demonstrate the effectiveness of our 3D mapping strategy.
Keywords :
SLAM (robots); aerospace computing; autonomous aerial vehicles; feature extraction; image texture; iterative methods; robot vision; stereo image processing; 3D mapping based VSLAM; 3D texture mapping; RANSAC; SIFT; SURF; UAV; feature detection methods; feature selection; landmark selection strategy; random sample consensus; scale invariant feature transform; speed up robust features; stereo views; unmanned aerial vehicle; visual simultaneous localization and mapping; visualized views; Cameras; Estimation; Feature extraction; Solid modeling; Surface texture; Vectors;
Conference_Titel :
Control & Automation (MED), 2012 20th Mediterranean Conference on
Conference_Location :
Barcelona
Print_ISBN :
978-1-4673-2530-1
Electronic_ISBN :
978-1-4673-2529-5
DOI :
10.1109/MED.2012.6265662
