SLAM, or simultaneous localization and mapping, is a key component in the development of truly independent robots. Vision-based SLAM utilising stereo vision is a promising approach to SLAM but it is computationally expensive and difficult to implement. New feature manipulation techniques are proposed which incorporate relational and positional information of the features into the extraction and data association steps.
Reference:
Joubert, D. 2010. Clustered features for use in stereo vision SLAM. Proceedings of the 25th International Conference of CAD/CAM, Robotics & Factories of the Future Conference, CSIR International Convention Centre, Pretoria, South Africa, 13-16 July 2010, pp 7
Joubert, D. (2010). Clustered features for use in stereo vision SLAM. http://hdl.handle.net/10204/4499
Joubert, D. "Clustered features for use in stereo vision SLAM." (2010): http://hdl.handle.net/10204/4499
Joubert D, Clustered features for use in stereo vision SLAM; 2010. http://hdl.handle.net/10204/4499 .
Proceedings of the 25th International Conference of CAD/CAM, Robotics & Factories of the Future Conference, CSIR International Convention Centre, Pretoria, South Africa, 13-16 July 2010
