Abstract
Recent biological evidence suggests that position and orientation can be estimated from an adequately compressed set of environment snapshots and their relationships. In this paper we present a pure appearance-based localisation method using an eigenspace representation of panoramic images. We first review several types of rotational invariant representation of panoramic images in terms of their efficiency for an eigenspace-based localisation problem. Then, for each set of images an eigenspace from 25 location snapshots is built and analyzed. We evaluated simple localisation of images not included in the training set. The results show good prospects for the panoramic eigenspace approach.
Cite
CITATION STYLE
Jogan, M., & Leonardis, A. (1999). Panoramic eigenimages for spatial localisation. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 1689, pp. 558–567). Springer Verlag. https://doi.org/10.1007/3-540-48375-6_67
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
