State of the art of vision-based self-localisation

This paper aims at giving an overview on state of the art tracking technology based on visual sensors. We cover different aspects of visual systems such as sensor technology, feature tracking, search complexity, and real-time behaviour. The presentation is structured into tracking of individual features, correspondence search, and pose estimation with an emphasis on algorithms. From our experience a recommended architecture for tracking would use CMOS sensors and extract a combination of features (corners, blobs, edges) to track.