Abstract

Visual data provides a rich source of information about our surroundings. It is therefore not surprising that computer vision techniques have been used for many years to help autonomous systems understand where they are in the world and how they can use that information to navigate. For example, techniques have been developed that are able to build detailed 3D models of small-scale environments and determine the 3D pose of the sensor within those models, all in real-time. At a larger scale, single images have been used to construct topological maps of large scale urban areas, enabling place recognition and navigation, even in changing lighting and weather conditions.

In this talk, Dr Calway will give an overview of his research in this area over the past 10 years. He will start with work on robust feature matching for some of the early monocular simultaneous localisation and mapping (SLAM) systems, to more recent work on fast relocalisation in maps and HDR reconstruction using RGB-D sensing and new work on semantic based localisation in 2D maps from images. The talk will assume an audience with a mix of backgrounds and will aim to provide an intuitive understanding of the challenges, the techniques that work and their capabilities.

Biography

Andrew Calway is a Reader in Computer Vision based in the Department of Computer Science at the University of Bristol. He is a member of the Bristol Vision Institute, the Visual Information Lab and the Bristol Robotics Lab. After obtaining his PhD from Warwick University in 1989, he spent two years as a visiting Royal Society Researcher in the Computer Vision Laboratory at Linkoping University in Sweden, prior to joining Bristol in 1995.

Dr Calway has worked in most areas of computer vision over a period spanning more than 25 years, with the last 10 years focused on work in visual SLAM and 3D tracking. As well as developing fundamental algorithms, he has led projects on the applications of the techniques to a range of areas including Wearable Computing, Augmented Reality and autonomous control of UAVs, funded by the EPSRC, EU and Innovate UK and in collaboration with numerous industrial partners.

http://people.cs.bris.ac.uk/~andrew/