A A A

Josh Veitch-Micahelis

Josh Veitch-Michaelis

Imaging Group

PhD Student

Tel: +44 (0)1483 204 927

Internal: 4927

Room: Boyd Centre E

Email: joshua.veitch-michaelis.12 {at} ucl.ac.uk

Research Interests

  • Fusion of stereo and LIDAR data

Qualifications

  • B.Sc. MPhys Physics (1st Class Hons), University of Warwick (2012)
  • Ph.D Space and Climate Physics, University College London (in progress)

Project

Fusion of LIDAR with stereo camera data - an assessment

Supervisors: Prof. Jan-Peter Muller and Dr Jonathan Storey (Innovative Small Instruments)

The main objective of this project is to examine and assess different methods that can be employed to merge stereo camera data with LIDAR, particularly imaging LIDAR, to provide 3D vision with high accuracy over a wide Field of View (FoV). Stereo imaging is a universal technique that provides dense 3D point cloud data and can be employed from microscopic to orbital distances. However, stereo suffers from high computational intensity and low range resolution [1].

Stereo range precision is limited by the base-to-height ratio (BHR) and the Instantaneous Field-of-View (IFoV) but covers a wide Field of View (FoV) whereas LIDAR is limited to a few spots or by the speed of the data acquisition system and the energy efficiency at large ranges or with imaging LIDAR by a narrow FoV. Traditionally, these techniques have been in competition, but ISI and UCL-MSSL have developed, as part of a TSB space feasibility study, an innovative approach to merging them to gain the advantages of each. Stereo vision can now benefit from advances in computing power (e.g. use of GPUs), whilst recent advances in optoelectronics and microelectronics allow reduction of complexity and cost of the LIDAR.

This PhD study will be based around developing and testing a design concept for a terrestrial guided vision system combining stereo-cameras with an imaging lidar. Issues concerned with measuring image-LIDAR geometry, co-registration of LIDAR and stereo camera measurements and the optimum methods for combining these two distinct types of range measurements will be studied alongside the implementation of such data fusion at near real-time video refresh-rates. This will include how best to improve the throughput of the stereo processing, how to employ LIDAR as seedpoints for region growing- based stereo matching as well as for stereo camera exterior orientation measurement on a per-frame basis and insertion of LIDAR points where there is insufficient image texture for matching.

The instrument is originally proposed as a ground-based system but design and processing should be directly transferable to spaceborne or rover systems enabling MSSL to build on its impressive track-record in developing advanced navigation systems.

Potential design concepts will be implemented in a simulation tool, partially developed by ISI to test their suitability for the underlying measurements required and laboratory and field trials preformed where possible (depending on availability of suitable equipment).

[1] J.-P. Muller, K.A. Collins, G.P. Otto, J.B.G. Roberts, Stereo Matching Using Transputer Arrays, in: S. Murai, (Ed), International Archives of Photogrammetry and Remote Sensing 27(B3), Kyoto, Japan, 1988, pp. 559-586.

Publications

Outreach

  • Working group leader at the International Astronomical Youth Camp (www.iayc.org)
  • Ordinary Member of the International Workshop for Astronomy e.v.

Activities

2013

  • RSPSoc Student Meeting, Glasgow (March)

One fact about me...

test

Page last modified on 19 aug 13 01:50