Introduction to SLAM and monocular SLAM

A brief guide to general and single-camera simultaneous location and mapping

SLAM stands for simultaneous localisation and mapping and is a concept which solves a very important problem in mobile robotics, made up of two parts: mapping, building a map of the environment which the robot is in, and localisation, navigating this environment using the map while keeping track of the robot’s relative position and orientation.

Tackling the SLAM problem effectively opens up many possibilities in autonomous mobile robots with applications from driverless cars through to robotic vacuum cleaners. While common methods of solving this... Read more

Image: a map of a 3D environment, created using SLAM techniques (2).

Next

Applications of monocular SLAM

A selection of real world uses of monocular SLAM

Monocular slam has many possible applications in the real world, using drones to map 3D environments, find survivors from disaster zones, or as a cheap way of giving environmental awareness to household robots... Read more

Image: Dyson's 360 Eye robotic vacuum cleaner, which employs the use of monocular SLAM (3).

Next

SLAM methods and algorithms

A couple of common methods and algorithms used to implement monocular SLAM, including:

and more about general monocular SLAM algorithms.

Original image

Next

About this website

This website forms a part of group 7's project for the Computing Topics course at Imperial College London.

Supervised by Dr. Stefan Leutenegger, our group has produced this website to present information we have gathered while researching the area of monocular simultaneous localisation and mapping (SLAM). Over the nine-week course, conducting collective and individual research on multiple sub-topics within our designated research area has given us substantial experience in group work and coordination, academic research and referencing, web development and report writing. As well as this, we have all been introduced to the fascinating topic of SLAM and research at the forefront of this area.

We also delivered a presentation on this topic:

This website was made using a free template called "Landed", from HTML5 UP. It is licensed under a Creative Commons Attribution 3.0 Unported License and contained HTML pages, CSS sheets and Javascript scripts, which we adapted (raw coding using text editors) and used to create this website. Every member of the group contributed to the coding of this website and coded at least one of the pages in full.

Maurice Yap
(Group Leader)


Web pages responsible for:

  • Home page
  • Introduction to SLAM
  • Introduction to monocular
  • Methods and algorithms

Alessandro Bonardi


Web pages responsible for:

  • MonoSLAM
  • Tools supporting algorithms

Contributed heavily towards the mathematics shown.

Abbie Howell


Web pages responsible for:

  • PTAM
  • LSD-SLAM

Paul Larsen


Web pages responsible for:

  • Direct vs. feature based methods
  • Applications
  • LSD-SLAM

Reference list

  1. Jan S. LSD-SLAM: Large-scale direct Monocular SLAM. [Online] Computer Vision Group, Technische Universität München. Available from: http://vision.in.tum.de/research/vslam/lsdslam [Accessed: 29th January 2016]
  2. Ryan Eustice. NGV mapping & localization. [Online] YouTube. YouTube; 2013. Available from: https://www.youtube.com/watch?v=G6ARrD601Qk [Accessed: 14th March 2016]
  3. Kinney C. Finally, Dyson delivers the Dyson 360 eye - their First robotic vacuum. [Online] Chip Chick. Chip Chick; Available from: http://www.chipchick.com/2014/09/dyson-360-eye-first-robotic-vacuum.html [Accessed: 14th March 2016]