Dense Visual Odometry from a Monocular RGB-D Camera

This repository contains an algorithm to estimate visual odometry with a monocular RGB-D camera. It has been tested on Matlab 2015b and Ubuntu 14.04.

This algorithm is closely related to the following papers:

• F. Steinbrucker, J. Sturm and D. Cremers, Real-Time Visual Odometry from Dense RGB-D Images, ICCV workshop 2011.
• C. Kerl, J. Sturm and D. Cremers, Robust Odometry Estimation for RGB-D Cameras, ICRA 2013.

Usage

• Exemplar frame sequences are listed in examples, which are randomly picked from TUM RGB-D SLAM Dataset. Please check the dataset configurations.

• Please test the dense visual odometry demo via running demo.m