Visual SLAM

This repo is based on the code introduced in the SLAM book: https://github.com/gaoxiang12/slambook2/tree/master

sphere_pose_graph.mp4

Gray points are initial guesses, while black points are optimized. An extension of this could be to visualize only the most recent optimization step.

Quickstart

Ensure bazel is installed, then run:

bazel run //:vslam -- $PWD/sphere.g2o

To-do

• profiling helper to call any function and evaluate time taken
• trajectory evaluation metrics
• camera calibration methods