Full test sequence - Githubissues

tinghuiz / SfMLearner

An unsupervised learning framework for depth and ego-motion estimation from monocular videos

MIT License

1.97k stars 557 forks source link

Full test sequence #114

Open BernardoTeixeira opened 5 years ago

BernardoTeixeira commented 5 years ago

how can you connect 5-frame snippets (with the format of tum, (timestamp tx ty tz qx qy qz qw) into a complete pose estimate in a global coordinate system?