How to run gradslam with the TUM dataset?
Closed this issue · 2 comments
How to run gradslam with the TUM dataset?
I want to run gradslam in TUM dataset. I try to run examples/icpslam.py
with the command:
python icpslam.py --dataset tum --dataset_path 'fr1's dataset'
But I just got two point clouds for visualization.
Can I run gradslam like orbslam2 or other slam method?
Thanks for your interest in GradSLAM!
Yes, you can definitely run gradslam for whole scene reconstructions.
This previous query is very related to yours: #12
I would actually point you to this demo code of gradslam PointFusion
on TUM/rgbd_dataset_freiburg1_xyz
:
https://colab.research.google.com/drive/1uxjEJI5B2yJJDG3SxoAagH9uEuT67rXC?usp=sharing
If you are interested in whole scene reconstructions, you can have a look at our gradslam documentation and the slam.step()
function and tutorials would be of high interest. Ideally, you would want to set up the gradslam tum dataloader
and run gradslam on batches to obtain whole scene reconstructions.
Closing, due to inactivity