Chenyu
The relative poses are obtained during feature matching. All you need to do is to rebuild a database. Actually, you can follow the pipeline in this [script](https://github.com/AIBluefisher/DAGSfM/blob/dev/scripts/shell/colmap_sfm.sh.in) but replace colmap's...
I'm not sure what causes the crash. You may need to debug that piece of code. But you can simply set `retriangulate` to false to avoid re-triangulation after merging: ```c++...
I think you aligned the sparse point clouds with the lidar scan to compute the F1 score. I wonder if it may relate to the SfM results' point density. Can...
> > I think you aligned the sparse point clouds with the lidar scan to compute the F1 score. I wonder if it may relate to the SfM results' point...
> Here are additional info:  The BA result:  upper one is global BA of distributed reconstruction, another is incremental reconstruction. The info of distributed reconstruction:  The scene...
> Here are additional info:  The BA result:  upper one is global BA of distributed reconstruction, another is incremental reconstruction. The info of distributed reconstruction:  Could you...
> > > Here are additional info:  The BA result:  upper one is global BA of distributed reconstruction, another is incremental reconstruction. The info of distributed reconstruction: ...
> > > Here are additional info:  The BA result:  upper one is global BA of distributed reconstruction, another is incremental reconstruction. The info of distributed reconstruction: ...
Yes. However, since COLMAP's database does not store relative poses, you need to modify the code that leverages relative poses to filter outliers. Let me know if you need further...
We can find a discussion [here](https://github.com/NVlabs/instant-ngp/discussions/367) for InstantNGP of how to use the LLFF dataset.