Combining SLAM poses with openMVG

openMVG / openMVG

open Multiple View Geometry library. Basis for 3D computer vision and Structure from Motion.

Mozilla Public License 2.0

5.56k stars 1.65k forks source link

Combining SLAM poses with openMVG #1887

Closed fightmoney closed 2 years ago

fightmoney commented 3 years ago

Hello, I'm trying to run openMVG dense reconstruction with camera poses pre-computed with ORB-SLAM3. How to put poses into the system?

pmoulon commented 3 years ago

I would suggest you to initialize a sfm_data.json from the output of ORB-SLAM3. You can use OpenMVG API to do that and take inspiration from the importer there https://github.com/openMVG/openMVG/tree/develop/src/software/SfM/import Let me know if it is helpful

RamadanAhmed commented 3 years ago

I think what you should use is a depth/disparity estimation using deep learning or any semi-global matching technique and apply transformation using camera pose to the point cloud you get from the image Just to keep the realtime behavouir of slam. In that case you need filterning techniques to filter overlapped points.

pmoulon commented 2 years ago

Closing due to lack of feedback. Feel free to reopen