build 3d map from multi depth images(infer outputs)

TRI-ML / packnet-sfm

TRI-ML Monocular Depth Estimation Repository

https://tri-ml.github.io/packnet-sfm/

MIT License

1.21k stars 241 forks source link

build 3d map from multi depth images(infer outputs) #209

Open sparkfax opened 2 years ago

sparkfax commented 2 years ago

packetnet-sfm is an excellent, thanks for your work. I have multi images for a building and get depth images from infer. I want to build 3d map for the building from sfm output, how to do that?

camviz do not fusion multi image depth together, am I right?

VitorGuizilini-TRI commented 2 years ago

You can display the pointcloud from multiple cameras simultaneously, is that what you mean by fusion? If you have the correct depth and pose, they should blend in seamlessly (that's what we show on our papers, we don't do any post-processing on the pointclouds).

sparkfax commented 2 years ago

I have 100 depth images, taken from different angles by one camera, for a building. How to get a 3D map of this building from these 100 depth images?