How could I make 3D reconstruction from just images, and what key technologies and research areas does it involves?

This is not like traditional photogrammetry technology, it depends on computer vision. 

Like PhotoScan software, load images, and directly obtain dense point cloud to conduct 3D reconstruction, without any other interactive operations. 

How could I implement it? what key technologies and research points are there? Thank you.

More Tian-Zhu Xiang's questions See All
Similar questions and discussions