Skip to content

swayamvaza/SFM-Reconstruct

 
 

Repository files navigation

Sparse 3D Reconstruction Project

A simple, multi-view 3D reconstruction pipeline using:

  • SIFT feature extraction
  • Feature matching
  • Epipolar geometry
  • Pose recovery
  • Triangulation
  • Incremental sparse reconstruction
  • Open3D visualization

Files

  • main.py — entry point
  • utils.py — image loading and camera intrinsics
  • features.py — SIFT extraction
  • matching.py — matching and image pair selection
  • geometry.py — essential matrix, pose recovery, PnP registration
  • triangulation.py — triangulation helpers
  • visualization.py — point cloud save + display
  • bundle_adjustment.py — placeholder / optional refinement hook
  • reconstruction.py — two-view and incremental reconstruction pipeline

Install

Install all the packages at once

pip install requirements.txt

Optional for bundle adjustment:

pip install scipy

Run

Two-view demo:

python main.py --images_dir path/to/images --mode two_view

Incremental mode:

python main.py --images_dir path/to/images --mode incremental

About

A pipeline that converts image sequences to a point cloud in the .ply extension

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages

  • Python 100.0%