Skip to content

Latest commit

 

History

History
15 lines (11 loc) · 678 Bytes

README.md

File metadata and controls

15 lines (11 loc) · 678 Bytes

structure-from-motion

Reconstructing the 3-D positions of a set of matching points in the images and inferring the camera extrinsic parameters

Steps for SFM Pipeline:

  • Keypoint Feature Extraction using SIFT
  • Feature Matching using BruteForceMatcher
  • Finding Essential Matrix using RANSAC global matching
  • Decomposing Essential Matrix into (R, t) components
  • Triangulation

Visualization

  • The 3D points can be visualized in softwares like Meshlab where you can easily upload the output.obj file that is generated after running python3 sfm.py
  • The program outputs the rotation, translation, and projection matrices for each pair of images in the data folder