/cv_challenge

Camera pose estimation using epipolar geometry

Primary LanguagePython

What's the challenge?

You are expected to find the 6DoF pose* of the camera in images img2 and img3 w.r.t. pose in reference image img1.

My Solution

Using SIFT feature tracking, FLANN matching and subsequent calculations of the fundamental and essential matrix. After decomposing the essential matrix finding the right R and T out of four possible combinations, by checking if points are in front of both cameras.