I have two views of a scene and I want to reconstruct the 3D points from the red lines that we can be seen from these two images:
the 3D reconstruction of the above red lines (using epipolar constrains) would be this:
the result is not correct. Left camera is “1” and the right camera is “2” (see the picture above). The blue points are constructed from left image to the right image (using epipolar constraint) and the red points are constructed from right image to left image (using epipolar constraint).
Why these points are extended spatually in direction of each 2D view? and the reconstructed scene is also wrong (upside down and left to right) as we see, the original red line in the original left image (camera 1) and right image (camera 2) above.
Is this maybe because I draw the red line wrong (something unphysical)?