3D reconstruction from a depth map (with camera parameters)

General introduction:

I have images A and B taken with the same camera (from which I know the internal parameters of K). Then I tried to find the matrix F and from it the essential matrix. From this I have reconstructed the full matrices of the P and P 'cameras.

Then I straightened out the two images and computed the depth map like the following: Imgur

My question

Now I would like to perform a dense 3D reconstruction using a depth map. What's not entirely clear is what the camera matrix should use to reprogram points, P or P '?

Or another transformation? P and P 'were camera matrices for A and B, but now I am considering a depth map created using straightened versions of A and B


source to share

1 answer

When you fix images, you define new properties for both. From this you get a homogeneous 4x4 Q matrix that associates the image coordinates and disparity [xyd] with the world coordinates [XYZ]. See the OpenCV tutorial for details.



All Articles