Supposed I have 2 projection matrices in the form of P1 (3x4 matrix) and P2 (3x4) matrix, is there a function in opencv that could compute the fundamental matrix?
Apparently there's a function called fundamentalFromProjections: [http://docs.opencv.org/trunk/d7/d15/g...]) but I couldn't find it in opencv in python.
The sfm module is difficult to compile with a lot of dependencies, but the code is still in the opencv_contrib repository. This method is simple enough.
https://github.com/opencv/opencv_contrib/blob/master/modules/sfm/src/fundamental.cpp#L109
Well, the vconcat function is the same as numpy vstack. The (i,j) access is the same as numpy's .item(i,j). And determinant is numpy.linalg.det.
I assume you know how to access one row of data, and create the 3x3 double or float output.
Array accesses. It makes an array of Mat_<t> called X and Y, and it's just like a python array. The Mat_<t> can be replaced with a numpy array of the appropriate type.
Asked: 2017-03-02 16:14:07 -0600
Seen: 2,626 times
Last updated: Mar 02 '17
