search

ducha-aiki / pymagsac

MAGSAC: marginalizing sample consensus, python version
Other
72 stars 14 forks source link

Essential Matrix estimation for equi-rectangular projection cameras. #11

Open rishchou opened 3 years ago

rishchou commented 3 years ago

Is there a way to estimate Essential Matrix using DefaultEssentialMatrixEstimator() for cameras which don't have camera intrinsic matrices (for eg, 360 degree cameras)

danini commented 3 years ago

For a new problem, you have to implement a solver and the residual functions.

You can copy class EssentialMatrixFivePointSteweniusSolver (https://github.com/danini/graph-cut-ransac/blob/master/src/pygcransac/include/solver_essential_matrix_five_point_stewenius.h) and implement your solver there.

I am not sure what is the point-to-model residual function there. If it is not the Sampson distance or the Symmetric Epipolar Distance, you should copy class EssentialMatrixEstimator (https://github.com/danini/graph-cut-ransac/blob/master/src/pygcransac/include/essential_estimator.h) and implement the residual function there.

Finally, you can define your own estimator as done in https://github.com/danini/graph-cut-ransac/blob/master/src/pygcransac/include/types.h.

I hope it helped.