Add option for minimal number of iterations

[Parskatt]

Hi! I'm having issues with pyrobustac terminating too early no matter how high high I set the confidence threshold. I'm basically having termination after 20 iterations (up to 50 iterations if I set confidence to 1-1e-15), which is too low.

I think these issues are due to affine correspondences needing fewer samples -> (1-p^2) or (1-p^3) instead of (1-p^5).

I think my concerns could easily be addressed by adding a minIter argument, to force ransac to run for longer.

[Parskatt]

This also shows in the variance of different ransac runs. With regular gcransac Im getting low variance between runs, while the affine estimates affine ransac gives me much higher variance.