Minimal sample size for GP3P is incorrect

laurentkneip / opengv

OpenGV is a collection of computer vision methods for solving geometric vision problems. It is hosted and maintained by the Mobile Perception Lab of ShanghaiTech.

Other

1.02k stars 353 forks source link

Minimal sample size for GP3P is incorrect #75

Closed eguendelman closed 6 years ago

eguendelman commented 6 years ago

Hi, It seems as though when gp3p is used for noncentral absolute pose using RANSAC, it thinks that the minimal sample size is 4 (see code below) whereas it only really uses/needs 3. Is this intentional? Thanks

https://github.com/laurentkneip/opengv/blob/0b2017d23bbbb9a39d3f508ad2cd7042a58be682/src/sac_problems/absolute_pose/AbsolutePoseSacProblem.cpp#L217

eguendelman commented 6 years ago

Sorry, now I see that the fourth point is used for selecting the right solution. Please ignore my comment. Thanks.