findFundamentalMat is returning wrong results with FM_8POINT on 3.0.0 beta (Bug #4186)

Added by Ujjwal Kumar about 10 years ago. Updated almost 10 years ago.

Status:Done Start date:2015-02-09
Priority:Normal Due date:
Assignee:- % Done:

0%
Category:calibration, 3d
Target version:3.0-beta
Affected version:branch 'master' (3.0-dev) Operating System:Windows
Difficulty:Medium HW Platform:x86
Pull request:

Description

It was working good in opencv 2.4.6,
I used some 3d points and projected on two images with fixed
pose change of 30 degrees around y axis.
all 3d points are non-planar. The projections are 100% accurate.
Then converted the resulting fundamental matrix to essential matrix.
and then recovered the rotation and translation.
none of the four rotation and translation values are close to actual pose.
The same experiment is working good with 7 point algorithm. So other maths is correct.

Related issues

duplicates Bug #3441: 8-point fundamental matrix estimation seems to be broken Done 2013-12-18

Associated revisions

Revision f7792667
Added by Alexander Alekhin over 9 years ago

Merge pull request #4186 from ultrafro:fix_setDepthGeneratorProperty_registration_flag_polarity

History

#1
Updated by Deanna Hood almost 10 years ago

  • Status changed from New to Done

Also available in: Atom PDF