Ask Your Question

How to handle outliers in calibration? [closed]

asked 2018-07-27 03:57:51 -0500

kevinkayaks gravatar image

I would like to calibrate cameras as well as possible. Among many calibration images, I can see by eye that certain chessboard corner locations have been affected by noise (water droplets). I would like to exclude these from the calibration.

I tried to set these imagepoints and their corresponding objectpoints to None: this did not work-- calibratecamera never converges.

What is the proper approach to mask outliers from the calibration computation?

edit retag flag offensive reopen merge delete

Closed for the following reason the question is answered, right answer was accepted by kevinkayaks
close date 2018-07-27 20:13:54.353533

1 answer

Sort by ยป oldest newest most voted

answered 2018-07-27 07:20:09 -0500

w33z gravatar image


i used the PCL library to remove outliers, here is an example

edit flag offensive delete link more


Actually it's not identification of outliers which is my problem, it's that I'm not sure how to input points arrays masked of outliers to calibrateCamera and stereoCalibrate.

Maybe I can study how/if this library does it

kevinkayaks gravatar imagekevinkayaks ( 2018-07-27 09:48:50 -0500 )edit

I obtained a solution. It's hard to explain. If anyone needs this, please email me

kevinkayaks gravatar imagekevinkayaks ( 2018-07-27 20:13:43 -0500 )edit

Question Tools

1 follower


Asked: 2018-07-27 03:57:51 -0500

Seen: 29 times

Last updated: Jul 27 '18