search

sourishg / stereo-calibration

:camera: :camera: Stereo camera calibration using OpenCV and C++
http://sourishghosh.com/2016/stereo-calibration-cpp-opencv/
612 stars 239 forks source link

Statistical analysis of calibration #8

Open sourishg opened 6 years ago

sourishg commented 6 years ago

  1. Generate insightful plots describing camera calibration accuracy. Here's an example: plot the graph of reprojection error vs number of images used for calibration.

  2. Compare multiple methods (AprilTags vs Checkerboard)

BarunKGP commented 6 years ago

How do I get the data? Do I run the code and store the data? Also, how do you want the comparison done?

sourishg commented 6 years ago

@BarunKGP yes you should collect the data by running the code. You can focus on solving (1) only for now since AprilTags calibration method for (2) isn't implemented yet.

I recommend using matplotlib for plotting.

pushkalkatara commented 6 years ago

The cam_left.yaml and the cam_stereo.yaml files are generated using opencv. Should i load these files using pyyaml in python and plot the graph in matplotlib or should i use matplotlib wrapper for c++?

sourishg commented 6 years ago

@pushkalkatara why are loading the yaml files? I think you only need to record the calibration error (which right now I'm echoing out in the terminal) and the number of images used for calibration (specified by the -n flag).

No need to write a C++ wrapper for matplotlib, you can use python.

pushkalkatara commented 6 years ago

@sourishg Please review the PR.

BarunKGP commented 6 years ago

@sourishg, should I embed python code in calib_intrinsic.cpp to use matplotlib and print out the graphs?