mrcal - camera calibrations and more!
mrcal is a photogrammetry toolkit that provides improved methods for making and using camera models (calibration, tracking, mapping, etc). Currently the focus is on calibration and uncertainty propagation, but more methods will be added with time. It was originally built at NASA/JPL to generate the high-accuracy calibrations demanded by long-range stereo, so it provides facilities to analyze calibration accuracy and to propagate and report uncertainties.
Why mrcal?
Because all other tools are terrible if you care about accuracy. They make basic questions like "how much data should I gather for a calibration?" and "how good is this calibration I just computed?" and "how different are these two models?" unanswerable.
This toolkit allows the user to confidently produce calibrations that are as good as possible. It provides lots of visualization capabilities to evaluate various properties of a model solution. And it provides powerful analysis methods, such as model differencing and quantification of projection and triangulation uncertainty. To fit the real-world lens behavior mrcal loosens many assumptions common in other toolkits: the calibration chessboard isn't assumed to be flat, lenses aren't assumed to follow simple parametric models, rectified stereo isn't done with a pinhole model, and so on.
Documentation index
Overview
Please see a tour of mrcal for a high-level overview of the capabilities of the toolkit.
Building, installing
First, the tools should be built or installed.
Background
Before using the tools, it is helpful to read about the terminology and conventions used in the sources and documentation.
At the core of a calibration routine is an optimization problem. Details about its formulation are useful to be able to interpret the results.
A lens can be represented by any of a number of lens models, and a model can be written to a file on disk.
A separate page describes various details about splined models.
Applications
A how-to-calibrate page describes this very common task.
Given a model, we can compute its projection uncertainties. Given more than one model, we can compare their intrinsics (projection behaviors).
If we have a calibrated pair of cameras, we can generate a depth map with dense stereo processing. Or we can compute sparse ranges by calling a triangulation routine.
Interfaces
Many common tasks can be accomplished by invoking one of a number of command-line tools.
If we need to do something more than what the pre-made tools can do, there're two sets programmatic interfaces available:
Releases
A version history, release notes, and the planned roadmap are available on the releases page.
Citing
To cite this work in a publication, use this bibtex stanza:
@misc{mrcal,
author = "Dima Kogan",
title = "mrcal",
howpublished = "\url{http://mrcal.secretsauce.net}",
}
Dev communication
For now let's use the github issue tracker for bug reporting and for communication in general. At some point I will probably set up a mailing list as well.
Author
Dima Kogan dima@secretsauce.net
License and copyright
These are listed on their own page.