rkwitt / polyran

Robust polynomial fitting using RANSAC

polyran

Overview

Robustly fit a N-th order polynomial to a 1D signal, using RANSAC.

Requirements

• Boost library (version 1.49 and higher), download from http://www.boost.org
• Eigen3, download from http://eigen.tuxfamily.org

Compilation

To compile, set the 'BOOST_ROOT' environment variable to the path where you installed Boost, e.g., /software/boost/, using

export BOOST_ROOT=/software/boost/

This should help cmake to do its job. Assuming that you checked out the code under /tmp/polyran, create a 'Build' directory, run 'cmake' and then 'make' to compile the code.

$ cd /tmp/polyran
$ mkdir Build
$ cd Build
$ cmake ..
$ make

Testing

To test the code, go to the test directory and run (assuming you built under 'Build')

$ ../Build/src/fit x.bin y.bin

which will estimate a 2-nd degree polynomial to the data in (x.bin, y.bin) and evaluate this polynomial at the positions in x.bin.