MASAT intial guess for SLAM

This repository contains the code for the MASAT algorithm: a robust and efficent approach for creating inital estimates prior to pose-graph optimization in off-line SLAM.

The paper describing the approach:

Károly Harsányi, Attila Kiss, Tamás Szirányi, András Majdik: MASAT: A fast and robust algorithm for pose-graph initialization

Requirements

• Eigen (http://eigen.tuxfamily.org)

Guide

Compile:

./make

Run:

./MASAT <PATH_TO_INPUT_FILE> <PATH_TO_OUTPUT_FILE>

The input files must be in .g2o format. There are some examples in the ./input_data folder.

Once the inital guess is finished, we suggest using the g2o framework for optimization and visualization.