GRAPH-SLAM(1)

💖 GRAPH-SLAM(1) 💖

okay bestie, buckle up, because we're about to talk about graph-slam — a command-line Graph-SLAM hub application, straight outta the Mobile Robot Programming Toolkit (MRPT). it's giving robotics. it's giving pose graphs. it's giving main character energy 🦋

---

✨ SYNOPSIS ✨

graph-slam  [--info] [--dijkstra] [--levmarq] [--no-span]
            [--initial-lambda <val>] [--max-iters <N>] [-q] [--view]
            [--3d] [--2d] [-o <result.graph>] -i <test.graph> [--]
            [--version] [-h]

---

🌸 USAGE EXAMPLES 🌸

like, here's the tea. three iconic ways to run her:

View Dijkstra-based spanning tree coordinates estimation of a 2D graph file

graph-slam --2d --dijkstra --view -i in.graph

Visualization of a 2D (or 3D) graph file

graph-slam --2d [or --3d] --view -i in.graph

Levenberg-Marquartd optimization of a 3D graph and visualize result

graph-slam --3d --levmarq --view -i in.graph

---

💅 DESCRIPTION 💅

so graph-slam is a command-line application that visualizes pose constraint graphs and then runs Graph-SLAM methods on them. she does both. she contains multitudes. ✨

and these are literally all the supported arguments and operations, no gatekeeping:

• --info — Op: loads the graph and shows you the stats and info about it. like a vibe check for your graph 💖
• --dijkstra — Op: runs CNetworkOfPoses::dijkstra_nodes_estimate() to estimate the global pose of nodes from a Dijkstra tree plus the edge relative poses. can be used together with: --view, --output
• --levmarq — Op: optimizes the graph with sparse Levenberg-Marquartd using global coordinates (via mrpt::graphslam::optimize_graph_spa_levmarq). can be used together with: --view, --output, --max-iters, --no-span, --initial-lambda 🔥
• --no-span — don't use the dijkstra initial spanning tree guess (optional)
• --initial-lambda <val> — initial lambda parameter (optional, lev-marq)
• --max-iters <N> — maximum number of iterations (optional)
• -q, --quiet — terse output. she's not gonna yap, periodt 🤫
• --view — Op: visualize the graph in a 3D view. if you use it alone, it represents VERTEX2 or VERTEX3 poses directly as stored in the input file. if you use it together with another operation, the final obtained graph after the operation will be shown — not the input original one. important!! 👑
• --3d — use 3D poses (must use exactly one of --2d and --3d)
• --2d — use 2D poses (must use exactly one of --2d and --3d)
• -o <result.graph>, --output <result.graph> — output file (optional) (*.graph, *.graphbin)
• -i <test.graph>, --input <test.graph> — (required) input file (required) (*.graph, *.graphbin). yeah it's required, she said what she said 💀
• --, --ignore_rest — ignores the rest of the labeled arguments following this flag
• --version — displays version information and exits
• -h, --help — displays usage information and exits

🫶 quick reminder bestie: you must use exactly one of --2d and --3d. not both. not neither. exactly one. the way that this is non-negotiable…

---

💀 BUGS 💀

please report bugs at https://github.com/MRPT/mrpt/issues — do NOT suffer in silence, that's not the vibe 🙈

---

👑 AUTHORS 👑

graph-slam is part of the Mobile Robot Programming Toolkit (MRPT), and was originally written by the MAPIR laboratory (University of Malaga). iconic institution behavior fr fr 🎓

this manual page was written by Jose Luis Blanco <joseluisblancoc@gmail.com>. slay, José 💜

---

🌈 COPYRIGHT 🌈

this program is free software; you can redistribute it and/or modify it under the terms of the BSD License.

on Debian GNU/Linux systems, the complete text of the BSD License can be found in /usr/share/common-licenses/BSD.

---

and that's the whole page, bestie! this was GRAPH-SLAM(1) from the Mobile Robot Programming Toolkit - MRPT, generated by perl v5.40.1 on 2025-03-21. go optimize some graphs, you absolute legend 🦋✨💖