{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/direct-sparse-visual-inertial-odometry-using","title":"Direct Sparse Visual-Inertial Odometry using Dynamic Marginalization","arxiv_id":"1804.05625","date":"2018-04-16","proceeding":null,"authors":["Lukas von Stumberg","Vladyslav Usenko","Daniel Cremers"],"abstract":"We present VI-DSO, a novel approach for visual-inertial odometry, which\njointly estimates camera poses and sparse scene geometry by minimizing\nphotometric and IMU measurement errors in a combined energy functional. The\nvisual part of the system performs a bundle-adjustment like optimization on a\nsparse set of points, but unlike key-point based systems it directly minimizes\na photometric error. This makes it possible for the system to track not only\ncorners, but any pixels with large enough intensity gradients. IMU information\nis accumulated between several frames using measurement preintegration, and is\ninserted into the optimization as an additional constraint between keyframes.\nWe explicitly include scale and gravity direction into our model and jointly\noptimize them together with other variables such as poses. As the scale is\noften not immediately observable using IMU data this allows us to initialize\nour visual-inertial system with an arbitrary scale instead of having to delay\nthe initialization until everything is observable. We perform partial\nmarginalization of old variables so that updates can be computed in a\nreasonable time. In order to keep the system consistent we propose a novel\nstrategy which we call \"dynamic marginalization\". This technique allows us to\nuse partial marginalization even in cases where the initial scale estimate is\nfar from the optimum. We evaluate our method on the challenging EuRoC dataset,\nshowing that VI-DSO outperforms the state of the art.","url_abs":"http://arxiv.org/abs/1804.05625v1","url_pdf":"http://arxiv.org/pdf/1804.05625v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"direct-sparse-visual-inertial-odometry-using","repo_url":"https://github.com/JakobEngel/dso","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok","spdx":"GPL-3.0"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1804.05625","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}