DPVO (Deep Patch Visual Odometry)

Software
Navigation and mapping
Library

Deep Patch Visual Odometry and SLAM estimate camera trajectories and point clouds from video or image directories using learned patch-based tracking, with an optional loop-closing backend.

Made by Princeton Vision & Learning Lab, United States (checked on github.com/princeton-vl (opens in a new tab)).

Source on GitHub: github.com/princeton-vl/DPVO (opens in a new tab). Facts read at commit 0ac95b6 (opens in a new tab).

Licenses

Code: MIT (opens in a new tab)
Read by Robotheca from LICENSE (opens in a new tab) (MIT) at commit 0ac95b6 (opens in a new tab), on 16 Oct 2026.
Commercial use: Yes

What it does

Runs visual odometry on any video or image directory with a single demo command.
Can save trajectories in TUM format, point clouds as PLY, and reconstructions in COLMAP text format.
Includes a real-time viewer (DPViewer) for visualizing reconstructions while they run.
Adds a SLAM backend (DPV-SLAM) for loop closure, and an optional classical backend using DBoW2 for very large loops.
Ships evaluation scripts for TartanAir, EuRoC, TUM-RGBD, ICL-NUIM and KITTI, plus training code on TartanAir.
Provides example calibration files for iPhone, TartanAir and EuRoC data.

Written by Robotheca from the project's README (opens in a new tab) at commit 0ac95b6.

Getting started

Difficulty: Advanced
Hardware it needs, as the project states it: Tested on Ubuntu 20/22 with CUDA 11/12; the Pangolin viewer needs CUDA 11 and cuDNN installed.
How to start: The project's README (opens in a new tab)

Maintenance

Status: Active: a commit in the last 90 days
Newest commit: 20 Sept 2026 on main
Latest release: None published on GitHub

Checked on GitHub on 16 Oct 2026.

Robots that can run it

No robot on Robotheca is linked to it yet. Links come from a robot's robot.yaml, from its dependency files read as text, or from Robotheca's editors.

Robotheca lists software as a reference: nothing is hosted or run here, and every fact links to where it was read.