SVO: Fast Semi-Direct Monocular Visual Odometry (TRO'17, ICRA'14)

We propose a semi-direct monocular visual odometry algorithm that is precise, robust, and faster than current state-of-the-art methods. The semi-direct approach eliminates the need of costly feature extraction and robust matching techniques for motion estimation. Our algorithm operates directly on pixel intensities, which results in subpixel precision at high frame-rates. A probabilistic mapping method that explicitly models outlier measurements is used to estimate 3D points, which results in fewer outliers and more reliable points. Precise and high frame-rate motion estimation brings increased robustness in scenes of little, repetitive, and high-frequency texture. The algorithm is applied to micro-aerial-vehicle state-estimation in GPS-denied environments and runs at 55 frames per second on the onboard embedded computer and at more than 300 frames per second on a consumer laptop. We call our approach SVO (Semi-direct Visual Odometry) and release our implementation as open-source software.

C. Forster, M. Pizzoli, D. Scaramuzza, SVO: Fast Semi-Direct Monocular Visual Odometry, IEEE International Conference on Robotics and Automation (ICRA), Hong Kong, 2014.
Paper: http://rpg.ifi.uzh.ch/docs/ICRA14_For...
Code: https://rpg.ifi.uzh.ch/svo_pro.html

Check out SVO 2:
Christian Forster, Zichao Zhang, Michael Gassner, Manuel Werlberger, Davide Scaramuzza
SVO: Semi-Direct Visual Odometry for Monocular and Multi-Camera Systems
IEEE Transactions on Robotics and Automation, to appear, 2016.
This paper includes comparison against ORB-SLAM, LSD-SLAM, and DSO and comparison among Dense, Semi-dense, and Sparse Direct Image Alignment.
PDF: http://rpg.ifi.uzh.ch/docs/TRO17_Fors...
Code: https://rpg.ifi.uzh.ch/svo_pro.html
Video:    • SVO 2.0: Semi-Direct Visual Odometry ...  


SVO has been used in a large number of projects since 2013. Here a non-comprehensive list organized by topic:
- Autonomous drone navigation:
1.    • Fully Loaded Quadcopter Achieves 20 m...  
2.    • Fast Autonomous Flight of a Vision-ba...  
3.    • Autonomous Quadrotor Landing using Co...  
4.    • How to Launch a Quadrotor  
5.    • How to Take a 3D Selfie with an Auton...  
6.    • Autonomous, Flying 3D Scanner  
7 .   • Aerial-guided Navigation of a Ground ...  
8.    • Autonomous Vision-based Flight over a...  
9.    • "On-the-spot" Terrain Classifier Trai...  
10.    • Collaborative Monocular SLAM with Mul...  
11.    • Air-Ground Localization and Map Augme...  
- Automotive
1.    • SVO with 4 Cameras and REMODE-CPU  
- 3D Scanning
1.    • REMODE: Probabilistic, Monocular Dens...  
2.    • SVO with 4 Cameras and REMODE-CPU  
3.    • Autonomous, Flying 3D Scanner  
- Commercial applications
1. Parrot-SenseFly Albris drone:    • albris Drone – Put the Future to Work  
2. Virtual reality:    • Видео  
3. iPhone app 3D around: https://3daround.dacuda.com/
4. iPhone app Staramba3D: https://itunes.apple.com/us/app/stara...
5. Zurich-Eye (now Facebook-Oculus VR): http://www.zurich-eye.com/

More on our research on visual odometry: http://rpg.ifi.uzh.ch/research_vo.html

Tutorial on visual odometry:
http://rpg.ifi.uzh.ch/docs/Visual_Odo...

Robotics and Perception Group, University of Zurich
http://rpg.ifi.uzh.ch/