For more info:

Links to the research papers:

Andrew J. Barry and Russ Tedrake.
Pushbroom Stereo for High-Speed Navigation in Cluttered Environments.
In Proceedings of the 2015 IEEE International Conference on Robotics and Automation (ICRA), May 2015.

Andrew J. Barry, Helen Oleynikova, Dominik Honegger, Marc Pollefeys, and Russ Tedrake.
FPGA vs. Pushbroom Stereo Vision for MAVs. Vision-based Control and Navigation of Small Lightweight UAVs,
IROS Workshop, 2015.


  1. terrible . i agree. filming is not evestable. a drone has way better options. what if this tech was appkied to a drone instead?

  2. The interesting point is not the AI like ability to avoid obstacles but the detection ability that can interpret an object correctly however it's done.

  3. So how long until the hobbyists will get their hands on this tech? Now that we're leaving the 8-bit Ardupilots and COMPILING KERNELS while we fly the things from around the world….(a demo that I saw) we should have collision avoidance and terrain mapping just any time now…

  4. Класно буржуям. Мне на одни комплектующие для одного такого аппарата нужно пол года работать.

  5. @Slides2006 Funny this guys video on Youtube looks like was edited by a ten year old. Yet he critiques this video. FailFish

  6. On Google search, Type in " Yun Parn Thomas Lee – Multlayer Neural Vision Model for Pattern Recognition and Stereo Vison", Chapter 3

Leave a Reply