CA Lecture Visual Odometry and SLAM // Studio

Today

• CA Lecture: Visual Odometry and SLAM
• Studio Time

For Next Time

• Work on your Machine Vision Project.
  • In-class demos will be on Monday November 11th, and code/write-ups are due on Tuesday November 12th at 7PM.
  • Note that prospective students will be joining us in class on the 11th!
• Consider whether there is feedback you’d like to share about the class

Visual Odometry and SLAM

At the top of the class, Ayush will be presenting on Visual Odometry and SLAM – a machine vision extension of our discussions during the particle filter and localization module. You can see a copy of his slides here.

Modern mobile robotics today is significantly underpinned by visual sensing and downstream tasks like odometry/localization, mapping, SLAM, image segmentation, and more. If you’d like to learn more, some resources to get started:

• Read up on papers by academics like:
  • Nicholas Roy and the Robust Robotics Group (e.g., Visual Odometry and Mapping for Autonomous Flight Using an RGB-D Camera)
  • Davide Scaramuzza (e.g., SVO: Fast Semi-Direct Monocular Visual Odometry)
• Check out some of these explainers:
  • Visual SLAM tutorial by Scaramuzza and Fraundorfer
  • A From Scratch blog post
  • Papers and tutorials from ETH Zurich
  • MatLab’s Explainer + Code
  • An Explainer with Python/OpenCV Code
• Try out some code
  • RTAB-Map’s Project Page
  • Open-source Github projects, like this
  • These resources for a ton more options and ideas!