Optical Motion Capture
Final Project can be found here
Project Goals
This project will focus on the computer vision aspects of motion capture systems. In particular, I plan to implement a very basic motion capture system that can be run from my home computer.
Background
I have worked with computer vision applications previously. Last quarter, I was involved with a project that estimated scene depth using video from a cell phone camera. This type of knowledge will be required to calculate 3D position of makers. I have also worked a fair amount with OpenCV and Matlab, two tools that are likely to come in handy for this project. If you are interested in working on this project please email me at: rroesler@stanford.edu
Topics that will be covered
Why use optical motion capture
- Compare with true animation and live action
- optical mocap vs other types of mocap
Required Hardware
cameras (Visible & Infrared, calibration issues)
Lights (Visible & Infrared)
markers (Passive & Active)
- studio setup (number of cameras, etc.)
Identifying Markers
- Finding markers in a video frame
- Correlating markers between frames
Tracking Markers
- probabilistically finding a marker's location (Bayesian Filtering)
- 3D-Triangulation
- Smoothing Issues
Turning Marker Locations into Usable Data
- True location vs. bone rotation
- Animation Using Mocap
Mocap in Industry & Academia
- Landmark movies in the history of mocap
- current mocap research