Assignment 6


Part A due 30 March
Part B due 16 April
Approval for Part B due by 23 March

Part A

For this assignment you are to collect a video (or several still images) moving a camera down a hallway or something else with significant 3-D structure (i.e. not UAV video where everything is flat...)  You are then to reconstruct from the movement of the camera and the locations of several points that are observed in the video.  Your video or set of still images should consist of at least 30 images.  You can do this in one of two ways:
  1. You can determine the calibration of the camera before taking the video.  Then, you can attempt to estimate the essential matrix between each set of cameras and compute the 3-D location of points from there.
  2. You can simply take the camera down the hallway and use the fundamental matrix and come up with a projectively-warped distortion of the space you have traversed with the camera.
Beyond the basic algorithms, you will have to determine the relative scale in movement between different sets of video frames so that you have the global movement from the first to the last frame of the camera.

There will be a written portion for Part A.

Part B

For this portion of the project, you are to propose and implement an improvement to your Part A.  Possible ideas include:
I am proposing that for Part B, there not be a written portion.  However, let's all meet during the final exam time, have some refreshments, and demonstrate to one another the cool projects you did.  What do you think?  (Final allocated time is April 24, 7-10pm.   Or, we could meet some other time that everyone is available.)