1. Proposed time line by Cathy:
Jan 2009 - test the new system at NH/Vermont
Nov 2009- Feb 2010 - place the equipment on the ship.
2. Possible thesis areas
Large scale + actual measurement
variations
- calibrated
- motion + uncalibrated / multiple view/ structure from motion using different constraints (Darthmouth paper http://www.cs.dartmouth.edu/farid/research/phototop/)
- projected grid: sparse but many point measurements
final result - measurement + error map
3. Website updates
- overview table
- sensitivity maps
----
Stability of different camera calibration methods: http://gias720.dis.ulpgc.es/Gias/Publications/waf2005-josep.pdf
The paper provides experimental results regarding the stability of the camera calibration (i.e variation in estimated parameters) for 8 calibration techniques.
Though we had noticed some of what he has reported, it is good to have a paper which does the experiments which we did not.
The calibration method we had finally started to use is not listed here, so we want to perform similar expts for our scale.
To do:
step 1: try the calibration technique we had used at 50meters range
step 2: test its stability.
step 3: record the effect on the final measurement
step 4: see the worst error in best case of capture - wrt lighting, speed of ship and texture.
but for any of the above we need step 0: get stereo cameras- so we should start with getting the lens repaired
Since literature seems to be so full of critical analysis, its best to apply it to our case, get the numbers and see if its acceptable for us.
