Hi,
I have just started "playing around" with reconstructing motion of fast objects using an a small array of cameras. I am imaging with light-sensitive monochrome cameras, which are more sensitive than color cameras in order to work at high-shutter speeds. I know that there is a wealth of knowledge here and there were two particular details about my set-up that I feel nervous about.
1) Between cameras, the grayscale levels are slightly different (due to sensor, but also a stepless aperture ring, which makes it hard to exactly match lens aperture). Are there clever ways to calibrate this easily? Am I worrying about nothing? The reconstructions from my first attempt look pretty solid already. I have just been eyeballing it, so far...
2) Should I be doing anything different with monochrome cameras to improve the reconstructions? I know color would have more detail, but it's not an option because I don't own those cameras. I am mostly interested in reconstructing the surface and less so the texture, so I have been texturing the surface of the moving objects with white and black color, to enhance the number of detectable features.
If there are tutorials, or great examples, I would love to see them. I have found the following:
AgiSoft 4D Timeline tools. Part 1 -
https://www.youtube.com/watch?v=N8OaQv-187oAgiSoft 4D Timeline tools. Part 2 -
https://www.youtube.com/watch?v=JoRhXFoGsXsMoving face - 3 perspectives
http://vimeo.com/43719708Hopefully, I'll eventually write my own tutorial, but I suspect that it will take time for me to feel sufficiently comfortable with the software and optimizations.