On Sunday, April 14, 2013 9:41:02 PM UTC-4, RocketMan wrote: > > Now, I guess I HAVE to see if it actually works… > > "I’m not even sure they’d need to pivot. It would be great not to use > moving parts for this if possible. Cheap webcams now have a 1280 pixel > horizontal resolution. If you separate two of those, facing straight ahead, > by a known distance then you could calculate the physical distance based on > the pixel separation of an image feature. > Doug, Do not bother doing the trigonometry. You will ultimately need to measure camera separation distance, focal length of each camera and lens distortion (barreling, pin cushion) which can be difficult to do accurately. I will dig out an algortihm that we used in the bad old days of photogrammetry called the Direct Linear Transform (DLT). See https://me363.byu.edu/sites/me363.byu.edu/files/userfiles/5/DLTNotes.pdf It inherently includes intrinsic camera paramters (focal length and lens distortion) and extrinsic camera parameters (focal point location and prinicipal axis orientation). To calibrate, you place a grid of points printed on a flat plate at several known ranges and measure pixel locations of the points. The algoithm then calculates binomial coefficients that provide simple computation of global locations for a point that is visible to both cameras. We used it for photogrammetry in biomechanics of human movement for scientific analyses similalry to multicamera motion capture for computer generated images in movies (e.g. Golem in LOTR). Joe
-- -- You are currently subscribed to the "R/C Tank Combat" group. To post a message, send email to [email protected] To unsubscribe, send email to [email protected] Visit the group at http://groups.google.com/group/rctankcombat --- You received this message because you are subscribed to the Google Groups "R/C Tank Combat" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
