Hi Joe - 



>>  measure camera separation distance, 

Sure. My back of the envelope run through assumed a 6" separation. 



It seems like most of the other camera parameters you mention go into 
calculating the camera field of view. That value is published for many webcams. 
For example, the one I use has a 53 degree horizontal FOV. If you know those 
values couldn't you calculate distances ? It does ignore lens distoration, and 
the fish-eye lenses on some webcams would certainly screw things up. 



I guess if the calculations turned out to be wrong, a system liek this could 
always be calibrated as you suggest. 



    - Doug 





----- Original Message -----


From: "Joe Sommer" <[email protected]> 
To: [email protected] 
Sent: Monday, April 15, 2013 8:13:09 AM 
Subject: Re: [TANKS] Re: cheap laser distance range finder with USB output 


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 . 
  
  

-- 
-- 
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.


Reply via email to