Hi Thibaud: I used Photoshop's image distortion, skew and perspective features extensively to "fix" many of the photos on the site so that it looks like the camera was centered in front of the center of the dial faces. I think most of them are pretty close to being correct.
I bet you're thinking how did I know if I stretched them correctly? Well, there are clues on many dials. For example, the original photo sent to me of Dial 40 (The Lullingstone Dial) showed the circle around the dial face as an oval with a shortened vertical axis because the photographer was looking up at the dial located high on the wall. Obviously, it is a circle and not an oval, so stretching the photo vertically fixes it. I did the same thing to distorted square sundial photos that looked rectangular instead of square (i.e.: The Buckleburry Dial). John John L. Carmichael Jr. 925 E. Foothills Dr. Tucson Arizona, USA Tel: 520-696-1709 Email: [EMAIL PROTECTED] Sundial Sculptures Website: http://www.sundialsculptures.com Stained Glass Sundials Website: http://advanceassociates.com/Sundials/Stained_Glass ----- Original Message ----- From: "Th. Taudin-Chabot" <[EMAIL PROTECTED]> To: <[email protected]> Sent: Wednesday, February 11, 2004 1:44 AM Subject: Re: Reverse Engineering > Hi fellow shadow watchers, > > There must be a possibility to cope this problem. > You might find two lines running parallel vertical. If you measure the > distance between them in the bottom and in the top you know pretty good > what de distortion is and you might be able to reconstruct the dial on a > piece of paper as if the picture is taken directly in front. > If it was taken a bit from aside, I hope there will be two parallel > horizontal lines in the picture and you can do them same for that distortion. > An other possibility is to print the picture on an overhead sheet and > project it on a surface that you can tilt in such a way that the distortion > is compensated for. > Thibaud Chabot > > At 17:15 10-02-2004, you wrote: > >One significant problem we have in reverse engineering many of these dials > >is that we cannot be sure the original photo was taken straight-on, normal > >to the dial face. Many of the photos submitted to John have been shot from > >the ground or other accessible vantage point. In orthogonlizing the > >images, errors are introduced, because the gnomon and dial face are not in > >the same plane. I don't think these can easily be corrected. > > --------------------------------------------------------------------- > Thibaud Taudin-Chabot > 52° 18' 19.85" North, 04° 51' 09.45" East, alt. -4.50 m > home email: [EMAIL PROTECTED] > > > - > -
