Loading ...
Sorry, an error occurred while loading the content.

Re: [PanoToolsNG] Re: Reverse distortion correction

Expand Messages
  • Milko Amorth
    Hi Tim, ... I see, the values are in mm relative to h from centre to top right corner of portrait orientation and expressed in percentage. ... rsrc is the
    Message 1 of 7 , Oct 19, 2007
    • 0 Attachment
      Hi Tim,
      > What I have is data sets which produce something like the graph Figure
      > 3 (A) on http://www.vanwalree.com/optics/distortion.html
      I see, the values are in mm relative to h from centre to top right corner
      of portrait orientation and expressed in percentage.

      > Looking at the Lens Correction Model details on
      > http://wiki.panotools.org/Lens_correction_model it seems I want to
      > convert my data to actual radial distances so rsrc and rdest are
      > specified (is src or dest the 'corrected' distance?) rather than a
      > percentage.

      rsrc is the source image and rdest is the destination, the corrected image.
      The radial distance r_src for each point in the source image is computed
      by the formula:
      r_src = (a * r_dest^3 + b * r_dest^2 + c * r_dest + d) *
      Units for r_src and r_dest is image width / 2, ie the edge of the image
      corresponds to r_src = 1.
      The internal unit used for r_src and r_dest _ is the smaller of the two
      image lengths divided by 2.

      > I presume I can then take my data and if say rsrc is 35 and rdest 38,
      > I can swap the source and destination and refit the polynomial - will
      > this give the the parameters needed to introduce distortion?
      You would need to convert your percentages of h in mm to positive or
      negativ pixel distance values of a,b and c and apply the PTcorrect tool
      with its reverted values.

      I did a short test of a lens correction and reverted a corrected duplicate
      with the PTcorrect tool with reverted values (positive to negativ and vice
      versa.) overlayed as layers in PS to show the difference. It is perfect,
      but cropped by dimensions lost in the original lens correction. PtGui will
      do the same with more canvas to play with.

      Here is the link to the paper from Prof. Dersch to explain its functions:

      Cheers, Milko

      Milko K. Amorth
      360° Immersive Imaging Productions

      fx: 604.909.5125

      Skype: VRdundee
      Adobe Photographers Directory: tinyurl.com/2e3tnq
      Flickr: flickr.com/photos/vrdundee/
      Youtube: youtube.com/user/VRDundee Member of IVRPA.org (International VR
      Photography Association): ivrpa.org/user/1133
      Member of the Panorama Tools Meeting Group Europe : panotools-meeting.com/
      Contributor to the WorldWidePanorama.com Project ( University of
      California Berkeley) : tinyurl.com/233lkg
      Member of 360cities.net network VR portal
    Your message has been successfully submitted and would be delivered to recipients shortly.