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

Re: Technical question on PT coordinate system

Expand Messages
  • Erik Krause
    ... As for Yaw, Roll and Pitch: Yes, as far as I know. The TrX, TrY, TrZ mosaic coordinate model is unique to hugin/nona. PTGui uses a different model with 5
    Message 1 of 5 , Feb 17, 2013
    • 0 Attachment
      Am 16.02.2013 15:54, schrieb Thomas:
      > 1] do all PT family stitchers use the same coordinate system as original
      > PanoTools?

      As for Yaw, Roll and Pitch: Yes, as far as I know. The TrX, TrY, TrZ
      mosaic coordinate model is unique to hugin/nona. PTGui uses a different
      model with 5 coordinates: VPx, VPy, VPz, VPpan and VPtilt. They are not
      documented (AFAIK) but Helmut Dersch found them out for himself and
      implemented them in PTStitcherNG.

      Furthermore PTGui uses PTStitcher o-line C-coordinates while hugin/nona
      use S-coordinates. That means hugin coordinates are relative to the
      complete source image while PTGui coordinates are relative to the
      cropped image specified by the C-coodinates (the crop rectangle/circle).
      See
      http://wiki.panotools.org/PTStitcher#o-line_options for details

      --
      Erik Krause
      http://www.erik-krause.de
    • Jim Watters
      ... From Helmut: PTGui uses a 2-axis pan/tilt + 3-d translation, in that order http://tech.groups.yahoo.com/group/PanoToolsNG/message/39331 Dev Ghosh who
      Message 2 of 5 , Feb 17, 2013
      • 0 Attachment
        On 2013-02-17 6:47 AM, Erik Krause wrote:
        > Am 16.02.2013 15:54, schrieb Thomas:
        >> 1] do all PT family stitchers use the same coordinate system as original
        >> PanoTools?
        > As for Yaw, Roll and Pitch: Yes, as far as I know. The TrX, TrY, TrZ
        > mosaic coordinate model is unique to hugin/nona. PTGui uses a different
        > model with 5 coordinates: VPx, VPy, VPz, VPpan and VPtilt. They are not
        > documented (AFAIK) but Helmut Dersch found them out for himself and
        > implemented them in PTStitcherNG.
        From Helmut: "PTGui uses a 2-axis pan/tilt + 3-d translation, in that order"
        http://tech.groups.yahoo.com/group/PanoToolsNG/message/39331

        Dev Ghosh who originally worked on the mosic mode for Hugin during a GSoC, but
        never got finished, has just posted on the hugin-ptx list, "mosaic mode with
        intuitive warp parameters", that he is implementing his original method to
        create orthogonal views.



        >
        > Furthermore PTGui uses PTStitcher o-line C-coordinates while hugin/nona
        > use S-coordinates. That means hugin coordinates are relative to the
        > complete source image while PTGui coordinates are relative to the
        > cropped image specified by the C-coodinates (the crop rectangle/circle).
        > See
        > http://wiki.panotools.org/PTStitcher#o-line_options for details
        Yes the GUIs of PTGui and Hugin use different cropping models.
        http://tech.groups.yahoo.com/group/PanoToolsNG/message/6212
        PTStitcher definitely understands both.

        --
        Jim Watters
        http://photocreations.ca
      • Thomas
        Thank you Jim and Erik. The references on PTStitcher are helpful. I am having a devil of a time figuring out how to apply PTGui y,p,r to a panosphere having a
        Message 3 of 5 , Feb 17, 2013
        • 0 Attachment
          Thank you Jim and Erik.

          The references on PTStitcher are helpful.

          I am having a devil of a time figuring out how to apply PTGui y,p,r to a panosphere having a right-handed coordinate system (meaning my X axis points left) Like PT, view is centered on (0,0,1) at y=p=r=0. Inverting the X axis in the rotation matrix seems to give correct alignments, however my projection mapping fns then fail.

          I need to invert the hand again without losing the alignment, this is surely possible but I can't find the right way to do it.

          Alternatively I could revert to left-handed coords, which Panini used for a long time, but I feel a bit stubborn since it took some effort to convert everything to right-handed.

          -- Tom

          --- In PanoToolsNG@yahoogroups.com, Jim Watters <jwatters@...> wrote:
          >
          > On 2013-02-17 6:47 AM, Erik Krause wrote:
          > > Am 16.02.2013 15:54, schrieb Thomas:
          > >> 1] do all PT family stitchers use the same coordinate system as original
          > >> PanoTools?
          > > As for Yaw, Roll and Pitch: Yes, as far as I know. The TrX, TrY, TrZ
          > > mosaic coordinate model is unique to hugin/nona. PTGui uses a different
          > > model with 5 coordinates: VPx, VPy, VPz, VPpan and VPtilt. They are not
          > > documented (AFAIK) but Helmut Dersch found them out for himself and
          > > implemented them in PTStitcherNG.
          > From Helmut: "PTGui uses a 2-axis pan/tilt + 3-d translation, in that order"
          > http://tech.groups.yahoo.com/group/PanoToolsNG/message/39331
          >
          > Dev Ghosh who originally worked on the mosic mode for Hugin during a GSoC, but
          > never got finished, has just posted on the hugin-ptx list, "mosaic mode with
          > intuitive warp parameters", that he is implementing his original method to
          > create orthogonal views.
          >
          >
          >
          > >
          > > Furthermore PTGui uses PTStitcher o-line C-coordinates while hugin/nona
          > > use S-coordinates. That means hugin coordinates are relative to the
          > > complete source image while PTGui coordinates are relative to the
          > > cropped image specified by the C-coodinates (the crop rectangle/circle).
          > > See
          > > http://wiki.panotools.org/PTStitcher#o-line_options for details
          > Yes the GUIs of PTGui and Hugin use different cropping models.
          > http://tech.groups.yahoo.com/group/PanoToolsNG/message/6212
          > PTStitcher definitely understands both.
          >
          > --
          > Jim Watters
          > http://photocreations.ca
          >
        • Erik Krause
          ... Read on Euler angles: http://en.wikipedia.org/wiki/Euler_angles (This is the exact point where my knowledge ends ;-) -- Erik Krause
          Message 4 of 5 , Feb 17, 2013
          • 0 Attachment
            Am 17.02.2013 21:07, schrieb Thomas:
            > I need to invert the hand again without losing the alignment, this is
            > surely possible but I can't find the right way to do it.

            Read on Euler angles: http://en.wikipedia.org/wiki/Euler_angles

            (This is the exact point where my knowledge ends ;-)

            --
            Erik Krause
            http://www.erik-krause.de
          Your message has been successfully submitted and would be delivered to recipients shortly.