Forum Replies Created

Viewing 4 posts - 1 through 4 (of 4 total)
  • Author
    Posts
  • in reply to: Physical Sandbox Design Question #111084

    Aussie3d
    Participant

    Thanks MaxD for the reply,
    I’ve actually made a lot of progress from this post, and a lot of lessons learned along the way.

    I’ll post a few here for anyone whom follows:

    How should the physical alignment of the projector area be with respect to the sandbox? Should it be aligned with the base area (my first assumption) or the top edge of the box?
    –If it aligns with the base level then their will be noticeable “blind spots” (in the upper perimeter) because of the short through of the projector refer to SKETCH
    –Where should the image be focused (top edge / bottom edge or average height)?

    As far as I can tell the excepted wisdom is that projector alignment/focus is with the average sand level seems to be the accepted convention people are using, and the “blind spots” will occur if kids build up near the outer perimeter.
    The shorter the throw on the projector (or the lower the sand surface) the worse the blind spots seem to be.

    Must the projected image fit the box precisely & align with the camera precisely or is this something that is calibrated later?
    –If calibrated can the image overflow the box a small amount? (eg just making things easier on the physical setup side)

    I’m still not 100% on this one but looks like there is straight forward calibration between the Kinect and the projected image (as nicely outline in the setup instructions), this is then calibrated to the inner corners of the box, so whilst a little overflow is plausible but not ideal and most people are setting projectors to be aligned to the inside edges of the box at the average sand level.

    Because…the chosen bottom edge projection line is on a rising gradient to the screen. This is causing a blank area where the projected image cannot hit the bottom of the sandbox (eg shadowing from the rear sanbox wall). Refer to photos and SKETCH.
    –Has anyone come across this before?
    –Does anyone have any ideas on how best to over come this? (yes a centreline projector would be lovely but on a budget please)

    I think the answer seems to be this is not an ideal choice of projector.
    -If a projector has a raising baseline then it needs a large keystone effect to allow the bottom cord to be vertical above the back edge of the box eg tilt the projector under and square the image with the keystone.
    If selecting a project best go for one with a horizontal base line or better yet a centreline projector, in my case suck it up and live with a blind spot or perhaps add a wedge (not yet tested)

    Extra notes for the unwise.
    Looks like the Kinect V2 (is not yet supported) is very different from the Kinect V1 / Kinect 360 but fortunately these are really cheap ($5 second hand)

    Good luck and thanks for the response

    in reply to: Calibrating with a mirror??? #111057

    Aussie3d
    Participant

    I’ve come across a similar problem with the Kinect raw image being mirror (left to right) with respect to the projected image, presumable due to the relative orientation of the Kinect camera vs the projector.

    Was there a final answer as to the correct calibration order to fix the mirroring?

    The recommended corner order is
    1. Bottom left
    2. Bottom right
    3. Top left
    4. Top Right

    But is this with respect to the camera view or the projected view (in my case these are mirrored left to right)

    I’m guessing my problem will go with changing the order of the corner inputs? Am I correct in this assumption

    Thanks

    in reply to: Vrui4.3 & Kinect 3.3 & SARndbox2.3 install error #111047

    Aussie3d
    Participant

    Thanks! Worked like a charm.

    in reply to: Feature wish list #111025

    Aussie3d
    Participant

    Hi,

    1) Not sure if anyone has suggested snow say above the top 10 % of the box eg above a specific altitude (depending on scale I guess)
    Maybe melt to water over time? (maybe whilst you are not playing with it)

    2) I always liked the idea of being able to build known land forms say the Everest region.
    –> go to google earth in contour mode and run a difference between the measured sand box and desired contour shape then project a gradient to either add or subtract sand to different parts of the box. (Perhaps a little complex for a museum environment but ok for an older class room.)

    I look forward to seeing any of the many great new ideas suggested come to life.

    Cheers
    B

Viewing 4 posts - 1 through 4 (of 4 total)