« CameraExport: PhotoSynth to Camera Projection in 3DS Max
Analogue Augmented Reality: HereIn Berlin »

Camperdown Graffiti Wall deep zoom

1 year, 5 months ago

I’ve made a “deep zoom collection” of the camperdown park graffiti wall, so that you can zoom into it and check the whole west-facing wall out in detail. This was made using PhotoSynth to reconstruct the wall, then my Meshlab to reconstruct a model, then the CameraExport app (I just made) to retexture the model in high-res in 3DS Max.

I can’t embed it, so you’ll have to link to the separate page. It requires the Silverlight plugin, and the wall starts at the very of top of the box. You can go fullscreen.

http://files.neonascent.net/camperdown/Test.html

If your browser setup is not compatible with the Silverlight plugin, you can alternative view a reduced image of the wall reconstruction on Flickr.

5 Responses to “Camperdown Graffiti Wall deep zoom”

  1. Nathan Craig says:
    February 6, 2011 at 5:09 am

    Very nice; this strikes me as a really intriguing solution to creating large planar panoramas. Thank you for sharing your work.

  2. Josh says:
    February 6, 2011 at 5:55 am

    @Nathan; Thanks! :)

    I’d probably suggest just using Microsoft Research Image Composite Editor (or similar) for panoramas, but if you want to document while also recording the geometry, this is probably the way to go.

  3. Jørn Anke says:
    February 9, 2011 at 11:42 pm

    Impressive work!
    I would be interested in doing something similar, but with related to other software then 3DS Max. For this I am primarily intested in the coordinates of the individual images as they are merged into the stitched image in photosynth. Any idea if this would be possible?

  4. Josh says:
    February 10, 2011 at 12:42 am

    @Jørn,

    So you are interested in the images as they are located/related in PhotoSynth’s 3D view, rather than the positions of the cameras? Once the camera positions are found, they can be set up in whatever software you’d like, but I don’t know how to get that particular info about the photosynth images. What is your application?

  5. Jørn Anke says:
    February 10, 2011 at 1:01 am

    @Josh,
    I think you understood my interest correct. What I am ultimatly after is to import the images into some GIS software and project the images onto a terrain surface, in order to make orthophotos. For this I would need either the image coordinates as projected onto the ground (e.g corners) -or camera; location, orientation, lens parameters, etc.

    What I have been doing so far is to write a tool for transforming the photosynth pointcloud coordinates into real word coordinats, (e.g. UTM). This is done by an affine-transformation. See: http://www.uavmapping.com

Leave a Reply