Recreating 3D buildings from "Axonometric" photos

Hi all,

I was wondering if anyone knows a clever way of recreating 3D models from axonometric(ish) photos.
I need to recreate a rough, but more or less proportionally accurate site model of a newly built housing development. I have airplane/drone photos which have been taken from very far away and could be considered axonometric.

My initial idea was to redraw/trace the building edges and their features (windows, balconies…) in 2D and then use FlowAlongSurface to match those features to a 3D shape.

Is there a more clever way of doing this? I have to do this for multiple buildings, which can become quite time consuming.

that seems like a pretty decent way of handling this…

you can use remap cplane as well if you don’t want any distortion and just want to flop flat objects onto another flat object.

draw your details in front or top view (or whatever view you prefer) and add a cplane to object on the face of the building, then remapcplane from the detail to the building…

very similar workflow

Hi,

ISO flowing you could set the image as picture and edit the controlpoints to adjust for the perspective.
Eliminating the need for the flow step.

-Willem

1 Like

Autodesk Recap Pro - from their website:
“ReCap Photo, a service included with ReCap Pro, processes drone photography to create 3D representations of current site conditions, objects, and more. It also supports the creation of point clouds, meshes, and ortho photos.”

1 Like

If you’re willing to indulge in a bit of Grasshopper, you can build up a tool that will help considerably with this process:

You can use it simply to get nice ortho-rectified images to trace off of your axon, or even use it to extract small pieces and build up a 3d model that way, depending on your needs. (never mind that in the video I got a facade upside down :upside_down_face:)

The script requires the Human and Wombat plugins for Grasshopper.
sample orthorectify setup.3dm (7.4 MB) Orthorectify from photo.gh (19.6 KB)

3 Likes

Wow, thanks to everyone for your replies and video tutorials :smiley:

This was exactly what I was hoping to learn.