3
$\begingroup$

I recently noticed that in some locations you could switch to a 3d mapping mode in Google Street View.
Which algorithm(s) did they use to generate the 3d models of the streets? I assume it was more than just capture, given the huge amount of data that would be necessary to produce an accurate representation of the streets.

Improve this question
$\endgroup$
2
  • 1
    $\begingroup$ There are many scanning techniques. Becuase they have a camera, and a accurate position sensor, that is basically a video you could use video tracking techniques to produce point clouds and build polygonmodels out of that. (for some ideas see ssontech.com/learning.html great resource even if you dont use their apps) $\endgroup$
    – joojaa
    Commented Sep 23, 2015 at 8:43
  • $\begingroup$ I have actually designed and implemented (with help of my colleague) an algorithm to create 3D mesh of the city from the data that is collected (for a competitor company). However, what you're asking would require breaking NDA and I don't believe any sane person would be willing to do that, so your question can only be truly answered by developers who have worked on something like this in their spare time and thus are not bound by legal paperwork (regardless of whether they still work for that company or not). $\endgroup$
    – 3D Coder
    Commented Oct 7, 2015 at 18:35

1 Answer 1

Reset to default
3
$\begingroup$

Google said how in 2012, avoiding explaining too many details

The never-ending quest for the perfect map

...new imagery rendering techniques and computer vision that let us automatically create 3D cityscapes, complete with buildings, terrain and even landscaping, from 45-degree aerial imagery

and here it's a video of the feature.

So we know that "Computer Vision" is used. I guess the edges are detected, along with the shadows, to build the elevation of the object. Then the segmented image is used as a texture on the built mesh.

However, i saw this paper linked, and i think a similar technique is used in Google to extract the 3d mesh from different 45° images.

There are a lot of papers (a book too) about "building extraction/detection".

src: How does the new google maps make buildings and cityscapes 3D?

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.