13
$\begingroup$

In terms of integrating Blender renders with existing photographs, is there a way to create a set of Blender mesh points using the parallax info between just a pair of stereo photographs?

I've gotten good at regular camera tracking with video, and also using BLAM for static images, but what I'd really like to do is take two very clear and stable stereo images and extract as many good 3D points from them as possible, without resorting to also videoing in order to reconstruct the scene.

Improve this question
$\endgroup$
5
  • 4
    $\begingroup$ Your question make me remind this article I saw wich somebody recreate a lightning in 3D from two picture: calculatedimages.blogspot.co.uk/2013/05/3d-lightning.html hope it can help $\endgroup$
    – lucblender
    Commented Apr 3, 2014 at 14:08
  • $\begingroup$ You could use external software and then import it into blender. This is one a quick google search brought up that doesn't look too bad. $\endgroup$
    – user2699
    Commented Apr 3, 2014 at 18:26
  • 1
    $\begingroup$ Why not just use the BLAM add-on both separate photos and combine the 3D data? Or am I not correctly understanding what you want to do? $\endgroup$
    – Mentalist
    Commented Sep 12, 2015 at 6:46
  • $\begingroup$ BLAM, being vanishing-point based, requires lines - specifically, edges of objects to share common X,Y, and Z axes. Parallax-based methods require only points. $\endgroup$
    – Gnaural
    Commented Dec 3, 2015 at 14:32
  • $\begingroup$ You may be interested in learning camera mapping from blender guru aka Andrew price's tutorial here on the Youtube $\endgroup$
    – hawkenfox
    Commented Dec 26, 2015 at 4:13

2 Answers 2

Reset to default
1
$\begingroup$

Or you could try point cloud reconstruction software which are free to download using Visual SFM and Meshlab. There are a few tutorials on the subject on youtube. As the question you are asking will require substantial write up, it's best to point you to some tutorials.

I learn this in 3 hours last night. However you may need more than 2 images.

https://www.youtube.com/watch?v=D6eqW6yk50k

Improve this answer
$\endgroup$
-1
$\begingroup$

Yes, there is a way, because there is an image processing package called Hugin (available for free), with tutorials from http://hugin.sourceforge.net/ which does exactly what you want.

I have downloaded it and tried to do exactly what you wish, but I've not yet mastered Hugin. I do perceive that it might be more doable with some images than with others. I have it, I'm trying to learn it, but I have to say that for me, it's been more difficult than Blender. .

Tutorial link: http://hugin.sourceforge.net/tutorials/perspective/

Improve this answer
$\endgroup$
3
  • $\begingroup$ Perhaps you could link to a tutorial on how to do this with hugin? $\endgroup$
    – user2699
    Commented May 9, 2014 at 5:05
  • 3
    $\begingroup$ This tutorial is focused on removing perspective from the image, not creating a 3d model from it. As hugin is for 2d stitching of panoramas, I would be very surprised if it can export 3d models. $\endgroup$
    – user2699
    Commented May 9, 2014 at 22:30
  • 2
    $\begingroup$ Thanks for the reply - problem is, in my experience with Hugin (which i think is an awesome program), it is sort of the opposite of the principle i'm looking for. Put simply, Hugin really wants zero parallax info; even as little as the offset of the camera-body on a tripod gives it a hard time for any sort of foreground-background setting. $\endgroup$
    – Gnaural
    Commented May 15, 2014 at 21:16

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .