Generation of scaled 3D model using photogrammetry and LEGO bricks

With the modern photogrammetric tools it is pretty easy to generate 3D models from almost any object. No matter if one is interested in small objects like sculptures or bones or larger objects like buildings or infrastructural objects. The workflow is more or less always the same. You need to take images from different positions making sure the overlap is big enough and everything is visible in at least 2 images (better in 3 or more). Then you process the images in a photogrammetric software solution like AgiSoft PhotoScan or MicMac. As a result you will get a nice 3D model.

Unscaled 3D model


3D model of a church near Munich

But going this way there is one important thing missing: the scale (and the spacial reference). That means your model is a nice thing to view, turning it around and looking at it from any side. But as mentioned before, the scale and spacial reference is missing, and so the model is useless if you like to use it for planning or analytical purposes. You cannot measure distances or volumes and it is hard to locate it on the right position on earth. To overcome this problem and giving the model a scale you need to do more than only take the images, you also need to provide some kind of reference data. In the case of larger outdoor objects you can use control points or, easier, the 3D Image Vector that automatically stores the exterior orientation of the camera together with every images. In the case of smaller objects and if you take the images indoor, the 3D image vector cannot be used and you need to build yourself some kind of coordinate frame.

Lego brick setup for the generation of a 3D model

Generation of 3D model using Agisoft PhotoScan

Now, you can use a professional 3D scale bar for this, what will be quite expensive. Or you are lucky and have little children playing with Lego. With Lego bricks you can build a 3D reference field in some minutes and use it as a reference for your project. The good thing with Lego is, that it helps you to define the three needed coordinate axis in an easy way. You just take the ground plate and add some towers on it, and then you take a ruler and measure some well visible points (e.g. corners). Now you have defined your reference frame and you can place your object inside the Lego build. With this setup you take images from your object from different positions and during the analysis you just have to measure your reference points (at least 3). As a result you will get a 3D model with a defined scale, from which you can easily measure distances or volumes. Sure the accuracy with the Lego bricks is not as good as with professional scale bars, but you can easily reach sub-millimeter accuracy.

The final 3D model of the dog cranial bone can be found on sketchfab:



Schreibe einen Kommentar

Deine E-Mail-Adresse wird nicht veröffentlicht. Erforderliche Felder sind mit * markiert.

Diese Website verwendet Akismet, um Spam zu reduzieren. Erfahre mehr darüber, wie deine Kommentardaten verarbeitet werden.