Image-Based 3D Reconstruction

During the first 18 months of my Ph.D., I worked on the topic of passive image-based surface 3D reconstruction (i.e., obtaining the 3D geometry and colour of an object/specimen/scene from regular pictures of it). At the time, the main focus of my research was the development of an easy-to-use solution that combined the necessary data acquisition and reconstruction steps into a single integrated framework. We called that system R2OBBIE-3D (in honour of Isaac Asimov’s short story). You can find a detailed description of our work here.

Since then, R2OBBIE-3D has been used for many applications, such as

I also worked on a strategy to extend Prof. Diego Nehab‘s awesome algorithm to more general cases (more on that soon).

Lastly, an INNOGAP ‘Proof-of-Principle’ grant (Spring 2015) allowed me to develop Nano-R2OBBIE-3D, a system similar to R2OBBIE-3D but adapted to the reconstruction of much smaller geometric details (down to ~1-2µm). You can find more details about it in my Ph.D. thesis (available soon).

Want me to make a coloured 3D reconstruction of an object/scene for you? Contact me!


fig_phelsuma.png
Left: a Madagascar giant day gecko (Phelsuma grandis). Photo by Brian Gratwicke. Right: digital 3D reconstruction of a similar specimen with (top) and without (bottom) colour texture. The scanning and reconstruction were performed using R2OBBIE-3D.