Imagine we have a set of images showing an object viewed through a magnifying glass. Which ways to reconstruct the shape of the magnifying glass do exist?
Possible assumptions: Index of refraction is known, calibrated camera setup, object shape and pose is either known or unknown.
I know a Team at Fraunhofeer IOSB in Karlsruhe that deals with image based inspection of transparent materials. The leader is Dr. Thomas Längle:
http://www.iosb.fraunhofer.de/servlet/is/2401/
Thank you for the quick answer and nice contact. This is an interesting starting point.
The methods proposed by IOSB seem to aim at highly controlled setups where special illumination from behind the objects is used to find flaws in transparent objects whose shape is known (unless I missed something). This is certainly mandatory for their very robust and fast inspection methods.
The setup I am looking for poses some additional problems : uncontrolled lighting not coming from behind the refracting element, unknown shape of the refracting element. I am aware that this problem will not be solved to the robustness and precision achieved by IOSB, but I wonder how far we can get in this setup.
I reviewed different types of reconstructions and their attributes in the following publication:
Conference Paper A REVIEW OF RECENT RANGE IMAGE RECONSTRUCTION ALGORITHMS
- Badges
- Science topic