Making photogrammetry better with spectral imaging

Three-dimensional scanning of teeth and bones has become a bigger and bigger part of morphological analysis. When it comes to teeth, microfocus CT is the most prominent approach. This is in part because the CT can reveal internal structures such as the enamel-dentin junction, and in part because the translucency of enamel makes it very difficult to get accurate scans from other techniques, including laser surface scanning, photogrammetry, and structured light scanning.

Researchers have long dealt with the translucency of enamel in photography by spraying the teeth with substances that temporarily render them opaque, like chalk dust. Nowadays we realize that many kinds of trace evidence can persist on fossill teeth, and subjecting them to repeated surface treatments for the purposes of scanning and photography is a bad idea.

A new paper by Aurore Mathys and coworkers tries something new: dividing the spectrum for scanning the surfaces of teeth: “Improving 3D photogrammetry models through spectral imaging: Tooth enamel as a case study”.

Spy 2B taken at many wavelengths, from Mathys et al. 2019

Enamel is relatively less translucent to ultraviolet wavelengths, and so Mathys and coworkers were able to obtain better scans by breaking down to ultraviolet scanning. I’m not sure this is the wave of the future, but it is a clever idea.