Hyperspectral imaging (SICOM SIGMA S9) - 5PMSHIA0

Goal(s)

Hyperspectral imaging is a non-conventional imaging mode that typically records for each pixel its reflectance in a large number of wavelengths, narrow and contiguous. This imaging mode provides a much finer discrimination of the imaged materials than the classical optical imaging sensors which operate only in three broad spectral ranges (red-green-blue). The objective of this course is to familiarize students with this imaging mode, its challenges and applications, and the specific questions raised by the high dimensionality of the data from a data science point of view.

Content(s)

The course will present an introduction to hyperspectral imaging, its challenges and the potential offered in applications as varied as material sciences, medical imaging, environmental observations, planetary exploration or astrophysics.
We will then review the different challenges to be solved on the whole processing chain, from acquisition, corrections (atmospheric and other), denoising, and data exploitation (spectral unmixing, super-resolution, classification, segmentation...). A part of the course will be devoted to astrophysical applications.

Basics in signal processing and image processing