Computational Hyperspectral Imaging: Math, Methods, and Mayhem
When
Noon, Nov. 16, 2018
Speaker
Abstract
Hyperspectral imaging has many applications, including food quality and safety, cancer screening, crop assessments in agriculture, and fake currency detection, though hyperspectral imaging systems can be excessively complex and expensive. To simplify these optical systems, computational methods are used to calculate hyperspectral data from specially encoded optical signals. We will break down the mathematical structuring of a patented optical encoding system, explore some deconvolution methods to calculate hyperspectral information from encoded signals, and discuss the implications of different real-world training data sets.