Investigating intensity and transversal drift in hyperspectral imaging data

When measuring data with hyperspectral cameras drift in the data distribution occurs over time and when the sensing device is changed. Frequently, this drift is characterized by intensity shift or wavelength shifts. In this contribution, we propose novel methods that reverse these shifts and demonstrate their capability to avoid the negative impact of drift on the classification performance. We show that our approaches perform on par or better in comparison to established methods. We also provide a theoretical motivation why one of the proposed methods can deal with both, intensity and wavelength shift provided bounds on the smoothness of the functional data are given.