Frank Sippel investigates multispectral and hyperspectrals cameras, which capture visual data that is invisible to the human eye. Such images can be applied in a large variety of applications in areas such as farming, medical imaging, or recycling.
When capturing hyperspectral data with a camera array, highly complex signal processing needs to be performed. Unfortunately, there is no data available helping to evaluate the quality of the reconstruction. To this end, Frank Sippel developed an approach to generate hyperspectral video data with state-of-the-art video rendering solutions. These hyperspectral videos can be generated for cameras at multiple positions, which allows to simulate data from our Camera Array for Multispectral Imaging (CAMSI). In addition, the generated dataset can be used for versatile applciations such as simulation and evaluation of different hyperspectral sensors or for the investigation of hyperspectral video compression methods.