High Spectral Spatial Resolution Synthetic HyperSpectral Dataset form multi-source fusion. (arXiv:2309.00005v1 [cs.CV])

This research paper introduces a synthetic hyperspectral dataset that
combines high spectral and spatial resolution imaging to achieve a
comprehensive, accurate, and detailed representation of observed scenes or
objects. Obtaining such desirable qualities is challenging when relying on a
single camera. The proposed dataset addresses this limitation by leveraging
three modalities: RGB, push-broom visible hyperspectral camera, and snapshot
infrared hyperspectral camera, each offering distinct spatial and spectral
resolutions. Different camera systems exhibit varying photometric properties,
resulting in a trade-off between spatial and spectral resolution. RGB cameras
typically offer high spatial resolution but limited spectral resolution, while
hyperspectral cameras possess high spectral resolution at the expense of
spatial resolution. Moreover, hyperspectral cameras themselves employ different
capturing techniques and spectral ranges, further complicating the acquisition
of comprehensive data. By integrating the photometric properties of these
modalities, a single synthetic hyperspectral image can be generated,
facilitating the exploration of broader spectral-spatial relationships for
improved analysis, monitoring, and decision-making across various fields. This
paper emphasizes the importance of multi-modal fusion in producing a
high-quality synthetic hyperspectral dataset with consistent spectral intervals
between bands.