FUsion MULTIspectral-imSPOC – FuMultiSPOC

The project FUsion MULTIspectral-imSPOC (FuMultiSPOC) aims to design
an innovative multi-sensor hyperspectral imaging system for generating
images with very high spectral and spatial resolution along with depth
information.

This project is based on the ImSPOC concept (IPAG-ONERA patent, 2018), a
novel technology generating snapshot hyperspectral images by a
miniaturized static interferometric spectrometer. The ImSPOC technology
is being explored focusing on the ultraviolet and infrared in specific
applications, such as the monitoring of atmospheric gasses. The
preliminary results have shown the incontestable great potential of this
technology due to its high spectral resolution and compactness but also
the difficulties linked to its use as general purpose technology for
hyperspectral imaging applications, such as those related to environment
monitoring by remote sensing. Specifically, the ImSPOC technology
suffers from difficulties due to a sensitive set-up, the need for a
precise characterization of the instrument and the essential role played
by the processing of the acquired data in order to access the spectral
information of a scene. A further limitation of ImSPOC is given by the
low spatial resolution of the acquisitions due to the snapshot
configuration, making this technology unsuitable for several civil
and military applications.

To overcome these problems \acro proposes to develop a multi-sensor
system by coupling an ImSPOC imager with a conventional multispectral
camera of higher spatial resolution. This configuration gives two
advantages: i) the use of the spectral information at low resolution
(multispectral image) as a reference for a more precise estimation of
spectra from the interferometric acquisitions of ImSPOC; and ii) the
generation of hyperspectral images with high spectral and spatial
resolution by data fusion. In addition, the ImSPOC configuration makes
it similar to a plenoptic camera. This will be used to estimate depth
(3D) information of a scene, which can be of high spatial resolution
after fusion with the multispectral image. The methodological
developments for performing the proposed processing by data fusion
require advanced skills in signal and image processing, optical design,
inverse problems and optimization, which are ensured by the recognized
experience of the partners.

The features of the proposed multisensor system for acquiring high
spectral-, spatial-resolution and 3D information of a scene
along with its compact volume is of great potential for the French
hyperspectral community with several expected scientific,
industrial, economic and societal outcomes.