Muses9-SnS Snapshot spectral imager

How would you define the ideal hyperspectral imager? you would agree that it shall acquire all spectral images instantly and at video rates and display all of them, together with the result of their analysis, i.e., spectral classification maps, in real-time!     

Pushing the limits of what is possible in imaging science, Spectricon made this happen! MUSES9-SnS is the world’s first hyperspectral imager acquiring hypercube image packs at video rate and displaying both spectral images and classification maps side-by side and in real-time.   There is no tradeoff between spectral and spatial resolution, with the later reaching 4K levels! 

The MUSES9-SnS Snapshot spectral map viewer has come to revolutionize the field of snapshot hyper/multi-spectral imaging, by effectively implementing a new concept that removes the historical tradeoff between spatial resolution, spectral resolution and processing time.  

State-of-the-art snapshot spectral imagers are increasingly attracting the interest of the imaging community because they can acquire hypercube data instantly. As such, they are suitable for live multi/hyper spectral inspection, supporting handheld set ups. However, current commercial snapshot spectral imagers offer high spectral resolution at the cost of spatial resolution. Typical spatial resolutions offered by these systems are quite low, being in the range of 400X400 pixels. This is because they employ either lens arrays of multiband mosaic filter arrays, projecting a plurality of filtered images onto an imaging sensor, thus dividing accordingly its effective area into several portions dedicated to specific spectral bands. This way, there is always trade-off between spectral and spatial resolution.

The MUSES9-SnS innovative technology combines the real-time acquisition of a sufficient set of 4K level spectral images, with an AI-based spectral estimation technique to compose a full hypercube cube, without compromising spectral or spatial resolution. Advanced ultrafast spectral classification runs is parallel with the video rate acquisition, ending up to displaying supervised or unsupervised classification maps in real time.