Spectrally Adaptable Compressive Sensing Imaging System
Abstract
Compressive Spectral Imaging is a revolutionary technique which senses the spatio-spectral information of a scene by using 2D coded projections. The underlying spectral 3D data cube is then recovered using compressed sensing (CS) reconstruction algorithms which assume that the underlying hyperspectral images are sparse in some representation basis. The great advantage of CSI is that the required number of measurements needed for reconstruction is far less than that required by traditional scanning methods. In practice, the Coded Aperture Snapshot Spectral Imaging (CASSI) systems efficiently implement CSI. The compressive CASSI measurements are often modeled as the summation of coded and shifted versions of the spectral voxels of the underlying scene. Thus, each CASSI measurement is a highly structured random projection of the underlying scene. The structure of these projections is dictated by the CASSI optical system whose only varying components are the coded aperture entries. The coded apertures are crucial inasmuch as they determine the quality of the image reconstruction as well as the required minimum number of FPA measurements.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 2014
- Accession Number
- ADA610295
Entities
People
- Dennis W. Prather
- Gonzalo R. Arce
- Javier Garcia-frias
Organizations
- University of Delaware