Coded Aperture Snapshot Imaging Based Spectral Classification
Sehoon Lim, Choongyeun Cho, Aveek Das, and Sek Chai
Computational Optical Sensing and Imaging (COSI) 2014 (Accepted), Jun 2014
Abstract
Coded aperture snapshot imaging (CASSI) is a new method that allows for multispectral imaging (MSI) in a compact form. A CASSI data cube is spectrally classified using a spectral classifier, adaptive cosine estimator (ACE), to detect target objects. In order to improve a signal-to-noise ratio (SNR) of a data cube, a multi-frame CASSI with frame stabilization is presented. The video-stabilized (VS) CASSI improves the SNR of a CASSI data cube without sacrificing the temporal resolution because the video-rated frames are stabilized and used. The VS-CASSI also reduces volume and weight of the optical system by simply using the video-rated frames.
The spectral classification quantitatively scores and classifies the spectra of image pixels resulting in the detection of target objects in the scene.
<<Go back to the previous page