Translator Disclaimer
3 June 2011 Anomaly recovery from compressed spectral imagery via low-rank matrix minimization
Author Affiliations +
This work describes a methodology for the recovery of anomalies and their spectral signatures from compressively sensed multi-spectral video using Principal Component Pursuit (PCP). In video surveillance, approaches based on PCP allow the anomaly detection in a cluttered background by modeling a sequence of video frames as a large data matrix composed by a low-rank matrix plus a sparse matrix. The low-rank matrix corresponds to the stationary background and the sparse matrix captures the anomalies in the foreground. The compressive spectral video frames are attained by the use of a Coded Aperture Snapshot Spectral Imaging (CASSI) system. The CASSI system allows the compressive measurement of spectrally rich video content by simply capturing a sequence of 2D coded aperture video frames. This paper describes improved procedures for the reconstruction of the video anomalies and their spectra based on the 2-D, aperture-coded, isolated anomalies.
© (2011) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Ana Ramirez, Henry Arguello, and Gonzalo R. Arce "Anomaly recovery from compressed spectral imagery via low-rank matrix minimization", Proc. SPIE 8058, Independent Component Analyses, Wavelets, Neural Networks, Biosystems, and Nanoengineering IX, 805806 (3 June 2011);

Back to Top