Compressive optical MONTAGE photography initiative: Noise and error analysis

Abstract

Image resolution is usually considered to be proportional to the focal plane sampling rate. However, as a result of Papoulis's generalized sampling theory [1], it is possible to reconstruct bandlimited images from sub-Nyquist sampled data provided multiple channels of such data are measured, with each set suitably transformed to permit the computation of the original image. The COMP-I program uses focal plane coding to set sub-band sampling characteristics. Analysis and simulation of this approach to errors and noise is presented. © 2005 Optical Society of America.

Year