Information capacity as a figure of merit for spectral imagers: the trade-off between resolution and coregistration

The performance of spectral imagers is customarily described by several characteristics including resolution, noise, and coregistration. These must be traded off against each other in a practical imager design. This paper proposes a way to use the information capacity, in an information-theoretic sense, as a figure of merit for spectral imagers. In particular, it is shown how a metric [Opt. Express 20, 918 (2012)] can be used to incorporate coregistration performance in a definition of total noise, which in turn can be used in the definition of information capacity. As an example, it is shown how the information capacity can be used to optimize the pixel size in a simple case that can be treated analytically. Generally, the information capacity is attractive as a fundamental, application-independent figure of merit for spectral imager optimization and benchmarking.