Information-theoretic assessment of sampled hyperspectral imagers

This work focuses on estimating the information conveyed to a user by hyper spectral image data. The goal is establishing the extent to which an increa se in spectral resolution enhances the amount of usable information. Indeed , a tradeoff exists between spatial and spectral resolution due to physical constraints of multiband sensors imaging with a prefixed SNR. After descri bing an original method developed for the automatic estimation of variance and correlation of the noise introduced by hyperspectral imagers, lossless interband data compression is exploited to measure the useful information c ontent of hyperspectral data. In fact, the bit rate achieved by the reversi ble compression process takes into account both the contribution of the "ob servation" noise (i.e., information regarded as statistical uncertainty, bu t whose relevance to a user is null) and the intrinsic information of radia nce sampled and digitized through an ideally noise-free process. An entropi c model of the decorrelated image source is defined and, once the parameter s of the noise, assumed to be Gaussian and stationary, have been measured, such a model is inverted to yield an estimate of the information content of the noise-free source from the code rate. Results are reported and discuss ed on both simulated and AVIRIS data.