The problem of estimating the coherence area, namely the maximum size of a set of data with one and the same texture signal, is addressed in this paper. Based upon the compound-Gaussian model for the received data, the problem is stated in terms of a composite binary hypothesis test. Two different strategies yielding two parameter-free tests have been envisaged for the cases where the in-phase and quadrature components of the signal or only its intensity are available respectively. For both cases, the performance of the tests has been investigated through its operating characteristics.
|Titolo:||Fitting the compound-Gaussian model to SAR sea images|
|Data di pubblicazione:||1993|
|Appare nelle tipologie:||4.1 Contributo in Atti di convegno|