A stochastic model for restoration of and edge extraction from synthetic aperture radar (SAR) images is presented. This model is based on an observation-prediction model which favors the restoration of piecewise-constant patches separated by long continuous edges. Speckle noise is filtered out by means of an a-priori fixed probability distribution dependent on the number of views. Results on synthetic and real images are reported. To reduce the computational cost, simulated annealing is replaced with a deterministic algorithm based on the weak membrane model and on the mean field equations adapted to SAR.
Time-invariant filtering and segmentation of SAR images by using mean-field annealing
Regazzoni, C.
1992-01-01
Abstract
A stochastic model for restoration of and edge extraction from synthetic aperture radar (SAR) images is presented. This model is based on an observation-prediction model which favors the restoration of piecewise-constant patches separated by long continuous edges. Speckle noise is filtered out by means of an a-priori fixed probability distribution dependent on the number of views. Results on synthetic and real images are reported. To reduce the computational cost, simulated annealing is replaced with a deterministic algorithm based on the weak membrane model and on the mean field equations adapted to SAR.
File in questo prodotto:
Non ci sono file associati a questo prodotto.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
