1. Despeckling of SAR Images Using Residual Twin CNN and Multi-Resolution Attention MechanismBlaž Pongrac, Dušan Gleich, 2023, original scientific article Abstract: The despeckling of synthetic aperture radar images using two different convolutional neural network architectures is presented in this paper. The first method presents a novel Siamese convolutional neural network with a dilated convolutional network in each branch. Recently, attention mechanisms have been introduced to convolutional networks to better model and recognize features. Therefore, we propose a novel design for a convolutional neural network using an attention mechanism for an encoder–decoder-type network. The framework consists of a multiscale spatial attention network to improve the modeling of semantic information at different spatial levels and an additional attention mechanism to optimize feature propagation. Both proposed methods are different in design but they provide comparable despeckling results in subjective and objective measurements in terms of correlated speckle noise. The experimental results are evaluated on both synthetically generated speckled images and real SAR images. The methods proposed in this paper are able to despeckle SAR images and preserve SAR features.
Keywords: synthetic aperture radar, speckle, speckle suppression, despeckling, deep learning, convolutional neural network
Published in DKUM: 21.02.2024; Views: 280; Downloads: 30
 Full text (13,70 MB)
This document has many files! More... |
2. |
3. |
4. MODELIRANJE SLIK SAR Z AVTO-BINOMSKIM MODELOMMarko Hebar, 2010, dissertation Abstract: Doktorska disertacija predstavlja odpravljanje pegastega šuma z modeliranjem in izločanjem informacij iz posnetkov SAR (ang. Synthetic Aperture Radar - SAR), ki je izvedeno z Bayesovim sklepanjem. Novost, ki jo predstavljam v doktorski disertaciji je uporaba avto-binomskega modela pri Bayesovem sklepanju prvega reda, kjer ga za apriorno verjetnost uporabimo pri modeliranju posnetka. Verjetje v Bayesovem sklepanju prvega reda modelira pegasti šum, ki ga opišemo z gama porazdelitvijo. Odpravljanje pegastega šuma je izvedeno s cenilko največje verjetnosti MAP (ang. maximum a posteriori - MAP), ki jo je analitično zelo težko rešiti, zato z avto-binomskim modelom uporabimo aproksimacijo z diferencialom. Parametri avto-binomskega modela se določijo z Bayesovim sklepanjem drugega reda. Robovi v posnetku se določijo z algoritmom rasti regij. Glede na koeficient variacije so ločene homogene od heterogenih regije in adaptivno prilagajanje okolice avto binomskega modela. Eksperimentalni rezultati so pokazali, da predstavljena metoda zelo dobro modelira teksture in ima dobre lastnosti odpravljanja pegastega šuma in zelo dobro modelira teksture v realnih posnetkih SAR.
Keywords: Avto-binomski model, izločevanje informacij, maksimum a posteriori (MAP) cenilka, Synthetic aperture radar (SAR), ohranjevanje tekstur.
Published in DKUM: 24.01.2011; Views: 2910; Downloads: 337
Full text (37,81 MB) |
