The paper presents an efficient unsupervised change mapping algorithm for 1-look amplitude SAR images. The method exploits the complementary characteristics of two change features and proposes a morphological non-parametric map combination to produce the final change map. The method is able to preserve the geometry of the changed regions without increasing the overall false alarm rate PFA . This is made possible by morphologically combining the binary change maps obtained from a low-PFA high-order statistics (CKLD) change feature and a detail-preserving multiscale ratio detector. A new thresholding method is also proposed for the CKLD feature. Experimental tests, performed on simulated 1-look SAR images, visually and objectively demonstrate the advantages of the proposed algorithm.
Garzelli, A., Zoppetti, C. (2019). Geometrically Accurate Change Mapping From VHR SAR Images. In Proc. IEEE IGARSS 2019 (pp.25-28). IEEE [10.1109/IGARSS.2019.8900419].
Geometrically Accurate Change Mapping From VHR SAR Images
Garzelli, Andrea
;Zoppetti, Claudia
2019-01-01
Abstract
The paper presents an efficient unsupervised change mapping algorithm for 1-look amplitude SAR images. The method exploits the complementary characteristics of two change features and proposes a morphological non-parametric map combination to produce the final change map. The method is able to preserve the geometry of the changed regions without increasing the overall false alarm rate PFA . This is made possible by morphologically combining the binary change maps obtained from a low-PFA high-order statistics (CKLD) change feature and a detail-preserving multiscale ratio detector. A new thresholding method is also proposed for the CKLD feature. Experimental tests, performed on simulated 1-look SAR images, visually and objectively demonstrate the advantages of the proposed algorithm.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/11365/1091778