Classification techniques for full-polarimetric (full-pol) Synthetic Aperture Radar (SAR) data utilize statistical and physical scattering properties. However, limited by polarimetric information, dual-polarimetric (dual-pol) SAR data classification often relies on backscatter intensity, leading to misclassification due to the inability of conventional descriptors to distinguish specific elementary targets. This study introduces a methodical unsupervised clustering technique for dual-pol SAR data, utilizing a target characteristic parameter that effectively distinguishes “dihedral-like” from “surface-like” targets within a scene. The proposed approach achieves enhanced land cover discrimination for dual-pol Single-Look Complex (SLC) and Ground Range Detected (GRD) SAR data.
DUCAT: Dual-pol Unsupervised Clustering and Analysis Technique
Bhattacharya, Avik;Gamba, Paolo
2025-01-01
Abstract
Classification techniques for full-polarimetric (full-pol) Synthetic Aperture Radar (SAR) data utilize statistical and physical scattering properties. However, limited by polarimetric information, dual-polarimetric (dual-pol) SAR data classification often relies on backscatter intensity, leading to misclassification due to the inability of conventional descriptors to distinguish specific elementary targets. This study introduces a methodical unsupervised clustering technique for dual-pol SAR data, utilizing a target characteristic parameter that effectively distinguishes “dihedral-like” from “surface-like” targets within a scene. The proposed approach achieves enhanced land cover discrimination for dual-pol Single-Look Complex (SLC) and Ground Range Detected (GRD) SAR data.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
