The paper presents a method for the classification of urban hyperspectral data with high spatial resolution, combining both spectral and spatial information. The approach extends previous methods by using Principal Components (PCs) from the hyperspectral data to build morphological profiles, which are then extended to create an Extended Morphological Profile (EMP). The EMP is fused with the original hyperspectral data through feature extraction techniques, such as Decision Boundary Feature Extraction (DBFE) and Nonparametric Weighted Feature Extraction (NWFE), to form a stacked vector. This vector is classified using a Support Vector Machine (SVM) classifier. The method is tested on ROSIS data from urban areas, showing significant improvements in accuracy compared to pixel-wise classification and morphological profile-based methods. The proposed approach also demonstrates excellent performance with limited training sets, making it suitable for practical applications.The paper presents a method for the classification of urban hyperspectral data with high spatial resolution, combining both spectral and spatial information. The approach extends previous methods by using Principal Components (PCs) from the hyperspectral data to build morphological profiles, which are then extended to create an Extended Morphological Profile (EMP). The EMP is fused with the original hyperspectral data through feature extraction techniques, such as Decision Boundary Feature Extraction (DBFE) and Nonparametric Weighted Feature Extraction (NWFE), to form a stacked vector. This vector is classified using a Support Vector Machine (SVM) classifier. The method is tested on ROSIS data from urban areas, showing significant improvements in accuracy compared to pixel-wise classification and morphological profile-based methods. The proposed approach also demonstrates excellent performance with limited training sets, making it suitable for practical applications.