Integrating spectral and textural features for urban land cover classification with hyperspectral data

Abstract

This paper presents a supervised classification framework that integrates discrete wavelet transform (DWT) based spectral and textural features for the urban land cover classification using hyperspectral data. Investigations involved application of 1-D DWT along the wavelength dimension of the hyperspectral data followed by 2-D DWT along spatial dimensions for spectral and texture feature extraction respectively. The combined spectral textural feature set is used for classification. The pixel wise classification on ROSIS data using SVM reveals that integration of spectral and textural information can better characterize the urban areas and statistically significantly improves the classification accuracy.