Land Classification using a novel Multispectral and SAR data Fusion in Doha area

Conference: EUSAR 2018 - 12th European Conference on Synthetic Aperture Radar
06/04/2018 - 06/07/2018 at Aachen, Germany

Proceedings: EUSAR 2018

Pages: 5Language: englishTyp: PDF

Personal VDE Members are entitled to a 10% discount on this title

Iervolino, Pasquale; Guida, Raffaella; Ayesh-Meagher, Amelia (Surrey Space Centre, University of Surrey, UK)

A new algorithm for land classification is presented in this paper and is based on the fusion of Multispectral, Panchromatic and Synthetic Aperture Radar (SAR) images. The novel approach relies on Generalized Intensity-Hue-Saturation (G-HIS) transform and the Wavelet Transform (WT). The fused image is derived by modulating the SAR texture with the high features details of the Panchromatic WT and by injecting this product at the place of G-HIS high feature details. Finally, a classification is performed on the fused product by using a Maximum Likelihood (ML) classifier. The algorithm has been tested on data acquired by Sentinel-1 (SAR) and Landsat-8 (Multispectral and Panchromatic) over the area of Greater Doha in Qatar in 2017. Results show an increment of 3% in the overall accuracy for the fused product compared to the Multispectral dataset.