A fuzzy inference system for multispectral image classification

Abstract

This work presents an approach for multispectral image classification that makes use of a Fuzzy Inference System (FIS). An IKONOS satellite sensor image of a neighborhood in Rio de Janeiro, Brazil has been used. The ground truth used in this work comprises six classes: trees, scrub, buildings, roads, water and shadows. Then, inputs sets, rules and outputs sets were defined. Four input variables, based on four indexes computed from the image spectral bands, have been considered: Normalized Difference Vegetation Index (NDVI), Buildings Index (BI), Water Index (WI) and Road Index (RI). Experimental results show that the proposed method outperforms other methods, achieving higher Overall and Average accuracies, providing a better representation of the classification process.