A Novel Filter Approach for Band Selection and Classification of Hyperspectral Remotely Sensed Images Using Normalized Mutual Information and Support Vector Machines

TL;DR

This paper presents a new filter-based method combining normalized mutual information and support vector machines for effective band selection and classification of hyperspectral images, demonstrating improved performance on benchmark datasets.

Contribution

Introduces a novel filter approach using normalized mutual information and SVMs for dimension reduction and classification of hyperspectral images, outperforming existing methods.

Findings

Achieves high classification accuracy with fewer bands.

Reduces computational time compared to state-of-the-art methods.

Effective on NASA AVIRIS benchmark datasets.

Abstract

Band selection is a great challenging task in the classification of hyperspectral remotely sensed images HSI. This is resulting from its high spectral resolution, the many class outputs and the limited number of training samples. For this purpose, this paper introduces a new filter approach for dimension reduction and classification of hyperspectral images using information theoretic (normalized mutual information) and support vector machines SVM. This method consists to select a minimal subset of the most informative and relevant bands from the input datasets for better classification efficiency. We applied our proposed algorithm on two well-known benchmark datasets gathered by the NASA's AVIRIS sensor over Indiana and Salinas valley in USA. The experimental results were assessed based on different evaluation metrics widely used in this area. The comparison with the state of the art methods proves that our method could produce good performance with reduced number of selected bands in a good timing. Keywords: Dimension reduction, Hyperspectral images, Band selection, Normalized mutual information, Classification, Support vector machines