New wrapper method based on normalized mutual information for dimension reduction and classification of hyperspectral images

TL;DR

This paper introduces a new wrapper feature selection method using normalized mutual information and SVM to enhance hyperspectral image classification, demonstrating improved accuracy on benchmark datasets.

Contribution

A novel wrapper feature selection approach combining NMI and error probability with SVM for hyperspectral image classification.

Findings

Improved classification accuracy on AVIRIS datasets

Effective reduction of redundant spectral bands

Enhanced thematic map quality

Abstract

Feature selection is one of the most important problems in hyperspectral images classification. It consists to choose the most informative bands from the entire set of input datasets and discard the noisy, redundant and irrelevant ones. In this context, we propose a new wrapper method based on normalized mutual information (NMI) and error probability (PE) using support vector machine (SVM) to reduce the dimensionality of the used hyperspectral images and increase the classification efficiency. The experiments have been performed on two challenging hyperspectral benchmarks datasets captured by the NASA's Airborne Visible/Infrared Imaging Spectrometer Sensor (AVIRIS). Several metrics had been calculated to evaluate the performance of the proposed algorithm. The obtained results prove that our method can increase the classification performance and provide an accurate thematic map in comparison with other reproduced algorithms. This method may be improved for more classification efficiency. Keywords-Feature selection, hyperspectral images, classification, wrapper, normalized mutual information, support vector machine.