Spatial Spectral Band Selection for Enhanced Hyperspectral Remote Sensing Classification Applications.

Ruben Moya Torres,Chris Mccullough,Changfeng Yuan,Johathan Piper,Peter Godfree,Peter W.T Yuen

doi:10.3390/jimaging6090087

Ruben Moya Torres, Chris Mccullough + Show 4 more

Open Access

https://doi.org/10.3390/jimaging6090087

Copy DOI

Abstract

Despite the numerous band selection (BS) algorithms reported in the field, most if not all have exhibited maximal accuracy when more spectral bands are utilized for classification. This apparently disagrees with the theoretical model of the ‘curse of dimensionality’ phenomenon, without apparent explanations. If it were true, then BS would be deemed as an academic piece of research without real benefits to practical applications. This paper presents a spatial spectral mutual information (SSMI) BS scheme that utilizes a spatial feature extraction technique as a preprocessing step, followed by the clustering of the mutual information (MI) of spectral bands for enhancing the efficiency of the BS. Through the SSMI BS scheme, a sharp ’bell’-shaped accuracy-dimensionality characteristic that peaks at about 20 bands has been observed for the very first time. The performance of the proposed SSMI BS scheme has been validated through 6 hyperspectral imaging (HSI) datasets (Indian Pines, Botswana, Barrax, Pavia University, Salinas, and Kennedy Space Center (KSC)), and its classification accuracy is shown to be approximately 10% better than seven state-of-the-art BS schemes (Saliency, HyperBS, SLN, OCF, FDPC, ISSC, and Convolution Neural Network (CNN)). The present result confirms that the high efficiency of the BS scheme is essentially important to observe and validate the Hughes’ phenomenon in the analysis of HSI data. Experiments also show that the classification accuracy can be affected by as much as approximately 10% when a single ‘crucial’ band is included or missed out for classification.

Highlights

Hyperspectral imaging (HSI) that exploits both spectral and spatial features of the scene [1,2], has made it a powerful technique for applications such as geographical mapping [3], classifications [4], and target detections [5,6], in multidisciplinary fields of agricultural [7], food industry [8], medical [9], and security [10], sectors
Band Selection (BS) Using Spectral Information Only classified by chance
According to the results presented in the last section, it is clear that the elimination of the nonAccording to the results presented inthe thelast last section, it is that clearthethat the elimination of the According to the results presented in it is the clear elimination discriminative bands are essentially important forsection, enhancing classification accuracy.ofItthe is nonalso non-discriminative bands are essentially important for enhancing the classification accuracy

Summary

Introduction

Hyperspectral imaging (HSI) that exploits both spectral and spatial features of the scene [1,2], has made it a powerful technique for applications such as geographical mapping [3], classifications [4], and target detections [5,6], in multidisciplinary fields of agricultural [7], food industry [8], medical [9], and security [10], sectors. The usefulness of HSI mainly stems from the very detailed spectral information of the scene that it provides; it is one of the drawbacks of HSI for achieving a high degree of classification or detection accuracy when it has high spectral dimension. This is so-called ‘curse of dimensionality’ that manifests itself by the presence of a ‘bell’

Objectives

Methods

Conclusion