A Novel Approach for Band Selection Using Virtual Dimensionality Estimate and Principal Component Analysis for Satellite Image Classification

Smriti Sehgal, Laxmi Ahuja, M. Hima Bindu

Source Title: International Journal of Intelligent Information Technologies (IJIIT)18(2)

ISSN: 1548-3657|EISSN: 1548-3665|EISBN13: 9781799893837|DOI: 10.4018/IJIIT.296272

MLA

Sehgal, Smriti, et al. "A Novel Approach for Band Selection Using Virtual Dimensionality Estimate and Principal Component Analysis for Satellite Image Classification." IJIIT vol.18, no.2 2022: pp.1-16. http://doi.org/10.4018/IJIIT.296272

APA

Sehgal, S., Ahuja, L., & Bindu, M. H. (2022). A Novel Approach for Band Selection Using Virtual Dimensionality Estimate and Principal Component Analysis for Satellite Image Classification. International Journal of Intelligent Information Technologies (IJIIT), 18(2), 1-16. http://doi.org/10.4018/IJIIT.296272

Chicago

Sehgal, Smriti, Laxmi Ahuja, and M. Hima Bindu. "A Novel Approach for Band Selection Using Virtual Dimensionality Estimate and Principal Component Analysis for Satellite Image Classification," International Journal of Intelligent Information Technologies (IJIIT) 18, no.2: 1-16. http://doi.org/10.4018/IJIIT.296272

Export Reference

Favorite Full-Issue Download

View Full Text HTML

View Full Text PDF

Abstract

Images, being around us in every aspect of life, have become an emerging field of research. Extensive image analysis has been done on binary as well as coloured images, which has led various researchers to explore images having deep spectral knowledge about a particular area of interest. High resolution images, having more than three spectral bands, capture minute details of an object in various spectral bands resulting in high computational complexity. In this paper, the authors have tried to reduce the complexity of multispectral image by selecting only the relevant bands need to reconstruct an image. Traditional principal component analysis technique is used for band selection of true color bands and classification-assessed results of both the images; original and dimensionality reduced images are compared using partitioning clustering technique. Experimental results show that compressed image after reduction of bands by PCA yields better classification results than the original image.